IIIIIIIIIIIIIIII'IIIIIIIII'

 

     
     

Illllgtlltllll/ltzll
U zutisirc

This is to certify that the

thesis entitled
COMPARISON OF AN ESTIMATE OF SKELETAL AGE
WITH CHRONOLOGICAL AGE
WHEN CLASSIFYING ADOLESCENT MALES
FOR MOTOR PROFICIENCY NORMS

presented by

Michael Grant Marshall

has been accepted towards fulfillment
of the requirements for

Ph- D. degeeinfiemmsical

Education, and Recreation

//éM’///Z///

Major profesg

 

Date April 19, 1978

 

0-7639

 

 

 

A T“
~ 0813*”‘973

© 1978

MICHAEL GRANT MARSHALL

ALL RI GHTS RESERVED

COMPARISON OF AN ESTIMATE OF SKELETAL AGE
WITH CHRONOLOGICAL AGE
WHEN CLASSIFYING ADOLESCENT MALES

FOR MOTOR PROFICIENCY NORMS

BY

Michael Grant Marshall

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Health, Physical Education
and Recreation

1978

ABSTRACT
COMPARISON OF AN ESTIMATE OF SKELETAL AGE WITH CHRONOLOGICAL

AGE WHEN CLASSIFYING ADOLESCENT MALES FOR MOTOR
PROFICIENCY NORMS

BY

Michael Grant Marshall

This investigation sought 1) to determine the proportion of the
variance in the skeletal development of adolescent males that is
explained by chronological age, height, weight, and a Sexual Maturation
Indicator Value (SMIV) estimate, and 2) to determine how an estimate of
skeletal age (BSA) that incorporated chronological age, height, weight,
and SMIV compared with chronological age when classifying adolescent
males for motor proficiency norms.

To answer part one required a sample of adolescent males who
represented a rectangular distribution of sexual maturation. Hand-wrist
X-rays, height and weight were obtained on 30 randomly selected boys
from East Lansing, Michigan who ranged in age from 134.5 to 180.0 months.
Four Physical Education teachers and the author assessed five primary
and secondary sexual characteristics on two separate occasions. The
sexual characteristics were; 1) facial hair, 2) axillary hair, 3) pubic
hair, 4) penis development and 5) scrotum and testes development. Via
the Greulich-Pyle Atlas, an experienced rater of hand-wrist X-rays
assessed the skeletal ages of the boys on two separate occasions.

The analysis determined that for this sample: (1) chronological
age explained 45 percent of the variance in the skeletal age of the
adolescent males; (2) height explained an additional 20 percent;

(3) weight contributed an additional 6 percent; and (4) the Sexual

Michael Grant Marshall

Maturation Indicator Value (SMIV) explained 12 percent more. Therefore,
the combination of chronological age, height, weight and an estimate

of sexual maturity explained 83 percent of the variance in the skeletal
age of the sample.

The regression analysis indicated that the most efficient combi-
nation of secondary sex characteristics for estimating SMIV was:

SMIV = 126.47 + 174(X(F)) + 1.20(X(A)) + 2.75(X(PH)) + 6.68(X(P))
where; F is the boys' facial hair rating, A is the boys' axillary hair
rating, PH is the boys' pubic hair rating, and P is the boys' rating
of penis development.

To answer part two required a large cross-section sample of
adolescent males. Therefore, at the author's request, the boys'
physical education teachers in the East Lansing and Mason, Michigan
Junior High Schools collected the required data on the entire male
populations of their schools for three years. The cross-sectional
sample consisted of 1578 boys on whom their teachers determined their~
chronological age, measured their height and weight, tested their ability
to perform sit-ups, pull—ups, standing long jumps, vertical jumps, and
shuttle runs, and assessed five primary and secondary sexual character-
istics.

The Multivariate Analysis of Variance computer program calculated
the raw regression coefficient and mean values for the chronological age,
height, weight, and SMIV estimate variables to create the estimate of
skeletal age (ESA) regression equation. By using identical classifi-
cation intervals for the ESA estimate and the chronological age (CA)

value, the analysis examined the within-cell variances for the five motor

Michael Grant Marshall

proficiency tests and the mean value progressions across the annual
increments.

Although the within cell variances for the classifications by CA
and BSA were not significantly different, the author concluded that
compared to CA, BSA classified adolescent males for motor proficiency
tests better because: 1) BSA is less likely to reward the competitors
of the basis of advanced biological maturation; and 2) whereas CA had
a preponderance of 13 and 14 year old subjects, SA dispersed the subjects
throughout the adolescent age range, indicating that BSA moved the
accelerated and delayed-maturers to groupings of similar biological
maturation. The regression analysis indicated that the most efficient
estimate of skeletal age was:

BSA = -16.15 + .277(X(CA)) + .212(X(H)) + .220(x(W)) = .594(X(SMIV))
where; CA is the boys' chronological age in months, H is the boy's
height in inches, W is the boys' weight in pounds, and SMIV is the

estimate of the boys' sexual maturation.

DEDICATION :

To Moose, Runt, and Pizwilly, three of the greatest kids that
a Dad could have.

And, to all of the boys whose athletic desires the present system
of athletic competition vanquished.

iii

ACKNOWLEDGEMENTS

At the East Lansing and Mason, Michigan Junior High Schools,
the author thanks the Boys Physical Education teachers; Craig Marsh,
Wally Juall, Ron Hornung, Dave Wills, Rich Wilson, and Marvin Pulver.
Special thanks goes to Jim Oestriech for his ideas, help, and encourage-
ment, to Sal DiFranco for his interest, and to the boys.

At Michigan State University, the author thanks his doctoral
guidance committee; chairman Vern Seefeldt, Glenn Hatton, Bill Heusner,
Rex Carrow, and Bill Schmidt. For his many hours of extra help, special

thanks goes to Bill Schmidt.

iv

TABLE OF CONTENTS

Chapter Page
I. INTRODUCTION . . . . . . . . . . . . . . 1
Need for the study . . . . . . . . . . . 6

Scope of the study . . . . . . . . . . . 7
Limitations of the study . . . . . . . . . 7

II. REVIEW OF LITERATURE . . . . . . . . . . . 9

History of the Assessment of Hand and Wrist Skeletal
Development . . . . . . . . . . . . . 10
Summary . . . . . . . . . . . . . . . 80

History of the Assessment of Adolescent Males' Sexual
ncharacteristic development . . . . . . . . 81
Summary 0 O O O O I O O O C O O O O O 147

Relationship Between Motor Proficiencies and Various
Classification Procedures for Adolescent Males . 152

Summary . . . . . . . . . . . . . . . 162

Chapter Summary . . . . . . . . . . . . 162

III. METHODS AND PROCEDURES . . . . . . . . . . . 164
Experimental Design #1 . . . . . . . . . . 164

Data Collection # 1 . . . . . . . . . . . 164
Measures #1 . . . . . . . . . . . . . 165
Experimental Design #2 . . . . . . . . . . 166

Data Collection #2 . . . . . . . . . . . 166
Measures #2 . . . . . . . . . . . . . 167
Treatment of the Data . . . . . . . . . . 168

IV. RESULTS AND DISCUSSION . . . . . . . . . . . 168
Question I . . . . . . . . . . . . . . 173
Question II . . . . . . . . . . . . . . 175
Question III. . . . . . . . . . . . . . 177
Question IV . . . . . . . . . . . . . . 188

V. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS . . . . 194
Smary O O O O O O O O O O O O O O O 194
Conclusions . . . . . . . . . . . . . . 195
Recommendations. . . . . . . . . . . . . 196
REFERENCES CITED 0 O O O O O O O O O O O O O O 197

APPENDICES

APPENDIX
A. Sexual Maturation Indicator Value (SMIV) Checklist .
B. The Procedure that the Radiologists Used to X-ray the
Hands and Wrists of the Subjects . . . . . . .
C. The Procedures that Professor Seefeldt Used to Assess
the Hand—Wrist X-rays and His Test-Retest Reliabili-
ties O O C O O O O O O O O O O O O O
D. The Procedures that the Radiologists Used to Measure
the Height and Weight of the Subjects . . . . .
B. The Procedures that the Four Teachers and the Author
Used to Assess the Sexual Maturation of the 30 Subj-
ects Who Represented a Rectangular Distribution of
Sexual Maturation . . . . . . . . . . . .
F. The Procedures that the Teachers Used to Measure the
Height and Weight of the Subjects . . . . . .
G. The Procedures that the Teachers Used to Test Their
Subjects of the AAHPER's Motor Proficiency Tests .
H. The Procedures that the Teachers Used to Assess the
Sexual Maturation of their Subjects . . . . . .
I. The Step-Wise Regressions for the Reordered Chronolog-
ical Age, Height, and Weight Independent Variables .
J. The Procedures for Determining a Sexual Maturation
Indicator Value (SMIV) Regression Equation . . .
K. The Step-Wise Regressions for the Reordered Chronolog-
ical Age, Height, Weight, and SMIV Independent Varia-
bles O O O O O O O O O O O O. O O O O
L. The Statistical Analysis to Determine Whether or Not
the 878 East Lansing Subjects Differed From the 700
Mason subjects 0 O O O I O O O O O O O O
M. The Proportion of This Sample's 1578 Boys Who Reached
the Five Stages of the Five Sexual Characteristics .
N. The Regression Equation for the Estimate of Skeletal
Age (BSA) 0 O I O O O O O O O O O O O
0.

An Additional Classification Procedure that the Results

Suggested . . . . . . . . . . . . . .

vi

Page

203

207

210

212

214

216

220

222

225

232

235

259

275

287

'\

5..

a..-_

‘I

3‘

LIST OF TABLES

Table Page
2.1 The Age by Sex of Carpal Appearance in Children

(Pryor, 1908) O O C O O O O O O O O O O O 13
2.2 The Frequency of the Carpal Appearance Order in 243

Children (Pryor, 1908) . . . . . . . . . . . 14
2.3 The Ages by Sex that the Hand-Wrist Bpiphysial Centers

Appeared (Pryor, 1925) . . . . . . . . . . . 15
2.4 Alphabetized Anatomic Stage Index (Rotch, 1908) . . . 16
2.5 Pryor's Anatomic Stage Index (Rotch, 1909) . . . . . 17
2.6 Academic Grade-Anatomic Stage (Rotch, 1910 . . . . . 19

2.7 The Descriptions of the Eight Bpiphyseal Development

Stages in Adolescent Males (Smith, 1913) . . . . . 21
2.8 Four Carpal Ossification Classes . . . . . . . . 22
2.9 Eleven Carpal Ossification Stages . . . . . . . . 23

2.10 Correlation Table for Chronological Age, Weight, Height
Mental Age, and Carpal Age (Baldwin, 1922) . . . . 25

2.11 X-ray Standard Descriptions (Lowell and Woodrow, 1922) . 26

2.12 The Nine Measurement Lines that Prescott Drew on the
Hand-Wrist X-rays (Prescott, 1923) . . . . . . . 29

2.13 The Correlations that Prescott Found for His Carpal
Ossification Ratios (Prescott, 1923) . . . . . . 30

2.14 The Number, Sex, Age Range, and Anatomical Indices of
Prescott's Sample (Prescott, 1923) . . . . . . . 31

2.15 The Four Bpiphyseal Union Phases that Stevenson Described
(Stevenson, 1924). . . . . . . . . . . . . 33

2.16 The Correlation Values Between McBlhannon's Inspectional
Method and the Freeman-Carter Method (McBlhannon,
1926) o o o o o o o o o o o O o o o o o 36

2.17 The Correlation Values Between Buchanan's Inspectional
Method and the Freeman-Carter Method (Carter, 1926) . 37

2.18 The Correlation Values Between the Inspectional Methods
of McElhannon and David and the Freeman-Carter Method
(carter, 1926) O O O O O O O O O O O O O O 38

vii

Table

2.19

2.22

2.23

2.24

2.25

2.26

2.30

2.31

2.32

2.33

2.34

2.35

2.36

Anatomic Index Values for the Boys and Girls Aged
1 to 17 Years (Baldwin, Busby, and Garside, 1928) . .

The Five Stages of Bpiphyseal Ossification . . . . .

The Percent of the Bpiphyses that Developed via
Different Stages (Hellman, 1928). . . . . . . .

Sawtell's Comparison of the Bureau of Educational
Experiments Sample and the Cleveland Sample
(Sawtell, 1929) . . . . . . . . . . . . .

The Nine Bpiphysial Maturation Stages that Todd
Described (Todd, 1930b). . . . . . . . . . .

The Female Advantage Over Males in Skeletal Differ-
entiation (Todd, 1931) . . . . . . . . . . .

The Female Advantage Over Males in the Age of Bpiphysial
Union (Todd, 1931) . . . . . . . . . . . .

The Twelve Hand-Wrist Measurements that Cattell Used
in Her Search for an Objective Assessment Method
(cattellp 1934) o o o o o o o o o o o o 0
Four "Extent of Ossification" Measurements (Kelly, 1937)

Five "Size of the Individual" Estimates (Kelly, 1937) .

Pyle's Product-Moment Correlation for Fifty 12-Month
Girls (Pyle and Menino, 1939). . . . . . . . .

Pyle's Product-Moment Correlation for Fifty 24-Month
Girls (Pyle and Menino, 1939). . . . . . . . .

Between and Within Rater Correlations (Bayley, 1940). .
Hand and Knee Correlation Table (Bayley, 1940). . . .

Suggested Procedure for x-ray Assessment According to
Greulich and PYle (Greulich and Pyle, 1950) . . . .

Variability in the Skeletal Age of the Boys and Girls
in the Brush Foundation Study (Greulich and Pyle,
1950) o o o o o o c o o o o o o o o o 0

Number of Maturity Indicator Stages for the Twenty-One
Hand-Wrist Bones (Greulich and Pyle, 1950) . . . .

Assumptions of the Human Ossification Process that
Provided the Bases for the NCHS Atlas of Hand-Wrist
Maturity (Pyle, Waterhouse and Greulich, 1971) . . .

viii

Page

42

44

44

46

49

51

52

53

59

59

62

63

64

64

71

72

73

77

Table

2.37

2.38

2.39

2.40

2.41

2.42

2.43

2.44

2.45

2.46

2.47

2.48

2.49

2.50

2.51

2.52

The Equivalent Hand—Wrist Skeletal Ages (months) for
Males and Females (Pyle, Waterhouse, and Greulich,
1971) O o o o o o o o o o o o o o o o

Pubic Hair Stages . . . . . . . . . . . .

Percentage of Boys at Each Age that Crampton Rated in
Each Stage (Crampton, 1904) . . . . . . . .

The "Puberty Index" of Boys with United States versus
German heritage (Crampton, 1904). . . . . . .

The Mean values for Height at Each Age for Each Pubic
Hair Stage (Crampton, 1904) . . . . . . . .

The Mean Values for Weight at Each Age for Each Pubic
Hair Stage (Crampton, 1904) . . . . . . . .

The Mean Values for Right Hand Grip Strength at Each
Age for Each Pubic Hair Stage (Crampton, 1904) . .

Percentage of Boys in the Three Pubic Hair Stages at
Six Month Intervals (Crampton, 1908a) . . . . .

The Number of Positive Spermatozoa Samples Found at
Each Age Level (Baldwin, 1928) . . . . . . .

The Sexual Development Checklist that Kubitschek
Developed (Kubitschek, 1932) . . . . . . . .

The Observed Frequency of Boys at Bach Age Level that
Were Retarded, Average or Late Maturing (Kubitsche,
1932) o o o o o o o o o o o o o o o o

The Description of the Sexual Development Changes at
Each Age Level (Kubitschek, 1937) . . . . . .

The Six Pubescent Classifications that Dimock
Developed (Dimock, 1935) . . . . . . . . .

Richey's Maturity Group Description Based on Time of
Axillary Hair Appearance (Richey, 1937) . . . .

The Amount of Androgenic and Estrogenic Activity Found
in the Urine of Adolescent Males (Dorfman, Greulich,

SOlomon' 1937) o o o o o o o o o o o o o

Genitalia Measurements on Males From Birth to Adult
(Schonfeld and Beebe, 1942) . . . . . . . .

ix

and

Page

79

81

82

82

83

83

83

84

86

87

88

89

90

91

93

96

Table Page

2.53 The Order of Appearance of the External Sexual
Characteristics in Males (Greulich, Dorfman, Catchpole
Solomon, and Culotta, 1942) . . . . . . . . .. 98

2.54 Descriptions of the Sexual Characteristic Changes During
Adolescence in Males (Greulich, Dorfman, Catchpole,
Solomon, and Culotta, 1942) . . . . . . . . . 99

2.55 The Number and Mean Skeletal Age Values of 476 Adoles-
cent Males (Greulich, Dorfman, Catchpole, Solomon
and CUlOtta, 1942) o o a o o o o o o o 0 ¢ 101

2.56 The Androgenic Material Level (I.U.) Classified By
Maturity Groups (Greulich, Dorfman, Catchpole, Solomon,
and Culotta, 1942) . . . . . . . . . . . . 102

2.57 The Estrogenic Level (I.U.) Classified by Maturity
Groups (Greulich, Dorfman, Catchpole, Solomon, and
Culotta, 1942). . . . . . . . . . . . . . 102

2.58 The Estrogenic Material Level (I.U.) Classified by
Chronological Age Groups (Greulich, Dorfman,
Catchpole, Solomon, and Culotta, 1942) . . . . . . 102

2.59 A Comparison of Gonadotrophin Classifications with
Chronological Age (Greulich, Dorfman, Catchpole,
Solomon, and Culotta, 1942). . . . . . . . . . 103
2.60 A Comparison of Gonadotrophin Classifications With
the Maturity Groups (Greulich, Dorfman, Catchpole,
Solomon, and Culotta, 1942). . . . . . . . . . 103

2.61 Schonfeld's Six Sexual Maturation Stages (Schonfeld,
1943a) I O O O O O O O O O O O O O O O O 104

2.62 The Size Range Within the Six Sexual Maturation Stages
for Penis Length, Penis Circumference, and Testes

volume (Schonfeld, 1943a) . . . . . . . . . . 105

2.63 The Proportion of Boys at Bach Age in the Six Sexual
Maturation Stages (Schonfeld, 1943a). . . . . . . 105

2.64 Pubic Hair Stages (Schonfeld, 1943b) . . . . . . . 107
2.65 Pubic Hair Stages (Ellis, 1946) . . . . . . . . . 109
2.66 Axillary Hair Stages (Ellis, 1946) . . . . . . . . 109

2.67 Three Sexual Maturation Classes for Adolescent Males
(Ellis, 1946) . . . . . . . . . . . . . . 111

Table Page

2.68 The Proportions of the Boys at Each Age in the Three
Sexual Maturation Classes (Ellis, 1946) . . . . . 111

2.69 The Proportion of the Boys that Ellis Rated as Pubescent
and Adolescent with Various Genitalia, Pubic Hair,

and Axillary Hair Ratings (Ellis, 1946) . . . . . 112

2.70 Five-Point Genitalia Maturation Scale (Reynolds and
Wines, 1948) o o o o o o o o o o o o o o 112

2.71 The PrOportion of the Boys at Each Age in the Five
Genitalia Stages (Reynolds and Wines, 1948) . . . . 113

2.72 The PrOportion of the Boys at Each Age in the Seven
Penis Size Stages (Reynolds and Wines, 1948). . . . 115

2.73 Pubic Hair Stages (Reynolds and Wines, 1948) . . . . 115

2.74 The Proportion of the Boys at Each Age in the Five
Pubic Hair Stages (Reynolds and Wines, 1948). . . . 116

2.75 The Most Common Order of Appearance of the Genitalia
and Pubic Hair Stages (Reynolds and Wines, 1948) . . 118

2.76 The Mean Age Value for Six Sexual Characteristics When
50 Percent of the Boys Displayed Childhood and 50
Percent Displayed Adult Characteristics (Hogben,
Waterhouse, and Hogben, 1948). . . . . . . . . 119

2.77 Five Sexual Maturation Groups Streamlined From Greulich's
Maturity Groups (Nicolson and Hanley, 1953) . . . . 120

2.78 The Proportions of the Boys at Bach Age in the Five
Sexual Maturation Stages (Nicolson and Hanley, 1953) . 121

2.79 Genitalia Development Stages for Adolescent Males
(Tanner, 1955). . . . . . . . . . . . . . 123

2.80 Pubic Hair Stages for Adolescent Males (Tanner, 1955) . 123

2.81 Genital Development Stages for Adolescent Males ,
(Falkner, 1961) . . . . . . . . . . . . . 126

2.82 Pubic Hair Development Stages for Adolescent Males
(Falkner, 1961) . . . . . . . . . . . . . 126

2.83 Four Classifications of Puberty for Adolescent Males
(Douglas and Ross, 1964) . . . . . . . . . . 127

xi

Table Page

2.84 The Proportion of the Boys in the Four Classifications
of Puberty (Douglas and Ross, 1964) . . . . . . . 127

2.85 Five Genitalia Development Stages for Adolescent Males
(Chang, Lee, and Chan, 1966) . . . . . . . . . 128

2.86 Five Pubic Hair Stages for Adolescent Males (Chang, Lee
and Chan, 1966) o o o o c o o o o o o o o o 128

2.87 The Ages at Which 50 Percent of the Boys Attained the
' Genitalia and Pubic Hair Stages (Chang, Lee, and

Chan, 1966) . . . . . . . . . . . . . . . 128

2.88 The Proportion of the Boys at Each Age in the Three
Genitalia Phases (Chang, Lee, and Chan, 1966). . - . 129

2.89 The PrOportion of boys at each Age in the Three Pubic

Hair Phases (Chang, Lee, and Chan, 1966) . . . . . 129
2.90 Facial Hair Development Checklist (Marshall, 1967). . . 132
2.91 Axillary Hair Development Checklist (Marshall, 1967) . . 132
2.92 Pubic Hair Development Checklist (Marshall, 1967) . . . 133
2.93 Penis Development Checklist (Marshall, 1967). . . . . 133
2.94 Scrotum and Testes Development Checklist (Marshall, 1967) 133
2.95 SMIV Formula (Marshall, 1967). . . . . . . . . . 134
2.96 Five Puberty Groups by the SMIV Breakdown (Marshall, 1967) 134

2.97 The Number, Mean Values and Standard Deviations of the
Boys in the Five Puberty Groups (Marshall, 1967). . . 135

2.98 Six Classifications of Pubertal Maturity for Adolescent
Males (Young, Greulich, Gallagher, and Heald, 1968). . 136

2.99 Pubic Hair Checklist (Young, Greulich, Gallagher, and
Heald’ 1968) o o o o o o o o o o o o o o o 136

2.100 Axillary Hair Checklist (Young, Greulich, Gallagher, and
Heald, 1968). . . . . . . . . . . . . . . 137

2.101 Facial Hair Checklist (Young, Greulich, Gallagher, and
Heald, 1968) o o o o o o o o . o o o o o o o 137

2.102 The Testicular volumnes, Penis Lengths, and Penis
Diameters Necessary for the Six Puberty Ratings (Young,
Greulich, Gallagher, and Heald, 1968) . . . . . . 138

xii

h)

h.)

A.)

A)

IKJ

 

Table

2.103

2.104

2.105

2.106

2.107

2.108

2.109

2.110

2.111

2.112

2.113

2.114

2.115

2.116

2.117

2.118

Breast Development Checklist for Adolescent Males
(Young, Greulich, Gallagher, and Heald, 1968). . . .

Eyebrow Development Checklist for Adolescent Males
(young, Greulich, Gallagher, and Heald, 1968). . . .

The Test-retest Correlations for the Five Observers
in This Study (Young, Greulich, Gallagher, and
Heald, 1968). . . . . . . . . . . . . . .

The Standard Partial Regression Coefficients for Five
Observers (Young, Greulich, Gallagher, and Heald,
1968) o o o o o o o o o o o o o c o o o

The Correlations Between Chronological Age, Skeletal
Age, a First Pubertal Maturity Rating, a Calculated
Pubertal Age, and a Second Pubertal Maturity Rating
(Young, Greulich, Gallagher, and Heald, 1968). . . .

The Average Age and Standard Deviation That the Boys
Attained the Genitalia and Pubic Hair Stages
(Marshall and Tanner, 1970). . . . . . . . . .
The Proportion of the Boys in the Five Genitalia Stages
Across the Four Pubic Hair Stages (Marshall and Tanner,
1970) o o o o o o o o o o o o o o o o o
The Proportion of the Boys in the Five Pubic Hair Stages
Across the Four Genitalia Stages (Marshall and Tanner,
1970) . . . . . . . . . . . . . . . . .
Pubic Hair Stages (Neyzi et a1., 1975). . . . . . .
Facial Hair Stages (Neyzi et a1, 1975). . . . . . .
Facial Hair Stages (Neyzi et a1., 1975) . . . . . .
Laryngeal Stages (Neyzi et all., 1975). . . . . . .
The Mean Attainment Age That the Boys Reached the Pubic
Hair, Axillary Hair, and Laryngeal Stages (Neyzi et a1.,
1975) o o o o o o o o o o o o o o o o o

The Proportion of the Boys at Each Age in the Five Pubic
Hair Stages (Neyzi et a1., 1975) . . . . . . . .

The Proportion of the Boys at Each Age in the Four
Axillary Hair Stages (Neyzi et a1., 1975) . . . . .

The Proportion of the Boys at Bach Age in the Four
Facial Hair Stages (Neyzi et a1., 1975). . . . . .

xii o,

Page

138

138

139

140

141

142

143

143

144

144

145

145

145

146

146

147

AEHM'N—mfl —--

Table

2.119

2.120

2.121

3.1

The Proportion of the Boys at Each Age in the Four

Laryngeal Stages (Neyzi et a1., 1975)

The Skeletal Age Mean Values and Standard Deviations

for the Boys Across Their Chronological Ages (Clarke

and Wickens, 1962b) . .

The Frequency of the Boys in Greulich's Five Maturity

Groups (Clarke and Dugutis, 1962d)

The SMIV Scores, Chronological Age, and Grade Levels of

the 30 Boys Selected to Represent a Rectangular

Distribution of Sexual Maturation.

The Number of East Lansing Subjects in Each Grade for

Each Test Period . . .

The Number of Mason Subjects in Each Grade for Each

Test Period . . . . .

The Step-Wise Regression for Chronological Age and

Height. . . . . . .

The Step-Wise Regression for Chronological Age, Height,

and Weight . . . . .

The Step-Wise Regression for Chronological Age, Height,

Weight, and SMIV . . .

The Within-Rater Correlation Coefficients for the Five
Raters Across the Five Sexual Characteristics and the
Last Column Shows the Mean Correlation Across Raters

for Each Variable . . .

Between-Raters Correlations
Paired Five Raters. . .

Between-Raters Correlations
Paired Five Raters. . .

Between-Raters Correlations
Paired Five Raters. . .

Between-Raters Correlations
Paired Five Raters . .

Between-raters Correlations of Scrotum and Testes

of Facial Hair for the

of Axillary Hair for the

of Pubic Hair for the

of Penis Development for

Development for the Paired Five Raters .

xiii

Page

147

156

157

165

167

167

174

175

177

177

179

179

179

179

180

. ..-.»._-—.—.~..—.—.-.—-mo_- *-

Table Page

4.10 Average Between-Raters Correlations of the Paired
Raters for the Five Sexual Characteristics. . . . . 180

4.11 Average of All of the Paired Between-Raters Correlations
for the Five Sexual Characteristics . . . . . . . 180

4.12 The Mean Squares Formulas for a 4-Way Anova Mixed Model
Where Raters(R), Time Points (P), and Subjects(S) are

Random and Sexual Characteristics(C) is fixed. . . . 183

4.13 The Estimates of the Variance Components Formulas in
this 4—Way Mixed Model . . . . . . . . . . . 183

4.14 The Mean Square Values for the Fifteen Sources of
Variance in This 4-way Mixed Model . . . . . . . 184

4.15 The Substituted Estimates of the Variance Components
Formulas O O O O O O O O O O O O O O O O 184

4.16 The Reliability Formulas for the Eight Sources of Error
in This 4-Way Mixed Model . . . . . . . . . . 185

4.17 The Substituted Reliability Formulas and Their Final
Value . . . . . . . . . . . . . . . . . 185

4.18 Adjusted Reliability Formulas. . . . . . . . . . 187

4.19 The Adjusted Substituted Reliabilities Formulas, and
Their Final Values. . . . . . . . . . . . . 187

4.20 The Month Ranges that Comparised the Annual Increment
Groups. . . . . . . . . . . . . . . . . 189

4.21 The BSA and CA Within Cell Variances for the Five Motor
Proficiency Tests . . . . . . . . . . . . . 189

4.22 The Subjects Distribution for the BSA and CA Groups . . 190

4.23 The Mean Value Progressions for Sit-Ups in the BSA and

CA Groups and the Between-Groups Inrements. . . . . 190
4.24 The Mean Value Progressions for Pull-Ups in the BSA and

CA Groups and the Between-Groups Increments . . . . 190
4.25 The Mean Value Progressions for Standing Long Jump in

the BSA and CA Groups and the Between-Groups

Increments . . . . . . . . . . . . . . . 191

xiv

 

llli lllll [’1'

Table

4.26

C.1

1.1

J.4

The Mean Value Progressions for Vertical Jump in the
BSA and CA Groups and the Between-Groups Increments. .

The Mean Value Progressions for Shuttle Run in the BSA
and CA Groups and the Between-Groups Increments . . .

Professor Seefeldt's Test-Retest Correlations for the
Twenty-Eight Sites. . . . . . . . . . . . .

After Height, the Percent of Additional Variance
Accounted for by Chronological Age and the Signifi-
cance Level . . . . . . . . . . . . . . .

After Chronological Age, the Percent of Additional
variance Accounted for by Weight and the Significance
Level . . . . . . . . . . . . . . . . .

After Weight, the Percent of Additional Variance
Accounted for by Chronological Age and the Signifi-
cance Level . . . . . . . . . . . . . . .

After Height, the Percent of Additional Variance ~

Accounted for by Weight and the Significance Level . .

After Weight, the Percent of Additional Variance
Accounted for by Height and the Significance Level . .

After Chronological Age and Weight, the Percent of
Additional Variance Accounted for by Height and the
Significance Level. . . . . . . . . . . . .

The Correlation Matrix for the Five Sexual Character-
istics O 0 O O O O I O O O O O O O O O

The Raw Regression Coefficients and Mean Values That the
Multivariate Regression Analysis Determined for the
SMIV Regression Equation. . . . . . . . . . .

Substituting Selected Ratings for the Five Sexual
Characteristics Values Into the SMIV Regression
Equation 0 O O O O I I O O O O O O O O 0

Raw Regression Coefficients for the Regression Analyses
Minus the Penis Development Variable and Minus the
Scrotum and Testes Development Variable. . . . . .

Substituting Selected Ratings for the Four Sexual Char-

Acteristics Values Into the (MSD) SMIV Regression
Equation 0 O O O C I C O O O O O O O O O

XV

Page

191

191

209

223

223

223

223

224

224

226

227

228

229

230

Table Page

J.6 Substituting Selected Ratings for the Four Sexual
Characteristics Values Into the (MPD) SMIV Regression

Equation . . . . . . . . . . . . . . . . 231
J.7 Comparing the SMIV, (MSD) SMIV, and (MPD) SMIV Values. . 231
K.1 The Step-Wise Regression for Chronological Age and SMIV . 233
K.2 The Step-Wise Regression for SMIV and Chronological Age . 233
K.3 The Step-Wise Regression for Height and SMIV. . . . . 233
K.4 The Step-Wise Regression for SMIV and Height. . . . . 233
K.5 The Step-Wise Regression for Weight and SMIV. . . . . 233
K.6 The Step-Wise Regression for SMIV and Weight. . . . . 234

L.1 East Lansing Data's Number of Subjects in the Total
Sample and in the Analysis Sample. . . . . . . . 237

L.2 East Lansing Data's Chronological Age Mean Values in the
Total Sample and in the Analysis Sample. . . . . . 237

L.3 East Lansing Data's Height Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 237

L.4 East Lansing Data's Height Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 238

L.5 East Lansing Data's Sit-Ups Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 238

L.6 East Lansing Data's Pull-Ups Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 238

L.7 East Lansing Data's Standing Long Jump Mean Values in

the Total Sample and in the Analysis Sample . . . . 238
L.8 East Lansing Data's Vertical Jump Mean Values in the
Total Sample and in the Analysis Sample. . . . . . 239

L.9 East Lansing Data's Shuttle Run Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 239

L.10 East Lansing Data's Facial Hair Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . . 239

L.1l East Lansing Data's Axillary Hair Mean Values in the
Total Sample and in the Analysis Sample. . . . . . 239

xvi

Table

L.12

L.13

L.18

L.22

Lo 23

East Lansing Data's Pubic Hair Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

East Lansing Data's Penis Development Mean Value in the
Total Sample and in the Analysis Sample. . . . . .

East Lansing Data's Scrotum and Testes Development Mean
Values in the Total Sample and in the Analysis Sample .

The Univariate F Values and Their Associated Significance
Levels for the East Lansing GRADE by YEAR Interaction .

Mason Data's Number of Subjects in the Total Sample and
in the Analysis Sample . . . . . . . . . . .

Mason Data's Chronological Age Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

Mason Data's Height Mean Values in the Total Sample and
in the AnaIYSis sa-Inple O O I O I O O O O I 0

Mason Data's Weight Mean Values in the Total Sample and
in the Analysis Sample . . . . . . . . . . .

Mason Data's Sit-Ups Mean Values in the Total Sample
and in the Analysis Sample . . . . . . . . . .

Mason Data's Pull-Ups Mean Values in the Total Sample
and in the Analysis Sample . . . . . . . . . .

Mason Data's Standing Long Jump Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

Mason Data's Vertical Jump Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

Mason Data's Shuttle Run Mean Values in the Total Sample
and in the Analysis Sample . . . . . . . . . .

Mason Data's Facial Hair Mean Values in the Total Sample
and in the Analysis Sample . . . . . . . . . .

Mason Data's Axillary Hair Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

Mason Data's Pubic Hair Mean Values in the Total Sample
and in the Analysis Sample . . . . . . . . . .

Mason Data's Penis Development Mean Values in the Total
Sample and in the Analysis Sample. . . . . . . .

xvii

Page

240

240

240

242

244

244

244

245

245

245

245

246

246

246

246

247

247

Table Page

L.29 Mason Data's Scrotum and Testes Development Mean Values
in the Total Sample and in the Analysis Sample . . . 247

L.30 The Univariate F Values and Their Associated Signifi-
cance Levels for the Mason GRADE by YEAR interaction . 247

L.31 East Lansing and Mason Data's Chronological Age Mean
Values in Their PrOportional Cell Samples . . . . . 249

L.32 East Lansing and Mason Data's Height Mean Values in
Their PrOportional Cell Samples . . . . . . . . 249

L.33 East Lansing and Mason Data's Weight Mean Values in
Their Proportional Cell Samples . . . . . . . . 249

L.34 East Lansing and Mason Data's Sit-Ups Mean Values in
Their Proportional Cell Samples . . . . . . . . 250

L.35 East Lansing and Mason Data's Pull-Ups Mean Values in
Their Proportional Cell Samples . . . . . . . . 250

L.36 East Lansing and Mason Data's Standing Long Jump Mean
Values in Their Proportional Cell Samples . . . . . 250

L.37 East Lansing and Mason Data's Vertical Jump Mean Values
in Their Proportional Cell Samples . . . . . . . 250

L.38 East Lansing and Mason Data's Shuttle Run Mean Values in
Their PrOportional Cell Samples . . . . . . . . 251

L.39 East Lansing and Mason Data's Facial Hair Mean Values in
Their PrOportional Cell Samples . . . . . . . . 251

L.4O East Lansing and Mason Data's Axillary Hair Mean Values in
Their Proportional Cell Samples . . . . . . . . 251

L.41 East Lansing and Mason Data's Pubic Hair Mean Values in
Their PrOportional Cell Samples . . . . . . . . 251

L.42 East Lansing and Mason Data's Penis Development Mean
Values in Their Proportional Cell Samples . . . . . 252

L.43 East Lansing and Mason Data's Scrotum and Testes Develop-
ment Mean Values in Their Proportional Cell Samples. . 252

L.44 The Univariate F Values and Their Associated Significance

Levels for the East Lansing and Mason SCHOOL by GRADE
by YEAR interaction . . . . . . . . . . . . 252

xviii

The Least Square Estimates, Standard Errors, and
Confidence Intervals of the Thirteen Variables . . .

The Observed Proportions for Facial Hair . . . . . .
The Observed Proportions for Axillary Hair . . . . .
The Observed Preportions for Pubic Hair . . . . . .
The Observed Proportions for Penis Development . . . .

The Observed Proportions for Scrotum and Testes
Development . . . . . . . . . . . . . . .

The Predicted Proportions for Facial Hair. . . . . .
The Predicted Proportions for Axillary Hair . . . . .
The Predicted Proportions for Pubic Hair . . . . . .
The Predicted PrOportions for Penis Development. . . .

The Predicted Pr0portions for Scrotum and Testes
Development . . . . . . . . . . . . . . .

The Sample Chronological Age, Height, Weight, and SMIV
Correlation Matrix. . . . . . . . . . . . .

The Raw Regression Coefficients and Mean Values that the
Multivariate Regression Analysis Determined for the
BSA Regression Equation . . . . . . . . . . .

Substituting Selected Chronological Age, Height, Weight,
and SMIV Scores Into BSA. . . . . . . . . . .

The Beta Weights, Ranges, and Maximum Impact Potentials
for the Four Independent Variables in the ESA
Regression Equation . . . . . . . . . . . .

The Raw Regression Coefficients and Standard Errors for
Chronological Age, Height, Weight, and SMIV . . . .

The Confidence Intervals for Chronological Age, Height,
Weight and SMIV. . . . . . . . . . . . . .

The Observed and Structural BSA Beta Weights for Chrono-
logical Age, Height, Weight, and SMIV . . . . . .

The Latent Regression Coefficients From the Structural
Analysis and the Mean Values From the Analysis of the
Observed mta O O O O O O O O O O O O O O

xix

Page

255
260
260
260

261

261
267
267
268

268

269

276

276

277

278

279

279

280

281

Table Page
N.9 The A Vector and the u Vector. . . . . . . . . . 282

N.10 Substituting Selected Chronological Age, Height, Weight,
and SMIV Scores into the (S)ESA Regression Equation. . 282

N.11 The Beta Weights, Ranges, and Maximum Impact Potentials
for the Four Independent Variables in the (S)ESA

Regression . . . . . . . . . . . . . . . 283
N.12 The (0)ESA and (S)ESA Maximum Impact Potentials. . . . 284
N.13 The Structural Analysis' PHI Matrix. . . . . . . . 284
N.14 The Structural Sample Correlation Matrix . . . . . . 284

N.15 The 02(X(i)) Equations for the Five Sexual Characteris—
tics and Their Solutions. . . . . . . . . . . 285

N.16 The. (x(i)) Equations for the Five Sexual Characteristics
and Their Solutions . . . . . . . . . . . . 286

0.1 The PrOportion of the Boys in This Sample in Bach Cell
When CA is Crossed with BSA. . . . . . . . . . 288

0.2 The Percentage of the Accelerated, Standard, and Delayed

Maturers Across the Six CA Intervals. . . . . . . 288
0.3 The Variances for the Motor Proficiency Tests When

Classifying by CA, BSA, and the Accelerated, Standard,

and Delayed Classifications. . . . . . . . . . 289

0.4 The Sit-Ups Variances Across the Six CA Intervals for the

Accelerated, Standard, and Delayed-Maturers . . . . 291
0.5 The Pull-Ups Variances Across the Six CA Intervals for
the Accelerated, Standard, and Delayed-Maturers . . . 291

0.6 The Standing Long Jump Variances Across the Six CA
Intervals for the Accelerated, Standard, and Delayed
Maturers . . . . . . . . . . . . . . . . 291

0.7 The Vertical Jump Across the Six CA Intervals for the

Accelerated, Standard, and Delayed-Maturers . . . . 291
0.8 The Shuttle Run Variances Across the Six CA Intervals

for the Accelerated, Standard, and Delayed-Maturers. . 292
0.9 The Sit-Ups Variances Across the Six CA Intervals for

the More Precise Accelerated, Standard, and Delayed-

Maturers Classifications. . . . . . . . . . . 295

Table

0.10

0.12

0.13

0.14

0.15

0.16

0.17

0.18

0.19

 

The Pull-Ups Variances Across the Six CA Intervals for
the More Precise Accelerated, Standard, and Delayed-
Maturers Classifications. . . . . . . . . . .

The Standing Long Jump Variances Across the Six CA
Intervals for the More Precise Accelerated, Standard,
and Delayed-Maturers Classifications. . . . . . .

The Vertical Jump Variances Across the Six CA Intervals
for the More Precise Accelerated, Standard, and
Delayed-Maturers Classifications . . . . . . . .

The Shuttle Run Variances Across the Six CA Intervals for
the More Precise Accelerated, Standard, and Delayed-
Maturers Classifications. . . . . . . . . . .

The Distribution of the 1578 Subjects Across the Six CA
Intervals by the More Precise Accelerated, Standard,
and Delayed-Maturers Classifications. . . . . . .

The Sit-Ups Mean Values Across the Six CA Intervals for
the More Precise Accelerated, Standard, and Delayed—
Maturers Classifications. . . . . . . . -. . .

The Pull-Ups Mean Values Across the Six CA Intervals
for the More Precise Accelerated, Standard, and Delayed-
Maturers Classifications. . . . . . . . . . .

The Standing Long Jump Mean Values Across the Six CA
Intervals for the More Precise Accelerated, Standard,
and Delayed-Maturers Classifications. . . . . . .

The Vertical Jump Mean Values Across the Six CA Intervals
for the More Precise Accelerated, Standard, and Delayed-
Maturers Classifications. . . . . . . . . . .

The Shuttle Run Mean Values Across the Six CA Intervals

for the More Precise Accelerated, Standard, and
Delayed-Maturers Classifications . . . . . . . .

xxi

Page

295

295

296

296

297

298

298

298

299

299

LIST OF FIGURES
Figures Page

2.1 Observed Genitalia Stages Proportions (Reynolds and
Wines, 1948). . . . . . . . . . . . . . 114

2.2 Observed Pubic Hair PrOportions (Reynolds and Wines,
1948) . . . . . . . . . . . . . . . . 117

2.3 Observed Sexual Maturation Stages Proportions
(Nicolson and Hanley, 1953). . . . . . . . . 122

2.4 Observed Genitalia Phase PrOportions (Chang, Lee, and
and Chan, 1966). . . . . . . . . . . . . 130

2.5 Observed Pubic Hair PrOportions (Chang, Lee, and Chan,
1966) . . . . . . . . . . . . . . . . 131

2.6. Observed Pubic Hair Stage Proportions (Neyzi et a1,
1975) . . . . . . . . . . . . . . . . 148

2.7 Observed Axillary Hair Stage Proportions (Neyzi
et a1, 1975). . . . . . . . . . . . . . 149

2.8 Observed Facial Hair Stage PrOportions (Neyzi et a1,
1975) O I O O O I O O O O O O O O O O 150

2.9 Observed Laryngial Stage Proportions (Neyzi et a1,
1975) . . . . . . . . . . . . . . . . 151

L.1 The East Lansing Data's Shuttle Run Mean Values GRADE
by YEAR Interaction . . . . . . . . . . . 242

L.2 The East Lansing Data's Facial Hair Mean Values GRADE
by YEAR Interaction . . . . . . . . . . . 243

L.3 The Mason Data's Sit-Ups Mean Values GRADE by YEAR
interaction 0 O O O O O O O O O O O C O 247

L.4 The East Lansing and Mason Data's Sit-Ups Mean Values
SCHOOL by GRADE Interactions for the First Year
Data Collection. . . . . . . . . . . . . 252

L.5 The East Lansing and Mason Data's Sit-Ups Mean Values
SCHOOL by GRADE Interactions for the Second Year
Data COlleCtion. O O O O O O O O O O C O 253

L.6 The East Lansing and Mason Data's Sit-Ups Mean Values
SCHOOL by GRADE Interactions for the Third Year Data
collection 0 O O O O O O O O 0 O O O O 253

M.1 Observed Facial Hair Proportions. . . . . . . . 262

xxiii

Figure Page

M.2 Observed Axillary Hair Proportions . . . . . . . 263
M.3 Observed Pubic Hair Proportions . . . . . . . . 264
M.4 Observed Penis Development Proportions. . . . . . 265
M.5 Observed Scrotum and Testes Development Proportions . 266
M.6 Predicted Facial Hair Proportions . . . . . . . 270
M.7 Predicted Axillary Hair Proportions . . . . . . 271
M.8 Predicted Pubic Hair Proportions . . . . . . . 272
M.9 Predicted Penis Development PrOportions . . . . . 273
M.10 Predicted Scrotum and Testes Development Proportion . 274

xxiv

CHAPTER I

INTRODUCTION

In the United States, competition is a way of life. In the
classroom, teachers use competition for grades to motivate students
to improve their academic skills. On the athletic fields, coaches
use competition for victory to motivate athletes to improve their
motor skills. Supporting the value of competition in education is
the principle that regardless of the competition's outcome, if the
students put forth their greatest effort, then everybody wins.
However, for competition to remain an ongoing motivation, it must
reward the quality of effort and not factors over which the compe-
titors have no control. In athletic competitions for adolescent
males, the participants have no control over their biological matura-
tion. Therefore, to reduce the influence that biological maturation
has on the outcome of the competition and to increase the likelihood
that the rewards match the quality of effort, the sponsors of ath-
letic competitions for adolescent males should: (1) equate the com-
petitors according to biological age, or (2) adjust the rules to
eliminate the advantage that early-maturing boys have over their
later-maturing peers.

When teaching physical education classes in East Lansing

Junior High School, the author encountered an athletic competition

for adolescent males for which he suspected that nobody had equated
the competitor's biological maturation. That competition was for

the Presidential Physical Fitness Award, which is part of the American
Alliance for Health, Physical Education, and Recreation's (AAHPBR)
Youth Fitness Program's motivational plan to increase the physical
fitness levels of 9 to 17 year old American children.

In 1966, President Lyndon B. Johnson established the Presi-
dential Physical Fitness Award. Added to the previously established
Achievement Award Certificate and the Embossed Gold Merit Seal, the
Presidential Physical Fitness Award provided strong incentives for
Physical Education Programs to add the AAHPBR Youth Fitness Tests
to their curricula. To win the Achievement Award Certificate, the
Embossed Gold Merit Seal or the Presidential Physical Fitness Award,
children had to score above the 50th, 80th, or the 85th percentiles,
respectively, on all of the tests.

Two reports by Hans Kraus and Ruth Hirshland created the
controversy from which the AAHPBR Youth Fitness Program evolved
(Kraus and Hirshland, 1953 and 1954). In their reports they ques-
tioned the physical fitness levels of American children. Using the
Kraus-Weber battery of six tests, Kraus and Hirshland measured: (1)
children's range of motion during straight and flexed leg sit-ups;
(2) the time that children sustained a supine double leg raise, a
prone double leg raise, and a prone upper trunk raise; and (3) while
maintaining a straight leg toe touch position, the distance above
or below the platform on which the children stood that children could

reach. The Low Back Clinic at Columbia Presbyterian Hospital developed

the Kraus-Weber battery of tests. Comparing American children with
Austrian, Italian and Swiss children, Kraus and Hirshland concluded
that the American children were not as physically fit as foreign
children.

Alerted by the Kraus—Hirshland studies, in 1956 President
Dwight D. Eisenhower summoned a national conference to discuss the
status of American children's physical fitness. President Eisenhower's
conference inspired the 1956 AAHPBR national convention to discuss
what physical education teachers might do to upgrade the physical
fitness of American children. As its first step, the AAHPBR created
the AAHPBR Youth Fitness Project. As its second step, the AAHPBR
appointed a special AAHPBR Research Council to direct the AAHPBR
Youth Fitness Project.

As their first directive, the AAHPBR Research Council decided
to conduct their own measurement of the physical fitness of American
children. To select their battery of physical fitness tests, the
AAHPBR Research Council proposed the following five criteria: (1)
the tests should be reasonably familiar to the students; (2) the
tests should require little or no equipment; (3) the tests should
apply equally to boys and girls; (4) the tests should apply to
children aged 9 through 17; and, (5) the tests must measure differ-
ent components of physical fitness. After considerable discussion
and preliminary testing, the AAHPBR Research Council chose the
following physical fitness tests: (1) the normal and modified pull-

up tests; (2) the straight leg sit-up test; (3) the shuttle run test;

(4) the standing long jum test; (5) the 50 yard dash; (6) the 600
yard run-walk; and (7) the softball throw for distance.

Directed by Professor Paul A. Hunsicker, in 1957 the Univer-
sity of Michigan's Survey Research Center coordinated the AAHPBR
Youth Fitness Program's first data collection. Eight thousand five
hundred children participated in the 1957 sampling.

When the AAHPBR Research Council compared those 1957 norms
with the norms of other countries, they confirmed the Kraus-Hirshland
finding that American children were not as physically fit as foreign
children. Therefore, the AAHPBR recommended that Physical Education
Departments across the United States add the AAHPBR Youth Fitness
Tests to their curriculum.

To determine whether or not the AAHPBR Youth Fitness Tests
had effected the physical fitness of American children, in 1963 the
AAHPBR commissioned its second national survey. Directed again by
Professor Paul A. Hunsicker, the University of Michigan's Survey
Research Center coordinated the measurement of a different sample of
9,200 children. In the battery of tests for the second sampling,
the AAHPBR substituted the timed flexed arm hang for the girl's
modified pull-up. When the AAHPBR compared the norms from their
1965 and 1957 surveys, with the single exception of the 17 year old
girl's softball throw for distance, for every age and every test,
they found that both the boys and girls had significantly improved
their scores.

In 1975, the AAHPBR commissioned its third national survey.

Directed jointly by Professors Paul A. Hunsicker and Guy A. Reiff,

the University of Michigan's Survey Research Center coordinated the
data collection. In their 1975 test battery, the AAHPBR Research
Council made three changes. They eliminated the softball throw for
distance, substituted the flexed leg sit-up for the straight leg sit-
up, and, in place of the 600 yard run-walk, for children aged 12 and
under, they provided the options of a one mile or nine minute run,
and for children aged 13 years and over, they provided the options

of the one and one-half mile to 12 minute run. Also, as their sixth
Youth Fitness Test criteria, the AAHPBR Research Council required
that the children should be able to assess their own performances.

When the AAHPBR compared their 1975 and 1965 norms, they
found that for both sexes across the eight chronological age classi-
fications and the six tests, the children improved in only six of
the 96 cells and they regressed in two cells. From examining those
results, researchers might conclude that the AAHPER's Youth Fitness
competition failed to stimulate American children to improve their
physical fitness.

When comparing the norms of American children with those of
foreign children or with those of American children from different
data collections, because random sampling negated the effect of the
variability in the bioloigical maturation, the AAHPBR could classify
the children by the same procedures the others had used. However,
when the AAHPBR established its Youth Fitness competition, the
situation changed. Because the children now competed individually
against each other, random sampling no longer negated the effect

of the variability in the biological maturation. Therefore, the

procedure by which the AAHPBR classified the competitive match-ups
should have eliminated the variability in biological maturation.
Research evidence suggests that if at the start of its competition,
AAHPBR had equated the children on the basis of biological maturation,
then they would have increased the likelihood of rewarding the quality
of the children's efforts.

In 1966, the AAHPBR used the following two procedures to
classify the children for their Youth Fitness competition: (1)
chronological age; and (2) the Nielson-Cozens Index. The Nielson-
Cozens Index combined chronological age, height and weight. In 1975,
because of‘a belief that height and weight added little to the
equality of classifying children for motor proficiency norms, the

AAHPBR eliminated the Nielson-Cozens Index.

Need for the Study

 

There is need for a study to determine the proportion of the
variability in adolescent males' biological maturation that is
explained by increments in their chronological age, height, weight
and a Sexual Maturation Indicator Value (SMIV) estimate. More
specifically, this investigation examined the following questions;

Question 1: What proportion of the variance in the skeletal

 

age of adolescent males is explained by increments in their chrono-
logical age, height and weight?

Question 2: Does an estimate of sexual maturation (SMIV)

 

explain a significant additional proportion of the variance in

adolescent males' skeletal age?

Question 3: Can boys' physical education teachers reliably

 

assess the sexual maturation of adolescent males?

Question 4: How does an estimate of skeletal age (SA) that

 

encompasses chronological age, height, weight and SMIV compare with
chronological age when classifying adolescent males for motor pro-

ficiency norms?

Scope of the Study

 

To examine questions 1, 2 and 3, the author randomly selected
30 East Lansing, Michigan sixth, seventh, eighth and ninth grade male
students to represent a rectangular distribution of sexual maturation.
And, to examine question 4, every six months for three years the boys'
physical education teachers assessed selected physical and motor
characteristics of the entire male student body of East Lansing
Junior High School and Mason Junior High School. Therefore, the
scope of this study extends to sixth, seventh, eighth and ninth

grade boys.

Limitations of the Study

 

The following factors may limit the interpretations of this
study:

(a) Because this study included only sixth, seventh, eighth
and ninth grade males, the findings apply only to males within the
age range of 11 to 17 years.

(b) Although the AAHPBR Youth Fitness Test Manual provided

specific instructions for the administration of its battery of tests,

the ability of the physical education teachers to conduct
proper testing methodology may limit the interpretations.

(c) Whenever researchers measure children's motor proficien-
cies, a variety of environmental influences may limit interpretations.
Some of those environmental influences are:

(l) the time of the day in which the children perform,

(2) the order in which the children perform,

(3) the presence of other children during the testing,

(4) the friction coefficient of the surface on which the
children perform, and

(5) the temperature and humidity of the testing site.

(d) The physical education teachers received no special
training. Presumably, they had no special competencies. Therefore,
their ability to observe the (SMIV) checklist variables in a scien-
tific manner may limit the interpretations.

(e) Except for the author's gratitude, the learning exper-
ience, and the advancement of their knowledge about the relationship
between sexual maturation and motor proficiencies, those teachers
received no reward for working in this study. Therefore, some might

argue that the use of volunteer personnel limited the interpretations.

CHAPTER II

REVIEW OF LITERATURE

For the purpose of classifying adolescent males for athletic
competitions, the author questions the use of chronological age.
Because adolescent males of the same chronological age may be years
apart in biological maturation, athletic competitions based on
chronological age potentially reward boys on the basis of advanced
biological maturation and not on the basis of effort or superior
athletic ability. While this dissertation examines only the American
Alliance of Health, Physical Education, and Recreation (AAHPBR) Youth
Fitness competition, the implications of the results extend to all
athletic competitions for adolescent males that use chronological
age to classify the boys.

One way that specialists in Child Growth and Deve10pment predict
biological maturation of adolescent males is by estimating skeletal
maturation from hand-wrist x-rays. In this study, Health, Physical
Education, and Recreation Child Growth and Development Professor Vern
Seefeldt used the Greulich-Pyle Radiographic Atlas of Skeletal Develope

ment of the Hand and Wrist to assess the skeletal age of 30 adolescent

 

boys. Therefore, the first portion of the review of literature covers

the history of the assessment of hand and wrist development.

10

Another way that specialists in Child Growth and Development
predict biological maturation in adolescent males is by estimating
the boys' sexual maturation. For this study, the author created a
sexual maturation checklist that the male physical education teachers
from East Lansing and Mason Junior High Schools used to assess the
sexual maturation of their students. Therefore, the second portion
of the review of literature covers the history of the assessment of
adolescent males' sexual development.

The purpose of this study is to examine the procedures by
which sponsors of athletic competitions classify adolescent males.
The AAHPER's use of chronological age to classify adolescent males
for motor proficiency norms is the focus of this investigation.
Therefore, the third portion of the review of literature covers the
relationship between the motor proficiency performances of adoles-
cent males and various classification indices.

History of the Assessment of
Hand and Wrist Skeletal Development

William K. Roentgen's December 1895 discovery of x-rays enabled
researchers to study the progressive changes in the human skeleton.
Prior to that discovery, knowledge of human skeletal development
resulted from anatomists dissecting cadavers. Because 1) they received
very few child cadavers and 2) they could not study their subjects
longitudinally, anatomists provided limited information on skeletal
development. However, by using x-rays, researchers could readily

follow the skeletal development in children.

11

Four months after Roentgen's finding, on April 2, 1896
Professor Sydney Rowland of Germany x-rayed clinical patients, and,
when describing the appearance of the bones of the children's hands
and wrists, he recognized that those bones indicated the children's
level of skeletal maturation (Pyle, Waterhouse, and Greulich, 1971).
Later in 1896, another German researcher, Joseph Ranke, reported
that x-rays presented a new method for determining children's bio-
logical age (Flory, 1936).

In the United States, Professor J.W. Pryor of the State
College of Kentucky initiated the research on the x-ray assessment
of hand and wrist skeletal development (Pryor, 1905). In 1905,
from 70 x-rays of children aged 44 days to 21 years, Professor Pryor
outlined the appearance order of the eight carpal ossification
centers, and the 21 metacarpal, phalangial, ulnar, and radial
epiphysial ossification centers. For boys and girls separately,
Pryor summarized the ages at which their ossification centers
appeared. He claimed that skeletal maturation occurred earlier in
girls' hands than in boys' hands. This female maturation advantage
was by days in early childhood, by months in early adolescence, and
by years at the completion of skeletal development. Also, Pryor
claimed that first born children's hands skeletally matured earlier
than non-first born children's hands.

The following year, Professor Pryor compared his findings
with the information on skeletal development contained in Cunningham's

Text Book of Anatomy, Morris' Human Anatomy, Holden's Human Osteology,

 

 

 

Gray's Anatomy and Poland's Skiagraphic Atlas (Pryor, 1906). He

 

12

concluded that the skeletal ossification process of children's

hands started earlier than reported in those references and that the
appearance order of the Children's epiphysial ossification centers
differed from the appearance order reported in those references.
Further, he emphasized that those references did not separate the
data on boys from the data of girls.

Then, by using x-rays of identical twin girls and boy-girl
twins, Pryor reaffirmed his claim that girls' hands skeletally
matured earlier than boys' hands. Additionally, he emphasized that
the distal ulnar and radial epiphyses developed simultaneously with
the carpals, metacarpals, and phalanges, and therefore, he now
included those sites with the skeletal maturation of the hand and
wrist.

Pryor's next research topic effort delved into the hereditary
nature of variations in carpal development (Pryor, 1907). After
viewing 360 x-rays of children from 225 families, Professor Pryor
observed that variation patterns in the appearance order of the
carpals followed family lines. By recording the appearance order
of the carpals for children from ten families, he showed that when a
child's carpal appearance order differed from the order typical in
the population, that child's brothers and sisters varied similarly.
Additionally, Pryor noted the bilateral symmetry of the hand-wrist
skeletal ossification process.

After expanding his sample of children aged 14 years and
under to 266 girls and 288 boys, Pryor reaffirmed that the skeletal

ossification process began earlier than previously reported (Table

13

2.1) and, that the carpal appearance order differed from those orders

given in previous reports (Table 2.2), (Pryor, 1908). From the

information in Table 2.1 he hypothesized that the children's carpal

appearance order to be: (1) Os Magnum; (2) Unciform; (3) Cuneiform;

(4) Semilunar; (5) Scaphoid; (6) Trapezoid; (7) Trapezium; and (8)

Pisiform. Additionally, Pryor reported the sex, age, order of birth,
carpal appearance order, and millimeter diameter size of each carpal

for 136 girls and 153 boys.

Tafle2J

The Age by Sex of Carpal Appearance in

Children (Pryor, 1908)

 

 

Carpal Sex Age of Appearance

Os Magnum Female Between 3-4 months

(Capitate) Male Between 7-8 months

Unciform Female Between 3-4 months

(Hamate) Male Between 7-8 months

Cuneiform Female 13th or 14th month

(Triquetral) Male early in the 3rd year

Semilunar Female Late in the 3rd year or early in the 4th year

(Lunate) Male Between the 4th and 5th years

Scaphoid Female About 4 years
Male About 6 years

Trapezoid Female Between the 4th and 5th year
Male Between the 5th and 6th year

Trapezium Female Between thd 4th and 5th year
Male Early in the 8th year

Pisiform Female Between the 9th and 10th year
Male Between the 12th and 13thyyear

 

In a 1925 report, Professor Pryor examined x-rays of 64 boys

aged 150 to 274 months and 81 girls aged 144 to 260 months. In

table form, Pryor recorded the ages, sex, and an epiphysial rating

of whether he rated Distinct or Indistinct for all 145 children.

l4

TafleZJ

The Frequency of the Carpal Appearance
Order in 243 Children (Pryor, 1908)

 

 

 

CARPALS lST , 2ND 3RD 4TH 5TH 6TH 7TH 8TH TOTAL
Os Magnum 238 5 0 0 0 0 0 0 243
Unciform 5 236 0 O 0 0 0 0 241
Cuneiform 0 O 219 10 0 0 0 0 229
Semilunar 0 0 12 176 8 2 6 0 204
Scaphoid 0 0 0 3 80 27 19 0 129
Trapezoid 0 0 0 8 36 59 21 0 124
Trapezium 0 0 0 7 30 31 51 0 119
Pisiform 0 0 0 0 0 v0 0 19 19

Pryor noted the ages at which the epiphysial centers appeared (Table
2.3). Responding to criticism that he had not compared actual bones
with x—rays of those same bones, Pryor countered that x-rays showed
the bones' texture and the very faint epiphysial lines more clearly
than the naked eye could detect (Pryor, 1925). In a 1928 report,
Professor Pryor presented an historical review of his 30 years of
research into the skeleton's ossification process.

Unaware of Pryor's reports, Boston Physician Thomas Morgan
Rotch pondered an anatomical age (Rotch, 1908). X-ray technician
A.W. George of Boston's Children's Hospital had alerted Dr. Rotch
to the anatomic age concept. From examining over 500 x-rays of
children's hands, wrists, elbows, shoulders, hips, knees, and ankles,
Dr. George decided that anatomic age indicated the biological age
differences between children better than chronologic age. And he
concluded that the carpal bones indicated anatomic age better than
any other bones. Working from 200 x-rays of children's carpal bones,

Dr. Rotch developed his Anatomic Stage Index. Because he wanted

15

Table 2.3

The ages by sex that the hand-wrist epiphysial centers appeared

(Pryor, 1925)

 

 

Bpiphysial Center Sex Time of Appearance
lst Metacarpal Female Early in the 3rd year
Male Early in the 3rd year
2nd Metacarpal Female Late in the lst year
Male Early in the 2nd year
3rd Metacarpal Female Early in the 2nd year
Male Late in the 2nd year
4th Metacarpal Female Early in the 2nd year
Male Late in the 2nd year
5th Metacarpal Female Early in the 2nd year
Male Late in the 2nd year
Middle Phalange:
First Row Female Late in the lst year
Male Early in the 2nd year
Second Row Female Early in the 2nd year
Male Late in the 2nd year
Third Row Female Late in the 2nd year
Male Late in the 3rd year
Ring Phalange:
First Row Female Early in the 2nd year
Male Late in the 2nd year
Second Row Female Early in the 2nd year
Male Early in the 3rd year
Third Row Fenale Late in the 2nd year
Male Late in the 3rd year
Index Phalange:
First Row Female Early in the 2nd year
Male Late in the 2nd year
Second Row Female Early in the 2nd year
Male Early in the 3rd year
Third Row Female Late in the 2nd year
Male Late in the 3rd year
Little Phalange:
First Row Female Early in the 2nd year
Male Late in the 2nd year
Second Row Female Late in the 2nd year
Male Early in the 3rd year
Third Row Female Late in the 2nd year
Male Late in the 3rd year
Thumb Phalange:
First ROW Female Late in the 2nd year
Male Early in the 3rd year
Second Row Female Early in the 2nd year
Male Late in the 2nd year
Radius Female About the 8th month
Male About the 15th month
Ulna Female Between the 6th and 7th year
Male Between the 7th and 8th_year

 

16

to separate anatomic from chronologic age, Rotch alphabetized his

Index (Table 2.4).

TafleZA

Alphabetized Anatomic Stage Index

(Rotch, 1908)

 

 

Chronologic Anatomic Description
y, Age Stage

6 months A Os Magnum and Unciform present

33 months B Distal Radial Bpiphysis appears

33 months C Cuneiform appears 1
27 months D Semilunar appears 1
42 months E Either the Trapezium or the Scaphoid appears
66 months F Both the Trapezium and the Scaphoid appear
78 months G Trapezoid appears

72 months H Distal Ulnar epiphysis appears

81 months I Total carpal development advances

99 months J Total carpal development advances

135 months K Pisiform appears

153 months L Pisiform is very clear and total carpal

development advances
162 months M Pisiform is almost as large as the Cuneifqgm

g

 

He hypothesized that when classifying children for school,

athletics, or child-labor, anatomic age defined children's readiness

better than chronologic age.

To support his claim, Rotch cited

Dr. C. Ward Crampton's work on the superiority of physiologic age

over chronologic age (Crampton, 1908a).

Because researchers could

use anatomic age from infancy to adulthood, Rotch contended that

anatomic age estimated biological age better than any other method.

After a series of communications with Professor Pryor, in

his next paper, Rotch deferred to Pryor's anatomic expertise and

corrected two errors in his first report (Rotch, 1909).

He reversed

the appearance order of the Trapezium and the Trapezoid, separated

17

the ossification processes of boys and girls, and included Pryor's
Anatomic Stage Index in his paper (Table 2.5). Rotch said that boys
and girls alike could use his Anatomic Stages, but to determine the
chronological age that corresponded with each Anatomic Stage, he had
to separate boys from girls. In a huge table containing data from
129 boys and girls aged one month to 13 years rated as intellectually
bright, Rotch provided their sex, height, weight, number of teeth,
and Anatomic Stage. From that table, he concluded that chronological
age, height, weight, number of teeth, and parental or guardian state-
ments insufficiently attested to the children's readiness for school,

athletics, and physical work.

Table 2.5

Pryor's Anatomic Stage Index (Rotch, 1909)

 

 

Sex Chronologic Anatomic Description
Age (months) Stage

Female 6 A Magnum and Unciform present
Male 12 A
Female 18 B Distal radial epiphysis
Male 24 B appears
Female 30 C Cuneiform appears
Male 30 C
Female 42 D Semilunar appears
Male 51 D
Female 51 E lst of the Scaphoid, Trapezoid,
Male 63 E or Trapezium appears
Female 57 F 2nd of the Scaphoid, Trapezoid,
Male 69 F or Trapezium appears
Female 63 G 3rd of the Scaphoid, Trapezoid,
Male 81 G and Trapezium appears

Female 75 H Distal Ulnar epiphysis appears
Male 87 H
Female 90 I Everything advances
Male 108 I
Female 102 J Everything advances
Male 120 J
Female 123 K Pisiform appears
Male 153 K

 

18

Although applauding Crampton's work on physiologic age,
Rotch reiterated that because Crampton's Index required someone to
view the boys' public hair, the schools and courts could not use his
procedure. Also, he said that because Crampton's Index applied only
to the adolescent aged boys, anatomical age estimated biological
maturation better than Crampton's Index. Nevertheless, Rotch credited
Crampton for disproving chronologic age as a classification procedure
for children's readiness for school, athletics, and child labor.
Rotch again called for child labor laws based on anatomic age and he
advised educators to consider epiphysial development as an educa-
bility indicator. To accentuate that point, Rotch suggested that
when educators found children whose mental development exceeded their
epiphysial development, to enable epiphysial development to catch up,
educators should withdraw those children from mental exercises.

In May of 1909, the Secretary of the Navy asked Dr. Rotch
to prepare an Anatomic Index that Naval officers could use to deter—
mine the anatomic age of Anapolis Cadets. From x-rays of Naval Cadets
aged 16 to 21 years old, Drs. Rotch and Harold Wellington Smith con-
cluded that after boys reached 13 years of age, researchers should
use the boys' epiphysical unions to describe their Anatomic Age.
They found that the epiphyses that appeared first also ossified first
(Rotch and Smith, 1910). Using the distal radial epiphysis as their
primary criteria, they proposed the following four levels of epi-
physial union: (1) distal radial epiphysial union had just begun;

(2) distal radial epiphysis had ossified one-quarter across the

19

radial shaft; (3) one-half distal radial epiphysis has ossified;
and, (4) distinct distal radial epiphysial lines.

Later that year, Harvard Pediatrics Professor T.M. Rotch
used his Anatomic Index to correlate biological age with school grade
level (Rotch, 1910). Rotch detailed the Anatomic Stage that he
believed the children in each grade should have attained (Table 2.6).
To demonstrate the concept, he provided hand-wrist x-rays of three
boys aged 7 to 10 years old. Because those boys' Anatomic Stages

were identical, he stated that they belong in the same school grade.

Table 2.6

Academic Grade-Anatomic Stage (Rotch, 1910)

 

 

Academic Grade Chronologic Age Anatomic Stage
K1 4 years E
K2 5 years F
lst 6 years G
2nd 7 years 6+ and H
3rd 8 years H+ and I
4th 9 years 1+ and J
5th 10 years J+
6th 11 years K
7th 12 years L
8th lBdears M

 

Under the auspices of the Committee on Physical Welfare of
School Children, New York City Physician C. Ward Crampton studied
the anatomic and physiologic ages of males (Crampton, 1910). He
said that Rotch's Anatomic Index was too irregular and subjective,
and, therefore, not reliable. Also, Crampton disagreed that the
skeletal development of the boys' hands indicated their total

skeleton's development.

20

Nevertheless, to test Rotch's Anatomic Stages, Dr. Crampton
gave permission and assistance to New York physicians Eli Long and
E.W. Caldwell to analyze hand-wrist x-rays of 200 New York City
Department of Education children aged 13 to 173 months (Long and
Caldwell, 1911). Although they did not compare carpal ossification
with total skeletal ossification, because in one-third of their
sample the right hand x-ray differed from that of the left hand,
Long and Caldwell doubted that carpal ossification accurately
reflected total skeletal development. Because Rotch's Anatomic
Stages I and J allowed x-ray readers to subjectively evaluate the
relationship between the children's increasing bone calcification
and their decreasing carpal interspace area, Long and Caldwell con-
demned Rotch's Index. Consequently, they graded their x-rays in
numerical order from least to greatest skeletal development. Long
and Caldwell concluded that although Rotch's Anatomic Stages unsatis-
factorily quantified skeletal development, they hoped that the
skeleton's ossification centers would eventuate such an index.

Continuing the work he and Dr. Rotch started in 1909, Boston
physician Harold Wellington Smith x-rayed 914 Naval midshipmen aged
17 to 23 years old (Smith, 1913). Using the 21 epiphyses of the
hand and wrist, Smith outlined their appearance order and age of
completed union. Within the epiphysial ossification process, Smith
differentiated eight stages (Table 2.7). For each hand-wrist epi-
physis, Smith plotted curves of the ages of the subjects at each of
his eight stages. Smith expected those curves to define a uniform

epiphysial development sequence. When those epiphysial deve10pment

21

Table 2.7

The Descriptions of the Eight Bpiphyseal

Development Stages in Adolescent Males
(Smith, 1913)

 

Stage Descriptions

 

1 No sign of beginning epiphysial union

2 A few imbrications have appeared in the clear epiphysial cartilage

3 The imbrications have interrupted the clear epiphysial cartilage
and have reached the proximal side

4 Except for the periphery, the imbrications have fused into inter-
lacing multiple lines that occupy the cartilage space

5 The interlacing multiple lines have contracted into an irregular
thick line with deep peripheral notches

6 The thick line has become a distinct, narrow line that extends the
full width of the bone except for a slight peripheral notch

7 Only portions of the line and traces of the peripheral notch remains

8 The bony matrix of the shaft is contiguous with the epiphysial end

curves did not define a uniform sequence. Based on those results, he con.
cluded that his stages did not determine epiphysial development
correctly.

For the first 25 years after Roentgen's discovery of x-rays,
individual researchers such as Professor J.W. Pryor and Dr. T.M.
Rotch did the majority of the research on children's hand and wrist
skeletal development. But, beginning around 1920, University
research centers became involved in the assessment of skeletal matura-
tion. While there were some one-time studies by individuals there-
after, from 1920 on, children research centers such as those at the
University of Iowa, Harvard University, the University of Chicago,
and Cleveland's Western Reserve University dominated the study of

hand and wrist skeletal development.

22

Directed by University of Wisconsin Professor C.R. Bardeen,
University of Wisconsin Graduate Student Earnest Donald requested
Roentgenology Instructor Dr. Howard Curl to x-ray 96 boys and 71
girls aged 42 to 138 months (Bardeen, 1921). Donald correlated the
children's carpal ossification with their chronological age, height,
and weight. After comparing Donald's results with Professor Pryor's
results, Bardeen concluded that the age differences in Donald's
sample were not as large as the differences in Pryor's sample. To
make finer subdivisions of carpal ossification than Pryor's Distinct
or Indistinct categories, Bardeen created four carpal ossification

classes (Table 2.8). Using those four ossification classes as his

Table 2.8

Four carpal ossification classes (Bardeen, 1921)

 

 

Class Description
if A Large, dense carpal ossification center
B Not as developed as Group A
C More developed than Group B
D Barely visible carpal ossification center

 

base, Bardeen divided the carpal ossification process into 11 stages
(Table 2.9). In only four of 96 boys and two of 71 girls did

Bardeen find significant differences in right and left hand ossifi—
cation stage. Because when he compared his sample's age of carpal
center appearances with height and weight, he found no relationship
between those variables, Bardeen said that based on present knowledge,
school officials should not delay children's education because their

carpal centers appeared late.

23

Table 2.9

Eleven Carpal Ossification Stages
(Bardeen, 1921)

 

 

Stage Description
1 Class C Capitate and Hamate
2 Class C Capitate, Hamate, and Triquetral
3 Class C Capitate, Hamate, Triquetral, and Lunate
4 Either Trapezium, Trapezoid, or Scaphoid appeared
5 Class D Trapezium or absent. Pisiform absent. All

other carpals class B or C.

6 Class C Trapezium and Trapezoid. Pisiform absent. All
other carpals class B.
7 Class D Pisiform or absent. All other carpals class B.
8 Class D Pisiform or absent. Class A Capitate and Hamate.
All other carpals class B.
9 Class D Pisiform. All other carpals class A.
10 ' Class B Pisiform. All other carpals class A.
11 A11 carpals class A.

 

In his three year study of 31 girls aged 10 to 13 years and
36 boys aged 11 to 13 years from the University of Iowa Junior High
Schools, University of Iowa Professor Bird T. Baldwin analyzed the
anatomical growth of children's carpal bones (Baldwin, 1921).
Baldwin searched for a completely objective hand-wrist x-ray eval-
uative method. First, he tried measuring the perimeters of the
carpals. Next, he tried tracing the outlines of the carpals on
millimeter graph paper. Using a planimeter, Baldwin eventually
decided to measure the total exposed area of the carpals. Because
the Pisiform was not present in all of his subjects, Baldwin excluded
the Pisiform from his measurement of the total exposed carpal area.

Using Pearsonian correlation coefficients to compare total
carpal area with height and weight, Baldwin found that the boys'

correlations were .879 and .755 respectively, and the girls'

24

correlations were .729 and .766 respectively. Although he found
that the carpal bones in the right and left hands of some children
had great individual differences, Baldwin concluded that on the
average the carpal areas for both hands were equivalent. For boys
and girls of the same chronological age, Baldwin found that the
girls' carpal area measurements exceeded the boys' carpal area
measurements.

Following Professor Pryor's example and because he had non-
identical twin sons, Baldwin studied four sets of twins. Baldwin
concurred with Pryor that in boy and girl twins, the girls' hands
matured earlier than the boys' hands. Additionally, Baldwin observed
that children's hand-wrist x-rays showed that their carpals had
distinct stages of growth. Therefore, Baldwin promised that he would
study the 280 University of Iowa laboratory school children to
standardize their carpal growth stages into male and female
anatomical calendars..

The next year Professor Baldwin reported on the mental growth
curves of normal and superior children (Baldwin, 1922). Among the
measures that he took on his subjects, Baldwin assessed an x-ray
or anatomic stage. Again, he determined the subjects' anatomic
stage by planimetering the total area of the carpal bones. Baldwin
correlated chronological age, weight, height, mental age, and total
exposed surface area of their carpal bones (Table 2.10). Baldwin
claimed that carpal age related closely to physiologic changes.

Frances Lowell and Herbert Woodrow examined hand-wrist

x-rays of 402 Minneapolis and St. Paul children aged 66 to 138

25

Table 2.10

Correlation Table for Chronological Age,
Weight, Height, Mental Age, and Carpal

Age (Baldwin, 1922)

 

 

Chronologic Mental Carpal
Age Weight Height Age Age
WT Chronologic Age 1.00
Weight .84 1.00
Height .98 .86 1.00
Mental Age .88 ' .71 .89 1.00

Carpal area .92 .88 .92 .83 1.00

months (Lowell and Woodrow, 1922). After separating the x-rays by
age and sex, they ranked the x-rays from best developed to least
developed. For that ranking, they considered the surface area of
the carpals, the shape of the carpals, the number of carpals, the
identity of the carpals present, and the carpal area relative to the
total wrist area. They selected the median x-ray as the standards.
Because this technique ultimately provided today's hand-wrist
skeletal development standards, researchers should recognize Lowell
and Woodrow as the pioneers of the concept. Lowell and Woodrow des-
cribed those median x-rays (Table 2.11).

In addition to those carpal measurements, Lowell and Woodrow
determined the number of permanent teeth. When they correlated the
carpal development with number of permanent teeth, they found a .20
correlation. Similarly, when they correlated the carpal development
with school grade, they found another low correlation. Because of
the low carpal development-school grade correlation, they concluded

that they could not agree with Rotch's concept of using anatomic age

26

Table 2.11

X-Ray Standard Descriptions (Lowell
and Woodrow, 1922)

 

Age

(Years) Description

 

Boys: 5 carpals present, one very small. No ulnar epiphysis. Carpal
55 area = 2.9 square centimeters.
Girls: 7 carpals. No ulnar epiphysis. Carpal area = 3.4 sq. cms.
Boys: 7 carpals, two very small. No ulnar epiphysis. Carpal area =
65 3.1 sq. cms.
Girls: 7 fair size carpals. Ulnar epiphysis extends one—quarter the
end of the ulna. Carpal area = 416 sq. cms.
Boys: Ulnar epiphysis extends one-sixth the end of the ulna. Carpal
75 area = 5.7 sq. cms.
Girls: Ulnar epiphysis extends sixty percent the end of the ulna.
Carpal area = 5.9 sq. cms.
Boys: Ulnar epiphysis extends sixty percent the end of the ulna.
85 Carpal area = 5.9 sq. cms.
Girls: Ulnar epiphysial-diaphysial contact. Trapezium-Trapezoid
overlaps. Carpal area = 7.8 sq. cms.
Boys: Oval ulnar epiphysis. Little or no space between the Os Magnum
95 and the Unciform. Carpal area = 6.1 sq. cms.
Girls: Radial epiphysis contacts shaft. Ulnar styloid begins. Ulnar
epiphysis better developed. Carpal area = 9.7 sq. cms.
Boys: Better ulnar epiphysial development. Better carpal develop-
105 ment. Carpal area = 8.9 sq. cms.
Girls: Well-developed ulnar styloid. Small pisiform. Carpal area =
9.7 sq. cms.
Boys: Ulnar styloid begins. Tiny pisiform. Carpal area = 9.6 sq.
115 cms.
Girls: Good-size Pisiform. Good Carpal compactness. Carpal area =
_Q 9.6 sq. cms.

 

to place children in school grades, nor could they agree with
Crampton's use of physiological age for the same purpose.

Under a Commonwealth Fund grant, Harvard University Professor
Walter F. Dearborn of the Harvard Graduate School of Education
examined the mental and physical development of over 5,000 children.
With Professor Dearborn's permission, Professor Daniel Alfred

Prescott of Harvard University's Psycho-Educational Clinic analyzed

27

3,000 hand-wrist x-rays (Prescott, 1923). Professor Prescott
wanted to develop an objective method for assessing children's hand
and wrist ossification stages, and using that objective method, he
wanted to establish objective hand-wrist skeletal development norms.

Because he believed that Pryor and Rotch did not discriminate
skeletal development increments in an objective manner, Prescott
rejected their indices. Although Prescott credited Baldwin with the
first attempt at establishing objective procedures, because Baldwin
failed to adjust his total carpal area index for the size of the
children's hands, he concluded that Baldwin's carpal area index
insufficiently estimated skeletal age. Prescott also praised
Lowell's and Woodrow's pioneer work, but he stated that because their
carpal development also failed to account for the variable sizes of
children's hands, Lowell's and Woodrow's procedure also insufficiently
estimated skeletal age.

For his initial experimental sample, Prescott selected 38
boys and girls aged 83 months to 114 months. He and an unidentified
co-worker independently ranked those 38 x-rays from least to most
developed. Because their rankings correlated at .98+, Prescott
contended that they had ranked those x-rays correctly. He found
bilateral symmetry in over 80 percent of the hands. Because in this
initial sample Prescott found that girls matured earlier than boys,
he repeated the experiment with a second sample of 39 boys aged 74
to 76 months. Even though he had eliminated the sex and age vari-
abilities, Prescott and his unidentified co—worker again ranked the

x-rays similarly. Therefore, Prescott believed that any objective

28

measurement that ranked the x-rays in the same order as their inspec-
tional ranking order would be the objective measurement method.

To find his objective method, the factors that Prescott con-
sidered were: (1) the number, size, and shape of the carpals; (2)
the quality of the carpal ossification; (3) the articulation between
the carpals; (4) the unossified area in the wrist, (5) the size of the
hand; (6) the number of the metacarpal, phalangial, ulnar, and radial
epiphysis present; and (7) the extent of their epiphysial-diaphysial
union. Prescott believed that only when he accounted for hand size.
could ossified carpal area demonstrate the extent of ossification.
Because planimetering the carpals directly on the x-ray proved
unreliable, Prescott projected the x-ray onto drawing paper. When
the projected image measured five times the x-ray original, he traced
the carpal bones.

In his search for the objective measurement method, Prescott
drew nine lines on each x-ray's tracing (Table 2.12).

In the beginning, Prescott used only Line A (the trapezium-
hamate diameter), Line B (the trapezium-triquetral diameter), and
Line C (the scaphoid-hamate diameter). When he found that those
three measurements adequately separated skeletal developmental levels,
he combined those measures into ratios. However, he found that
those ratios also lacked discriminative ability. At this point
Prescott realized that he had to account for the varying hand sizes.
Therefore, by connecting the following four points, Prescott created
the last five of his lines; (1) the medial-most point on the proximal

end of the shaft of the first metacarpal, (2) the medial-most point

29

Table 2.12

The Nine Measurement Lines that Prescott
Drew on the Hand-Wrist X-Rays
(Prescott, 1923)

 

 

QLine Description
TIA The widest diameter from the outside of the trapezium to
the outside of the hamate
B The widest diameter from the outside of the trapezium to the
outside of the triquetral
C The widest diameter from the outside of the scaphoid to the
outside of the hamate
D The distance between the first metacarpal's landmark and the
fifth metacarpal's landmark
E The distance between the fifth metacarpal's landmark and the
radius' landmark
F The distance between the first metacarpal's landmark and the
radius' landmark
G The distance between the radius' landmark and the ulna's
landmark
H The distance between the ulna's landmark and the first meta-
carpal's landmark
Q K The widest diameter between the hamate and the triquetral

 

on the proximal end of the shaft of the fifth metacarpal, (3) the
lateral-most point on the distal end of the shaft of the radius,
and (4) the medial-most point on the distal end of the shaft of the
ulna.

Theorizing that he needed a ratio of the yet-to-ossify to
the already-ossified area, Prescott hypothesized that Line D times
Line E represented the yet-to-ossify area and Line A times Line K
represented the already-ossified area. Because 1) this DE/AK ratio
decreased sytemmatically with increasing age from 5.10 down to 1.58

and 2) the DB/AK ratio correlated .85 with their inspectional

ranking order, Prescott believed that he was on the right path to

30

determining the objective measurement method. However, when he found
that Baldwin's, and Lowell's and Woodrow's, carpal area measures cor-
related .93 with Prescott's inspectional ranking order, Prescott
decided to search for a better ratio.

Consequently, Prescott added the following three measures to
his x-ray tracings: (l) D = the sum of the widest diameters of all
of the carpals; (2) W = the diameter of the wrist (Line D); and (3)
inter-c = the sum of the shortest distances between the carpals.

From those new criteria Prescott created three ratio combinations.
When he correlated those new ratios with their inspectional ranking

he found all of the correlations exceeded .94 (Table 2.13). In

Table 2.13

The Correlations that Prescott Found for His
Carpal Ossification Ratios
(Prescott, 1923)

 

 

Experimental Groups 2_ inter-c D
W W inter-c
38 boys and girls aged 88-114 months .97 .98 .99-
39 boys aged 74-76 months .96+ .94 .96-
46 boys aged 77-79 months .95 .96 .97+
84 boys aged 86-88 months .95 .95 .97

 

addition to his first two experimental groups, Prescott used two

more experimental samples to correlate his findings. Therefore,
Prescott contended that all three ratios provided a reliable, signifi-
cant, and discriminating measure of carpal development. However,
because D/W required only nine x—ray tracing measurements and less
than three minutes per x-ray, whereas the inter-c and D/ inter-c
ratios required five and eight minutes, respectively, Prescott

adopted the D/W ratio as the objective measure of assessing anatomical

age.

31

Using the D/W ratio, Prescott measured two more experimental
groups to further test his objective measurements' ranking order
with their inspectional ranking order. From the x-rays of 92 boys
aged 83-86 months and 59 girls aged 83-86 months, he found that the
D/W ratio correlated 98+ and 96+, respectively. Prescott called
his D/W ratio values the Anatomical Index. By determining the
Anatomic Index for the median x-ray of each of his age groups,
Prescott indexed the anatomic ages. To create Table 2.14, the author
interpreted the charts that Prescott provided. Prescott claimed that

Table 2.14

The number, sex, age range, and anatomical indices of Prescott's
sample (Prescott, 1923)

 

 

 

 

Number of Sex of Age range Anatomic
.Sub'ects Sub'ects of Sub'ects Index
114 mam. months 1. /o
137 Female 75-80 months 1.82
135 Female 81—86 months 1.92
158 Female 87-92 months 2,03
86 Female 93-98 months 2.13
38 Female 108-119 months 2.50
37 Female 120-131 months 2,68
40 Female 132-143 months 2,90
26 Female 144—155 months 3.06
52 Female 156—167 months 3,14
46 Female 168—179 months 3,18
52 Female 180—191 months 3.18
28 Female 192—203 months 3.26
36 Female 204—215 months 3,18
_4g, 15 Femele 216—227 months 3.20
112 Male 69-74 months 1.23
162 Male 75—80 months 1.43
175 Male 81—86 months 1.54
149 Male 87-92 months 1.70
108 Male 93-98 months 1-84
17 Male 108-119 months 2-00
29 Male 120—131 months 2.38
24 Male 132-143 months 2.55
34 Male 144—155 months 2.69
36 Male 156—167 months 2-98
41 Male 168-179 months _ 3-10
50 Male 180-191 months 3.18
35 Male 192-203 months 3-10
28 Male 204—215 months 3.28

15 Male 216-227 months 3.24

32

the uniform distribution of the Anatomic Indices established the
validity and merit of his D/W ratio.

Supervised by Western Reserve University Anthropology Pro-
fessor T. Wingate Todd, Masters' Degree Candidate Paul H. Stevenson
examinedllO skeletons from persons who died at ages 15 to 28 years
old (Stevenson, 1924). Stevenson wanted to determine the sequence
and duration of epiphysial unions. To acquire a general understand-
ing of epiphysial union, Stevenson developed four epiphysial union
phases (Table 2.15). For the 28 epiphyses associated with the
humerus, radius, ulna, femus, tibia, scapula, hip, ribs, and clavicle,
Stevenson discussed the age of union of the diaphyses with their
epiphyses. Considering the ulna and radius together, Stevenson
noted that their proximal epiphyses united earlier than their distal
epiphyses. He determined that the distal epiphyses of the radius
and ulna united at age 19 years.

In 1921, the University of Chicago Department of Education's
Laboratory Elementary School staff began x-raying the hands and
wrists of their sample. After a year of collecting those x-rays,
University of Chicago doctoral candidate Thomas M. Carter analyzed
those x-rays. After two years of work, Professor Frank N. Freeman
and Mr. Carter reported their carpal ossification ratio (Freeman and

Carter, 1924). While they referenced the Rotch, Woodrow, and

33

Table 2.15

The Four Bpiphyseal Union Phases that
Stevenson Described

(Stevenson, 1924)

 

 

Phase Description
The epiphysial and diaphysial margins
1 project saw-tooth-like external margins.
(No union) During macenation, the epiphysis may
separate from its diaphysis.
2 The epiphysial line develops and occasionally
(Beginning Union) the two margins knit together.
Active bone union completed, but, a fine
3 reddish demarcation line continues. Most
(Recent union) difficult phase to assess.
Completed epiphysial union. Do not mistake
4 ' the capsular attachment line with the
(Complete union) epiphysial line.

 

Baldwin articles, they did not review Prescott's work. However, in
a footnote, they said that they now had Prescott's article, but they
had completed their work prior to receiving it.

Freeman and Carter believed that carpal development indicated
the children's immediate position in the skeletal maturation process.
Because they found carpal diameter measures too irregular, they
decided that Baldwin's total carpal area measure might be more
reliable. Because general skeletal size influenced total carpal
area, they concluded that their measurement procedure had to account
for the varying hand sizes. Therefore, they conceived a carpal
ossification ratio.

In their carpal ossification ratio, their numerator factor
was the ossified carpal area measure, and their denominator was a

carpal quadrilateral area measure. They hypothesized that their

34

carpal quadrilateral area measure adjusted for the general skeletal
size. Although later in their article they said that they would
specify the points of their carpal quadrilateral, their only descrip-
tion was that "points on the bones of the extremities of the arm and
of the hand" formed the carpus quadrilateral.

After analyzing the hand-wrist x-rays of 20 girls and 20
boys at each age from five years through 17 years, Freeman and Carter
provided separate tables for: (1) the carpal quadrilateral; (2)
the ossified carpal area; and (3) the carpal ossification ratio.
For the boys, the ossified carpal area ranged from 45 mm. at age five
to 355 mm. at age 17. Therefore, their carpal ossification ratio
ranged from 45/160 = 0.28 at age five to 355/330 = 1.07 at age 17.

Four months after Freeman and Carter reported their method,
Arthur 1. Gates reported on his 1922-23 experimental study at the
Horace Mann School (Gates, 1924). Because Gates needed a measure of
anatomical development, he reviewed the literature for an objective
procedure. After reading Baldwin's work, Gates decided that Baldwin
failed to account for varying hand sizes. Therefore, Gates hypothe-
sized an ossified carpal area to a carpal quadrilateral area ratio.

To represent his carpal quadrilateral area, Gates selected
points on the extreme outer corners of the first metacarpal, fifth
metacarpal, ulna, and radius. By combining both hands and dividing
the ossified areas, Gates computed his Ossification Percentage. For
30 boys and 28 girls with an average of 68 months and an average
deviation of 12 months, Gates' Ossification Percentage ranged from

11 to 44 percent, with a mean of 25 percent. For 26 boys and 30

35

girls with an average age of 115 months and an average deviation of
10 months, Gates' Ossification Percentage ranged from 41 to 72 percent
with a mean of 55 percent.

To show that when compared to the Freeman-Carter Ossification
method any inspectional method inadequately estimated skeletal age,
Joseph C. McElhannon reported on a reliability study that he partici-
pated in at the University of Chicago (McBlhannon, 1926). For 20
boys and 20 girls at each age from seven years to ten years, McElhannon
first inspected the x-rays and ranked them from least to most mature.
McElhannon explained that the landmarks of the Freeman-Carter carpal
quadrilateral were the "proximal points of the radius, ulna, fifth
metacarpal, and the epiphysis proximum of the first metacarpal,"
and that the ossified carpal area included the Multangulum Majus,
Multangulum Minus, Naviculare, Lunatum, Triquetrum, Capitatum,
Hamatum, Pisiforme, and the distal radial and ulna epiphyses. In
addition to measuring the area of the carpals and the carpal quad-
rilateral by the Freeman-Carter method, he observed the shape,
number, and identity of the carpals present. Then, to determine
which inspectional method ranked the x-rays the best, McElhannon
used the following six different criteria; (1) the development of
the Multangulum Majus, Multangulum Minus, and Novaculare; (2) the
development of the Lunatum and Triquadrum; (3) the development of
the Capitatum and Hamatum; (4) the development of the distal radial
epiphyses and the distal ulnar epiphysis; (5) the fusion of the

distal ulnar epiphysis; and (6) general observation.

36

Because a composite of all of those inspectional criteria
correlated almost perfectly with his initial general inspectional
ranking, McElhannon determined that all inspectional ranking methods
were equivalent. Therefore, he concluded that having several inspec-
tional criteria failed to increase the rater's ability to discern
the skeletal maturation.

Next, he correlated his inspectional method with the Freeman-

Carter method (Table 2.16). McElhannon concluded that as the carpal

Table 2.16

The Correlation Values Between McBlhannon's
Inspectional Method and the Freeman-

Carter Method (McBlhannon, 1926)

 

f'7 year 7 year 8 year 8 year 9 year 9 year 10 year 14 year
boys Qgirls boys girls boys girls girls boys
.85 .91 .82 .73 .75 .62 .41 .53

 

bones began to overlap and fill the carpal quadrilateral, the relia-
bility of inspectional methods decreased. Therefore, because as
carpals grew, the inspectional estimate decreased, McElhannon advised
that workers reject inspectional methods and use a scientific mea-
surement method (McBlhannon, 1926).

Later that same year, Thomas M. Carter also reported on the
reliability of the Freeman-Carter method (Carter, 1926). In this
article, Carter described the following four points as the Freeman-

Carter carpal quadrilateral landmarks: (l) the extreme outside

37

point of the distal ulnar shaft; (2) the most proximal point of the
fifth metacarpal; (3) the nearest point on the first metacarpal
epiphysis; and (4) the extreme outside point on the distal radial
shaft. To verify the Freeman-Carter measurement method, Carter and
fellow graduate student E.J. Brown independently measured 15 children's
x-rays by the Freeman-Carter procedure. Then they x-rayed those
children a second time and again independently measured their x-rays.
Their test-retest correlations were .97 and .96, respectively.
Carter concluded that those correlations verified the reliability
of the Freeman-Carter method.

Additionally, Carter reported that three other University of
Chicago graduate students also tested the Freeman—Carter method.
Using erays of 20 boys and 20 girls at each age of 9, 10 and 11
years old, W.H. Buchanan correlated an inspectional method with the

Freeman—Carter method (Table 2.17). J.G. McElhannon and Everett

Table 2.17

The Correlation Values Between Buchanan's
Inspectional Method and the Freeman-

Carter Method (Carter, 1926)

 

9 yr old 9 yr old 10 yr 10 yr 11 yr 11 yr
boys girls boys girls boys girls

Buchanan's .85 .90 .78 .61 .69 .58
Correlations

Davis studied x-rays of 20 boys and 20 girls at each age of seven,
eight and nine years, and 20 girls aged ten years and 20 boys aged

14 years (Table 2.18)

38

Table 2.18

The Correlation Values Between the Inspectional Methods
of McElhannon and Davis and the Freeman-Carter
Method (Carter, 1926).

7 yr 7 yr 8 yr 8 yr 9 yr 9 yr 14 yr 10 yr

1 boys girls boys girls boys girls bovs girls
McBlhannon's

within-rater .85 .91 .82 .73 .75 .62 .41 .53
Davis'
within-rater .86 .62 .83 .73 .78 .73 .87 .92
McElhannon-Davis

;between-raters .49 .81 .91 .91 .88 .69 .51 .69

 

Table 2.17 shows why McBlhannon concluded that the accuracy
of his inspectional method decreased as the carpals filled the
Freeman-Carter carpal quadrilateral. However, McBlhannon's inter-
pretation failed to hold for Davis. In Fact, Davis' two highest
values were for the oldest two groups.

In 1927, University of Minnesota Professors Richard E.
Scammon and Gordon H. Scott studied the accuracy of the 7-inch
Amsler planimeter that Baldwin and Freeman and Carter had used
(Scammon and Scott, 1927). Unaware of the readings, an observer
planimetered the area of several circles which ranged from one to
four centimeters in diameter 25 consecutive times. A second observer
recorded each of the 25 readings. They determined that the Amsler
planimeter failed to provide reliable readings. Therefore, they

examined the Hammer area-by-weight method of determining ratios.

39

In the Hammer area-by-weight method, workers trace the
desired area on special Hammer paper, meticulously cut out those
areas and determine the desired ratio by weighting the two areas.

To compute the Freeman-Carter ratio by the Hammer method, the workers
would trace the carpal quadrilateral and the carpals. Next, they
would cut out and weigh the carpal quadrilateral area, and then

cut out and weigh the ossified carpal areas. By dividing the weight
of the ossified carpal areas by the weight of the carpal quadrilateral
area, the workers would determine the Freeman-Carter ratio.

.. However, Scammon and Scott realized that the Hammer method
had three primary sources of error. (1) The ability of the workers to
precisely cut out the elaborate outlines of the various structures
varied. However, they found that careful operators minimized that
source of error. (2) The paper's moisture content also varied, which
caused the weight to vary depending on the ambient humidity. After
much experimenting, Scammon and Scott decided that workers should
use Eastman Kodaloid No. 3 celluloid in place of the Hammer paper.
(3) The thickness of the Hammer paper also varied, which caused the
weight to vary. With regard to uniform thickness, they determined
that celluloid sheets varied 50 percent less than the Hammer
paper. Because of the transparency of the celluloid, workers traced
directly from the x-rays and since their sharp stilus easily cut
through the celluloid, the Scammon-Scott method eliminated two steps
from the Hammer method.

After Professor Bird T. Baldwin's death, Laura M. Bushby and

Helen V. Garside wrote the final draft of Baldwin's ten year study on

40

anatomic growth (Baldwin, Busby, and Garside, 1928). Iowa Child
Welfare Research Station research associate S. Idell Pyle assisted
Busby and Garside with the report's bibliography. Although Baldwin's
study began in 1918, he collected the majority of his data between
1923 and 1926. Baldwin searched for basic anatomic growth principles,
and if he uncovered a pattern, then he wanted to develop a general
anatomic growth calendar.

After unsuccessfully experimenting with roller and polar
planimeters, Baldwin et a1. purchased Coradi's circumscribing disc
planimeter to measure their area values. By remeasuring 56 carpals
of eight ten-year-old boys, they determined that the reliability
correlation coefficient of their planimeter measurements was .9996.
To obtain their wrist rectangle measurement, they measured the width
of the wrist from the most exterior points of the first and fourth
metacarpals and the length of the wrist from the midpoint of the third
metacarpal's proximal end to the midpoint of the radius' distal end,
then they multiplied those two values. For the wrist rectangle mea—
surements, they used a millimeter graduated sliding calipers.

Baldwin et a1. conducted an experiment on the effect of x-ray
hand position. By purposefully mispositioning their subjects' hands
they found that the greatest difference in total carpal area was
only 2.4 percent. Therefore, they concluded that children's ability
to maintain the x-ray hand position had a negligible effect on the
Anatomic Index value. Using several grades of paper and linen trac-
ing cloth, they tried measuring carpal area and wrist rectangle by

the Hammer weight method. However, they found that besides taking

41

longer, the weight method decreased their accuracy. 0n the premise
that the ratio between a carpal's area and its longest diameter
remained constant, they measured each carpal's longest diameter.
However, they found that the ratio of carpal area to longest diameter
did not remain constant.

To compare the hand and wrist x-rays of boys and girls,
Baldwin et a1. recorded the age and order that they first observed
each epiphyses, each carpal bone and each sesamoid bone, the age at
the onset of ossification of the epiphyses, carpals and sesamoids,
and the age that the distal ulnar and radial epiphysial fused. In
every case, the girls' hands skeletally matured earlier than the
boys' hands.

From the right and left hand x-rays of the 634 boys and 538
girls, Baldwin et a1. recorded the number of carpal bones present,
the area of each carpal and the total carpal area. From 200 older
children's right and left hand x-rays, they determined: (1) the
number of epiphyses present; (2) if those epiphyses had fused; (3)
the number of sesamoid bones present; and (4) two wrist diameter
measurements. To account for the relationship between carpal size
and general skeletal size, they searched for an anatomic index. By
dividing the total carpal area by the wrist rectangle, they computed
an Anatomic Index. For both sexes at each age level, they determined
a regression equation with carpal area the dependent variable and
height and shoulder width the independent variables.

Using their Anatomic Index procedure, Baldwin et al. developed

an Anatomic Index Table (Table 2.19). Except between the one to two

42

Table 2.19

Anatomic Index Values for the Boys and
Girls Aged l to 17 Years (Baldwin,
Busby, and Garside, 1928)

 

 

 

__y Boys Girls
:_AGE (years) N Mean N Mean
1 19 .078 17 .082
16 .092 15 .108
3 22 .116 25 .144
4 26 .152 31 .180
5 31 .180 40 .248
6 40 .220 42 .314
7 45 .226 37 .373
8 48 .342 30 .401
9 31 .382 29 .474
10 ‘ 30 .461 30 .529
11 40 .514 35 .598
12 42 .550 25 .630
13 35 .570 24 .689
14 37 .618 29 .685
15 38 .664 36 .707
16 24 .666 25 .706
17 18 .697 17 .712

 

year and 15 to 16 year old males, the 13 to 14, 15 to 16 and the

16 to 17 year old females, they found significant differences between

every age level. Whereas the girls reached their maximum mean

Anatomic Index by age 15, the boys still had not reached their maximum

mean Anatomic Index by age 17.

In December of 1920, Physical Anthropology research associate

Milo Hellman of New York's American Museum of Natural History x-rayed

the right hand of 60 New York City Hebrew Orphan Asylum girls aged

123-147 months (Hellman, 1928).

In the subsequent three years of

his four year longitudinal study, he measured 56, 47 and 35 of those

same girls. Hellman desired knowledge of the ages at which the hand

43

epiphyses of girls ossified. Because he found its bones readily
accessible and easily observable, he chose the hand as his skeletal
development indicator. Each year he painstakingly recorded any
morphological changes associated with the hand's ossification process.
After several years of intense examination, he realized that he had
to (1) delineate the various phases of epiphysial fusion, (2) arrange
those phases in their manifested order, and (3) based on the girls'
chronologic age, measure the onsets of those phases.

Using the median line, Hellman measured the millimeter length
of each digit from the distal end of each first phalanx to the proxi-
mal end of its metacarpal. He also measured the millimeter length
between the extreme ends of each phalanx, and between the end of the
shaft and its diaphysis. As a result of his intense x-ray examina-
tions, Hellman discovered that epiphysial ossification proceeded
irregularly. Instead, he found that epiphysial ossification passed
through five fundamental morphologic phases. Hellman labeled those
phases Stage A through E and he stated that his five epiphysial
ossification stages did not proceed sequentially (Table 2.20).
Hellman also detailed the percent of the epiphyses that developed via
different stages (Table 2.21).

Hellman concluded that epiphysial ossification started
between 143 months and 155 months and terminated between 171 months
and 180 months. Therefore, once started the entire hand epiphysial
ossification process required less than 36 months.

Using the first three years data of a five year study by the

New York City's Bureau of Educational Experiments, Anthropology

44

Table 2.20
The Five Stages of Bpiphyseal Ossification

That Hellman Determined (Hellman, 1928)

 

Stages

Description

The epiphysis and diaphysis are distinctly separate.
The epiphysial cartilage narrows until the separation
disappears.

A bud-like protrusion extends from the margins of the
epiphysis and diaphysis. It can be one bud in the center
and spread to the periphery, or, one or two buds at the
periphery and spread towards the middle.

The actual fusion of the epiphysis with its diaphysis
which shows on the X—ray as a dark line.

The epiphysis obtains the trabecular struction
continuous with its shaft.

Table 2.21
The Percent of the Bpiphyses that Developed

Via Different Stages (Hellman, 1928)

 

 

 

Transitions Number Percent
From A to B 419 30.6
From A to C 241 17.6
From A to D 26 1.0
From A to E 0 .0
From B to C 145 10.6
From B to D 127 9.3
From B to E 42 3.1
From C to D 195 14.2
From C to B 69 5.0
From 0 to E 107 7.8
totals 1371 100.1

 

45

Professor Ruth Otis Sawtell surveyed hand and wrist x-rays of chil-
dren aged six months to seven years (Sawtell, 1929). She reported
that whereas the children's carpal ossification was erratic during
early childhood, their hand and wrist epiphysial ossification was
orderly. She found that the ratio of the maximum transverse diameters
of the epiphyses to its diaphysis provided consistent age and sex
differences for children aged one to eight years old, and that the
distal radial epiphysis to diaphysis ratio didnot exceed 1.00 until
after eight years of age. Therefore, by multiplying the radial
epiphysis' maximum transverse diameter x 100, and then dividing by
the distal radial diaphysis' maximum transverse diameter, Sawtell
obtained her Radius Index measurement. Sawtell remarked that both
she and Western Reserve University Professor T. Wingate Todd had
selected the distal radial epiphyses and the metacarpal and phalangial
epiphyses as the most promising aspects of the hand with which to
solve the separate problems of differentiation uniformity and qual-
itative trait variability.

In her next report, with permission from Professor Todd,
Professor Sawtell used her Radius Index to assess Western Reserve
University's x-rays (Sawtell, 1929b). She compared her Bureau of
Educational Experiment's children with Todd's Cleveland children
(Table 2.22).

In 1930 Western Reserve University Professor T. Wingate Todd
actively entered the research field of skeletal development. The
first volume of the 1930 Child Deve10pment periodical contained

three of Todd's articles on skeletal development.

46

Table 2.22

Sawtell's Comparison of the Bureau of Educational Experiments
Sample and the Cleveland Sample (Sawtell, 1929)

 

Bureau of Educational Experiments Sample

 

 

 

 

 

 

 

Age(yrs) Boys Girls _4

N Mean N Mean

1 10 17.4 7 24.9

2 15 43.4 11 55.7

3 16 59.2 14 70.3

4 21 66.2 24 79.4

5 22 76.7 24 87.5

6. 20 83.0 21 91.0

7 11 91.6 15 95.5

8 10 95.0 10 99.8

Cleveland sample
Age(yrs) Boys Girls

N Mean N Mean 7

5 9 79.3 9 89.1

6 40 83.8 31 89.3

7 40 89.3 32 92.9
y 8 44 94.0 30 99.3 #

 

Todd's first report contended that skeletal development specialists
based their estimates of children's age of epiphysial union on insuffi-
cient valid evidence (Todd, 1930a). He said that basically their esti-
mates came from Henle's 1891 report. In spite of impressive sample sizes,
because x-ray researchers had never compared children's actual bones with
x-rays of those bones, Todd questioned their observations. Additionally,
Todd stated that because those x—ray workers failed to learn the skeletal
maturation patterns of primates, they could not comprehend the skeletal

maturation patterns of humans. Because Western Reserve University's

47

Hamann Museum housed more quantifiable human skeletons than any

other institution, Todd said that no one could successfully refute
his observations. The Hamann Museum possessed 200 children and young
people skeletons, 150 anthropoid skeletons, and more than 850
mammalian skeletons.

Todd explained that although developing epiphyses form a
nearly perfect cap, when this cap was large and complicated or deeply
embedded in soft tissue, its x-ray easily confused the untrained
worker. Therefore, he re-emphasized that except for those few x-ray
workers that he had trained with actual skeletons, x-ray researchers
could not relate the x-ray to the actual.

To evaluate the several stages of epiphysial union throughout
the body, Todd examined the Hamann skeletons. After careful analysis
he decided to label epiphysial unions as: (0) no union; (1) com-
mencing union; (3) recent union; and (4) completed union. If the
x—ray showed an indefinitely outlined whitish margin, then the actual
bones showed slight bone tissue condensation between the diaphysis
and epiphyses. When Todd compared the x-rays with the actual bones
in his "commencing union" stage, he said that the actual bones showed
the progressive narrowing of the diaphysoepiphysial gap. And he
said that for a period of six months after epiphysial-diaphysial
fusion, a fine red line marked the site of the union and this post—
fusion red line eventuated to a white scar. Todd labeled the white

scar stage as his "completed union" stage. By averaging the scores

48

he assigned to every epiphyses in the skeleton, Todd estimated the
skeleton's epiphysial maturation. Through careful dissections, Todd
found that those epiphysial ossification stages occurred in other
mammals as well as man.

Because he questioned the reliability of epiphysial union
data of the standard anatomical textbooks, in 1923 Todd had asked his
graduate student, P.H. Stevenson, to trace the origins of those data.
Stevenson reported that except for Dwight's data in Piersol's Human
Anatomy textbook, the standard anatomical textbooks based their data
on Henle's 1871 report. Therefore, by examining the Hamann Museum's
collection of over 200 children and adolescent skeletons, Todd compared
the actual bones and the x-rays of those bones to find the facts
about epiphysial union (Todd, 1930b). Because Todd discovered the
epiphysial development proceeded in a complicated manner, and
because previous skeletal maturation research lacked fundamental
comparitive information, in spite of adverse reactions to Stevenson's
1924 report, Todd delayed his rebuttal until he clinically applied
his observations. After analyzing the x-rays of over 1,000 living
children and clinically applying his observations, Todd answered
those opposed to Stevenson's 1924 report.

The x—ray sites where Todd said that workers could readily
ascertain the successive stages of epiphysial development were the
distal radial epiphysis and the distal humeral epiphysis. Todd
described nine stages of epiphysial maturation (Table 2.23).

In his third article in the 1930 volume of Child Development,

Todd discussed the principles of skeletal maturation and differentiation

49

Table 2.23

The nine epiphysial maturation stages that Todd described (Todd, 1930b)

 

yStages

V'—

1

Description

While a spherical epiphysial nodule progressively
ossifies to reproduce the outline of the growing carti-
laginous epiphysis, the diaphysis likewise ossifies.

Adjacent to the delimiting epiphysial and diaphysial
surfaces, the bone texture changes. On the X-ray stage 2
shows a hazy ribbon between the surfaces.

The bony outlines develop a thin coating of more
condenced osseous tissue. On the X-ray, Stage 3 shows a
fine delimiting line of bone between the surfaces.

The bony outlines parallel each other. On the X-ray
stage 4 shows a closing of the epiphysial-diaphysial gap.

The bony outlines parallel each other. On the x-ray
stage 5 shows a definite parallelism in the clear billowy
bony outlines.

The bony outlines narrow. Todd corresponded his stage
6 to Hellman's Stage B.

The bony outlines erratically extend across the epiphy-
sial-diaphysial gap from the deeper surface towards the
outer bone surfaces. On the X-ray, stage 7 shows as patchy
irregularity of Opposing surfaces with isolated double
fine parallel lines. Todd corresponded his stage 7 to
Hellman's Stage C.

On the outer surface of the bone the bony surfaces
appear as a thin red line. On the X-ray, stage 8 shows a
thicker single white line. Todd corresponded his stage 8
to Hellman's Stage D.

The osseous trabeculae extends from the shaft through
the epiphysis and the bone loses all traces of epiphysial
differentiation. Todd corresponded his stage 9 to Hellman's
Stage E.

in the knees, elbows, and hands (Todd, 1930c). He noted that whereas

epiphysial growth merely increased the bones' dimensions, when the

shaft of the bones grew, the bones changed their proportion as well

as their dimensions. As evidence that skeletal maturation disabilities

involved only dimensional stunting and not differentiational stunting,

Todd pointed to the number of short people versus the number of people

with inhibited skeletal differentiation.

50

To select the X-ray standards for his skeletal maturation
atlas, Todd catalogued his 1,000 hand—wrist X-rays by sex into six
month age intervals. Then, within each group he ordered the X-rays
from least to most differentiated. And finally, he selected the
X-ray nearest the mean value as each group's standard. Thereafter,
whenever they evaluated the skeletal age of their sample, Western
Reserve University's child growth and development research team used
Todd's atlas.

Todd acknowledged that Professor Sawtell had checked her
Radial Index against his qualitative study, and that she had found
both methods equally dependable. Therefore, when Todd discussed the
principles of differentiation, he concluded that while the maturation
process of the short bone shafts in the hand, the ulnar and radial
epiphyses, and the carpals each uniquely described children's hand
and wrist skeletal development, the distal radial epiphysis described
hand-wrist skeletal development better than the rest.

In addition to his previous complaints about X—ray workers
lacking training in reading X-rays properly and in understanding
mammalian skeletal development, Todd argued that when those workers
prOposed a natural variability in skeletal development, they incor-
rectly analyzed the conditions under which those alleged variabili-
ties occurred (Todd, 1931). When workers argued that the male and
female skeletal maturation pattern differed, Todd countered that man
and woman manifested a single genetic plan. While Todd found that
female hands matured earlier than male hands, he also found that the

female precosity did not proceed linearly (Table 2.24). Each group

51

Table 2.24

The female advantage over males in skeletal differentiation (Todd, 1931)

 

 

 

Ages Skeletal Differentiation Advantage (months)

(Years) , Knee Hand Elbow-—
5-0 to 5-5 6 0 0 ‘
5-6 to 5-11 6 0 0
6-0 to 6-5 6 6 0
6-6 to 6-11 12 6 0
7-0 to 7-5 15 6 0
7-6 to 7-11 12 6 0
8-0 to 8-5 9 6 15
8-6 to 8—11 12 4 9
9-0 to 9-5 6 0 6
9-6 to 9-11 18 O 6

10-0 to 10-5 18 12 24 :
10-6 to 10-11 24 15 24 i
11-0 to 11-5 18 18 18 ;
11-6 to 11-11 12 12 12 .
12-0 to 12-5 12 12 9
12-6 to 12-11 6 4 24

 

had 20 to 25 subjects. Todd noted that at about nine years of age,
the females' skeletal differentiation rate slowed until the sex dif-
ference disappeared for the hand and almost disappeared in the knee

and elbow joints, but that after the tenth year, the females' skeletal
differentiation rate accelerated again.

From 13 years of age upward, Todd evaluated children's skeletal
development by epiphysial union (Table 2.25).

Studying the relation of race to skeletal maturation, Todd
found that while white males and Negro males "punched the time clock
together" and probably white and Negro girls also, American boys and
girls skeletally differentiated a couple of months earlier than Italian
boys and girls.

After examining the relationship between children's

bodily forms and rate of skeletal differentiation, Todd concluded that

52

Table 2.25

The female advantage over males in the
age of epiphysial union (Todd, 1931)

 

 

--- AGE (years) Bpiphysial Union Advantage (months)
12-6 to 12-11 18
13-0 to 13-5 18
13-6 to 13-11 18
14-0 to 14-5 15
14-6 to 14-11 12

15-0 to 15-5
15-6 to 15-11
16-0 to 16-5
16-6 to 16-11

oumm

children with advanced height and weight displayed advanced skeletal
development.

Eleven years after Prescott's monograph, Harvard University
Professor Psyche Cattell re-examined Prescott's D/W ratio (Cattell,
1934). Cattell concluded that because both the numerator (D) and
the denominator (W) increased with age, Prescott's ratio and any
similarly designed index failed to provide satisfactory results.

To determine an objective ossification measurement, Cattell
selected 54 annual X-ray plates of 9 feeble—minded girls from 12 to
18 years old. After inspecting those X-rays, Cattell decided on the
following four measurement criteria: 1) the measurements should be
objective, 2) the measurements should not be influenced by hand posi-
tion changes during the X-raying procedure, 3) the measurements should
be taken only on those bones that showed a prolonged growth period,
and 4) the measurements should give equal weighting to the short bones,
the long bones, and the epiphyses. Using those criteria, Cattell

chose twelve measurements to begin her study (Table 2.26).

53

Table 2.26

The twelve hand-wrist measurements that Cattell used in her
search for an objective assessment method (Cattell, 1934)

 

Number Description of the measurement
1-5 The widest diameters of the Capitate, the Triangulate,
the Lunate, the Havicular, and the Greater Multangular.
6 7 The lengths of the first and fifth metacarpal shafts.
8-9 The maximum transverse widths of the first and fifth
metacarpal shaft and the proximal fifth metacarpal shaft.
10-11 The maximum transverse widths of the distal first
metacarpal shaft and the proximal fifth metacarpal shaft.
12 The maximum transverse diameter of the distal radial
epiphysis.

W

Because independent measurements by two regular Harvard
Growth Study research assistants correlated at .95+, Cattell concluded
that her twelve measurements not only compensated for random experi-
mental errors, but also provided total annual increments sufficient
to be reliable from birth through early adolescence.

Supported by a General Education Child Development Research
Fellowship, University of Chicago's Charles D. Flory searched through
anthropometric measurements and osseous indices to find a measure from
which he could predict the age at which females reached puberty (Flory,
1935a). From the University of Chicago Laboratory Schools' Cumulative
Records System, Flory selected 80 girls on whom they had taken yearly
measures beginning with their tenth birthday. Flory chose to study
females because the Child Research Center had more complete records
on girls than on boys and because he could use the girls' first date
of menstruation as a criterion of onset of puberty. Flory chose the
following anthropometric measurements: 1) head width, 2) head length,

3) chest girth, 4) chest depth, 5) chest width, 6) lung capacity,

54

7) iliac width, and 8) trochanter width. The osseous indices that
Flory chose were: 1) the Freeman-Carter Carpal Ossification ratio,
2) the age at Pisiforme appearance, 3) the age at the appearance of
the sesamoid at the distal end of the first metacarpal, and 4) Flory's
undefined skeletal age value. Flory determined that all of the osseous
indices predicted the onset of female puberty better than any anthro-
pometric measure, and that the girls' age at the time their first
metacarpal appeared, coupled with their age when their Freeman-Carter
ratio exceeded 1.00 provided the best predictor of the age at onset
of female puberty.

From examining the right hand x-rays of 100 newborns, 300 one
to four year old children, 6000 University of Chicago Laboratory
School children, and 200 University of Chicago students, Charles D.
Flory determined sex differences in the rate of skeletal growth and
degree of skeletal maturation difference by the following four pro-
cedures (Flory, 1935b). 1) At annual intervals for each sex, Flory
selected the X-ray typical of that group's skeletal maturation. When
he compared the male and female standards, Flory found that the five
year old female standard equalled the six year old male standard and
the twelve year old female standard equalled the fourteen year old
male standard. When he used epiphysial closure in the hand as his
skeletal maturity standard, Flory found that fifty percent of the
girls reached completed skeletal maturity by age seventeen, whereas
fifty percent of the boys did not reach completed skeletal maturity
until they were eighteen and one-half years of age. 2) Because the

appearance order of the Naviculare, Lunatum, and Triquetrum carpals

55

related inversely to their eventual size, their area growth curves
had to cross. Therefore, by using the age at which those three area
growth curves crossed as a measure of skeletal maturity, Flory found
that the girls' Naviculare, Lunatum, and Triquetrum growth curves
crossed at age eight and one-half years whereas the same curves in
his male samples did not cross until the boys reached ten years of
age. 3) By using total carpal area as a measure of skeletal maturity,
Flory found that the girls neared their maximum total carpal area at
fifteen years, whereas the boys did not near their maximum total
carpal area until they were seventeen years of age. 4) By using the
children's Freeman-Carter method as a measure of skeletal maturity,
Flory found that the girls reached the 1.00 ratio just prior to age
twelve years, whereas the boys did not reach the 1.00 ratio until
almost fourteen years of age. Flory concluded that those four findings
supported Pryor's contention that female hands matured earlier than
male hands, first by days, then by months, and eventually by years.

In his 1936 Monograph, Charles D. Flory summarized the litera-
ture on children's hand-wrist development (Flory, 1936). In his
opening statement, he presented: 1) the University of Chicago Labora-
tory School children's skeletal development norms, 2) his evaluation
of the Freeman-Carter method, 3) his own skeletal development rating
technique, and 4) a discussion of the predictability of several skeletal
development measures.

As early as five years after the University of Chicago study
that started in December of 1921, University of Chicago master degree

candidate Paul M. Cook questioned the validity of the Freeman-Carter

56

method. Flory cited the following three reasons why the Freeman—
Carter method failed: 1) the slightest positional change during the
X-raying procedure altered the resulting ratio, 2) because the first
metacarpal epiphysis appears after one year of age for infant
measures workers had to guess at that carpal quadrilateral landmark,
and 3) when the children's skeletal maturity approached adult status,
workers had problems locating the quadrilateral landmarks on the
distal radial shaft and the distal ulnar shaft.

Because he determined that no ratio index could provide a
completed skeletal deve10pment value, Flory developed an inspectional
procedure for evaluating carpal development. As early as 1931, Flory
supplemented the quantitative methods with his inspectional method.
While Todd used the wrist as but one of several areas for which
researchers could assess skeletal maturation, Flory explained that
because he found X-raying several different joint areas and assessing
those X-rays expensive and time consuming, he assessed only the wrist.
Flory said that not only does the wrist X-ray easily, but complete
skeletal maturation occurred in the wrist. Flory admitted that he had
utilized many of Todd's suggestions.

For both sexes at annual increments between eight and seventeen,
Flory selected 100 high quality X-rays. After arranging each group's
X-rays from least to most skeletally mature, Flory chose the median
X-ray as each group's standard. Then, Flory requested twenty graduate
students and professors to independently order those standards from
least skeletally mature to most skeletally mature. In every case,

those untrained persons arranged Flory's standards as Flory had

57

selected them. Along with the X-ray standard for each group and sex,
Flory provided a written description of the skeletal maturation
changes since the preceding standard.

One year after Flory's monograph, Professor T. Wingate Todd
published his own Atlas of Skeletal Maturation (Hand) (Todd, 1937).
Although Todd criticized Flory's work because Flory had studied only
the hand and only hands of healthy youngsters, Todd did credit Flory
with a useful research contribution to the study of skeletal matura-
tion (Todd, 1937). Todd noted that during March of 1934, while he
struggled with the verbal descriptions for his own standards, Dr.
Flory had visited Todd's laboratory.

By 1937, Todd had already developed four separate sets of
standards. He created his earlier 1927 and 1931 standards from
chance selection of school children. But, because 1) Todd wanted
age standard deviations of no more than three months for each standard,
2) his random sample selection had provided too few X-rays showing
equal skeletal maturity throughout the skeleton, and 3) Todd had
inadequate serial X-rays of children, Todd refused to publish either
his 1927 or his 1931 standards. While his 1934 standards did have
enough serial X-rays and sufficiently low age standard deviations,
Todd still had reservations about the number of X-rays that showed
unequal skeletal maturity throughout the skeleton. Even though his
1935 standards still had a few X-rays which displayed incongruent
total body skeletal maturity, Todd decided to footnote his concern

about those X—rays and publish his 1935 standards.

58

To create his 1935 standards, Todd and Dr. Carl E. Francis
assessed over 4000 children's X-rays at least twelve times each, or
ahmxt one-half million separate X-ray readings. To select his age by
sex standards, Todd divided the X-rays by sex into six month age
groups. Within each of those groups, in addition to Dr. Francis and
himself, Todd asked various assistants and Dr. H. V. Morley to use
Todd's skeletal maturity indicators to order the X-rays from least to
most skeletally mature. From the central mode of each group's X-rays,
Todd chose the X-ray on which he had the least concern about unequal
skeletal development. Todd said that because he had children of
greater constitutional health in his 1927 and 1931 X-ray samples than
in his 1935 X-ray sample, those early children's advanced skeletal
maturation skewed his 1927 and 1931 standards towards accelerated
skeletal maturity. Because of that skewness problem, as each group's
X-ray standard, Todd selected each group's median X-ray as its
standard. Only the time required by the trial and error method of
practical experience delayed his presentation of those standards.
Each of Todd's sex by age groups contained approximately 125 X-rays.
Additionally, Todd argued that carpal and epiphysial appearance
related to nutrition; therefore, they were not maturity indicators.

After her historical review of the measurement procedures for
assessing hand and wrist skeletal maturation proved unrewarding,
University of Iowa Child Welfare Research Station Professor Harriet
J. Kelly decided to create her own skeletal maturation measurement
procedure (Kelly, 1937). To select the most reliable and practical

anatomic index, Kelly evaluated "four extent of ossification"

59

measurements (Table 2.27) and five "size of the individual" estimates
(Table 2.28) from the right hand and wrist X-rays of 703 Iowa City

white males aged 69 to 219 months.

Table 4.27

Four "extent of ossification" measurements (Kelly, 1937)

 

 

Symbol Description
CA The disc planimetered sum of the carpal areas
CD The sum of the carpal bones greatest diameters
Ca The sum of Cattell's twelve measurements: the Capitate

Hamate, Triangular, Lunate, Navicular, and Greater
Multangular greatest diameters; the first and fifth
metacarpal epiphyses broadest widths; the distal first
metacarpal and proximal fifth metacarpal shafts broadest
widths; and; the distal radial epiphysis widest diameter.

CM The sum of the eight carpal bones greatest diameters, the
distal radial and ulnar epiphyses widest diameters, and
the first four metacarpal epiphyses broadest widths

Table 4.28

Five "size of the individual" estimates (Kelly, 1937)

 

Symbol Description

 

WD 8(A + B); the average wrist diameter

QA 8(A + B) X 8(0 + C); the quadrilateral area
DW E; the diagonal of the wrist

QF where s = 8(A+c+s) and P = B(B+D+E),

QF = S(S-A)(S-C)(S-E) + P(P-B)(P-D)(P-E);
determing the quadrilateral by formula
LM The average length of the second, third, and fourth
metacarpal bones
To calculate the estimates of the size of the individual,
Kelly specified the following four points as her carpal quadrilateral
landmarks: 1) the medial-most point on the first metacarpal shaft,

2) the medial-most point on the fifth metacarpal shaft, 3) the lateral-

most point of the distal ulnar shaft, and 4) the medial—most point of

60

the distal radial shaft. She defined Line A as the points between the
first and fifth metacarpal shaft landmarks, Line B as the line between
the distal radial and ulnar shaft's landmarks, Line C as the line
between the fifth metacarpal shaft landmark and the distal ulnar

shaft landmark, Line D as the line between the distal radial shaft
landmark and first metacarpal shaft landmark, and Line E as the
diagonal line between the first metacarpal shaft landmark and the
distal ulnar shaft landmark. By combining those "four extent of
ossification" measurements and five "size of the individual" esti-
mates, Kelly created twenty indices.

When Kelly related those indices to chronological age, while
the actual extent of the ossification increased, nine of their values
decreased; Therefore, Kelly discarded those nine indices. Because
she considered carpal area (CA) as a valid extent of ossification
measure, and, because she had shown her wrist quadrilateral value as
computed by her (QF) formula to be a valid estimate of the size of
children's wrists, Kelly chose CA/CF as her evaluation criterion for
the other indices. Kelly sought indices that required less measure-
ment time; therefore, she eliminated four indices because they required
the time-consuming task of planimetering carpal area. Because she
believed that ratios of like quantities were more practical than
ratios of unlike quantities, Kelly eliminated all of the mixed quantity
ratios. Through those criteria, Kelly eliminated thirteen of her
original twenty indices.

After applying the remaining seven indices to consecutive

X-rays from twelve individuals, Kelly found that while the actual

61

extent of ossification increased, only her Ca/QF and CM/WD ratio
values did not decrease. Because 1) CM/WD correlated highly with

her CA/QF criterion index and 2) to measure CM/WD utilized the highly
reliable sliding caliper instead of the questionable reliability of

a planimeter, Kelly selected the CM/WD ratio as her estimate of
anatomic growth.

To test the CM/WD reliability, Kelly selected and assessed
the following two experimental groups: 1) Group A contained 35
children aged 91 to 98 months and 2) Group B contained 35 children
aged 177 to 182 months. After delaying six months, Kelly reassessed
Groups A and B. By correlating her first and second CM/WD values,
she found that their Pearson product moment coefficients were .99 and
.98, respectively.

Using 710 X-rays of 151 girls and boys aged 3 to 60 months,
Antioch, Ohio Fels Research Institute workers S. Idell Pyle and
Camelle Menino compared the Flory monograph standards with the Todd
Atlas standards (Pyle and Menino, 1939). To reveal their possible
prejudices, the authors admitted that during the initial phase of
Todd's work on his Atlas, Pyle had spent several months in Todd's
laboratory, and after Todd had completed his Atlas, Menino had
studied with Todd. Neither worker had met Dr. Flory.

On 3x5 notecards, Pyle and Menino independently rated the
children's metacarpal shafts and bases; their proximal, middle, and
distal phalanges; their carpals; their distal radial epiphysis; and
their distal ulnar epiphysis. They always used Flory's monograph to

assess the entire group first. To determine how nearly they agreed

62

in their assessments, they randomly selected the skeletal age ratings
for 75 children aged 36 months. Using Todd's Atlas and Flory's
monograph for their assessments, they reported between raters Product-
Moment correlation coefficients of .97 and .98, respectively. They
attributed their higher Todd Atlas correlation to the fact that they
practiced by rating with Flory's Atlas before they rated with Todd's
Atlas, and to the fact that instead of Flory's annual increments,
Todd's Atlas had standards every six months. On 50 children aged
36 months, they reported identical within-rater Product-Moment corre-
lation coefficients of .87.

Because of the questionable value of appearance age of ossi-
fication centers to assess skeletal age, Pyle randomly selected and

rated 50 girls aged 12 months (Table 2.29). With substitutes for six

Table 2.29

Pyle's Product-Moment correlation for fifty
lZ-month girls (Pyle and Menino, 1939)

%

 

 

:Tfactors Correlation
Todd's Skeletal Age and number of ossification centers .82
Flory's Skeletal Age and number of ossification centers .72
Todd's Skeletal Age and Flory's Skeletal Age .91
With the number of ossification centers held constant,

‘L_Todd's Skeletal Age and Flory's Skeletal Age .81

 

girls they could not re-examine, Pyle assessed those same girls at
age twenty-four months (Table 2.30). Tables 2.28 and 2.29 contain
Pyle's correlation for those girls when they were 12 months old and
when they were 24 months old. Because Pyle's 12-month to 24-month

evaluations of Todd's skeletal age and number of ossification centers

63

Table 2.30

Pyle's Product-Moment correlation for fifty
24-month girls (Pyle and Menino, 1939)

 

 

'_Factors Correlation
Todd's Skeletal Age and number of ossification centers .67
Flory's Skeletal Age and number of ossification centers .83
Todd's Skeletal Age and Flory's Skeletal Age .90
With the number of ossification centers held constant,

Todd's Skeletal Age and Floryjs Skeletal Age .83

 

correlations decreased from .82 to .67, they concluded that after
toddlers' epiphyses differentiate, the number of ossification centers

no longer influenced the Todd Atlas assessment. When they compared
their ratings using the Flory monograph with ratings using the Todd
Atlas, their Todd Atlas rating provided lower skeletal ages than

their Flory monograph ratings. Pyle and Menino concluded that

because Todd's Atlas had six month intervals for children aged six years
or less, Todd's Atlas was superior to Flory's monograph.

Under the direction of Harold E. Jones, University of Cali-
fornia Institute of Child Welfare research associate Nancy Bayley
compared the central tendencies and standard deviations of her
California sample with the norms of Todd's Atlas and Flory's mono-
graph (Bayley, 1940). Except for the first three years of her study
when she annually X-rayed the children, for eight years Bayley semi-
annually X-rayed the left hand and knee of children beginning at age
eleven. Her California norms corresponded closely to Todd's and
Flory's norms.

To test her measurement reliability, Bayley and another worker

independently assessed the California children's X-rays twice. For

64

those assessments of the California children's X-rays Bayley provided
between and within raters correlations (Table 2.31). Because except
for one .85 value she found that all of the correlations exceeded .90,
Bayley hypothesized that their assessments reliably estimated skeletal
age. To determine whether or not the hand and knee ratings differed,

Bayley correlated her hand and knee ratings (Table 2.32). The high

F"

Table 2.31

Between and within raters correlations (Bayley, 1940)

 

 

 

Date N Hands N Knees
Correlations Correlations

Within-Bayley Spring 1935 90 .92 91 .92
Fall 1935 87 .97 86 .94

Spring 1936 86 .97 86 .96

Fall 1936 82 .98 82 .94

Spring 1937 78 .97 79 .94

Between Bayley Fall 1937 79 .92 79 .91
and Rater "T" Spring 1938 76 .94 76 .85
Within-Rater "T" Spring 1938 76 .94 76 .90

Table 2.32

Hand and knee correlation table (Bayley, 1940)

 

 

 

Boys Girls
Date N Correlations N Correlations
Spring 1935 90 .89 87 .91
Fall 1935 86 .89 84 .90
Spring 1936 86 .91 78 .93
Fall 1936 82 .87 because of the increased
Spring 1937 78 .87 number of completed skeletal
Fall 1937 79 .90 development ratings, Bayley
Spring 1938 76 .91 stopped computing these

correlation values.

correlations between the hand and knee reinforced Bayley's concept that
skeletal development proceeded similarly throughout the skeleton.

Nevertheless, Bayley concluded that hand assessments estimated skeletal

65

age better than knee assessments. However, when studying the relation-
ship between skeletal development and other data, she advocated that

a maturity score from the combined hand and knee assessments estimated
skeletal age better than any assessment of either area alone.

Although in most instances her cross-sectional assessments
remained unchanged, to check the accuracy of isolated cross-sectional
assessments, Bayley suggested that workers serially order the children's
X—rays and use those longitudinal assessments to compare with their
cross-sectional assessments. Bayley contended that when researchers

‘

examine children's X-rays longitudinally, those researchers more readily
notice peculiarities in individual bone contours and chance variations
in bone positions, and thereby they make fewer assessment errors.
Influenced by the late T. Wingate Todd, the Brush Foundation
averaged the hand and wrist, elbow, shoulder, hip, knee, and foot and
ankle assessments. Advised by Dr. W. W. Greulich, Western Reserve
University Brush Foundation workers Clara C. Beuhl and S. Idell Pyle
searched for the minimum number of those six areas that would provide
a reliable estimate of skeletal status (Beuhl and Pyle, 1942). By
comparing the ages at first appearance of the distal ulnar epiphysis,
the thumb's first sesamoid, and the iliac's crest, they tested the
value of onset of ossification as a skeletal maturity determinator.
To reduce the required number of X-rays, they substituted the onset
of epiphysial fusion of the index finger's distal phalanx for the
onset of ossification in the iliac crest. Because onset of ossifica-

tion in the iliac crest of females occurred at the time of menarch,

they studied whether or not the onset of ossification in the iliac

66

crest of males also indicated the beginning of their reproductive
ability. Todd believed that the onset of epiphysial calcification
was particularly sensitive to environmental factors, and therefore,
except for extremely healthy children, he advised against using onset
of epiphysial calcification as a maturity indicator. However, from
their more recent evidence, Beuhl and Pyle concluded that epiphyses
were no more sensitive to their environment at onset of calcification
than at any other period in their deve10pment.

Brush Foundation workers X-rayed their skeletal development
sample every three or six months until five years of age and annually
thereafter.i Because 1) the distal ulnar epiphysis, the humeral
olecranon process, the humeral medial epicondyle, and the calcaneus
epiphysis begin to calcify at about five years of age, 2) they wanted
a skeletal deve10pment measure at the five year old age level, and
3) they wanted to decrease the number of required X-rays, Beuhl and
Pyle selected the distal ulnar epiphysis as their skeletal develop-
ment criterion for the 5% to 7% year olds. They labeled this age
range as the Ulnar Age. Because the thumb's first sesamoid bone begins
to calcify before the onset of adolescence, Beuhl and Pyle selected
the thumb's first sesamoid to represent the 10 to 12 year old age
level. They labeled this age range as the Sesamoid Age. Then, they
selected the iliac crest to represent the 13 to 14 year old age level.
They labeled this age range as the Iliac Age.

Using 10 annual X-rays of 30 boys and 30 girls starting at
age five, Beuhl and Pyle interpolated the ages at ossification onset

for the distal ulnar epiphysis, the thumb's first sesamoid, and the

67

iliac crest. By comparing Ulnar Age, Sesamoid Age, and Iliac Age with
total skeletal age, except for the girls' Sesamoid Age, they found no
significant differences between the mean total skeletal age measures
and the Ulnar Age, Sesamoid Age, and Iliac Age measures. And, they
found that whereas the mean total skeletal age correlated .24 with
menarchial age, the distal ulnar epiphysis correlated .62 with
menarchial age. Therefore, they concluded that while a correlation

of .62 insufficiently predicted female puberty, considering the Ulnar
Age preceded menarch by seven years, the age at ossification for the
distal ulnar epiphysis significantly exceeded the predictive ability
of mean total skeletal age. Also, they found that whereas the mean
total skeletal age correlated .58 with menarchial age, Sesamoid Age
correlated .71 with menarchial age, and whereas the mean total skeletal
age correlated .50 with menarchial age, Iliac Age correlated .83 with
menarchial age. In their conclusions, they stated that Ulnar Age,
Sesamoid Age, and Iliac Age were as reliable as mean total skeletal
age and, Ulnar Age, Sesamoid Age, and Iliac Age correlated with
menarchial age better than mean total skeletal age.

Comparing the age at onset of epiphysial fusion in the index
finger's distal phalanx with the onset of ossification in the iliac
crest, Beuhl and Pyle found that whereas for boys the onset of ossifi-
cation in the iliac crest started before epiphysial fusion in the
index finger's distal phalanx, for girls the epiphysial fusion of the
index finger's distal phalanx started before the onset of ossification
in the iliac crest. Nevertheless, because they wanted to decrease the

number of required X-rays, Beuhl and Pyle substituted the age of onset

68

of epiphysial fusion in the index finger's distal phalanx for the
onset of ossification in the iliac crest.

At the Symposium on the Concept of Maturity from the Anatomical,
Physiological, and Psychological Points of View sponsored by the
Society for Research in Child Development and held in New York City
on December 28, 1949, Stanford University Professor William W.
Greulich presented his rationale for using hand and wrist X-rays to
assess children's developmental status (Greulich, 1950). Greulich
said that because the genetic heritage of American children varied
considerably, he found that chronological age, height, and weight
were not reliable indicators of growth and development.

To illustrate the relationship between skeletal deve10pment
and sexual maturation, Greulich provided both hand-wrist X-rays and
photographs of a girl with precocious puberty aged 23 to 72 months,
and of a twenty-two year old hypogonadal male. Greulich determined
that at 72 months of age the girl with precocious puberty had a
skeletal age and sexual maturation equal to a normal girl aged 172
months. Contrarily, the twenty-two year old hypogonadal male had the
skeletal age and sexual maturation equal to that of a normal sixteen
year old male.

By arranging the mean annual height gains of seven menarchial
age groups of normal girls so that their menarchial age point touched
the same vertical line, Greulich demonstrated that the female's
maximum annual height increment occurred during the year preceding
menarch. Greulich said that because healthy, adequately nourished

children's skeletons develop as a unit, the skeletal maturation of

69
their hands and wrists are valid indicators of total skeletal matura-
tion. Because of that relationship between the maturational changes
in the skeletal and reproductive systems, Greulich used hand-wrist
X-rays to predict menarch. He found that menarch occurred between
the onset and completion of phalangial union, usually immediately
after distal phalangial epiphysial fusion.

Greulich also used X-rays to predict early versus late
maturing children. For normal boys, he found that changes in their
primary and secondary sex characteristics accompanied or followed
their maximum annual height gains. Greulich provided photographs of
10 boys aged 168 to 170 months. While he observed that the develop-
ment of their primary and secondary sex characteristics ranged from
early prepubertal to slightly sub-adult, Greulich determined that
from 148 months to 200 months their skeletal ages ranged in direct
relation to their sexual maturation. Greulich contended that the
hand—wrist X-rays of those 10 boys demonstrated the close relation-
ship between the developmental status of the reproductive and
skeletal systems. Therefore, the degrees of development of the
reproductive and skeletal systems reliably indicated the boys' general
bodily maturity. However, because the skeleton reflected the body's
general level of maturation from birth to completed skeletal develop-
ment, Greulich said that the hand—wrist X—ray provided a better
biological maturation estimate than primary and secondary sex
characteristics.

Greulich provided the following five reasons why he believed

that hand—wrist X-rays determined developmental status better than

7O

sexual maturation: l) hand-wrist X-rays readily showed progressive
maturity changes, 2) hand-wrist skeletal maturation occurred in an
orderly sequence, 3) hand-wrist skeletal maturation covered the entire
period from birth to early childhood, 4) without regard to genetic
differences, hand-wrist X—rays permitted direct comparisons between
children, and 5) hand-wrist skeletal maturation was intimately

related to sexual development.

As a sequal to Todd's 1937 Atlas of Skeletal Maturation (Hand),
Stanford University Anatomy Professor William W. Greulich (former
Brush Foundation Director) and Brush Foundation research associate
S. Idell Pyle had the Stanford University Press publish their Radio-
graphic Atlas of Skeletal Development of the Hand and wrist. In the
book's preface they explained how in 1931 T. Wingate Todd had started
the Brush Foundation's twelve year investigation of human growth and
deve10pment (Greulich and Pyle, 1950). They noted that pediatricians
had recommended the early Brush Foundation children, and that those
children were of white North European ancestry with above average
economic and educational status. To complete his 1937 Atlas, they
said that Todd included children from various Cleveland public schools
and social agencies. The Greulich and Pyle Atlas included only
children in the Brush Foundation sample. Other than the different
sample, the Greulich-Pyle Atlas differed from Todd's Atlas only in
that they provided annual standards from age five to age thirteen in
girls and from age five to age fifteen in boys. However, to accommodate

the rapid change in skeletal age during the adolescent years, they did

71

include a standard for the thirteen and one-half year old female and
a standard for the fifteen and one-half year old male.

Because they discovered that children suffering repeated
illnesses displayed incongruencies in their skeletal maturation, Pyle
developed a graphic method for describing skeletal status. By plotting
the children's serial ages of individual hand-wrist bones on graphs,
Pyle's procedure disclosed the fluctuations between the least and
most skeletally mature bones. Therefore, until the X-ray readers
devised their own method for assessing the X-rays, Greulich and Pyle

suggested a six-step procedure (Table 2.33).

Table 2.33

Suggested procedure for X-ray assessment according to
Greulich and Pyle (Greulich and Pyle, 1950)

 

 

Step Description

1 Compare the X-ray with the Atlas' standard of the same sex
and nearest chronological age.

2 Compare the X-ray with the Atlas' next oldest and youngest
standard.

3 For a more detailed comparison, select the standard that
generally resembles the X-ray.

4a If the X-ray is of infant to early childhood children, then
the presence of carpal and/or epiphysial centers provide the
guidelines.

4b If the X-ray is of later childhood to pre-pubertal children,

then the shape of the carpals and our list of maturity
indicators provide the guidelines.

5 To prevent overlooking features, develop a regular order for
assessing the individual bones and epiphyses; such as; distal
radius, distal ulna, carpals, metacarpals, and phalanges.
View the carpals in a regular order; such as; Capitate, Hamate,
Triquetral, Lunate, Navicular, Greater Multangular, Lesser
Multangular, and Pisiform.

6 For each bone assign the age of the standard selected that.
resembled it. If the bone's development falls between two
standards, estimate the age at which the bone has progressed
from the lower standard to the higher standard.

72

To enable the physician to determine the extent to which
previously ill children had corrected their imbalanced skeletal
development, they suggested that the X-ray readers plot the least
and most mature hand—wrist osseous center. The smaller the difference
between the skeletal maturation of those osseous centers, the closer
the children were to recovery from their illness. During their dis-
cussion of what constituted a significant deviation from normal in
skeletal age relative to chronologic age, they provided a table of
the variability in skeletal age of the boys and girls in the Brush

Foundation study (Table 2.34). If Greulich and Pyle had used their

Table 2.34

Variability in the skeletal age of the boys and girls in
the Brush Foundation Study (Greulich and Pyle, 1950)

 

 

Number Chronological Age (months) Skeletal Age (months)
121 3 3.01
129 6 6.09
137 9 9.56
130 12 12.74
106 18 19.36
105 24 24.97
107 30 32.40
127 36 38.21
138 42 43.89
170 48 49.04
176 54 56.00
191 60 62.43
186 72 75.46
182 84 88.20
168 96 101.35
160 108 113.90
177 120 125.68
154 132 137.32
165 144 148.82
175 156 158.39
163 168 170.02
124 180 182.72

99 192 195.35

68 204 206.21

73

Atlas to assess the skeletal ages of the Brush Foundation sample, then
the mean skeletal ages should have equalled the chronological ages.
However, because they used Todd's Atlas, the Greulich-Pyle skeletal
ages exceeded those that Todd reported. Therefore, because Greulich
and Pyle used the Brush Foundation children exclusively for their
1950 Atlas, their standards exceeded those in Todd's 1937 Atlas.

In the back of their Atlas, Greulich and Pyle provided
detailed descriptions and schematics of the maturity indicators for

their twenty-one hand-wrist bones (Table 2.35). For each hand-wrist

Table 2.35

Number of maturity indicator stages for the twenty—one
hand—wrist bones (Greulich and Pyle, 1950)

 

 

Number Hand-Wrist Bones Number Maturity Indicator Stages
1 ’ Distal Radial Bpiphysis l3
2 Distal Ulnar Bpiphysis 10
3 Capitate 8
4 Hamate 10
5 Triquetral 8
6 Lunate 7
7 Navicular 8
8 Greater Multangular 10
9 Lesser Multangular 8

10 1st Metacarpal Bpiphysis 9

11 2nd Metacarpal Bpiphysis 9

12 2nd and 3rd Metacarpal Bpiphyses 8

13 4th and 5th Metacarpal Bpiphyses 8

l4 Thumb's Proximal Phalanx Bpiphysis 8

15 2nd and 3rd Fingers Proximal Phalange's Bpiphyses 9

16 4th and 5th Fingers Proximal Phalange's Bpiphyses 9

17 '2nd and 3rd Fingers Middle Phalange's Bpiphyses 8

18 4th and 5th Fingers Middle Phalange's Bpiphyses 8

l9 Thunb's Distal Phalanx Bpiphysis 8

20 2nd and 3rd Fingers Distal Phalange's Bpiphyses 8

21 4th and 5th Fingers Distal Phalange's Bpiphyses 8

 

bone, the number of maturity indicator stages varied. They warned

that because 1) their maturity indicators were not equivalent or of

74

the same relative value and 2) hand-wrist X-ray position could dis-

tort the maturity indicators' appearance, assessors of X-rays should

only use the maturity indicator sketches as auxiliary guidelines.
University of California Institute of Child Welfare Professors

Nancy Bayley and Samuel R. Pinneau found that their X-ray readings

from the 1937 Todd Atlas and the 1950 Greulich-Pyle Atlas differed

(Bayley and Pinneau, 1952). Therefore, because they expected the

1950 Greulich-Pyle Atlas to replace the 1937 Todd Atlas, they revised

Bayley's 1946 Tables for Predicting Adult Height from Skeletal Age

that she had based on the 1950 Greulich-Pyle Atlas. When they com-

pleted their new tables and compared them with their 1946 tables,

they concluded that the two sets of tables compared favorably.
University of Oxford Professor Roy M. Acheson provided the

following four reasons why he objected to the inspectional techniques

offered by the Todd 1937 Atlas and the Greulich-Pyle 1950 Atlas:

1) whereas considerable evidence has shown a wide variability in

ossification patterns, the Todd and Greulich-Pyle Atlases presupposed

a fixed pattern of appearance time and order and subsequent develop-

ment of the hand-wrist carpals and epiphyses; 2) because when the

bones fit between two standards, the raters of X-rays had to interpolate

the bone age, the six month intervals between the Todd and Greulich—Pyle

standards were too large, and, therefore, the raters of X-rays unavoid-

ably made assessment errors; 3) because the Todd and Greulich-Pyle

Atlases had separated the standards for each sex, raters of X-rays

had to further qualify terms like "skeletal age 30 months" as to

whether the subject was male or female; and 4) because the skeletal

75

maturation process does not follow the same schedule as chronological
age, just as height and weight have their own dimensional units,
skeletal maturation should have its own dimensional unit (Acheson,
1954).

To devise his method of assessing skeletal maturation, Acheson .
set up the following three criteria: 1) regardless of the children's
ossification pattern, because to each X-ray assessment all round bones
and epiphyses make their own contribution, the raters of X-rays can
evaluate all X-rays equally; 2) raters of X-rays should record even
the small increases in skeletal maturation; and 3) because raters of
X-rays will record skeletal maturation in Oxford Maturity Units,
they will use the same standards for both sexes.

Acheson decided that as each bone's maturity indicator stages
appeared, the raters would assign that stage as the bone's Oxford
Maturity Unit. Therefore, the sum of all of the hand—wrist bones'
Oxford Maturity Units represented an exact measure of skeletal
maturity. As his basis for determining the hand-wrist maturity
indicator stages, Acheson selected the descriptions and illustrations
that Greulich and Pyle reported.

In 1959 Greulich and Pyle revised their 1950 Atlas (Greulich
and Pyle, 1959). They improved the reproduction of their standards,
and where they determined that the intervals between standards had
been too long, they added new standards. Additionally, they revised
their descriptions of the maturity indicators and had their maturity
indicator schematics redrawn. In a new appendix section, Greulich
and Pyle added the Bayley and Pinneau 1952 Tables for Predicting

Adult Height from Skeletal Age.

76

For the purpose of collecting information on the health of
the U.S. population, in 1956 the Congress authorized the U.S. Public
Health Service to conduct a National Health Survey on a continuing
basis. Because they determined that children's growth and development
were legitimate health data, in 1963 the staff of the National Center
for Health Statistics (NCHS) started assessing children's skeletal
maturation. In addition to being an important basis for classifying
children's health data, they believed that a representative current
sample of U.S. children provided reliable normative data on children's
skeletal development.

To determine the available hand-wrist X-ray assessment pro-
cedures and the uses thereof, the NCHS staff requested advice from
clinicians and directors of long-term skeletal maturation studies.
Among the directors that the NCHS staff contacted were Harvard Growth
Study Director Harold C. Stuart and former Brush Foundation Director
William W. Greulich. Stuart and Greulich recommended that the NCHS
staff ask Case Western Reserve University Developmental Anatomy
research associate S. Idell Pyle to prepare the National Health
Survey's hand-wrist X—ray standards. By 1962 specialists in child
growth and development research had determined that they needed only
one series of hand—wrist X-rays for both sexes, provided they assigned
male and female chronological age equivalents to each standard.

NCHS's Chief Medical Advisor Alice M. Waterhouse and co-worker
Tavia Gordon found Professor Pyle already preparing standard X-ray
references for children's shoulder, elbow, and hand-wrist joints.

Using subjects when she determined had moderate skeletal maturation

77

rates, Pyle had assigned male—female age equivalents to her standards.
As Professor Todd had done, Pyle decided to use semi—annual standards
between the ages of three and thirteen years.

Between July 1963 and the end of 1965, by using four specially
constructed and equipped tractor trailers, the NCHS staff examined
7,100 children ranging in age from 6 to 12 years (Pyle, Waterhouse,
and Greulich, 1971). They parked the trailers in forty locations
across the United States. Then, from 1966 to 1970, the NCHS staff
examined children aged 12 to 18 years, thirty percent of whom they
had assessed in their 1963-1965 tour. To design a hand-wrist X-ray
reference Atlas, the NCHS staff of Paul T. Bruyere, Arthur J.
McDowell, Alice M. Waterhouse, Tavia Gordon, Elizabeth S. Johnson,
Peter V. Hamill, and S. Idell Pyle made the five assumptions about

the ossification process (Table 2.36).

Table 2.36

Assumptions of the human ossification process that provided
the bases for the NCHS Atlas of hand~wrist maturity
(Pyle, Waterhouse and Greulich, 1971)

 

Number Description
i 1. In the cortexes of growing bones, the transitions, ossifying
features and articular facts are the same for both sexes
and across all races. ‘
2. During both childhood and adolescence, the chronological
intervals of the phases through which growing bones
develop vary between very brief and prolonged.

3. In males and females on the same skeletal development
schedule, the ossification process occurs more slowly in males.
4. For direct skeletal age readings, a single standard X-ray
series must show male and female bone ages.
5. When using the single standard X-ray bone age series for only

one sex. although the skeletal age reading for the opposite
sex is an indirect reading, nevertheless, we can translate

that indirect reading into its proper sex-related skeletal age.

 

78

The clinicians and directors that the NCHS staff consulted
proclaimed that they had tested the NCHS's first three assumptions
and had found those assumptions to be correct. With regard to the
NCHS's fourth assumption, the NCHS staff found that Professor Pyle
and her associates had tested that assumption twice, and based upon
the accuracy of that assumption, they had published their 1962
Radiographic Atlas of Skeletal Development of the Foot and Ankle
and their 1969 Radiographic Standard of Reference fer the Growing
Knee. The clinicians and directors said that they did not know if
anyone had tested the NCHS's fifth assumption.

In addition to the Flory Monograph, the Todd Atlas, and the
Greulich-Pyle Atlas, the NCHS staff examined B. Speijer's 1950 Atlas
that he based on hand-wrist X-rays of Dutch children, D. W. MacKay's
1952 Atlas that he based on the hand-wrist X-rays of East African
children, and W. W. Sutow's and K. Ohwada's 1953 Atlas that they
based on the hand-wrist X-rays of Japanese children. When construct-
ing those Atlases, Greulich and Pyle, Speijers, MacKay, and Sutow
and Ohwada admitted that they referred to Flory's Monograph and Todd's
Atlas. Consequently, although the six Atlases provided separate X-ray
standards for males and females, nevertheless, of those workers, all
designed their X-ray standards to enable the assessors to determine
the skeletal age for individual hand-wrist bones without regard to
sex differences.

After agreeing that sufficient empirical and experimental
evidence supported their first four assumptions, in 1965 the NCHS

staff gave their hand-wrist X-ray assessors the following three

79

directives: 1) you will not know the name, age, or sex of the person
whose X-ray you are rating; 2) using the 1959 Greulich-Pyle male
standards only, you will assign a bone age to each hand-wrist bone;
and 3) when a bone does not match a standard perfectly, you will
interpolate its bone age mathematically.

By March 1971, eleven Case Western Reserve University medical
school students that S. Idell Pyle had trained in hand-wrist X—ray
rating had assessed the skeletal ages of 377,800 hand-wrist bones
from 13,496 X-rays. From that work, the NCHS provided the equivalent

hand-wrist skeletal ages for males and females (Table 2.37).

Table 2.37

The equivalent hand-wrist skeletal ages (months) for males
and females (Pyle, Waterhouse, and Greulich, 1971)

 

 

AFemale _Ma1e- Female Male, Female Male; Female Malg__
77—50 36 91 78 143 120 193 162

55 42 97 84 153 126 198 168

61 48 104 90 163 132 203 174

68 54 111 96 170 138 207 180

74 60 117 102 176 144 211 186

80 66 123 108 182 150 216 192

85 72 133 114 187 156 228 206

 

While having twelve standards for females with skeletal ages
less than 5 years, the NCHS Standard of Reference for the National
Health Examination Survey provided twenty-six hand-wrist standards
for children ranging in age from 3 to 228 months. On the page beside
each plate, the NCHS Standard of Reference listed the sex age equivalents
for each bone, and it discussed the bone ossification that had occurred

since the preceding plate.

80

Summary

In the history of the assessment of hand and wrist skeletal
development, the author discussed the research that American specialists
in skeletal development had conducted beginning with Professor Pryor's
work in 1905. The discussion traced the research from early works of
Pryor and Rotch through the University related children research
centers phase that started around 1920 up to the domination of the
field by T. Wingate Todd beginning in 1930. The early hand-wrist
skeletal development research encompassed many different inspectional
and measurement methods. However, when Professor T. Wingate Todd
entered the skeletal maturation research field, the concept of assess-
ing skeletal maturation by non-inspectional methods diminished.
Although Todd rushed his 1937 Atlas into print, his concepts have
continued. Because of Todd's unexpected death in 1938, William W.
Greulich and S. Idell Pyle have carried on Todd's work.

From all of this work, some of the skeletal maturation prin-
ciples pertinent to this investigation are: 1) the skeletal maturation
phenomena represent children's true biological age better than any
other measurable phenomenon of growth and development; 2) because of
the presence of two long bone epiphyses, eight carpal bones, five
metacarpal epiphyses, and fourteen phalangial epiphyses, skeletal
maturation estimates based on the hand and wrist reliably represent
the total skeleton; and 3) the 1959 Greulich—Pyle hand-wrist Atlas
provides a reliable criterion upon which assessors of hand-wrist

X-rays can base their estimates of children's skeletal age.

81

HISTORY OF THE ASSESSMENT OF ADOLESCENT MALES'
SEXUAL CHARACTERISTIC DEVELOPMENT

 

 

In 1900 New York City Physician C. Ward Crampton measured
the physical and mental development of adolescent males in an attempt
to compare the onset of puberty in males with the onset of menarch
in females (Crampton, 1904). From his vast measurement experience,
Crampton believed that the best available indicator of the onset of
male puberty was the boys' pubic hair development. After considerable
cataloging, Crampton decided that pubic hair development had three dis-
tinct stages (Table 2.38). Because of insufficient data, Crampton initi-

ally divided the pubescent stages into two sub-groups.

Table 2.38

Pubic hair stages (Crampton, 1904)

 

 

Classification Description
1. Pre-pubescent No pubic hair
2. Pubescent a. straight, unpigmented pubic hair

b. straight, pigmented pubic hair
3. Post-pubescent Kinked and curled, pigmented pubic hair

From data he collected on 1,200 male students aged 11 to 18
years from New York City High School of Commerce, Crampton determined
the percentage of boys that he classified in his three stages of
pubic hair development. Including only the boys aged 13%, 14%, and
158 years, he provided the percentage of boys at each age that he
rated in each stage (Table 2.39).

Crampton noted the orderly pattern of the boys' pubic hair
development. By multiplying the percentage of boys at a specific age

by the number he had assigned to the pubescent stage, Crampton created

82

Table 2.39

Percentage of boys at each age that Crampton
rated in each stage (Crampton, 1904)

 

 

 

Age

Pubescent Stage 13% 145 15%
1 51% 23% 12%
2 29% 34% 22%
3 18% 42% 65%

 

his "Puberty Index". For example, by using the values shown in Table
2.38 for the 13% year old boys, by multiplying (51 x l, 29 x 2, and
18 x 3), we can determine that the "Puberty Index" for the 13% year
old boys is 51+58+54 = 163. Using his Puberty Index, Crampton com-
pared boys whose parents were born in the United States with boys
whose parents were born in Germany (Table 2.40). Crampton concluded
that boys whose parents were born in Germany entered pubescence later

than boys whose parents were born in the United States.

Table 2.40

The "Puberty Index" of boys with United States versus
German heritage (Crampton, 1904)

 

 

 

Heritage Ages

13% 14% 155
U.S. born parents 165 219 248
Germanyiborn parents 150 195 219

 

To study variability in the boys' height, weight, and right
hand grip strength, Crampton divided the boys by his Pubic Hair Index
within each chronologic age level (Tables 2.41, 2.42, and 2.43).
Crampton noted that at every age level the taller, heavier, and

stronger boys had the more advanced pubic hair deve10pment.

83

Table 2.41

The mean values for height at each age for

each pubic hair stage (Crampton, 1904)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aqe Pubescent Stage N Mean
_1, 1 109 145.52
3 40 155.25
1 77 146.26
145 2 107 148.17
3 130 159.38
1 27 148.79
155 2 44 152.61
3 131 161.31
Table 2.42
The mean values for weight at each age for
each pubic hair stage (Crampton, 1904)
Age Pubescent Stage N Mean
1 116 39.92
135 2 53 39.32
3 49 41.48
1 76 35.53
145 2 108 39.02
3 132 47.50
1 28 38.38
155 2 45 39.38
3 132 49.54
Table 2.43
The mean values for right hand grip strength at each
age for each pubic hair stage (Crampton, 1904)
Age Pubescent Stage N Mean (Kilos)
l 93 28.15
135 2 53 32.24
3 34 35.88
1 67 28.28
145 2 94 31.28
3 115 41.44
1 23 29.78
15% 2 40 31.75
3 125 43.24

 

84

After collecting data for four additional years, C. Ward
Crampton reported his findings on 3,835 New York City High School
boys aged 11 to 18 years old (Crampton, 1908a). In half year intervals,
he provided the percentage of boys that he rated in each of his three

pubic stages (Table 2.44). Because he did not have any subjects less

Table 2.44

Percentage of boys in the three pubic hair stages
at six month intervals (Crampton, 1908a)

 

 

 

Pubertal Age (months)
Stage I47 153 159 165 171 177 183 189 195 201 207 213
1 (81) 69 55 41 26 16 9 5 2 1 0 0
2 (16) 25 26 28 28 24 20 10 4 4 2 . 0
3 (3) 6 18 31 46 60 70 85 93 95 98 100

 

than twelve years old, Crampton used parentheses to denote that he

had estimated the percent in the 147 month age group. Because the
largest increments of maturation occurred between ages of 165 and 171
months and 183 and 189 months, and the smallest increments occurred
between the ages of 177 and 183 months, Crampton concluded that as
measured by his Pubic Hair Index, adolescent males' physiological
maturation had a bimodal distribution centered on the age of 174
months. By observing the rate at which the 69 percent of pre-pubescent
boys at age 153 months advanced to their pubescent pubic hair stage,
Crampton examined the variability in adolescent males' rate of physio—
logical maturation. He noted that while 14 percent of those boys
became pubescent within six months, he still rated 17 percent of them
pre—pubescent twenty-four months later. In addition to his use of

presence and type of pubic hair as his criteria for pubescence,

85

Crampton suggested that the amount of pubic hair, the appearance and
amount of axillary hair, and the appearance and amount of hair on the
upper lip could also serve as useful criteria for assessing
pubescence.

From an unsolicited source, Crampton received research evidence
of well-formed spermatozoa in the secretion of the boys he rated as
pubescent. After inspecting the spermatozoa-pubic hair relationship,
Crampton determined that those pubescent boys with urine containing
well-formed spermatozoa were within a few months of his post-pubescent
rating. Crampton recommended that someone should conduct a thorough
investigation on the relationship of his pubic hair stages to the
presence of well-formed spermatozoa.

In a second 1908 article, Assistant Director of New York
Public Schools, C. Ward Crampton explained that he created his second
stage of pubic hair development because raters of pubic hair develop-
ment had some trouble differentiating between those boys with only
some pubic hair and those boys with an abundance of pubic hair
(Crampton, 1908). Therefore, whenever raters were uncertain about
the boys' pubic hair stage Crampton suggested that they rate those
boys as pubescent. As had Thomas Morgan Rotch with his Anatomic
Index, so Crampton used his Pubic Hair Index to advise educators and
employers of the importance that children's biological age rather than
their chronological age had on their readiness to meet educational or
vocational responsibilities.

Twenty years later at the University of Iowa Child Welfare

Laboratory, University of Iowa Professor Bird T. Baldwin studied the

86

relationship between the development of sex characteristics in males
and their production of spermatozoa (Baldwin, 1928). Between the
dates of May 23, 1923, and April 5, 1924, Baldwin collected 1,135
6:00 a.m. urine samples from 123 state reform boys ranging in age
from 9 to 17 years. The University laboratory tested those samples
for the presence of spermatozoa. If they found a positive slide,
they stained the spermatozoa with carbol fuchsin, and photographed
the slide. Baldwin provided the frequency of spermatozoa that they

found at each annual age level (Table 2.45). They found spermatozoa

Table 2.45

The number of positive spermatozoa samples found
at each age level (Baldwin, 1928)

 

 

 

Descriptive Agesngears)
Data *9 10 11 12 13 14 15 16 17
Number of

Urine Samples 10 30 64 231 206 188 305 75 27
Number of
Positive 2 6 10 29 21 20 25 8 2
Specimens
in 29 of the 123 state reform school boys. Baldwin examined 24 of the
29 boys with positive spermatozoa specimens for axillary and pubic
hair development. He found 18 boys had axillary hair pigmentation
and pubescence and 23 boys had pubic hair pigmentation and pubescence.
To establish more definite standards for the physiological
development of primary and secondary sexual characteristics in
adolescent males, and to establish the relationship between their
testicular development and the development of the secondary sexual

characteristics, University of Pennsylvania Graduate School of Medicine

Physician Paul E. Kubitschek assessed the primary and secondary sexual

87

characteristics of 730 boys aged 9 to 18 years (Kubitschek, 1932).
Although he excluded all Hebrew boys from the study, Kubitschek
included 125 Negro boys in his cross-sectional data collection.
Kubitschek developed a sexual developmental checklist that he used

to assess the boys' sexual development (Table 2.46). Kubitschek

Table 2.46

The sexual development checklist that Kubitschek
developed (Kubitschek, 1932)

k

 

 

 

Sexual Evaluation
__§hacteristic
Penis - Length cm. Thickness cm.
Scrotum
large large large
Testicles Peanut medium Almond medium Walnut medium
- small small small
Pubic hair 0 1 2 3 4
Axillary hair 0 l 2 3 4
Chest hair 0 l 2 3 4
Facial hair 0 1 2 3 4
Voice

 

neglected to describe the procedure that he followed to measure penis
length and thickness, scrotum, and voice. But, he did report that he
determined the boys' testicular size by palpation. When rating pubic
hair, axillary hair, chest hair, and facial hair, Kubitschek used 0 to
represent complete absence and 4 to represent adult growth. After he
assessed all of the boys, he returned to their sexual development
assessment tables for each annual age level, and be classified the

boys' sexual development as average, retarded, or advanced (Table 2.47).
For clarification, the author has placed the proportion value in
parentheses beside the observed value. Although Kubitschek announced

that he expected the Negro boys to sexually mature earlier than the

Table 2.47

The observed frequency of boys at each age level that were retarded,
average or late maturing (Kubitschek, 1932)

 

 

 

Age Kubitschek's Classifications
Groups N Retarded Average Advanced
9 50 8(l6%) 36(72z) 6(12%)
10 50 8(16%) 33(66%) 9(l8%)
ll 65 14(21.5%) 38(58.5%) 13(20%)
12 80 26(32.5%) 40(50%) 14(17.5%)
13 100 30(30%) 52(52%) 18(18%)
14 110 40(36.4%) 47(42.7%) 23(20.9%)
15 105 36(34.3%) 46(43.8%) 23(21.9%)
16 100 22(22%) 51(51%) 27(27%)
17 50 12(24%) 25(50%) 13(26%)
18 20 5(25%) 12(60%) 3(15%)

 

‘

white boys, he found no difference between the Negro and white boys

in onset and rate of sexual maturation.

When he discussed the results of his sexual development

assessments, Kubitschek divided his sample into boys 12 years old

and under and boys 13 years old and older.

Whereas he classified

' the younger boys solely on the size of their penis, scrotum, and

testes, he classified the older boys chiefly on the development of

their secondary sexual characteristics (Table 2.48).

Kubitschek

concluded that the variation in the degree of sexual development in

the annual age levels showed that chronological age indexed sexual

development very poorly.

In 1935, George Williams College Professor Henley S. Dimock

reported on his longitudinal survey of 200 adolescent males (Dimock,

1935). As his pubescence assessment procedure, Dimock used the

"Crampton criteria."

Dimock credited W. B. West's Masters thesis at

the International YMCA College in Springfield, Massachusetts, for

89

Table 2.48

The description of the sexual development changes

at each age level (Kubitschek, 1937)

 

Age

10

11

12

13

14

15

Description

Small peanut size testicles mainly. A few testicles
slightly larger. The degree of variation betweens boys was
very small. Penis length was 3 to 4 cms.

Very little increase in testes size. The degree of
variation between boys was very slight.

Same as the 10 year old group.

 

Testicles begin a fairly rapid size increase—to the size of
a small almond. The degree of variation between boys was
considerable, as many boys still had peanut size testicles.
Penis length was 3 to 7 cms. Pubic hair was frequently
encountered (25% of the cases).

The majority of the boys initiated marked sexual development.
The outstanding sexual development features of this age group

‘were pubic hair growth, voice deepening, and ma8culine body

configuration. Also, these boys displayed a marked testes
size increase. No axillary hair present. The degree of
testes size variation between boys was extensive.

Marked sexual development has continued. The degree in
sexual development variation between boys was still extensive.
Pubic hair was moderately long and dense and covered a 3 cm.
diameter area above and lateral to the base of the penis.
Some axillary hair present in the average and advanced boys.
Husky and uneven vocal quality in some boys. The degree of
testicular development variation was greatest for this age
level than any other age level. Also, the degree of penis
size development variation was equally as great for this age
level.

The sexual development rate has decreased. However, the
degree of sexual development variation is still very marked.
The secondary sexual characteristics develop more than the
primary sexual chacteristics. Long, dense, pubic hairs
extend laterally and cephally. Thin, but definite, axillary
hair present in almost all of the boys. Some boys have
mustache hair and sideburn hair. The Average boys had large
almond or small walnut sized testicles.

90

Table 2.48 (continued)

 

Age

16

17

Description

The sexual development has slowed considerably. The out-
standing sexual development feature of this age group was

the continued secondary sexual characteristics development
and slowed primary sexual characteristics development. Pubic
hairs extended 4 to 5 cms. lateral of the penis. Moderately
long, but thin axillary hair present. Definite facial hair
present. No chest hair present. Testes and penis size only
slightly larger than the 15 year old group.

Greater degree of testes and penis size uniformity. The out-
standing sexual development feature of this age group was the
pubic hair growth up the linea alba. Moderately long, thick
axillary hairs present. Prominent facial hair present. No
chest hair present, except in several of the Advanced rated
boys. Penis and testes size was more uniform.

determining that the pubic hair development index more reliably esti-

mated sexual development than indices based on either axillary or

mammary hair development. For his longitudinal study, Dimock created

six classifications of pubescence status and change (Table 2.49).

Table 2.49

The six pubescent classifications that Dimock
developed (Dimock, 1935)

 

Classification Percent Description

 

Pl-Pl 8% Boys he rated pre-pubescent both years

Pl—P2 13% Boys he rated pre-pubescent the first year,
and, pubescent the second year

P2-P2 1% Boys he rated pubescent both years

Pl-P3 36% Boys he rated pre-pubescent the first year,
and, post-pubescent the second year

P2-P3 26% Boys he rated pubescent the first year,
and, post-pubescent the second year

P3-P3 16% Boys he rated post-pubescent both years

 

With those six classifications, Dimock analyzed the boys'

height and weight. For both height and weight, Dimock found the

smallest to largest change in the boys who went from Pl-Pl, Pl-P2,

91

P2-P2, P3-P3, Pl-P3, and P2-P3. During the year the boys entered
post-pubescence (P2-P3), Dimock determined that they added twice as
much height and weight as the boys who remained pre-pubescent (Pl-Pl)
for the same year period. After conducting several strength measure-
ments on the boys, Dimock concluded that the most rapid strength
development occurred after the boys reached post-pubescence (P3-P3).
Based on the Brace motor ability test results, Dimock reported that
except for his Pl-P2 classification, the boys' motor ability increased
throughout adolescence, and that substantial motor coordination
improvements accompanied the boys' rapid growth in height and weight.
Because the Pl-P2 classification failed to improve their motor
coordination, he concluded that adolescent awkwardness matched the
P1—P2 boys' sudden growth onset rather than the later rapid growth
period.

To indicate whether or not the University of Chicago Labora-
tory School boys had attained puberty, University of Chicago Professor
Herman G. Richey classified boys according to the age when their

axillary hair appeared (Table 2.50) (Richey, 1937). Richey found that

Table 2.50

Richey's maturity group descriptions based on time of
axillary hair appearance (Richey, 1937)

 

 

'VGroup Description
1 Boys with axillary hair present on or before 14th birthday
2 Boys with no axillary hair present on their 14th birthday,

but, who had axillary hair present on their 15th birthday.
3 Boys with no axillagy hair present bygtheir 15th birthday

92

boys in Maturity Group 1 were heavier at all ages than those in
Maturity Groups 2 and 3, and boys in Maturity Group 2 were heavier
at all ages than those in Maturity Group 3.

To determine the activity of androgenic and estrogenic hor-
mones extracted from the urine of adolescent males, Yale University
Professors Ralph I. Dorfman, William Walter Greulich, and Charles 1.
Solomon collected the total daily urinary excretion of 18 boys aged
93 to 195 months (Dorfman, Greulich, and Solomon, 1937). By dissolv-
ing the androgenic material of the urine in olive oil, subcutaneously
injecting that solution into day old white leghorn chickens, and then
determining the increase in the weight of the chickens' combs, they
measured the amount of androgenic activity. They measured the estro-
genic activity of the urine by D'Armour's and Gustavson's adult spayed
rat vaginal smear technique. Table 2.51 contains the results of those
assays for the eighteen boys that they measured. They noted that
Cases 5 and 14 were physically more mature than the other boys. They
concluded that androgenic and estrogenic activities varied considerably
for boys of the same chronological age. Therefore, as their next
investigation they proposed to examine the relationship between boys'
sex hormone secretions and degree of physical maturity.

Using 24 hour urine specimens from 18 prepubertal boys aged 11
years and under and 15 adolescent boys aged 12 to 16 years, Yale
University Professors H. R. Catchpole, W. W. Greulich, and R. T.
Sollenberger measured the urinary excretion of follicle stimulating
hormone (FSH). They found that none of the 18 prepubertal boys showed

any FSH, but eleven of the 15 adolescent boys showed adult amounts of

93

Table 2.51

The amount of androgenic and estrogenic activity found in the urine of
adolescent males (Dorfman, Greulich, and Solomon, 1937)

 

 

Case Chronologic Androgenic Estrogenic
Number Age Activity Activity

1 93 months 1.1 9.0

2 108 months 2.0 5.0

3 132 months 2.0 15.0

4 132 months 2.0 10.6

5 158 months 15.0 48.0

6 158 months 1.2 9.0

7 158 months 4.0 12.0

8 159 months 5.3 24.0

9 160 months 8.0 13.0
10 162 months 8.0 10.0
11 171 months 7.0 --
12 ‘ 176 months 16.5 38.0
13 181 months 6.0 . 18.0
14 184 months 32.0 67.0
15 189 months 20.0 65.0
16 190 months 8.4 24.0
17 194 months 14.0 95.0
18 195 months 19.0 68.0

 

 

FSH. They concluded that the urinary FSH of adolescent males related
to their developmental status rather than their chronological age
(Catchpole, Greulich, and Sollenberger, 1938).

To gain more information about the normal variation in the
developmental sizes of male genitalia, New York Physicians William A.
Schonfeld and Gilbert W. Beebe used a centimeter ruler, a circumference
gauge, and an orchidometer to measure the penis length and circumfer-
ence, and testicular volumes of males (Schonfeld and Beebe, 1942).
They hypothesized that the true physiological length of the penis was
its erect length. However, because they believed they could not
feasibly obtain measurements on the erect penis of their subjects,

they substituted a measure of the fully stretched length of the flaccid

94

penis for the erect penis length. They measured from the mons veneris
to the tip of the glans for the penis length measurement. To test
the validity of their penis length measurement, they asked 150 males
of different ages to outline their erect penis on cardboard for four
successive mornings. By correlating their measurements of the sub-
jects' drawings with their measurements of the fully stretched but
flaccid penis length, they obtained a coefficient of .983. Based on
that result, they stated that their measurement of the fully stretched
but flaccid penis was a valid estimate of the erect penis length.

To obtain the penis circumference measure, they used a set
of circumference gauges. After they slipped a circumference gauge
over the glans, they determined whether or not that gauge would easily
slide over the shaft of the fully stretched penis. Therefore, they
measured the largest circumference of the fully stretched but flaccid
penis. To test the validity of the penis circumference measurement,
they asked the same 150 subjects to insert their erect penis through
measured holes punched into cardboard for four successive mornings.
The subjects reported the smallest hole through which their erect
penis fit with ease. By correlating the subjects' findings with their
circumference gauge measurement, they obtained a coefficient of .965.
Based on that result, they stated that with the aid of the regression
equation Y = .1362 + 1.316X, where Y‘is the predicted penis circum-
ference and X is the observed circumference of the stretched but flaccid
penis circumference, their measurement of the flaccid penis circumfer-

ence was a valid estimate of the erect penis circumference.

95

To obtain the testicular volume, they simultaneously palpated
their standard set of graded orchidometer models and the testes.

They recorded the volume of the orchidometer model that they determined
was the same size as the boys' testes. To test the validity of the
testicular volume measurements, they used this orchidometer to esti-
mate the volume of 400 testicular clay models and to estimate the
volume of 35 postmortem excised human testes. They found that they
frequently overestimated the volumes of the smaller testicular clay
models and postmortem testes and that they frequently underestimated
the volumes of the larger testicular clay models and postmortem testes.
However, beCause they determined that their average error was ten
percent or less, they stated that their orchidometer model measurements
validly estimated the boys' testicular volume (Table 2.52).

They concluded that male genitalia growth proceeded so rapidly
that for any pubescent age the genitalia measurements skewed towards
the higher values, and that pubescent boys' testicular volume measure-
ments were bi-modally distributed. Because the testicular volume
increased so rapidly at pubescence, Schonfeld and Beebe hypothesized
that researchers could use this rapid testicular size increase as an
indicator of the onset of male pubescence. Because the pubescent
rate of genitalia growth far outstripped the rate of growth in height
and weight, they concluded that genitalia growth was a phenomenon of
pubescence. After their subjects reached ten years of age, Schonfeld
and Beebe argued that chronological age no longer reliably estimated

genital development.

L;

Genitalia measurements on males from birth to adult

96

Table 2.52

(Schonfeld and Beebe, 1942)

 

 

 

Age N Penis Relaxed Penis Erect Penis Testicular
_(months) Length Circumference Circumference Volume
0-5 125 3.75cm 3.63cm 4.9lcm .52cc
6-11 51 4.04 3.70 5.01 .72
12-23 50 4.59 3.89 5.26 .74
34-35 50 5.04 4.06 5.48 .83
36-47 50 5.43 4.06 5.47 .80
48-59 50 5.75 4.10 5.53 .82
60-71 50 6.00 4.19 5.65 .79
72-83 50 6.02 4.27 5.76 .75
84-95 50 6.17 4.46 6.01 .82
96-107 50 6.21 4.23 5.70 .83
108-119 50 6.32 4.45 5.99 .94
120-131 82 6.20 4.57 6.15 .95
132-143 91 6.56 4.74 6.37 1.48 g
144-155 93 7.13 5.05 6.78 2.22 c
156-167 104 8.73 5.79 7.76 4.86 2
168-179 121 9.77 6.88 9.19 7.72 8
180-191 101 11.81 7.62 10.16 11.76 3
192-203 76 12.50 7.99 10.65 12.61 g
204-215 61 13.26 8.43 11.23 15.12 m
216-239 71 13.11 8.61 11.47 16.06
240-300 54 13.02 8.55 11.39 16.47

 

At the Adolescent Study Unit and Institute of Human Relations

of Yale University's School of Medicine, Professors William Walter

Greulich, Ralph I. Dorfman, and Hubert R. Catchpole and Physicians

Charles I. Solomon and Charles S. Culotta studied selected changes in

somatic growth and endocrine function that occurred during the male

adolescence period (Greulich, Dorfman, Catchpole, Solomon, and

Culotta,

1942).

Professor Dorfman reported on the urinary excretion

of estrogens and androgens, Professor Catchpole reported on the urine

levels of pituitary gonadotrophic hormone, and Professor Greulich

reported on skeletal and other somatic changes.

Physicians Solomon

.3

‘—

-04

m:

. fix

.5

«4.

3

97

and Culotta assisted the three professors on each of their reports.

Seven hundred fifty-seven white males ranging in age from 10-15 years
comprised their sample. One hundred ninety-four of the boys were in

the Cannecticut School for Boys, 32 resided in the New Haven Community
Center orphanage, 476 attended the large, expensive Andover, Massachusetts,
Phillips Academy private school, and 55 of the boys lived at home. To
direct special attention to the nature and sequence of external bodily
changes associated with adolescent males' sexual maturation, they

obtained nude photographs of the boys. In addition to the photographs,

the investigators described each boy's external sexual development.

By examining the photographs and the written descriptions,
Professor Greulich outlined the development of the boys' primary and
secondary sexual characteristics. Greulich reported on 1) the growth
and deve10pment of the boys' external genitalia, 2) the appearance and
development of the boys' head hair, facial hair, pubic hair, scrotal,
perineal, and cirumanal hair, axillary hair, and limb and trunk hair,
3) the activity of the boys' axillary sweat glands, 4) the appearance
and activity of the boys' mammary gland, and 5) the boys' voice
changes. Without providing ages, Greulich listed the appearance order
of the external sexual characteristic changes in his boys (Table 2.53).
He cautioned that the order of appearance designated only the onset
of the development of those sex characteristics, but that each sex
characteristic developed at its own rate thereafter. Greulich stated
that because of the adolescent boy's great variability in sexual

development, specialists in child growth should consider chronological

age only as a measure of how long the boys had lived and not an index

98

Table 2.53

The order of appearance of the external sexual characteristics in males
(Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

Order Description

1 Accelerated testes growth followed by accelerated penis
growth.

2 Long, downy pubic hairs appear at the base of the penis

3 Long, coarse, straight, pigmented pubic hairs appear at
the base of the penis.

4 Axillary perspiration markedly increases.

5 The downy vellus at the corners of.the upper lip becomes
longer, coarser, and darker.

6 Long, coarse, downy hairs appear on proximal third of their
forearm's extensor surface and on the distal fourth of their
arm's lateral surface.

7 Coarse, moderately pigmented hair appears on the distal half
,of their legs‘ extensor surfaces and on the distal third of
the legs' flexor surfaces. ‘

8 Long, downy, hairs appear at their sideburns areas.

9 Circumanal hairs appear.

10 Moderate to dense pubic hair appear and covers the pubic
triangle region.

11 Short, fine, pigmented hairs appear in their axilla.

12 Their voice deepens.

13 When present, their subareolar masses attain their maximum
sizes.

14 Their pubic hairs spread totheir thighs' medial surfaces and
up the linea alba.

15 Their penis and testes attain their full adult dimensions.

16 A few terminal hairs appear on their chins and sideburns
areas, and, their upper lip hairs become coarses and darker.

17 A few terminal hairs appear around their areolae and on
their sternum.

18 Their head hairlines differentiate toward adult configura-
tion.

19 Adult hair quality and quantity present on their forearms,

arms, legs, and thighsL__

of the amount of progress that they had made toward their adult maturity.

Therefore, he suggested a procedure to classify adolescent boys according

to sexual development. Greulich proposed five categories that encompassed

the successive developmental stages of adolescent boys' external primary

and secondary sexual characteristics (Table 2.54). From his categories,

99

Table 2.54

Descriptions of the sexual characteristic changes during adolescence
in males (Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

u

 

 

 

 

 

 

‘_§roup Description
1 a. Early childhood penis, testes, and scrotum size and
configuration.
b. Pubic hair no better developed than abdominal hair
c. Early childhood facial hair
d. No definitive Circumanal, perineal, or axillary hair
2 a. Definitely larger testes
b. Some penis growth 7
c. Conspicuous, long, lightly pigmented, downy hairs at
the base of the penis
d. Early childhood facial hair
e. No definitive Circumanal, perineal, or axillary hair
3 a. Testes somewhat larger
b. Penis length has increased
c. Pubic hairs have increased in length, coarseness, and
amount, and, occasional long, coarse, straight
pigmented hair intersperse the pubic region
d. Half of the boys have short, lightly pigmented, upper
lip hair and some similar sideburn hair
e. A few boys had some short, coarse, pigmented circum-
anal hair
f. A few boys had some short, fine, lightly pigmented
axillary hair
4 a. Testes somewhat to considerably larger
b. Penis diameter has increased
c. Adult type pubic hair covers the pubic triangle
d. Small to moderate amount of short, lightly pigmented
coarse upper lip hairs, and a similar amount of long,
fine, unpigmented cheek hair ‘
e. A few moderately long, coarse, pigmented, chin hair
f. No throat or sub-mandibular hair
g. Well-differentiated Circumanal hair
h. Axillary hair is less dense and undifferentiated than
adult axillary hair
5 a. Adult proportion testes
b. Adult configuration and proportion penis, may appear
disproportionately large
c. Adult quality and quantity pubic hair and extends
toward the thighs and linea alba
d. Adolescent mustache
e. Relatively long, moderately coarse, lightly pigmented
sideburn hairs
f. Moderate growth of chin hair

Well-developed Circumanal and axillary hair

100

researchers could describe the adolescent boys' sexual development status.

Greulich cited research on both animals and humans that indi-
cated skeletal maturation and sexual maturation related closely to
each other. The research clearly established that normal skeletal
maturation did not progress without adequately functioning gonads.

As evidence, he reported that hypogonadal or castrated adolescent

males displayed delayed epiphysial fusion. Greulich compared adolescent
boys' skeletal maturation with their primary and secondary sexual
characteristics development. He x-rayed the hand-wrists and elbows

of the 476 private school boys in the sample. Using Todd's 1937
Maturation Atlas, Brush Foundation worker Clara Buehl assessed the
boys' skeletal age (Table 2.55). Because he had insufficient sample
sizes in Maturity Groups 1 and 2, the mean skeletal age values fluc-
tuated erratically. However, beginning with Maturity Group 3, the

mean skeletal ages increased systematically. The mean skeletal age
from Maturity Group 3 to Maturity Group 3-4 increased 3.5 months,

from Maturity Group 3-4 to Maturity Group 4 5.7 months, from Maturity
Group 4 to Maturity Group 4-5 7.7 months, and from Maturity Group 4-5
to Maturity Group 5 10.7 months. Greulich noted that within a single
maturity group, the older boys tended to be more skeletally mature than
the younger boys. However, the older boys were less skeletally mature
than their chronological ages indicated. Within the same maturity
group, he noted that the younger boys skeletally matured earlier than

their chronological ages indicated.

101

Table 2.55

The number and mean skeletal age values of 476 adolescent males
(Greulich, Dorfman, Catchpole, Solomon and Culotta, 1942)

 

Chronological Age (months)

 

 

Groups 144.0 156.0 168.0 180.0 192.0 204 N x
1 §_ 4 0 0 1 0 0 5 156.4
x 154.2 xx x 165.0 xx xx
1-2 3_ 0 0 4 0 0 0 4 157.2
x xx xx 157.2 xx xx xx
2 §_ 0 1 6 1 0 0 8 166.5
x xx 164.0 167.5 163.0 xx xx
2-3 §_ 1 1 1 1 0 0 4 162.8
x 165.0 165.0 148.0 173.0 xx xx
3 §_ 0 8 21 7 0 1 37 166.8
x xx 161.8 167.0 170.8 xx 179.0
3-4 §_ 0 3 14 10 0 0 27 170.6
x xx 168.7 168.9 173.4 xx xx
4 §_ 1 7 41 35 5 0 89 176.3
x 163.0 167.3 175.3 179.1 180.4 xx
4-5 §_ 1 10 42 43 13 3 112 184.7'
x 176.0 180.4 181.1 186.6 191.1 197.7
5 §_ 1 4 45 89 46 5 190 195.4
x 181.0 180.2 189.8 196.2 199.2 209.0
TOTAL ‘g 4 38 174 187 64 9
x 171.2 169.5 178.1 188.2 196.1 201.9

 

For their determinations of the estrogenic and androgenic

levels in the urine samples from 89 boys aged 4 days to 22 years with

most of the boys aged 93 months to 206 months, they used vaginal

responses of adult spayed female mice and comb growth response of

day—old chicks, respectively. Professor Dorfman provided the andro-
genic material level (I.U.) and estrogenic material level (I.U.) by
Greulich's Maturity Groups, and estrogenic material level (I.U.) by
chronological age groups (Tables 2.56, 2.57, and 2.58).

For their determinations of the amount of pituitary gonado-

trOphic hormone in the adolescent boys' urine, they used the same

urine samples that Dorfman used for his estrogenic and androgenic

102

Table 2.56

The androgenic material level (I.U.) classified by Maturity Groups
(Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

 

Descriptive Maturity Groups
Data 1 2 3 4 5
g. 11 19 22 20 20
X 4.09 6.63 10.04 19.70 28.80
S.D. 2.50 3.36 6.36 10.25 15.30
Range 1-8 2-16 2-26 7-53 9-57

 

Table 2.57

The estrogenic material level (I.U.) classified by Maturity Groups
(Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

 

 

Descriptive Maturity Groups
Data 1 2 3 4 5
§_ 6 ll l6 17 20
X 7.00 11.36 16.69 37.76 45.85
S.D. 3.06 7.56 11.49 15.44 21.80
Range 2-10 3-30 3-45 15-75 17-95
Table 2.58

The estrogenic material level (I.U.) classified by chronological age
groups (Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

 

Descriptive Chronological Age Groups
Data 99-155 156-167 168-179 180-191 192-204
E_ 9 15 12 19 15
X 16.56 13.87 25.58 39.68 38.33
S.D. 13.52 13.03 18.13 21.66 20.96
Range 2-48 3-50 5-75 3-75 8-95

 

material study. To bio-assay for pituitary gonadotrophic hormone, they

used 21-23 day old female mice. Professor Catchpole compared gonado-
trophin classifications with chronological age groups and Greulich's

five Maturity Groups (Tables 2.59 and 2.60). Catchpole concluded that

as the adolescent males' sexual development advanced, their urine

103

Table 2.59

A comparison of gonadotrophin classifications with chronological age
(Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

 

 

‘—Chronological Gonadotrophin Classifications
Age (months) 3 3-7 7
109-144 21 O 0
145-162 8 3 0
163-180 12 8 3
181-198 10 10 10
Table 2.60

A comparison of gonadotrophin classifications with the Maturity Groups
(Greulich, Dorfman, Catchpole, Solomon, and Culotta, 1942)

 

 

 

Maturity Gonadotrophin Classifications

Groups 3 3-7 47*
l 5 0 O
2 7 1 O
3 4 3 0
4 6 7 4
5 2 6 8

 

gonadotrophin levels generally tended to increase from early child—
hood's undetectable levels to levels characteristic of adults.

To determine the chronological age range in the onset of male
pubescence, Physician William A. Schonfeld measured the penis length
and circumference, and testes volume of 1500 randomly selected males
aged 1 day to 25 years (Schonfeld, 1943a). And, from the adolescent
males within that sample, he assessed the degree of maturation in
secondary sexual characteristics. To subdivide those boys into sexual
maturation groups, Schonfeld incorporated Crampton's Pubic Hair Index
and created six sexual maturation stages (Table 2.61). In addition
to noting the development in boys' secondary sex characteristics,

Schonfeld determined the statistical range of their genitalia growth

104

Table 2.61

Schonfeld's six sexual maturation stages (Schonfeld, 1943a)

 

 

Stage Description
1 Prepubescent boys, no sexual development present
2 Active growth of the penis and testes, but no pubic hair
present
3 Crampton's Pubic Hair Index Stage 1 = prepubescence
4 Crampton's Pubic Hair Index Stage 2 = pubescence
5 Crampton's Pubic Hair Index Stage 3 = post-pubescence
6 Adult sexual maturation

 

measures. Based on data that he displayed in a graph showing the
relationship between testes volume and age, Schonfeld concluded that
male pubescence began at any age between 10 and 16 years.

Much of Schonfeld's report discussed the theory and practice
of endocrine therapy for hypogonadal boys. .Basically, his treatise
was that physicians must be certain that they have properly diagnosed
the problem before they treat boys for hypogonadism.

Without an accompanying written description, Schonfeld pro-
vided a schematic table of the stages of sexual development. In that
schematic table, Schonfeld had drawings of hairline, facial hair,
chin, larynx, breasts, axillary hair, body configuration, body hair,
pubic hair and prostate development. Also, he provided the size range
in the six sexual maturation stages for penis length, penis circum-
ference, and testes volume. By extracting the data for penis length
and circumference, and testes volume from that schematic table, the
author created Table 2.62. Inspection of this table demonstrates the
tremendous overlap in the size ranges for each of Schonfeld's stages.
For example, a boy with a penis length of 8 cm. could belong in any

of Schonfeld's first four stages.

105

Table 2.62

The size range within the six sexual maturation stages for penis length,
penis circumference, and testes volume (Schonfeld, 1943a)

Schonfeld's Six Sexual Maturation Stages

 

 

l 2 3 4 ‘ 5 6
Penis length (mm.) 3-8 4.5-9 4.5-12 8-15 9-15 10.5-18
Penis circum. (cm.) 3-5 4-6 4-8 4.5-10 6-10 6-10.5
Testes volume (cc.) .3-1.5 1.75-6 1.75-13 2-20 6-20 8-25 __

 

Without an accompanying written explanation, Schonfeld pro-
vided a table that gave the frequency of the boys in annual age incre-

ments that he rated in his six sexual maturation stages (Table 2.63).

Table 2.63

The proportion of boys at each age in the six sexual
maturation stages (Schonfeld, 1943a)

 

 

 

Age Schonfeld's Sexual Maturation Stages

(Years) 1 2 3 4 5 6
9 100.0 0.0 0.0 0.0 0.0 0.0
10 94.0 0.0 0.0 0.0 0.0 0.0
11 76.0 12.0 12.0 0.0 0.0 0.0
12 44.0 14.0 32.0 10.0 0.0 0.0
13 15.0 18.0 38.0 21 0 8.0 0.0
14 6.0 15.0 26 0 26 0 27.0 0.0
15 0.0 2.0 16.0 22 0 53.0 7.0
16 0.0 1.0 9.0 11.0 59.0 20.0
17 0.0 0.0 3.0 7.0 39.0 51.0
18 0.0 0.0 0.0 7.0 30.0 63.0
19 0.0 0.0 0.0 0.0 26.0 74.0
20-21 0.0 0.0 0.0 0.0 17.0 83.0
22-25 0.0 0.0 0.0 0.0 0.0 100.0

 

Pubescence began as early as age 10 for four percent of the boys and
by age 14 ninety-four percent of the boys had started pubescence.
Also, seven percent of the boys completed their sexual maturation by
age 15 and only fifty-one percent of the boys had completed their

sexual maturation by age 17.

106

To evaluate the variation in sexual maturation of normal boys,
William A. Schonfeld also assessed the genital status of his subjects
(Schonfeld, 1943b). Schonfeld excluded any boys with known or sus-
pected endocrinopathy or other associated abnormality in genitalia
development. As primary sexual characteristics, Schonfeld listed the
testes, their epididymides, the penis, the scrotum, and the prostate.
Schonfeld advised that when considering testes growth, in addition to
testicular volume increases, researchers should consider the relation-
ship of the testes to the epididymides. He found that while at birth
the testes and their epididymides are about equal in size, during
early pubescence because adolescent males' testes grow at a more rapid
rate, their epididymides are only about one-quarter of their testes
size, and in adults the epididymides are about one-ninth the size of
their testes. Further, during the adolescent males' rapid growth
period, epididymides bind more closely to their testes. Regarding
the relative size of the boys' right and left testicles, Schonfeld
determined that in 23 percent of the boys the size of the right
testicle exceeded the left, in 22 percent of the boys the size of the
left testicle exceeded the right, and the remaining 55 percent of the
boys had testicles equal in size. In his discussion of the scrotal
changes associated with the onset of male puberty, Schonfeld observed
that to accommodate the testes enlargement, the scrotal sac distended
at its distal end while its proximal end became relatively narrower.

In his discussion of the prostate gland changes associated
with the onset of male puberty, Schonfeld observed that as the result

of its five undifferentiated lobes enlarging, the prostate's monolobar

107

organ progressively increased in size. At puberty, he noted that

the two lateral lobes enlarged relatively greater than the other
prostate lobes, thereby creating a midline groove. The posterior
lobe capped that midline groove, the anterior lobe shrunk into insig-
nificance, and the middle lobe enlarged similar to the lateral lobes.
After the boys reached that degree of prostate differentiation,
Schonfeld reported that physicians could massage the prostate and
cause the prostate to secrete its contents. Schonfeld reported that
he had massaged such prostates, and, that he was studying the micro-
scopic constituents of those secretions.

As secondary sexual characteristics, Schonfeld listed hair
deve10pment, breast changes, and voice or larynx growth. Despite the
wide variation in the degree of hair development among individuals,
families, and races, he considered hair development an important
evaluative factor in assessing the sexual development of adolescent
males. In his discussion of pubic hair development, Schonfeld

described four stages (Table 2.64).

Table 2.64

Pubic hair stages (Schonfeld, 1943b)

 

 

Stage Description
1 Darkening of the villous hair at the base of the penis
2 Extension of the pubic hair laterally to the inguinal
region
3 Upward growth of the pubic hair to cover the mons pubis
triangle
4 Extension of the pubic hair to the umbilicus

 

Schonfeld combined the hair growth in the axillary, on the

upper lip, on the face, body, and perineal areas into one discussion.

108

He said that adolescent boys' axillary hair developed after their
pubic hair completed stage two, mustache hair developed concurrent
with axillary hair, chin and sideburn hair developed shortly after
axillary and mustache hair, and body and perineal hair developed
after pubic, axillary, mustache, chin, and sideburn hair, but well
before the facial hair. Schonfeld found forehead hairline change one
of the last secondary sexual characteristics to occur. He described
that immature boys' uninterrupted, bowlike forehead hairline changed
to the adult double lateral wedge-shaped indententation, interrupted
forehead hairline.

Regarding adolescent males' breast development, Schonfeld
reported that unilateral or bilateral subareolar nodules occurred
and receded at varying pubescent stages and the degree of adolescent
male breast deve10pment varied from small, painless, subareolar
nodules to pronounced, painful subareolar nodules that secreted a
fluid.

Regarding the larynx development of adolescent males, Schonfeld
determined that in addition to rapid growth, the boys' larynx changed
its form, and the enlarged thyroid cartilage's ventrodorsal diameters
created the adolescent males' laryngeal prominence. Concurrent with
laryngeal prominence, the vocal folds lengthen and thicken, thereby
dropping the pitch in the voice by one octave.

To determine the mean height and weight differences of adolescent
males when grouped by sexual maturity, Professor Richard W. B. Ellis
of the University of Edinburgh Department of Child Life and Health

reviewed descriptions of physical growth and primary and secondary

109

sexual characteristics that trained personnel at two residential
schools made annually on 208 boys aged 11-16 years (Ellis, 1946).
At both schools, the medical personnel recorded presence and extent
of pubic, axillary, facial, and body hair, and they described the
boys' genitalia development. Also, by paying special attention to
acne and sebaceous and apocrine sweat gland development, they assessed
the boys' skin appearance.

Without knowledge of Greulich and his associates' report,
from those primary and secondary sexual characteristics deve10pment
descriptions, Ellis created pubic hair and axillary hair sexual matura-

tion stages (Tables 2.65 and 2.66). Ellis found facial hair difficult

Table 2.65

Pubic hair stages (Ellis, 1946)

 

 

Stage Description
PO Early childhood fine body hair or slightly longer, non-
pigmented vellus present
P1 One or more coarse, pigmented hairs appear at the base or
sides of the penis
P2 Pigmented hairs extend across the root of the penis
P3 Pigmented hairs cover approximately half of the pubic
triangle
P4 Pigmented hairs cover the entire pubic triangle

 

Table 2.66

Axillary hair stages (Ellis, 1946)

 

 

Stage Description ##4
A0 Complete absence of hair

Al One or more pigmented hairs present

A2 One centimeter long, pigmented, hair present and they cover

a two centimeter diameter area
A3 Profuse axillary hair present

110

to asses, and therefore he rated facial hair as either present or
absent.

Regarding genitalia development, Ellis described the growth
of the penis, testes, and scrotum separately. First the penis
lengthened considerably, then because of corpora cavernosa develop-
ment, the penis thickened considerably. Because the testes of
adolescent males grew more rapidly than their epididymides, the rela-
tive testes-epididymides prOportion differed markedly from childhood,
and that concurrent with the spurt in testicular size, the testes
softened. Not only did the scrotum become pendulous with the enlarged
testicles, but it also changed from being wider at the proximal end
to being wider at the distal end. Whereas the testicles and scrotum
developed synchronously, their testicles and penis did not necessarily
develop in synchrony. Rather than create specific stages for those
three components of genitalia, Ellis simply rated the boys' genitals
as prepubescent, pubescent, and adolescent.

Because he found that observations of facial acne failed to
develop sequentially, Ellis did not use the skin observations as
criteria for sexual maturation. He considered the observations on
the development of the apocrine and sebaceous glands as confirmatory
and not diagnostic. Using those criteria of secondary sexual charac-
teristics and genitalia development as his basis, Ellis created three
classes of sexual maturation in adolescent males (Table 2.67).

To verify the reliability of his adolescent male sexual matura-
tion classes, Ellis rated the 132 boys for a second time one hundred

days later. Because in the second rating he did not place any boys

111

Table 2.67

Three sexual maturation classes for adolescent males (Ellis, 1946)

 

 

Class Description

Prepubescent No deve10pment present

Pubescent Pubic hair present and/or pubescent genital development
present

Adolescent Profuse pubic hair present and considerable growth in

ppenis length and girth and testes volume

in a class lower than he had in the first rating, Ellis regarded his
classification procedure to be reliable. Ellis gave the proportion
of the boys aged 11 to 16 that he rated as prepubescent, pubescent,

and adolescent (Table 2.68).

Table 2.68

The proportions of the boys at each age in the three sexual
maturation classes (Ellis, 1946)

 

 

 

Age Ellis' Sexual Maturation Classes
(Years) N Prepgbescent Pubescent Adolescent -
11 to 12 26 88.5 11.5 0.0
12 to 13 39 64.1 35.9 0.0
13 to 14 73 42.5 46.6 11.0
14 to 15 44 13.6 25.0 61.4
15 to 16 26 0.0 30.8 69.2

 

Although he used cross-sectional data, Ellis concluded that
pubic hair appeared when genital development advanced to pubescent
and that axillary hair appeared when genital development advanced to
adolescent. From Ellis' discussion regarding the extent of genitalia,
pubic hair and axillary hair development of boys that he rated as
pubescent and adolescent, the author created Table 2.69.

Beginning when the boys were nine years old, for twelve suc-

cessive years Fels Research Institute for the Study of Human Development

112

Table 2.69

The proportion of the boys that Ellis rated as pubescent and
adolescent with various genitalia, pubic hair, and
axillary hair ratings (Ellis, 1946)

 

 

 

Ellis' Percent Genitalia Pubic Hair Axillary Hair
Classification Rating Rating Rating
10% Prepubescent P1 to P3 A0
23% Pubescent P0 A0
Pubescent 52% Pubescent P1 to P3 A0
12% Pubescent P1 to P3 Al or A2
3% Prepubescent P0 A3
48%* Adolescent P2 to P5 AO
Adolescent 48%* Adolescent P2 to P5 Al or A2
4%* Adolescent P1 to P3 A3

 

(*Those percentages are estimates)

workers semi-annually photographed, measured, X-rayed, and wrote
sexual characteristic development descriptions on 59 boys (Reynolds
and Wines, 1948). Professor Earle L. Reynolds and Janet V. Wines

constructed a five-point genitalia maturation scale (Table 2.70).

Table 2.70

Five-point genitalia maturation scale (Reynolds and Wines, 1948)

 

 

Rating Description
1 Infantile development
2 Scrotum enlarges, reddens, and changes texture
3 Penis enlarges and sculptures
4 Pronounced penis sculpturing and darkening
5 Adult reddish brown color, with lose penile skin and loss

of sharp sculpturing

They provided photographs that coincided with each of their five
genitalia stages. Reynolds and Wines calculated the prOportion of
the boys that they rated in each of their five genitalia stages

(Table 2.71). The average boy required less than six years to complete

113

Table 2.71

The proportion of the boys at each age in the five genitalia
stages (Reynolds and Wines, 1948)

 

 

 

Ages Reynolds-Wines Genitalia Maturation Stages
(Years) N T l 2 3 4 5
9.0 59 100.0 0.0 0.0 0.0 0.0
9.5 59 97.0 3.0 0.0 0.0 0.0
10.0 57 95.0 5.0 0.0 0.0 0.0
10.5 52 83.0 17.0 0.0 0.0 0.0
11.0 49 65.0 33.0 0.0 0.0 0.0
11.5 47 36.0 55.0 2.0 4.0 0.0
12.0 43 23.0 63.0 4.0 5.0 0.0
12.5 43 9.0 42.0 9.0 7.0 0.0
13.0 39 3.0 18.0 42.0 38.0 0.0
13.5 39 0.0 13.0 41.0 64.0 0.0
14.0 37 0.0 8.0 23.0 87.0 0.0
14.5 35 0.0 3.0 5.0 94.0 0.0
15.0 34 0.0 0.0 3.0 97 0.0
15.5 31 0.0 0.0 3.0 97 3.0
16.0 27 0.0 0.0 0.0 96 4.0
16.5 26 0.0 0.0 0.0 88 12.0
17.0 22 0.0 0.0 0.0 77 23.0
17.5 16 0.0 0.0 0.0 56 44.0
18.0 16 0.0 0.0 0.0 25 75.0

 

his genitalia maturation. From those proportions, the author con-
structed Figure 2.1.

Apart from their genitalia ratings, Reynolds and Wines rated
the boys' penis size in stages from one through seven. They provided
photographs and gave the prOportion of the boys that they rated in
each of their seven penis size stages (Table 2.72).

Reynolds and Wines provided five point pubic hair stages
(Table 2.73). After rating the pubic hair development of their sample,
they calculated the proportion of those boys that they rated in each
of their five pubic hair stages (Table 2.74), and the author used

those proportions to construct Figure 2.2.

114

mam

AmSuCOEv m0¢ AflUHUOAOZOZEU
OHN VON mmH NaH wma omH vhH mod NwH

P n I b p n I

no
W
A,
a.
no

6‘1““6 3

Amvma .mmaaz 6:6 mwaocswmo

mcoauuomoum mommum «Hamuficmo pw>uomno

H.N oughflm \\\\
.\

me omH vva wMH NMH

6 53b

53““6 \

51 2?

“on
.00
r:
vow

 

SNOIIHOdOHd

115

Table 2.72

The proportion of the boys at each age in the seven penis
size stages (Reynolds and Wines, 1948)

 

 

 

 

 

 

FfiAgeS Reynolds-Wines Penis Size Stages
(Years) N 1 2 3 4 5 6 7
10.0 58 ‘“ 28.0 53.0 14.0 5.0 0.0 0.0 0.0
10.5 54 26.0 54.0 15.0 5.0 0.0 0.0 0.0
11.0 50 24.0 54.0 16.0 6.0 0.0 0.0 0.0
11.5 48 25.0 48.0 17.0 8.0 2.0 0.0 0.0
12.0 43 16.0 56.0 16.0 7.0 0.0 5.0 0.0
12.5 42 12.0 36.0 40.0 5.0 2.0 2.0 2.0
13.0 39 8.0 10.0 36.0 26.0 15.0 3.0 3.0
13.5 39 3.0 13.0 8.0 36.0 20.0 15.0 5.0
14.0 37 3.0 8.0 5.0 11.0 35.0 22.0 16.0
14.5 33 0.0 3.0 3.0 18.0 27.0 33.0 15.0
15.0 31 0.0 0.0 3.0 16.0 23.0 42.0 16.0
15.5 30 0.0 0.0 0.0 20.0 20.0 40.0 20.0
16.0 ‘25 0.0 0.0 0.0 16.0 20.0 48.0 16.0
16.5 26 0.0 0.0 0.0 15.0 23.0 42.0 19.0
17.0 22 0.0 0.0 0.0 14.0 36.0 41.0 9.0
17.5 17 0.0 0.0 0.0 18.0 35.0 41.0 6.0
18.0 15 0.0 0.0 0.0 27.0 40.0 27.0 7.0
Table 2.73
Pubic hair stages (Reynolds and Wines, 1948)

Stage Description

1 No hair _1

2 Pigmented, straight, sparse PUbiC hair appears at the base of

the penis
3 Pubic hair has curled slightly, spread slightly, and darkened
4 Pubic hair has curled, increased moderately and spread

moderately, but not to the thighs
__¥5 Adult pubic hair covers the pubic triangle

116

Table 2.74

The prOportion of the boys at each age in the five pubic
hair stages (Reynolds and Wines, 1948)

 

Reynolds-Wines Pubic Hair Maturation Stages

 

 

Ages N l 2 3 4 5
8.5 59 100.0 0.0 0.0 0.0 0.0
9.0 59 98.0 2.0 0.0 0.0 0.0
9.5 59 98.0 2.0 0.0 0.0 0.0
10.0 59 98.0 2.0 0.0 0.0 0.0
10.5 54 91.0 9.0 0.0 0.0 0.0
11.0 50 86.0 14.0 0.0 0.0 0.0
11.5 48 69.0 29.0 2.0 0.0 0.0
12.0 45 60.0 31.0 7.0 2.0 0.0
12.5 42 48.0 43.0 7.0 2.0 0.0
13.0 40 28.0 38.0 20.0 12.0 2.0
13.5 39 10.0 33.0 20.0 33.0 3.0
14.0 37 3.0 14.0 22.0 59.0 3.0
14.5 34 3.0 3.0 9.0 74.0 12.0
15.0 33 3.0 0.0 3.0 76.0 18.0
15.5 30 0.0 0.0 0.0 73.0 27.0
16.0 25 0.0 0.0 0.0 56.0 44.0
16.5 26 0.0 0.0 0.0 46.0 54.0
17.0 22 0.0 0.0 0.0 32.0 68.0
17.5 16 0.0 0.0 0.0 12.0 88.0
18.0 15 0.0 0.0 0.0 0.0 4100.0

 

Reynolds and Wines noted that after the first appearance of
pubic hair, the boys reached the adult stage within four years. By
combining the findings on genitalia and pubic hair development, they
determined the most common sequence of appearance in genitalia and
pubic hair stages (Table 2.75). Although this order occurred most
frequently, nevertheless forty—one percent of the boys reached
Genitalia Stage 3 prior to Pubic Hair Stage 2, thirty-two percent
reached Genitalia Stage 4 prior to Pubic Hair Stage 3, and seventy
percent of their boys reached Genitalia Stage 2 prior to reaching
Pubic Hair Stage 2. However, all boys in their sample reached Pubic

Hair Stage 5 prior to reaching Genitalia Stage 5. Therefore, they

117

mam

OHN

vow

mmH

(-

vama .mmcaS paw mwaockmmv mcofluuomoum Ham: Danna uw>uowno N.N musmwm

NmH

I

AmAHCOE. mum A¢UHwoqozom=U

mma

P

omH

whH

Goa «ma

I’ 99m:

omH

omH

vva

t

wMH

NMH

b-

O
H

I

“om
.oo
fies
vow

 

SNOILHOdOHd

118

Table 2.75

The most common order of appearance of the genitalia and
pubic hair stages (Reynolds and Wines, 1948)

 

Order Genitalia or Pubic Hair Stage

Genitalia Stage 2

Pubic Hair Stage 2

Genitalia Stage 3

Pubic Hair Stage 3 and Genitalia Stage 3
Pubic Hair Stage 4

Pubic Hair Stage 5

Genitalia Stage 5

\JO‘m-waH

concluded that the genitalia maturation of adolescent males started
earlier, but finished later than pubic hair maturation.

To provide a picture of the sequence of changes associated
with puberty, and to examine the variability in the onset and duration
of each sequential phase, Hamilton Hogben, J. A. H. Waterhouse, and
Lancelot Hogben examined 642 England secondary school boys from the
Borough of Tottenham (Hogben, Waterhouse, and Hogben, 1948). To
estimate the boys' sexual development, they assessed their pubic,
axillary, chin and mustache hair growth, pubic hair contour, testes
position, voice quality, seminal activity and urine albumin content.
They recorded the development of pubic, axillary, chin, and mustache
hair development as: 1) absent, 2) short, or 3) copious, and pubic
hair contour as: 1) apical, 2) intermediate, or 3) horizontal. For
testes position or inguinal canal closure, if the boys failed to
voluntarily retract the testicles into the inguinal canal, then the
examiner gently pushed the testicles and recorded whether or not they
entered the inguinal canals. For voice quality, the recorded response

was: 1) unbroken, 2) breaking, or 3) broken. For seminal activity,

119

they recorded either: 1) already or 2) not yet. And for urine albumin
content the recorded response was: 1) with albumin or 2) without
albumin. They determined the ages at which fifty percent of the boys
still displayed childhood development, and the ages at which fifty

percent of the boys displayed adult development (Table 2.76).

Table 2.76

The mean age value for six sexual characteristics when 50 percent
of the boys displayed childhood and 50 percent displayed adult
characteristics (Hogben, Waterhouse, and Hogben, 1948)

 

 

Criteria 50% Childhood (years) 50% Adult (years)
Pubic Hair 12.9 15.0
Inguinal Canal 13.5 13.6
Voice 13.6 14.5
Axillary Hair 14.8 16.0
Mustache 15.5 unavailable
Chin Hair 16.0 unavailable

 

To describe a sexual maturation assessment procedure for
adolescent males suitable for cross-sectional studies, University of
North Carolina Professors A. Hughes Bryan and B. G. Greenberg asked a
school physician to use Ellis' pubic hair scale and axillary hair
scale and Greulich's genitalia scale (Bryan and Greenberg, 1952).
That physician easily followed Ellis' pubic and axillary hair scales;
however, he had difficulty distinguishing between the stages in
Greulich's genitalia scale. Therefore, after a year's experience,
the physician adopted his own three—point genitalia classifications.
Based on arbitrary pubescent norms of 5.5 centimeters for penis
length and 1.5 centimeters for penis width, the physician recorded the

boys as: l) immature, 2) pubescent, or 3) adolescent.‘ To compare the

120

boys' penis sizes with the arbitrary pubescent penis size, the
physician used a penis model.

In their study of the derivation and interrelationship of
adolescent males' physiological maturity indices, University of
California Institute of Child Welfare Professors Arline C. Nicolson
and Charles Hanley streamlined Greulich's five sexual maturation

groups (Table 2.77).

Table 2.77

Five sexual maturation groups streamlined from Greulich's
Maturity Groups (Nicolson and Hanley, 1953)

 

 

Stage Description . -

1 Childhood penis, testes, and scrotum development

2 Penis and testes enlarge, and, lightly pigmented, downy
pubic hairs appear .

3 Penis lengthens, and, straight, coarse, pigmented pubic
hairs intersperce with stage two pubic hair

4 Penis thickens, testes enlarges, and curly, pigmented, adult-
like, pubic hairs cover most of the pubic triangle

5 'Adult penis and testes, and, pubic hairs fill the pubic

triangle and more

Beginning when the boys were 21 months old, the University of
California Guidance Study staff annually measured 92 Berkeley,
California, boys for six years, then semi-annually measured them for
another ten years. Nicolson and Hanley provided the percent of their
cases that reached each stage at annual age increments (Table 2.78),
and from Table 2.78 the author constructed Figure 2.3.

Because he found that detailed descriptions of human bodily
changes during adolescence were either widely scattered through medical
journals or not described from a biological vieWpoint, St. Thomas'

Hospital Medical School Senior Physiology Lecturer James M. Tanner from

121

Table 2.78

The proportions of the boys at each age in the five sexual
maturation stages (Nicolson and Hanley, 1953)

 

 

 

3995 Greulich-Streamlined Sexual Maturation Stages
(Years) (1) 2 3 4 5

9.0 (99.0) 1.0 0.0 0.0 0.0

9.5 (98.0) 2.0 0.0 0.0 0.0

10.0 (92.0) 8.0 0.0 0.0 0.0
10.5 (86.0) 12.0 2.0 0.0 0.0
11.0 (72.0) 23.0 3.0 2.0 0.0

11.5 (70.0) 18.0 10.0 2.0 0.0
12.0 (58.0) 18.0 20.0 5.0 0.0

12.5 (59.0) 9.0 16.0 15.0 1.0

13.0 (50.0) 3.0 22.0 23 0 1.0 (2.0)
13.5 (49.0) 3.0 16.0 31.0 9.0 (11.0)
14.0 (40.0) 2.0 9.0 16.0 22.0 (33.0)
14.5 (39.0) 2.0 4.0 10.0 12.0 (45.0)
15.0 (20.0) 0.0 3.0 7.0 25.0 (70.0)
15.5 (10.0) 0.0 3.0 3.0 14.0 (84.0)
16.0 (9.0) 0.0 0.0 2.0 5.0 (89.0)
16.5 (0.0) 0.0 0.0 3.0 6.0 (97.0
17.0 (0.0) 0.0 0.0 0.0 2.0 (100.0)
17.5 (0.0) 0.0 0.0 0.0 1.0 (100.0)
18.0 (0.0? 0.0 0.0 0.0 0.0 (100.0)

 

the University of London wrote Growth at Adolescence (Tanner, 1955).
In his second chapter, Tanner discussed the development of the
adolescent males' reproductive system. By combining the 1951 Stolz
and Stolz text, the 1951 Reynolds and Wines report, and the 1953
Nicolson and Hanley report, Tanner created his sequence and timing
figures and discussions on genitalia and pubic hair development. For
his index of genitalia stages, Tanner compiled five stages and pro-
vided a pictorial guide for each of those stages (Table 2.79).
Because Tanner did not explain that he observed adolescent
males to create those five genitalia stages, the author assumes that
he compiled his genitalia stages from the references that he cited.

For his index of pubic hair, Tanner compiled five stages and provided

122

Ammma S3c§bmfsemaonofizv mcofiuomoum mmumum cowumusuma Hmnxom omzowno m.~ gunman

Amspcosv mom gaoHooaozommo
cam eon mmH Nod owH omH «6H mod «ma mma omH wva mma «ma

I I I F I (P I I I P n I I I
I

~.m “on

3k.» . 8
.os
.06

 

SNOICLHOdOH d

123

Table 2.79

Genitalia development stages for adolescent males (Tanner, 1955)

 

 

Stage Description

1 Early childhood testes, scrotum, and penis size and pro-
portion

2 Testes enlarged. Scrotum reddens and changes texture. Little
or no penis enlargement.

3 Penis lengthens. Testes and scrotum grew more.

4 Penis widened and its glans developed. Testes and scrotum
further enlarged. Scrotum darkened.

5 Adult size and shape genitalia.

 

a pictorial guide for each of those stages (Table 2.80). Again, the

author assumes that Tanner compiled those pubic hair stages from

indices that others created. Tanner did not provide indices for

axillary hair or facial hair.

Table 2.80

Pubic hair stages for adolescent males (Tanner, 1955)

 

 

_pStage Description

1 The pubes vellus is identical to the abdominal wall vellus,
or, no pubic hair development

2 A sparse growth of long, slightly pigmented, straight or
slightly curled, downy, pubic hairs appeared at the base
of the penis

3 Considerably darker, coarse, curled pubic hairs spread over
the pubic junction

4 Adult-type pubic hairs cover only the pubic triangle

5 Adult-type and guantity pubic hair

 

Using the data Phillips Academy Physician J. Roswell Gallagher

helped collect for W. W. Greulich's classic 1942 report, Harvard

Medical School Pediatrics Department Professor Edward E. Hunt, Boston's

Forsythe Children's Dental Infirmary worker Grace Cooke, and Harvard

University Anthropology Professor J. Roswell Gallagher compared various

124

researchers' data on sexual maturation (Hunt, Cooke, and Gallagher,
1958). From Greulich's five Maturity Groups and four intermediate
groupings, they classified boys into nine sexual maturation groups.
To create their sexual maturation stages, Hunt, Cooke, and Gallagher
streamlined Greulich's five Maturity Groups.

By statistically analyzing Reynolds' and Wines' data, Tanner's
data, and their own data, Hunt et a1. tested the reliability of
Greulich's five Maturity Groups. They concluded that on a cross-
sectional basis, observers judged pubic hair initiation more accurately
than they judged the advent of testes growth. They hypothesized that
the difficulty that the observers had in recognizing the advent of
testes growth resulted from the absence of comparative photographs
when the boys were pre—pubescent. When the observers assessed
Greulich's Maturity Group three, they found satisfactory agreement.
Because 1) Greulich's fifth stage does not describe sexual maturation
and 2) the phenomenon of penis shrinkage in late adolescence confused
some observers, they concluded that on a cross-sectional basis Greulich's
Maturity Group five was very difficult to assess.

To study the future growth of short-statured 14 year old males,
Harvard Medical School Pediatrics Professors J. Roswell Gallagher and
Carl C. Seltzer X-rayed the hand-wrist area and used Greulich's
Maturity Groups to assess the sexual maturation of 101 boys aged 14
years old annually for three years (Gallagher and Seltzer, 1961). They
used the Bayley-Pinneau skeletal maturation tables for predicting the

boys' adult height, and compared their sexual maturation to their pre-

dicted adult height.

125

They determined that 14 year old boys less than 63 inches tall
differed in rate of adolescent growth from 14 year old boys more than
63 inches tall. For the nineteen 14 year old boys less than 63 inches
tall, they found an average height gain of six and one-half inches,
whereas for the eighty-two 14 year old boys more than 63 inches tall,
they found an average height gain of only four inches. Of those
nineteen 14 year old boys less than 63 inches tall, the seven rated
at Maturity Group two averaged 7.1 inches in actual height gain, with
a range of 6.0 to 8.3 inches, the seven rated at Maturity Group
three averaged 6.8 inches in actual height gain, with a range of 5.7
to 8.3 inches, the four rated at Maturity Group four averaged 5.5
inches in actual height gain, with a range of 5.1 to 5.9 inches, and
the one rated at Maturity Group five grew only 3.2 inches more. When
physicians 1) cannot obtain a hand—wrist X-ray, 2) do not know how to
read the hand-wrist X—ray, or 3) do not have the time to read the
X-ray and work with the Bayley-Pinneau tables for predicting adult
height, they advised that Greulich's Maturity Groups may prove helpful
in predicting the future height gain potential of adolescent males.

As one of the several measurements that he advised physicians
to take on their adolescent male patients, Physician Frank Falkner
included a test for pubertal status (Falkner, 1961). In the reference
section of his paper, Falkner cited Tanner's Growth at Adolescence.
However, within the paper, Falkner did not credit anyone for the crea-
tion of his genital and pubic hair development stages (Tables 2.81 and
2.82). By combining the boys' genital and pubic hair development

ratings, Falkner created his "Pubertal Status" score. Therefore, if

126

Table 2.81

Genital deve10pment stages for adolescent males (Falkner, 1961)

 

 

Stage Description
1 Childhood penis and testes size
2 Testes enlarge and scrotal skin reddens and coarsens
3 Penis lengthens
4 Penis general size enlargens and scrotal skin darkens
pp 5 Adultpgenitalia

 

Table 2.82

Pubic hair development stages for adolescent males (Falkner, 1961)

 

 

Stage Description
1 No true pubic hair
2 sparse, downy pubic hair at the base of the penis
3 Pigmented, coarse, curly, pubic hair increases in amount
4 Adult pubic hair limited to the pubic triangle
5 Adult pubic hair spread to the thighs
6 Adult pubic hair spread up to the linea alba

 

he rated a boy's genital stage as 3 and his pubic hair stage as 2,
then he recorded the boy's pubertal status as 3:2.

To determine whether or not adolescent boys with advanced
sexual maturation had an academic advantage over adolescent boys with
delayed sexual maturation, J. W. B. Douglas and J. M. Ross reviewed
1,872 school medical examinations of 15 year old boys wherein the
school physician recorded the boys' genitalia maturation, pubic hair
development, axillary hair development, and vocal quality (Douglas
and Ross, 1964). Based upon those primary and secondary sexual char-
acteristics, Douglas and Ross created four classifications of puberty
(Table 2.83). They reported the number of the 1,872 fifteen year old

males that they rated in each of their four classifications of

127

Table 2.83

Four classifications of puberty for adolescent males
(Douglas and Ross, 1964)

 

 

Classification Description

Infantile Infantile genitalia, no pubic hair, no axillary
hair, and an unbroken voice

Early Signs Early adolescent genitalia, some pubic hair, and
some axillary hair

Advanced More maturation than the Early Signs category,
but not adult maturation

Mature Genitalia completely developed: penis and glans

substantially enlarged, testes enlarged, scrotum
pendulous and rugose, pubic hair profuse, axillary
hair profuse, and. a broken voice

puberty (Table 2.84). They did not explain why they placed 177 boys

in an unknown puberty classification.

Table 2.84

The proportion of the boys in the four classifications
of puberty (Douglas and Ross, 1964)

 

 

Classification Actual Number Proportion
Infantile 177 9.5
Early Signs 597 31.9
Advanced 504 26.9
Mature 417 22.3

_p Unknown 177 9.5

 

To determine the sexual maturation process of Hong Kong
Chinese adolescent boys, University of Hong Kong Anatomist K. 8.
Frances Chang, Physician Marjorie M. M. C. Lee, and S. J. Chan assessed
the penis growth, pubic hair development, axillary hair deve10pment,
and vocal quality of 3,658 Southern Chinese boys aged 6 to 18 years
old (Chang, Lee, and Chan, 1966). They created five genital develop-

ment stages (Table 2.85) and five pubic hair development stages

128

Table 2.85

Five genitalia deve10pment stages for adolescent males
(Chang, Lee, and Chan, 1966)

¥

 

 

Stages Descriptions
1 Childhood size penis
2 Penis a little larger than childhood size
3 Penis has lengthened and widened
4 Penis has enlarged further and its skin has darkened
5 Adult size and shape penis

 

(Table 2.86). They determined the age of the boys at the point where
fifty percent of the boys achieved each genitalia and pubic hair
stage. As a result of their discussion, the author created Table

2.87, which shows the ages at which 50 percent of the boys attained

Table 2.86

Five pubic hair stages for adolescent males
(Chang , Lee, and Chan, 1966)

 

 

._Stages Description

1 No pubic hair present

2 Sparse, straight or slightly curled, downy pubic hair
present only at the base of the penis

3 Numerous, dark, coarse, and curly pubic hairs spread
slightly beyond the base of the penis

4 Adult-type pubic hairs present, but, limited in distribution

5 Adult-type pubic hairs spread to the thighs, up the linea

alba, and down to the scrotum

Table 2.87

The ages at which 50 percent of the boys attained the genitalia
and pubic hair stages (Chang , Lee, and Chan, 1966)

 

 

Genital Stages 50% Age Pubic Hair Stages 50% Age
2 13.19 2 13.31
3 13.75 3 13.70
4 14.60 4 14.60
5 16.47 5 16.47

 

129

the genitalia and pubic hair stages. They concluded that the ages

of onset and duration of genital and pubic hair stages for their
subjects were about the same. By calling stage 1 non-pubescent, by
combining stages 2, 3, and 4 and calling them pubescent, and by

calling stage 5 post-pubescent, they constructed graphs to show the
relationship between those three sexual maturation phases in adolescent
males. 'From the graphs of Chang, Lee, and Chan, the author estimated
the prOportion of boys in each phase from ages 11 through 18 for
genitalia and pubic hair (Tables 2.88 and 2.89). Figures 2.4 and

2.5 graphically illustrate Tables 2.88 and 2.89.

Table 2.88

The proportion of the boys at each age in the three
genitalia phases (Chang, Lee, and Chan, 1966)

 

 

 

Age Non-Pubescent Pubescent Post-Pubescent
711.0 99.0 1.0 0.0

12.0 88.0 12.0 0.0

13.0 57.0 42.0 1.0

14.0 13.0 82.0 5.0

15.0 3.0 85.0 13.0

16.0 1.0 63.0 36.0

17.0 -0.0 33.0 67.0

18.0 0.0 13.0 87.0

Table 2.89

The proportion of boys at each age in the three pubic
hair phases (Chang, Lee, and Chan, 1966)

 

 

Age Non-pubescent Pubescent Post-pubescent
11.0 99.0 1.0 0.0

12.0 90.0 10.0 0.0"
13.0 60.0 39.0 1.0

14.0 17.0 80.0 3.0

15.0 4.0 83.0 13.0

16.0 1.0 60.0 39.0

17.0 0.0 32.0 68.0

18.0 0.0 16.0 84.0

 

130

mam

.OHN

b

VON

”mama .gmsu cam .moq .mcmsov mCOAuHomoum mmmnm mwamuwcom wo>wmmno v.m musmfim

AmquOEv mum AdUHUOAozommo
mmH NmH mmH omH vha mod mod mmH omH vva mma «MA

I I I I I I I I P P I

$4976
d

"om
.00
.0h
.0m

 

 

SNOILHOdOHd

131

wan

OHN

vom

I

Aomma .cmno Ugo .moA .mcmnov mcofiuuomowm “Hun chasm Um>uomno m.N gunman

Am£H:OEv Hut AdUHooqozommu
mmH NmH 0mH omH vhH med Nwa omH omH gvd QMH NMH

(I I I I I I I P I I

Mans-59%

.OM

A“cm
.om
.05
tom

 

 

SNOIIHOdOHd

132

Chang et al. found that axillary hair appeared when genitalia
and pubic hair reached their fourth stages. They found it difficult
to determine vocal changes from cross-sectional sampling, but when
they did assess vocal change they determined that those changes
occurred when genitalia and pubic hair reached their third stage.

Relying upon the pioneer works of Greulich et a1., Ellis, and
Reynolds and Wines to classify the sexual maturation of adolescent
boys, Michigan State University Masters degree candidate Michael G.
Marshall created five checklists for the primary and secondary
sexual characteristics of facial hair development, axillary hair
development, pubic hair development, penis development, and scrotum

development (Tables 2.90 through 2.94) (Marshall, 1967).

Table 2.90

Facial hair development checklist (Marshall, 1967)

 

 

 

Stage Description (Ask the boys if they shaved todav)
1 No facial hairs on mustache, sideburns, or, chin
2 Unpigmented sideburn hair present
3 Downy, lightly pigmented hairs present at the corners of
the upper lip
4 Lightly pigmented, coarse, mustache hairs present
5 Adult mustache hairs present
Table 2.91

Axillary hair development checklist (Marshall, 1967)

 

 

Stage Description
‘—' 1 No axillary hairs present
2 Slight, unpigmented axillary hairs present
3 Downy, lightly pigmented axillayr hairs in one or both axilla
4 Pigmented, coarse, axillary hairs cover 8 to 1 inch area
5 Coarse, curly, well—developed axillary hairs fill both axillg

 

133

Table 2.92

Pubic hair development checklist (Marshall, 1967)

 

 

Stage Description

1 No pubic hairs present

2 Downy, unpigmented, pubic hairs present

3 Sparse, pigmented, straight, pubic hairs present

4 Curly, coarse, pigmented, pubic hairs present, but not
extensive

5 Adult-type pubic hairs extends to the medial surface of the
thighs

 

Table 2.93

Penis development checklist (Marshall, 1967)

 

 

Stage Description
1 Childhood sized penis
2‘ Slight penis growth
3 Penis shaft lengthens, with the glans proportionately small
4 Penis shaft lengthens and thickens
5 Maximum sizeppenis

 

Table 2.94

Scrotum and testes development checklist (Marshall, 1967)

 

 

__§tage Description
1 Childhood scrotum and testes
2 Scrotum and testes visibly enlarged, one testicle may show
lowered
3 One testicle definitely has lowered
4 Scrotum darkening beginning
5 Maximum size testes and maximum scrotum darkening

 

Because he wanted Junior High School Boys Physical Education
teachers to assess the students' sexual maturation on a cross-
sectional basis for each characteristic, Marshall described each stage
independent of the other stages. After inspecting the data obtained

from the 671 East Lansing Junior High School boys that the physical

134

education teachers had assessed, Marshall combined the five sexual
characteristics scores into a single Sexual Maturation Indicator
Value (SMIV). To calculate the SMIV score, he used the boys' pubic
hair development score as the primary indicator and used the facial
hair, axillary hair, penis and scrotum and testes scores to add to

or subtract from the pubic hair rating (Table 2.95). Once he obtained

Table 2.95

SMIV formula (Marshall, 1967)

 

SMIV = (Facial hair rating + Axillary hair rating + Penis develop-
ment rating + Scrotum and testes development rating) -
(4 X Pubic hair rating)/ 10 + Pubic hair rating

#7

the boys' SMIV score, he subdivided the SMIV scores into five puberty
groups (Table 2.96). Marshall reported the number of subjects, their
average age, and the standard deviation for each of his puberty groups

(Table 2.97).

Table 2.96

Five puberty groups by the SMIV breakdown (Marshall, 1967)

 

 

Puberty Group SMIV Breakdown
A 1.0 to 1.5
B 1.6 to 2.5
C 2.6 to 3.5
D 3.6 to 4.5
__7 E 4.5 to 5.0

 

To determine the extent to which the pubertal ratings of
adolescent males were associated with each primary and secondary sexual
characteristics and to determine their between and within rater relia-

bilities, Harvard School of Public Health Professor H. Bontourline

135

Table 2.97

The number, mean values and standard deviations of the boys
in the five puberty groups (Marshall, 1967)

 

 

 

Descriptive Marshall's Pubertinroups
Data A B C D E
Number 118 209 206 132 11
Mean(months) 154.17 156.61 161.20 165.21 172.09
__§tandard Deviation 7.16 6.66 6.90 7.71 6.95

 

Young, Stanford University Anatomist W. W. Greulich, Harvard Medical
School Professors J. R. Gallagher and T. Cone, and Washington, D.C.,
Children's Hospital Physician F. Heald assessed eleven primary and
secondary sexual characteristics of 70 Boston area adolescent males
three separate times at two month intervals (Young, Greulich,
Gallagher, and Heald, 1968). Young selected those 70 boys to repre-
sent a rectangular distribution of sexual development rather than a
normal bell-shaped distribution. Using the descriptions developed
by Greulich (1942), Schonfeld (1943), and Stolz and Stolz (1951),
they created six classifications of pubertal maturity (Table 2.98).

To assess the eleven primary and secondary sexual characteris-
tics necessary to classify the boys in their six pubertal maturity
stages, Young et al. used Tanner's genitalia pictorial guide and
several checklists. Tables 2.99, 2.100 and 2.101 contain their pubic
hair, axillary hair, and facial hair checklists. On their checklists,
they rated the boys at plus, minus, or right at each stage. Thereby,
they had sixteen ratings for each checklist.

In addition to their pubic hair checklist, they categorized
the pubic hair configuration as: 1) concave, 2) straight, and 3) convex.

In addition to their facial hair checklist, they asked the boys if they

136

Table 2.98

Six classifications of pubertal maturity for adolescent
males (Young, Greulich, Gallagher, and Heald, 1968)

 

Classification

 

 

 

 

 

 

Description
1 Testicular volume 1 _ H
(Childhood) No secondary sexual characteristics present
Testicular volume 2
a - Pubic hair 1 1
2 Axillary and body hair 0
(Prepuberal) Slight increase in penis length and diameter
Eyebrows, line of cheek, hairline, and voice 1
Testicular volume 3 or 4
Pubic hair 2
3 Axillary hair 0
(Puberty lst. Penis definitely enlarged
Stage) Facial and bodv hair 1 and 2
Testicular volume 4
4 Pubic hair 3 or 4
(Puberty 2nd Axillary hair 2 or 3
Stage) Penis further enlarged
Testicular volume 5
5 Pubic hair 3 or 4
(Puberty 3rd Axillary hair 2 or 3
Stage) Penis further enlarged
Testicular volume 6
6 Adult voice

(Adult form)

Pubic hair

Pubic hair 4
Axillarv hair 3 or 4

Table 2.99

checklist (Young, Greulich, Gallagher, and Heald, 1968)

 

 

Stage Description
0 No pubic hairs present
1 Downy, unpigmented, fine, straight, pubic hairs present
2 Pigmented, coarse, crinkled, pubic hair cover a small area
3 Pigmented, coarse, crinkled, pubic hairs cover a moderate
' area
4 Pigmented, coarse, crinkled, pubic hiars cover a considerable

area

re:

the

P}.

137

Table 2.100

Axillary hair checklist (Young, Greulich, Gallagher,
and Heald, 1968)

 

 

Stage Description

0 No axillary hairs present

1 Fine, straight, axillary hairs cover a small area

2 Pigmented, coarse, curly or crinkled, axillary hairs cover
a small area

3 Pigmented, coarse, curly or crinkled, axillary hairs cover
a moderate area

4 Pigmented, coarse, curly or crinkled, axillary hair cover a

considerable area

Table 2.101

Facial hair checklist (Young, Greulich, Gallagher, and Heald, 1968)

 

 

Stace Description
0 No facial hairs present
1 Downy, straight, facial hairs present
2 Some pigmented, coarse, facial hairs present
3 Pigmented, coarse, facial hairs present in moderate amounts
4 Pigmented, coarse, facial hairs present in considerable

amounts

shaved. Upon receiving an affirmative response, they asked how many
times per week the boys shaved. For their assessment of penile length
and diameter, they measured the penis with a millimeter metal tape.
Because they referred to the 1943 report by Schonfeld, the author
assumes that they measured the stretched but flaccid penis. For their
testicular volume measurement, they adopted Schonfeld's technique of
comparing the testicle with the known volume of wooden models. They
reported testicular volumes, penis length, and penis diameters that
they considered necessary to rate their boys in stages 1 through 6

(Table 2.102). Additionally, they provided checklists for rating the

138

Table 2.102

The testicular volumes, penis lengths, and penis diameters necessary

for the six puberty ratings (Young, Greulich,
Gallagher, and Heald, 1968)

 

 

Pubertal Testicular Volume Penis Length Penis Diameter
Rating (ccm) (cm) (cm)
1 1.5 1.6 1.3
2 3.0 2.1 1.6
3 6.5 2.8 2.1
4 10.5 3.2 2.5
5 15.5 3.5 2.9
6 21.5 4.0 3.2

 

breast development (Table 2.103) and the eyebrow development (Table

 

 

 

 

 

2.104).
Table 2.103
Breast development checklist for adolescent males
(Young, Greulich, Gallagher, and Heald, 1968)
Stage Description
0 No visible or palpable mammary gland
1 Slight, but definite palpable mammary gland
2 More marked mammary gland enlargement
3 Moderate to considerable mammary gland enlargement
4 Considerable bilateral mammary gland enlargement
Table 2.104
Eyebrow development checklist for adolescent males
(Young, Greulich, Gallagher, and Heald, 1968)
Stage Description
1 Infantile eyebrows present
2 Eyebrows thicken
3 Eyebrow hairs coarsen,aand, the eyebrows thicken and grow

towards the midline
4 Adult-type eyebrows present

139

By comparing the scalp hairline to four hairline diagrams
that ranged from bowlike to double apical, they rated the hairline
on a four-point scale. Similarly, by comparing the facial configura-
tion to four jawline diagrams that ranged from round—chin to V-shaped
chin jawlines, they rated the facial configuration on a four-point
scale. For their measurement of vocal quality, they categorized the
voices as: l) childish, 2) broken, 3) adult.

They reported the test-retest correlations for the five raters
on the eleven primary and secondary sexual characteristics assessments

(Table 2.105). The average test—retest correlations for facial hair,

Table 2.105

The test-retest correlations for the five observers in
this study (Young, Greulich, Gallagher, and Heald, 1968)

 

 

 

Sexual Observers
Characteristics 1 2 3 4 5 AVE
Eyebrows .80 .80 .82 .83 .81 .812
Voice .86 .85 .86 .87 .86 .870
Facial hair .86 .87 .88 .87 .85 .864
Axillary hair .91 .90 .91 .91 .89 .904
Body hair .79 .81 .80 .80 .77 .794
Pubic hair .96 .95 .96 .96 .96 .958
Hair line .77 .77 .78 .78 .77 .774
Cheek line .85 .86 .86 .87 .87 .862
Pubic hair shape .65 .64 .64 .64 .61 .636
Penis length .83 .81 .82 .84 .84 .828
Testicular volume .95 .94 .93 .97 .96 .954

 

axillary hair, pubic hair, and testicular volume of .864, .904, and
.954, respectively, suggest that specialists in child growth and
development should be able to assess those sexual characteristics
with acceptable reliability.

From those test-retest correlations, Young et a1. concluded

that testicular volume, pubic hair, and axillary hair demonstrated

140

sufficient assessment reliability for them to use as major variables
in assessing sexual development. To determine the beta weight that
each observer attributed to those three major variables, they computed
separate multiple regression analyses of each observer's assessments.
They reported each observer's standardized partial regression coef-

ficients for their selected three major variables (Table 2.106).

Table 2.106

The standardized partial regression coefficients for five observers
(Young, Greulich, Gallagher, and Heald, 1968)

 

 

 

Sexual Observers
Characteristic l 2 3 4 5 AVE
Axillary hair .282 .283 .309 .264 .180 .267
Pubic hair .449 .478 .467 .387 .403 .434
Testicular volume .295 .252 .245 .378 .434 .331

 

By multiplying the averaged standardized partial regression coefficients
by the standard deviation of the pubertal maturity ratings and then
dividing the value by the standard deviation of each major variable,
they determined a regression equation for pubertal age. The resulting
regression equation for predicting the Pubertal Age of Adolescent

Males was: (PAAM) = (.5 + .5[pubic hair] +- 4[testicular volume] +
.3[axillary hair]).

At this point they stated that to evaluate the pubertal age
regression equation, they needed a criterion separate from sexual
development that correlated highly with biological maturity. They
chose the Greulich-Pyle Atlas to assess skeletal age.

From 120 adolescent males from Palermo, Italy, Young selected

42 boys whose skeletal ages ranged from 11 to 16 years on whom they

141

had the data from which they could determine pubertal age. They
reported the correlations between the boys' chronological age,
skeletal age, first pubertal maturity rating, calculated puberty

age, and second pubertal maturity rating (Table 2.107). The author

Table 2.107

The correlations between chronological age, skeletal age, a first
pubertal maturity rating, a calculated pubertal age, and a
second pubertal maturity rating (Young, Greulich,
Gallagher, and Heald, 1968)

 

Chrono. Skeletal lst deertal Cal. Pub. 2nd.

 

Age Age Mpturitv Rat. Age P.H.T.
Chronological Age 1.00
Skeletal Age .53 1.00
lst Pubertal Mat. Rating .54 .92 1.00
Calc. Pubertal Age .52 .92 .99 1.00
2nd Pubertal Mat. Rating .53 .93 .99 .99 1.00

 

noted that the correlations between chronological age and their
estimates of pubertal age were low (.54 and less), but because they
derived their calculated pubertal age from their pubertal maturity
ratings, those correlations were high (.99). Although their specially-
selected rectangular sample partially explains their high correlation
(.92) between skeletal age and pubertal age, they concluded that their
equation for pubertal age provided a precise method for calculating

the pubertal maturity of adolescent boys.

To determine the chronological age variation at which adolescent
boys start and pass through Tanner's genitalia and pubic hair stages,
University of London's Institute of Child Health, Growth, and Develop-
ment Professors W. A. Marshall and J. M. Tanner viewed the nude, full

body photographs of 228 British boys who were subjects in the Harpenden

142

Growth Study (Marshall and Tanner, 1970). The Institute collected
those data every three months during the adolescent age period.

By sequentially viewing each boy's genitalia and pubic hair changes,
and by using Tanner's genitalia and pubic hair stages, they classi-
fied the maturational stages of the boys. They considered their
pubic hair stage 2 as an unreliable rating stage because unpigmented
hairs did not show well on the photographs. If the pubic hair had
spread up the linea alba, they assessed those boys as stage 6. They
calculated the average age and standard deviation at which the boys

reached each stage (Table 2.108). They reported the proportion of

Table 2.108

The average age and standard deviation that the boys attained the
genitalia and pubic hair stages (Marshall and Tanner, 1970)

 

 

 

Descriptive Genitalia Stages Pubic Hair Stages
Data 2 3 4 5 2* 3 4 5
Mean 11.64 12.85 13.77 14.92 13.44 13 90 14.36 15.18
S.D. 1.07 1.04 1.02 1.10 1.09 1.04 1.08 1.07

 

(*They considered this stage unreliable.)

boys in the five genitalia stages while also in each of the pubic hair
stages (Table 2.109), and conversely, the proportion of boys in the
five pubic hair stages while also in each of the genitalia stages
(Table 2.110). They noted that forty-one percent of their boys

reached genitalia stage 4 before their pubic hair development began.

If the boys' genitalia deve10pment paralleled their pubic hair develop-
ment, then the proportion of the boys would be greatest when the geni-
talia and pubic hair stages were equal. However, the cell containing

the genitalia and pubic hair stage 2 boys provided only thirteen percent

143

Table 2.109

The proportion of the boys in the five genitalia stages across
the four pubic hair stages (Marshall and Tanner, 1970)

 

 

 

Pubic Hair Tanner's Genitalia Stages
Stages 2 3 4 5
1 98.0 72.0 16.0 0.0
2 2.0 17.0 37.0 0.0
3 0.0 9.0 36.0 8.0
4 0.0 2.0 11.0 54.0
5 0.0 0.0 0.0 38.0

 

Table 2.110

The proportion of the boys in the five pubic hair stages across
the four genitalia stages (Marshall and Tanner, 1970)

 

 

 

Genitalia Tanner's Pubic Hair Stages
Stages 2 3 4 5
1 1.0 0.0 0.0 0.0
2 13.0 4.0 0.0 0.0
3 45.0 17.0 6.0 0.0
_ 4 41.0 75.0 65.0 10.0
5 0.0 4.0 29.0 90.0

 

of the sample. Likewise, the cell containing the genitalia and pubic
hair stage 3 boys provided only seventeen percent of the sample.
However, the genitalia and pubic hair stage 4 cell provided sixty-five
percent of the sample, and the cell for the genitalia and pubic hair
stage 5 boys provided ninety percent of the boys. Relative to rate
of genitalia development after reaching stage 3, the boys' rate of
pubic hair development increased.

To establish the ages at which Istanbul adolescent males
attained their secondary sexual development stages, Istanbul University
Pediatrics and Administrative Services Professors Olcay Neyzi, Hulpa

Alp, Ayfer Yalcindag, Suleyman Yakacikli, and Alper Orphon studied

A‘

Pu

.L

f
L”

Q.
t

W
O

“C
.C

43
a

ti

144

1,530 Turkish boys aged 9 to 17 years (Neyzi, Alp, Yalcindag, Yakacikli,
and Orphon, 1975). All of the boys were within three months of their
birthday. By visually inspecting the boys' penis, facial hair, axillary
hair, and pubic hair and by palpating the boys' testes and breast
tissue, Suleyman Yakacikli assessed the boys' sexual development.

After reviewing the Reynolds and Wines 1951 report, the Tanner 1962
text, and the Marshall and Tanner 1970 report, they adopted their

own rating scales for assessing pubic hair development, facial hair
deve10pment, axillary hair deve10pment, and laryngeal deve10pment of
adolescent males (Tables 2.111 through 2.114). They reported the mean
attainment age that their boys reached their four sexual characteris-

tics' development stages (Table 2.115).

Table 2.111

Pubic hair stages (Neyzi et a1., 1975)

 

 

Stage Description

PHl No pubic hair present

PH2 Sparse, straight or slightly curly, pubic hairs present at
the base of the penis

PH3 Darker, coarser, curlier, pubic hairs spread over the pubis
junction

PH4 Adult-type pubic hairs cover only the pubis, not the inner
thigh area

PHS Adult-type pubic hairs cover adult or near-adult area

 

Table 2.112

Facial hair stages (Neyzi et a1., 1975)

 

 

Stage Description

Axl No axillary hairs present

Ax2 Sparse axillary hairs present

Ax3 Moderate amount of curly axillary hairs present

Ax4 'Abundant amount of curly axillary hairs present

145

Table 2.113

Facial hair stages (Neyzi et a1., 1975)

 

 

Stages Description

FHl No facial hairs present

FHZ Slightly long, pigmented facial hairs present at the corners
of the upper lip

FH3 Complete mustache present and some facial hairs on the cheeks
and chin

FH4 Adult or near-adult facial hairs distribution and quantity
present.

 

Table 2.114

Laryngeal stages (Neyzi et a1., 1975)

 

 

Stages Description
V1 Infantile vocal quality present
V2 Some voice cracking present
V3 Noticeable male quality voice present
V4 Definite male qualipy voice present

 

Table 2.115

The mean attainment age that the boys reached the pubic hair,
axillary hair, and laryngeal stages (Neyzi et a1., 1975)

 

 

Pubic Hair Axillary Hair Facial Hair Larynx
Stages Stages Stages Stages
Means PH2 PH3 PH4 AX2 Ax3 FHZ FH3 V2 V3

 

Ages 12.28 13.81 16.07 13.55 15.78 14.72 X 13.71 16.22

In their appendix, Neyzi et al. reported the proportion of
the boys at each age that attained each stage of the four sexual
characteristics (Tables 2.116 through 2.119). In parentheses, the
author placed the proportion value. From those tables, the author
constructed Figures 2.6 through 2.9 (Neyzi, Alp, Yalcindag, Yakacikli,

and Orphon, 1975).

146

Table 2.116

The proportion of the boys at each age in the five
pubic hair stages (Neyzi et a1., 1975)

 

f

 

 

 

 

 

 

AGE Pubic Hair Stages
(Years) N PHl 9112 PH3 PH4 PHS
9 167 l60(95.8) 7(4.2) 0(0.0) 0(0.0) 0(0.0)
10 175 l60(9l.4) l4(8.0) 1(0.6) 0(0.0) 0(0.0)
11 195 l62(83.l) 30(15.4) 3(l.5) O(0.0) 0(0.0)
12 163 100(6l.3) 51(31.2) ll(6.7) 1(0.6) 0(0.0)
13 164 60(36.6) 62(37.8) 34(20.7) 8(4.9) 0(0.0)
14 162 35(21.6) 46(28.4) 63(38.9) l7(lO.5) 1(0.6)
15 162 7(4.3) 24(14.8) 81(50.0) 38(23.5) 12(7.4)
16 177 0(0.0) 14(7.9) 64(36.2) 78(44.l) 21(11.9)
17 165 0(0.0) 4(2.4) 54(32.7) 70(42.4) 37(22.4)
_' Table 2.117
The proportion of the boys at each age in the four
axillary hair stages (Neyzi et a1., 1975)
AGE Axillary Hair Stages
LYears) N AXl AXZ Ax3 Ax4
9 167 l67(100.0) O(0.0) 0(o.0) 0(0.0)
10 175 l74(99.4) 1(0.6) 0(0.0) 0(0.0)
11 195 191(97.9) 3(l.5) 1(0.5) O(0.0)
12 163 l44(88.3) 16(9.8) 3(l.8) 0(0.0)
13 164 117(7l.3) 40(23.4) 7(4.3) 0(0.0)
14 162 66(40.7) 77(47.S) 18(ll.l) 1(0.6)
15 162 20(12.3) 89(54.9) 42(25.9) 11(6.8)
16 177 2(l.l) 74(41.1) 78(44.l) 23(13.0)
17 165 O(0.0) 48(29.1) 79(47.9) 38(23.0)

 

147

Table 2.118

The proportion of the boys at each age in the four facial
hair stages (Neyzi et a1., 1975)

 

 

 

 

 

 

 

 

 

AGE Facial Hair Stages
(Years) N FHl FH2 FH3 FH4 111
9 167 167(100.0) 0(0.0) 0(0.0) O(0.0)
10 175 175(100.0) 0(0.0) 0(0.0) O(0.0)
11 192 191(99.5) 1(0.5) O(0.0) 0(0.0)
12 163 l60(98.2) 'V3(l.8) O(0.0) 0(0.0)
13 164 143(87.2) 20(12.3) 1(0.6) 0(0.0)
14 162 108(66.7) 51(31.5) 2(1.2) 1(0.6)
15 162 67(41.4) 76(46.9) 19(ll.7) 0(0.0)
16 177 41(23.2) 88(49.7) 46(26.0) 2(l.l)
17 165 13(7.9) 62(37.6) 83(50.3) 7(4.2)
Table 2.119
The proportion of the boys at each age in the four
laryngeal stages (Neyzi et a1., 1975)
AGE - Laryngeal Stages
(Years) N V1 V2 V3 V4
9 E7 16771000) 0(0.0) 0(0.0) 0(0.0)
10 175 173(98.9) 2(l.1) O(0.0) 0(0.0)
11 195 191(97.9) 4(2.l) 0(0.0) O(0.0)
12 163 147(90.2) 15(9.2) 1(0.6) 0(0.0)
13 164 l20(73.2) 40(24.4) 4(2.4) O(0.0)
14 162 80(49.4) 75(46.5) 7(4.3) 0(0.0)
15 162 34(21.0) 90(55.6) 36(22.2) 2(1.2)
16 177 10(5.6) 83(46.9) 82(46.3) 2(l.l)
17 165 2(1.2) 44(26.7) 102(61.8) l7(10.3) 44
Summary

principles of sexual development that are pertinent to this investiga-

tion:

they can determine the biological age of adolescent males, the next
best predictor of their biological age is an assessment of the boys'

sexual deve10pment, 2) the adolescent males' primary sexual

From the preceding review, the author has drawn the following

1) when researchers cannot obtain hand-wrist X-rays from which

148

Amhma .Hm no Huaozv mcofiuuomoum mwmum Ham: Danna poImeno o.~ musmflm

Amsucosv mwa gaoHooaozommo
mHN OHN voN mad «ma ems omH aha moa «ma mma oma vva QMH and

I I I P I I I I )I I I I I

vOH

"cm
.09
.on
.om

 

SNOILHOdOHd

149

OHN

OHN

b

VON

Amhma .Hm um wuhozv mg0fluwomoum woman new: mumaawxm pm>ummno n.m onsmwm

mmH

NmH

p

Amaucoso mac gaoHooaozommo
«S

D

OOH

OOH

med

 

NOH

D

mma

L

OmH

I

va

L

OMH

1.

NMH

.OH

uON

10v

vOm
'00
.Oh
yom

 

SNOILHOdOHd

150

3an .Hm um Hummzv mcowunomoum ommum “Hm: H303 cwIAomno m.N wusmfim

A mnucopnv 8d SHUQHOZOM EU
«h

00H

omH

H

OOH

NOH

 

.OH

"om
.09
.05
wow

 

SNO ICLHOdOH d

AmhmH ~Hm um Huhozv mcoHpHomoum wmmum HmHm:>NMH pm>u0mno m.N whamwm

“maugOEV mug aduHooaozommo
me OHN vow mmH NmH mmH omH «pH mmH NmH omH omH vVH mmH NMH

I I I I F I I I I I I I I I

 

2 wish
vOH
.9
34
.9
.9
.ON
a \

1 J” 4 .Om

A ..

... “a V
«.3
and
‘1‘)
3
>4 .ov
\ v

."om
fom
,.Oh
vom

 

SNOILHOdOHd

152

characteristics of penis, and testes and scrotum develop in a manner
so orderly that many researchers have thoroughly described the stages
through which those characteristics pass in their progress toward
maturity, and 3) the adolescent male's secondary sexual characteristics
of facial hair, axillary hair, and pubic hair also develop in a manner
so orderly that many researchers have thoroughly described the stages
through which those secondary sexual characteristics pass in their
progress toward maturity.

RELATIONSHIP BETWEEN MOTOR PROFICIENCIES AND VARIOUS

CLASSIFICATION PROCEDURES FOR ADOLESCENT MALES

The purpose of athletic competition is to reward the competi-
tors putting forth the greatest effort and those with the higher
athletic ability. Therefore, to prevent a biasing effect that favors
those adolescent males with early onset and accelerated rates of bio-
logical maturation, sponsors of adolescent male competitions should
equate the boys according to biological age. For sixty years,
specialists in child growth and development have attempted to equate
athletic competitions for adolescent males by using various indices
to classify them.

To determine an administratively feasible index by which pri-
mary and secondary schools could homogeneously classify their boys for
physical education classes, Indiana University Instructor Karl W.
Bookwalter intensively surveyed the literature on various classifica-
tion indices (Bookwalter, 1939). After examining McCloy's Classifi-
cation Index, a Force Index, Cozen's General Athletic Ability Index,

and Individual Athletic Events Index, Brace's Motor Ability Index,

153

MacCurdey's Physical Capacity Index, Roger's Physical Fitness Index,
and Roger's Strength Index, Bookwalter concluded that elementary
schools should classify boys for athletic competition either by
Roger's Strength Index or McCloy's Classification Index, and Junior
High Schools should classify boys either by Roger's Strength Index
or by MacCurdy's Physical Capacity Index.

To study the relationship between motor performance and the
classification indices of skeletal age, physiological maturity, and
the anthropometric measurements of height, weight, stem length/height,
and breadth/length, University of California Institute of Child
Welfare worker Anna S. Espenschade studied adolescent boys from the
University of California's Adolescent Study (Espenschade, 1940).
Using skeletal age values assessed by Professor Nancy Bayley,
Espenschade found positive correlations between the skeletal maturity
and motor performances of adolescent males. Using Professor Stolz's
three sexual development phases for adolescent males, Espenschade
found a direct relationship between motor progression and sexual
deve10pment. In several of the motor performance events, she found
that the motor performance and weight advanced with increasing age.
She found that whereas the relationship between weight and ability
to throw a ball for distance decreased somewhat over time, the rela-
tionship between weight and standing long jump scores increased
steadily over time.

To contrast maturational, structural, strength, and motor
traits of fifth through ninth grade boys, University of Oregon Pro—

fessor H. Harrison Clarke and Kay H. Peterson studied 202 boys from

154

the 1956-1957 Medford, Oregon, Boy's Growth Study (Clarke and
Peterson, 1961a). They divided the boys into athletic participants
and non-participant groups. Within the athletic participant group,
they asked the coaches to rate the boys as: l) exceptional athletes,
2) good athletes, 3) "out for the sport" athletes. Clarke and
Peterson did not describe how they determined skeletal age. Never-
theless, they found that the outstanding elementary school athletes
were more skeletally mature than their lesser skilled and non-
participating peers. They found that the outstanding elementary
school athletes exceeded their lesser skilled and non-participating
peers in height, weight, and lung capacity measurements. They found
that the outstanding junior high school athletes also had signifi-
cantly higher mean skeletal ages than their lesser skilled and non-
participating peers. And, in addition to greater mean values for
height, weight, and lung capacity, they found that the outstanding
junior high school athletes also exceeded their lesser skilled and
non-participating peers in arm girth, chest girth, and calf girth
measurements. In their conclusion, Clarke and Peterson stated the
outstanding elementary and junior high school athletes were definitely
more skeletally mature, taller and heavier, and stronger than their
lesser skilled and non—participating peers.

To examine the differences in physical and motor traits of
boys with advanced, normal, and retarded skeletal maturity, University
of Oregon Professor H. Harrison Clarke and James C. E. Harrison
studied 273 boys from the 1956—1957 Medford, Oregon, Boys' Growth

Study aged 9, 12, and 15 years old (Clarke and Harrison, 1962a). As

155

their measure of biological maturity they assessed hand-wrist X-rays
by a Clarke-Hayman short method. For the three maturity groups at

all three ages, they found the greatest mean differences in average
body weight. When they compared the three groups on the basis of

grip strength, back lift strength, leg lift strength, pushup ability,
pullup ability, Roger's arm strength, Roger's Strength Index, Roger's
Physical Fitness Index, mean cable-tension strength, elbow flexion
strength, shoulder flexion strength, shoulder inward rotation strength,
and standing long jump ability, the only variables for which they did
not find significant differences in favor of the boys with advanced
skeletal maturation over those with retarded skeletal maturation were
pushups and pullups. However, because they knew that both body weight
and muscular strength increased with advancing skeletal maturity, thus
equalizing each variable's gain, they concluded that they should have
expected the pushups and pullups results.

To construct maturity, structural, strength, and motor ability
development curves for adolescent boys, University of Oregon Professor
H. Harrison Clarke and J. Stuart Wickens studied 280 boys from the
1956—1957 Medford, Oregon, Boys' Growth Study. There were forty boys
per group at the seven annual ages from 9 through 15 years old (Clarke
and Wickens, 1962b). While they did not precisely describe the method
that they used to assess skeletal age, they did reference the Greulich
and Pyle hand-wrist Atlas. They reported the average skeletal age at
each annual age for their sample (Table 2.120).

By classifying their sample in annual chronological age groups,.

they found that the mean 60-yard shuttle run development curve plateaued

156

Table 2.120

The skeletal age mean values and standard deviations for the boys
across their chronological ages (Clarke and Wickens, 1962b)

 

 

 

Descriptive Chronological Age

Data 9 10 11 12 13 14 15
Mean (m0-) 106.12 114.43 131.75 -146.25 160.52 173.62 184.25
5.0- (moo) 1.82, 1-89 2108 .1191. 2-04 1.77 1-;1__,

 

between their 9 to 10 year old boys, rose sharply between their 10 to
12 year old boys, dipped between their 12 to 13 year old boys, rose
sharply between their 13 to 14 year old boys, and plateaued between
their 14 to 15 year old boys. Moreover, the mean standing long jump
development curve had a nearly straight line rise between their 9 to
12 year old boys, dipped between their 12 to 13 year old boys, rose
slightly between their 13 to 14 year old boys, and decreased between
their 14 to 15 year old boys.

To determine the relationship between reaction time, movement
time, and task completion time to adolescent boys' motor performances,
strength scores, anthropometric measurements, and pubescent sexual
deve10pment stages, University of Oregon Professor H. HarriSon Clarke
and Don Glines studied data on 65 boys within two months of their)
thirteenth birthday (Clarke and Glines, 1962c). As their measure of
sexual development, they used Greulich's 1942 pubic hair and genitalia
development indices. They found that their estimate of the boys'
sexual development had no significant correlation with their total
body reaction, arm reaction, movement time, or task completion times.

To compare adolescent males' skeletal age, various anthropo-

metric measurements, and various motor performance abilities with their

157

sexual development status, University of Oregon Professor H. Harrison
Clarke and Earnest W. Degutis studied 236 boys from the 1956-1957
Medford, Oregon, Boys' Growth Study aged 10, 13, and 16 years old
(Clarke and Degutis, 1962d). They used the Clarke-Hayman short
method for their skeletal age measure. By using the criteria pro-
posed by Greulich (1942), they assessed sexual development. They
reported the observed frequency of boys in Greulich's five sexual
maturation stages (Table 2.121). In parentheses, the author added

the proportions that represented the frequency at each chronologic age.

Table 2.121

The frequency of the boys in Greulich's five Maturity Groups
(Clarke and Degutis, 1962d)

 

Greulich Maturity Groups

 

 

Ages N 1 2 3 4 5
10 86 72(83.7) l4(l6.3) O(0.0) 0(0.0) O(0.0)
13 65 2(3.1) 25(38.5) 26(40.0) 8(12.3) 4(6.2)
16 86 O(0.0) 0(0.0) O(0.0) 12(14.0) 78(86.0)

 

For 10, 13, and 16 year old boys, they found that when sexual develop-
ment advanced, skeletal ages also advanced. For their 13 year old
boys, they found that the skeletal age of the Maturity Group 2 boys
was six months delayed from their average chronological age, that the
skeletal age of the Maturity Group 3 boys was more than three months
advanced over their average chronological age, and that the skeletal
ages of the Maturity Groups 4 and 5 boys were more than fifteen and
one-half months advanced over their average chronological age. The
data for the 16 year old boys revealed that the skeletal ages of the

Maturity Group 4 boys were twelve months delayed from their average

158

chronological age, and those of Maturity Group 5 boys were seven
months advanced over the average chronological age. Clarke and
Degutis concluded that although Greulich's Maturity Group Index was
not as sensitive to maturation changes as skeletal age for the pre-
pubescent 10 and post-pubescent 16 year old boys, it was effective
for the pubescent 13 year old boys.

To determine the value of age, height, and weight as classi-
fication indexes for secondary school students performing the AAHPER's
Youth Fitness tests, Pennsylvania State University Professors Elmer
A. Gross and Jerome A. Casciani analyzed data that physical education
teachers in 82 Pennsylvanian secondary schools collected on over
13,000 junior and senior high school boys and girls (Gross and
Casciani, 1962). They found that with the exception of junior high
school boys performing the softball throw test, the standing long
jump test, and the 50—yard dash, whether singly or in combination,
age, height, and weight measures had little value in classifying
students for motor proficiency tests.

To determine the relationships between age, height, and weight
to motor performances of adolescent boys, University of California
Professor Anna S. Espenschade studied the 50-yard dash, standing long
jump, throw for distance, situp, and pullup performances of 3,800
California boys aged 10 to 18 years old (Espenschade, 1963). By
placing the 12 to 15 year old boys into one group, Espenschade found
that age, height, and weight accounted for twenty-five percent of the
variance in their 50-yard dash, standing long jump, and throw for

distance performances. With the single exception of situps, she found

159

that all test performances increased steadily at all annual age
increments.

Except to say that undoubtedly physical size was more similar
within its groupings, Espenschade did not elaborate on the California
Classification Plan. However, she reported that when she grouped
boys according to physical size, she found the California Classifi-
cation Plan superior to chronological age alone. Nevertheless,
because 1) the California Classification Plan evolved thirty years
ago and 2) today's boys are taller and heavier, she advised that
someone update the California Classification Plan's grouping criteria.

Although she found that an Index created by combining the
boys' age, height, and weight predicted the performances of a few
adolescent boys more accurately than their age alone, Espenschade
believed that the amount of prediction improvement was not sufficient
to justify the required added time and effort. Therefore, Espenschade
recommended that researchers use chronological age alone as their
basis for developing motor performance test norms.

To examine the stability of dynometric muscular strength and
motor performance ability from childhood to adolescence and to
determine the predictive ability of motor performances at adolescence
as predicted from their childhood body size, muscular strength, and
motor performance measurements, University of Wisconsin Professors
G. Lawrence Rarick and Frank L. Smoll annually measured 25 boys from
age 7 to 12 years old (Rarick and Smoll, 1967). To obtain their
adolescent values, they measured the boys a seventh time five years

later. Their anthrOpometric data included height and weight from

160

which they determined the reciprocal ponderal index. Their dynometric
muscular strength measure included wrist flexion, elbow flexion,
shoulder abduction, shoulder medial rotation, ankle extension, knee
extension, hip flexion, and hip extension. Measures of gross motor
skills included the 30-yard dash, overarm throw for velocity, and
standing long jump.

Rarick and Smoll found that from ages 7 to 12 years and from
childhood to age 17, height, weight, physique, standing long jump,
and 30-yard dash deve10pments showed relatively stable growth trends.
However, upper extremity dynometric muscular strength and the overarm
throw for velocity developed in an unstable manner. The bestcombina-
tion of physical growth, muscular strength, and motor performance
measurements that they took on the boys when they were 10 years old
accounted for seventy-eight percent of the standing long jump scores
and fifty-nine percent of the 30-yard dash scores of the boys when
they reached 17 years old. They concluded that boys with good per-
formances in selected motor abilities during their childhood did not
necessarily retain those performance levels during adolescence.
However, if boys were strong and skilled in running and jumping during
their childhood, then they were more likely to rank high in those
skills during adolescence.

In 1971 University of Oregon Physical Education Research
Professor H. Harrison Clarke compiled all of the studies that he and
his graduate assistants had reported on the 1956 to 1968.Medford,
Oregon, Boys' Growth Study and published it in a book entitled

Physical and Mbtor Tests in the Medford Boys Growth Study (Clarke,

161

1971). From their extensive research Clarke drew the following con-
clusions: l) exceptional adolescent male athletes are more skele-
tally mature, taller, heavier, stronger, more powerful, and more
highly skilled than their peers; 2) although the characteristics
differ at the three school levels, some fundamental characteristics

of 12 year old elementary school athletes, 15 year old junior high
school athletes, and 17 year old senior high school athletes evidenced
themselves when the boys were younger: 3) the longitudinal signifi-
cance of exceptional athletes' physical characteristics may extend

to specific sports, and for various sports and at different school
levels, those characteristics may vary; 4) as measured by their
relative strength, especially for elementary and junior high school
boys, general physical fitness indirectly contributed to a higher
physical fitness of the athletes; 5) at all school levels, athletes
who participated in all interscholastic sports scored consistently
high in the gross strength measures; 6) at all school levels, athletes
who participated in all interscholastic sports scored consistently
high in the explosive power measures; 7) although successful athletes
generally possessed common characteristics, the quality of each of
those characteristics varied between athletes, such that, if athletes
were weak in some characteristics, then by being very strong in other
characteristics they compensated for weak characteristics; and 8) out-
standing elementary school athletes did not necessarily become the
outstanding junior high school athletes, and the outstanding junior
high school athletes did not necessarily become the outstanding senior

high school athletes.

162

Summary

From a review of the research literature, some of the prin-
ciples concerning the relationships between motor proficiency levels
in adolescent males and various classification indices pertinent to
his investigation were: 1) biologically advanced adolescent males
generally possess an advantage in motor performance over their bio-
logically normal and delayed peers; 2) researchers have not found
a classification index that is more desirable than chronological
age from which they can predict the motor proficiencies of adolescent
males on whom they have collected cross-sectional data; and 3) to
identify a Classification index that predicts the future motor pro-
ficiency of adolescent males requires longitudinal data.

The review of literature contained in Chapter II encompassed
the research histories of the assessment of hand-wrist skeletal age,
the assessment of sexual maturation in adolescent males, and the
relationship between the motor proficiency ability of adolescent
males and various classification procedures. In the review of the
assessment of hand-wrist skeletal age, specialists demonstrated that
estimates of skeletal age from hand—wrist X-rays represented the best
predictor of biological age that researchers in child growth and
deve10pment can use. Moreover, the X—ray standards contained in the
Greulich and Pyle hand-wrist Atlas are reliable guides for the
assessment of skeletal maturation.

In the review of the assessment of sexual maturation of
adolescent males, the specialists showed that raters can reliably

assess the sexual characteristics of facial hair, axillary hair, pubic

163

hair, penis development, and scrotum and testes development. The
estimate of sexual maturation represented the second best predictor
of biological age in adolescent males that is currently available.

The review of the relationship between the motor proficiency
ability of adolescent males and various classification procedures
demonstrated that l) the use of chronological age does not take into
account the vast differences in the biological maturity of adolescent
males and that 2) researchers need to collect longitudinal data to
investigate predicting future motor proficiency.

The review of the literature provided the following findings
that are pertinent to this investigation: 1) hand-wrist X—ray skeletal
age estimates are the best estimates of the biological age of adolescent
males; 2) the X-ray standards contained in the Greulich and Pyle hand-
wrist Atlas are reliable and valid guides for assessing skeletal age;
3) the primary and secondary sex characteristics of facial hair,
acillary hair, pubic hair, penis development, and scrotum and testes
deve10pment provide reliable criteria for the sexual maturation of
adolescent males; and 4) specialists in classifying adolescent males
for motor proficiency have tried but rejected the use of biological
maturation as a means of classifying participants for athletic

competition.

CHAPTER III

METHODS AND PROCEDURES

Because finding answers to the four questions in this desserta-

tion required two samples, chapter three provides two explanations of the

experimental designs, data collections, and measures taken.

Experimental Design #1

 

Questions one, two, and three needed a sample of adolescent males
who represented a rectangular distribution of sexual maturation. In order
to randomly select such a sample, it was necessary to know the sexual
maturation level of the potential subjects. Therefore, in conjunction
with the Fall AAHPBR Youth Fitness testing, the boys physical education
teachers at East Lansing Michigan's Macdonald Middle School used the
SMIV checklist to assess the sexual maturation of their students. Also,
the author assessed the sexual maturation of the ninth grade boys in East
Lansing's High School. A letter requesting permission for their son to
participate in this study was sent to the parents. Forty-eight percent
of the parents returned those forms and forty—seven percent were affirm-
ative. From groupings based on sexual maturation levels, 30 subjects were
randomly selected. Table 3.1 provides the SMIV score, chronological age,

and grade of the boys.

Data Collection #1
All of the data were collected during a single Saturday morning

session.

164

165

Table 3.1

The SMIV scores, chronological age, and grade levels of the
30 boys selected to represent a rectangular
distribution of sexual maturation

 

 

Subject SMIV Chronological Grade
Age (months)
JB 1.1 134.5 7
JH 1.2 150.0 6
FF 1.3 145.5 6
EM 1.6 137.5 6
HS 1.7 143.0 7
RS 1.8 155.5 7
TB 1.9 143.5 6
JC 2.0 152.5 8
DD 2.1 170.0 9
ST 2.2 173.0 9
CM 2.3 179.0 9
_BO 2.6 150.0 6
DF 2.7 138.0 6
JS 2.8 146.0 6
KB 2.9 163.0 8
JP 3.0 161.5 8
JL 3.1 170.5 9
TS 3.2 156.5 8
DB 3.3 173.0 9
CG 3.6 173.0 9
JP 3.7 153.0 7
CS 3.8 174.0 9
PW 3.9 164.0 8
PS 4.0 162.0 8
P2 4.1 169.5 9
RH 4.2 154.0 7
DT 4.3 160.0 7
CO 4.6 175.5 8
UK 4.7 169.5 9
DP 4.8 180.0 8

 

Measures #1

 

Skeletal Age

 

Michigan State University Olin Health Center Radiologists X-rayed
both hands and wrists of the subjects. (Appendix B contains a detailed
description of the methodology.)

Using the Greulich-Pyle hand-wrist atlas, Michigan State Univer-

sity Professor Vernal D. Seefeldt viewed hand-wrist X-rays of the subjects

166

and on two separate occasions determined the skeletal age (Appendix C).
For each subject's skeletal age, the computer averaged the two ratings.

Physical Growth

 

Michigan State University 01in Health Center Radiologists measured
the height and weight of the subjects. They recorded height to the nearest
one-quarter inch and the analysis rounded .25 to .3 and .75 to .8. They
recorded weight to the nearest one-half pound (Appendix D).

Sexual Maturation

 

Using the SMIV checklist, two East Lansing, Michigan and two Mason,
Michigan boys physical education teachers and the author assessed sexual
maturation. Where the data collection required only a single score for
the five sexual characteristics, the analysis averaged the ten scores
assigned each trait (Appendix E).

Chronological Age

 

School records provided the age of the subjects on the date of the

data collection. Their age was recorded to the nearest one-half month.

Experimental Design #2

 

Question 4 needed a large, cross-sectional sample of adolescent

males.

Data Collection #2

 

Beginning with the second week in April, every six months for
three years, the East Lansing teachers assessed the seventh, and eighth
grade male populations and the Mason teachers assessed the seventh, eighth,
and ninth grade male population. Additionally, in the final test period,
the East Lansing teachers assessed the sixth grade boys in East Lansing's

new Macdonald Middle School. For those subjects on whom the teachers

167

collected data for more than one test period, because the sample needed

more subjects at higher SMIV scores, the analysis included their last

complete data set. Table 3.2 contains the number of East Lansing subjects

in each grade for each test period. Of the 878 East Lansing boys in this
Table 3.2

The number of East Lansing subjects in each grade for each
test period

 

 

 

 

TEST GRADE LEVEL
PERIOD SIXTH SEVENTH EIGHTH NINTH
l O 48 76 O
2 O 68 156 0
3 O 41 48 O
4 O 49 57 0
5 0 74 4O 0
k; 6 ‘ 83 96 42 0
TOTALS 83 376 419 O 878

 

study, 83 were sixth graders, 373 were seventh graders, and 418 were

eighth graders. Table 3.3 contains the number of Mason subjects in each

grade for each test period. Of the 700 Mason boys in this study, 207
Table 3.3

The number of Mason subjects in each grade for each test period.

 

 

 

 

TEST GRADE LEVEL

PERIOD SIXTH SEVENTH EIGHTH .NINTH

1 o 31 30 12

2 o 29 26 11

3 o 28 36 115

4 o 31 32 19

s o 60 51 89

6 o 28 3o 42
v—TOTALS o 207 205 288 700

 

were seventh graders, 205 were eighth graders, and 288 were ninth graders.

Measures #2

 

School

The boys attended either East Lansing or Mason..

168

Test Period

 

Test period was recorded as; first, second, third, fourth, fifth,
or sixth.

@1999.

Grade was recorded as; sixth grade in the fall, seventh grade in
the fall, seventh grade in the spring, eighth grade in the fall, eighth
grade in the spring, ninth grade in the fall, or ninth grade in the spring.

Chronological Age

 

School records provided the age of the subjects on the date of
the data collection. Age was recorded to the nearest one-half month.

1 Physical Growth

 

At each school, the teachers measured the height and weight of the
boys. They recorded height to the nearest one-quarter inch and the
analysis rounded .25 to .3 and .75 to .8. They recorded weight to the
nearest one-half pound (Appendix F).

AAHPBR Youth Fitness Tests

 

The teachers tested the following motor proficiency skills;
1) sit—ups with a maximum of 99, 2) pull-ups, 3) standing long jump to
the nearest inch, 4) vertical jump, to the nearest inch, and 5) shuttle
run, to the nearest one-tenth of a second (Appendix G).

Sexual Maturation

 

Using the SMIV checklist, the teachers assessed the sexual matura-
tion of the boys. For each sexual characteristic, the analysis used the

number of the stage that the teachers checked (Appendix H).

Treatment of the Data

 

a) Question one: What proportion of the variance in the skeletal

169

age of adolescent males is explained by increments in their chronological
age, height, and weight?

The Multivariate Analysis of variance computer program determined
the square of the multiple r values and the step-wise regression analysis.
The square of the multiple r provided.the proportion of the dependent
variable (skeletal age) that is explained by the increments in the inde-
pendent variables (chronological age, height, weight, and SMIV). VThe
step—wise regression analysis explained the contribution to the variance
in the dependent variable that is unique to each independent variable.
Therefore, question one became a five-step statistical analysis.

1) What is the square of the multiple r for estimating skeletal
age from chronological age? I

2) What is the square of the multiple r value for estimating
skeletal age from chronological age and height?

3) Was the additional contribution unique to height statiStically
significant?

4) What is the square of the multiple r value for estimating
skeletal age from chronological age, height, and weight?

5) Was the additional contribution unique to weight statistically
significant?

Educational researchers typically choose the .05 significance level.
Therefore, in this study if the analysis falsely accepts that height and
weight do contribute significantly, then the estimate of skeletal age
(ESA) will incorrectly include those variables. However, if the analysis
falsely rejects the significant contributions of height and weight, then

ESA will incorrectly exclude those variables.

170

b) Question two: Does SMIV explain a significant additional pro-
portion of the variance in skeletal age? By using the same statistical
procedures as in question one, question two became a two-step statistical
analysis.

1) What is the square of the multiple r value for estimating
skeletal age from chronological age, height, weight, and SMIV?

2) Was the additional contribution unique to SMIV statistically
significant?

If the analysis falsely accepts that SMIV does contribute signifi-
cantly, then ESA will incorrectly include the SMIV variable. However, if
the analysis falsely rejects the significant contribution of SMIV, then
ESA will incorrectly exclude the SMIV variable.

c) Question three: Can Boys Physical Education teachers reliably
assess the sexual maturation of adolescent males?

The correlation analysis determined the within-rater reliabilities
and the between-rater reliabilities. However, between-rater correlations
combine the errors between raters. across sexual characteristics, and
over time which results in an underestimation of the between-rater relia-
bility. Therefore, an extension of the intraclass correlation method of
estimating reliabilities that bases its procedure on analysis of variance
calculated the individual reliabilities for between raters, across sexual
characteristics, and over time.

By conceptualizing this analysis in the following two ways, the
intraclass correlation method provided eight separate reliabilities;

1) researchers using the five sexual manifestations individually and
equally as indicants of sexual maturation and 2) researchers using the

average score of the five sexual manifestations as the indicant of sexual

171

maturation.

1) p WITHIN, or for a fixed rater, a fixed sexual characteristic,
and a fixed time point, what was the reliability coefficient?

2) p STABILITY OVER TIME, or for a fixed rater and a fixed sexual
characteristic, what was the reliability coefficient?

3) p EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS, or for a fixed
rater and a fixed time point, what was the reliability coefficient?

4) p STABILITY and EQUIVALENCE, or for a fixed rater, what was
the reliability coefficient?

5) p RATERS, or for a fixed sexual characteristic and a fixed
time point, what was the reliability coefficient?

6) p RATERS STABILITY, or for a fixed sexual characteristic, what
was the reliability coefficient?

7) p RATERS EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS, or for a
fixed time point, what was the reliability coefficient?

8) p RATERS STABILITY and EQUIVALENCE, or across raters, across
sexual Characteristics, and over time, what was the reliability coeffi-
cient?

The reliability coefficient that directly answers question three
is the p RATERS when the analysis used the average of the five sexual
manifestations as the indicant of sexual maturation.

d) Question four: How does an estimate of skeletal age (ESA) that
encompasses chronological age, height, weight, and SMIV compare with
chronological age when Classifying adolescent males for motor proficiency
norms?

The following statistics examined this question:

1) within cell variances

172

Smaller within-cell variances indicate more homogeneous groups.
2) mean value increments
Larger mean value differences across the annual classifications

indicate more homogeneous groups.

CHAPTER IV

RESULTS AND DISCUSSION

Question I. What pr0portion of the variance in the skeletal age

 

of adolescent males is explained by increments in their chronological
age, height, and weight?

The Multivariate Analysis of Variance computer program statisti-
cally analyzed the data (Schiefley and Schmidt, 1973). The square of the
multiple r estimated the percent of the variance in the dependent var-
iable explained by the independent variables.

In the first part of question I, the dependent variable was skeletal
age and the independent variable was chronological age. When chronologi-
cal age was the only independent variable, the square of the multiple r
was .440 or approximately 45 percent. Therefore, the pr0portion of the
variance in skeletal age explained by chronological age was 45 percent.

In the second part of question I, while the dependent variable
remained skeletal age, the independent variables became chronological age
and height. When chronological age and height were the independent var-
iables, the square of the multiple r was .648 or 65 percent. Therefore,
the proportion of the variance in skeletal age explained by the contribu-
tion of chronological age and height was 65 percent. The inclusion of
height uniquely explained an additional 20 percent of the variance.

As a result of part two, the third part of question I became; Was
the additional 20 percent of the variance in skeletal age unique to height
statistically significant? Step-wise regressions determine the signifi-

173

174

cance levels of additional contributions by each independent variable.
To interpret step-wise regressions, when the contribution of the last
variable is significant, it is not possible to discuss the significance
level of the preceding variable. However, when the contribution of the
last variable is not significant, it is possible to discuss the signifi—
cance level of the preceding variable.

Table 4.1 contains the step-wise regression for chronological age
and height. Because the significance level for height was less than .05,

Table 4.1

The step-wise regression for chronological age and height

 

 

Percent of Additional Significance
Variance Level
Chronological Age 44.937
Height 19,903 P less than .0006

 

the additional contribution unique to height was significant.

In the fourth part of question I, while the dependent variable
remained skeletal age, the independent variables became chronological age,
height, and weight. When chronological age, height, and weight were the
independent variables, the square of the multiple r was .708 or approxi-
mately 71 percent. Therefore the proportion of the variance in skeletal
age explained by the combination of chronological age, height, and weight
was 71 percent. The inclusion of weight uniquely explained an additional
6 percent of the variance.

As a result of part four, the fifth part of question I became; Was
the additional 6 percent of the variance in skeletal age unique to weight
statistically significant? Table 4.2 contains the step-wise regression
for chronological age, height, and weight. Because the significance

level for weight was less than .05, the additional contribution unique

175

Table 4.2

The step-wise regression for chronological age, height, and weight

 

 

Percent of Additional Significance
Variance Level
Chronological Age 44.937
Height 19.903 —
Weight 5.932 P less than .0300

 

to weight was significant.

To summarize, the answers to question I's five sub-questions are;

1) What is the square of the multiple r for estimating skeletal
age from chronological age? For this sample, that square of the multiple
r value was .449.

2) What is the square of the multiple r for estimating skeletal
age from chronological age and height? For this sample, that square
multiple r value was .648.

3) Was the additional contribution unique to height statistically
significant? For this sample, height did uniquely explain a significant
additional proportion of the variance.

4) What is the square of the multiple r value for estimating skeletal
age from chronological age, height, and weight? For this sample, that
square of the multiple r value was .708.

5) Was the additional contribution unique to weight statistically
significant? For this sample, weight did uniquely explain a significant
additional prOportion of the variance.

At this point some readers might ask, if the order of the inde-
pendent variables reversed, then what would happen to the step-wise
regressions (Appendix I)?

Question II. Does SMIV explain a significant additional propor-

 

tion of the variance in skeletal age?

176

In the first part of question II, the dependent variable remained
skeletal age while the independent variables included chronological age,
height, weight, and the five sexual characteristics of; facial hair
axillary hair, pubic hair, penis development, and scrotum and testes
development. The square of the multiple r was .839 or approximately 84
percent. Therefore, the proportion of the variance in skeletal age
explained by the combination of Chronological age, height, weight, and
the five sexual characteristics was 84 percent. The inclusion of the
five sexual characteristics variables uniquely explained an additional
13 percent of the variance.

Because doing a step-wise regression analysis with chronological
age, height, weight, and the five sexual characteristics would give us
the significance levels for the five sexual characteristics individually,
a SMIV regression equation that combined the five sexual characteristics
was determined (Appendix J).

When the independent variables were chronological age, height,
weight, and SMIV, the square of the multiple r was .828 or approximately
83 percent. The inclusion of SMIV uniquely explained an additional 12
percent of the variance.

AS a result of part one, the second part of question II became;
Was the additional 12 percent of the variance in skeletal age unique to
SMIV statistically significant? Table 4.32 contains the step-wise regres-
sions for chronological age, height, weight, and SMIV. Because the signi-
ficance level for SMIV was less than .05, the additional contribution
unique to SMIV was significant.

To summarize, the answers to question II's two sub-questions are:

1) What is the square of the multiple r value for estimating

177

Table 4.3

The step-wise regression for chronological age, height, weight

 

 

and SMIV
Percent of Additional Significance
Q Variance Levels
Chronological Age 44.937
Height 19.850
Weight 5.966
_SMIV 12:014 P less than .0004

 

skeletal age from chronological age, height, weight, and SMIV? For this
sample, that square of the multiple r value was .828.

2) Was the additional contribution unique to SMIV statistically
Significant? For this sample, SMIV did uniquely explain a significant
additional proportion of the variance.

At this point some readers might again ask, if the order of the
independent variables reversed, then what would happen to the step-wise
regressions (Appendix K)?

Question III. Can Boys Physical Education teachers reliably
assess the sexual maturation of adolescent males?

Correlation coefficients estimated the within rater reliability
between the first and second assessments of the five raters on each of
the five sexual characteristics (Table 4.4).

Table 4.4

The within-rater correlation coefficients for the five raters across
the five sexual Characteristics and the last column shows the
mean correlation across raters for each variable

 

 

 

Sexual RATERS Average across
Characteristics 1 2 3 4 5 Raters
Facial hair .789 .683 .949 .492 .649 .772
Axillary hair .777 .915 .926 .694 .877 .838
Pubic hair .762 .940 .838 .883 .904 .865
Penis .874 .931 .954 .851 .885 .899
Scrotum and .859 .916 .818 .710 .827 .826

Tastes

178

The within-rater correlations for facial hair ranged from .949
to .492 while the last column shows that the mean correlation across
raters was .77. For axillary hair, the within-rater correlations ranged
from .962 to .694 and the average correlation across raters was .838.

For pubic hair, the within-rater correlations ranged from .940 to .762
with the .865 averaged correlation across raters. For penis develop-
ment, theawithin-rater correlations ranged from .954 to .851 with an
average across raters correlation of .899. And, for scrotum and testes
development, the within-rater correlations ranged from .916 to .710
with an average correlation across raters of .826. From highest to
lowest, the order of the average correlation across raters was; 1) penis
development, 2) pubic hair, 3) axillary hair, 4) scrotum and testes
development, and 5) facial hair.

Because correlation statistics describe how well two values
approximate each other, those within-rater correlations told us only how
well each rater agreed with himself. However, just because a rater's
correlation shows that he agreed highly with himself, that did not neces-
sarily indicate that that rater correctly assessed the boys' sexual char-
acteristics. To determine how well the raters agreed with each other,
each rater was paired with every other rater, and then those raters' first
and second assessments were correlated (Tables 4.5 through 4.9) . Table
4.10 contains each raters' average paired betwee-raters correlations for
the five sexual Characteristics. And, Table 4.11 contains the average

of all of the paired between raters correlations for the five sexual

characteristics.

179

Table 4.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Between-raters correlations of facial hair for the paired five raters
Facial Hair Raters
Ratings 1/2 1/3 1/4 1/5 2/3 2/4 2/5 3/4 3/5 4/5
1/1 .618 .778 .490 .633 .576 .458 .578 .307 .612 .462
1/2 .734 .796 .706 .795 .728 .542 .627 .515 .681 .556
2/1 .724 .851 .443 .629 .698 .365 .620 .337 .690 .608
2/2 .710 .895 .577 .760 .781 .667 .738 .622 .760 .633
Average .696 .830 .554 .704 .696 .508 .591 .445 .696 .565
Table 4.6
Between-raters correlations of axillary hair for the paired five raters
Axillary Hair Raters
Ratings ‘ 1/2 1/3 1/4 1/5 2/3 2/4 2/5 3/4 3/5 4/5
1/1 .733 .734 .716 .770 .829 .811 .809 .714 .851 .907
1/2 .721 .653 .716 .785 .787 .827 .862 .781 .853 .796
2/1 .843 .737 .744 .770 .828 .787 .786 .705 .853 .747
2/2 .848 .740 .734 .816 .821 .802 .821 .723 .840 .800
Average .786 .716 .728 .785 .816 .807 .820 .731 .849 .813
Table 4.7
Between-raters correlations of pubic hair for the paired five raters
Pubic hair Raters
Ratings 1/2 1/3 1/4 1/5 2/3 2/4 2/5 3/4 3/5 4/5
1/1 .797 .756 .818 .819 .795 .913 .935 .866 .861 .890
1/2 .833 .763 .800 .780 .917 .861 .891 .828 1748 .848
2/1 .837 .763 .767 .875 .830 .886 .945 .869 .874 .854
2/2 .895 .824 .797 .833 .909 .874 .882 .805 .784 .911
Average .841 .771 .796 .827 .863 .884 .913 .842 .817 .876
Table 4.8
Between-raters correlations of penis development for the paired
five raters
Penis dev. Raters
Ratings 1/2 1/3 1/4 1/5 2/3 2/4 2/5 ~3/4 3/5 4/5
1/1 .904 .828 .782 .765 .925 .904 .894 .876 .880 .841
1/2 .876 .818 .761 .861 .890 .815 .901 .820 .837 .859
2/1 .914 .877 .920 .840 .871 .909 .843 .824 .883 .755
2/2 .932 .823 .823 .914 .840 .827 .862 .796 .815 .811
Average .907 .837 .824 .845 .882 .889 .875 .854 .817

.829

 

180
Table 4.9

Between-raters correlations of scrotum and testes development for
the paired five raters

 

 

 

 

 

Scrotum dev. Raters
Datings 1/2 1/3 1/4 1/5 2/3 2/4 2/5 3/4 3/5 4/5
1/1 .854 .516 .701 .682 .603 .826 .801 .621 .571- .760
1/2 .835 .647 .764 .755 .721 .800 .889 .403 .571 .792
2/1 .881 .516 .811 .761 .671 .875 .792 .688 .651 .711
2/2 .887 .728 .764 .787 {741 .813 .840 .528 .627 .821
Average .864 .602 .760 .746 .687 .836 .831 .556 .605 .776
Table 4.10

Average between-raters correlations of the paired raters for the
five sexual characteristics

 

 

 

 

 

Sexual Raters
Characteristics 1 2 3 4 5 Avg;
Fac1al hair .696 .623 .667 .518 .639 .629
Axillary hair .754 .807 .778 .770 .817 .785
Pubic hair .809 .875 .823 .850 .858 .843
Penis dev. .853 .888 .851 .840 .848 .856
§grotum dev. .743 .805 .613 .732 .740 .727

Average .771 .800 .746 .742 .780 .768

Table 4.11

Average of all of the paired between-raters correlations for
the five sexual characteristics

 

Sexual Characteristics

 

 

Facial Axillary Pubic Penis ScrOtum
Hair Hair Hair Development Development
.629 .785 .843 .856 .727

 

For facial hair, rater 1'5 .696 average paired between-raters
correlation was the highest of the five raters. The .692 average of all
of the paired between-raters correlations placed facial hair fifth in
this list of five characteristics.

For axillary hair, rater 5'3 .817 average paired between-raters

181

correlation was the highest and rater 1'3 .754 average correlation was
the lowest. The average of all of the paired between-raters correlations
of .785 placed axillary hair third.

For pubic hair, rater 2'5 .875 average paired between-raters
correlation was the highest and rater 1'3 .809 average correlation was
the lowest. The average of all of the paired between-raters correlations
of .843 placed pubic hair second.

For penis deve10pment, rater 2'5 .888 average paired between-
raters correlation was the highest and rater 4'5 .840 average paired
between-raters correlation was the lowest. The average of all of the
paired between-raters correlations of .856 placed penis development first.

For scrotum and testes development, rater 2'3 .805 average paired
between-raters correlation was the highest and rater 3'5 .613 was the
lowest. The average of all of the paired between-raters correlation of
.727 placed scrotum and testes development fourth.

The averages of the average paired between-raters correlations
for the five sexual Characteristics contained in Table 4.11 demonstrate
the consistency of the five raters. From highest to lowest ability of
those five raters, the order of the raters was; rater 2 (.800), rater 5
(.750), rater l (.771), rater 3 (.746), and rater 4 (.742). The averages
of all of the paired between-raters correlations suggest the agreement of
the five raters. The order of the averages of all of the paired between-
raters correlations for the five sexual characteristics were; penis
development (.856), pubic hair (.843), axillary hair (.785), scrotum and

testes development (.727), and facial hair (.629).

182

As the author's doctoral guidance committee's statistics advisor,
Michigan State University College of Education Professor William H. Schmidt
provided the following extension of the intraclass correlation method of
estimating reliabilities based on analysis of variance techniques. The
four variables in this analysis were; 1) Raters (R), 2) Sexual Character-
istics (C), 3) Time points (T), and 4) Subjects (S). Because the Raters,
Time points, and Subjects were random variables and the combined Sexual
Characteristics was a fixed variable, this 4-way ANOVA analysis became a
mixed model. By using Time as a variable rather than within-cell repli-
cations, each cell contained only one value.

As its first step, this analysis required the determination;of
the expected Mean Square (MS) formulas for a 4-way Analysis of Variance
mixed model. By using the Millman-Glass Rules of Thumb for Writing the

Anova Table, the Mean Square formulas contained in Table 4.12 were

 

determined (Millman and Glass, 1967).

By assuming that those four variables did not interact, the 02(RCTS)
value became the error estimate (02(e)). Using the Table 4.12 Mean Square
formulas, the formulas for the estimate of the variance components were
isolated (Table 4.13).

After determining those formulas, an N-way analysis of variance
computer program calculated the Mean Square values (Table 4.14) (Edon,
1976). By substituting those Mean Square values into the estimates of
the variance components formulas, the estimates of the variance components
were calculated (Table 4.15).

For the first series of reliabilities, the analysis assumed that
the five primary and secondary sexual characteristics were manifestations

of the underlying trait of sexual maturation. Therefore, those reliabil-

183

Table 4.12

The Mean Squares formulas for a 4-way Anova mixed model where Raters(R),
Time points(T), and Subjects(S) are random and Sexual

 

 

 

Characteristics(C) is fixed
Mean Squares Formula
1. RCTS = 64(RCTS)
2. CTS = 02 RCTS) + r02(CTS)
3. RTS = c (RTS)
4. RC5 = g (RCTS) + t02(RCS)
5. RCT = g £RCTS) + so (RCT)
6. TS = C3 (RTS) + rcg“(TS)
7. CS = g (RCTS) + r02(CTS) t02(RCS) rto§(cs)
8. CT = 02(RCTS) + r02(CTS) 502(RCT) rso‘(CT)
9. Rs Co2(RTS) + ct02(RS)
10. RT = cg (RTS) + C502(RT)
11. RC = C0 (RTS) + c502(RT)
12. S = c02(RTS) + csg (RT)
13. T = C3 (RTS) + rcgz(TS) cng(RT) rcsgz(T)
14. C = g (RCTS) + r02(CTS) t02(RCS) 502(RCT)=+ rtO (CS) +
rs 2(CT)
15. R = co“(RTS) + ct02(RS) csoz(RT) ctsoz(R)
Table 4.13

variance components formulas in this 4-way
mixed model

The estimates of the

 

Estimate of the

 

 

Variance Components Formulas

1. d§(e) = MS(RCTS)

2. Cfl(CTS) = MS(CTS) - MS(RCTS)/r

3. 6‘(RTS) = MS(RTS) - MS(RCTS)/c

4. 62(RCS) = MS(RCS) - MS(RCTS) t

5. 6§(RCT) = MS(RCT) - MS(RCTS) s

6. 62(TS) = MS(TS) - MS(RCTS) - c62(RTS)/rc

7. 6 (CS) = MS(CS) - MS(RCTS) - t62(RCS) - r62(CTS)/rt

8. 62(CT) MS(CT) - MS(RCTS) - 562(RCT) - r62(CTS)/rs

9. 62(Rs) = MS(RS) - MS(RCTS) - c52(RTS)/ct

10. 62(RT) = MS(RT) - MS(RCTS) - c62(RTS)/cs

11. 62(RC) = MS(RC) - MS(RCTS) - sézéRCT) - t62(RCS)/tx

12. 6:(S) = MS(S) - MS(RCTS) - eta (RS) - rc62(TS) - c62(RTS)/rct
13. 6 (T) = MS(T) - MS(RCTS) - cs62(RT) - rc62(TS) - c62(RTS)/ rcs
14. 62(C) MS(C) - MS(RCTS) - tst(RC) - rsdz(CT) - rt62(CS) -

2 562(RCT) - t62(RCS) - r62(CTs)/rts

15. 6 (R) = MS(R) — MS(RCTS) - C552(RT) - ct62(Rs) - c62(RTs)/cts

184

Table 4.14

The Mean Square values for the fifteen sources of variance in

this 4-way mixed model

Sources of Variance

Mean Scuares

 

Raters-Sexual characteristics-Time-Subjects (RCTS) .1798
Raters-Sexual characteristics-Subjects (RC5) .3348
Raters-Time-Subjects (RTS) .3476
Sexual characteristics - Time - Subjects (CTS) .1903
Raters - Sexual characteristics - Time (RCT) .1244
Raters - Subjects (RS) .7641
Sexual characteristics - Subjects (CS) 2.4635
Time - Subjects (TS) .4430
Raters - Sexual characteristics (RC) 6.6689
Raters - Time (RT) .2190
Sexual characteristics - Time (CT) .2323
Subjects (S) 44.7968
Raters (R) 16.3810
Sexual characteristics (C) 97.4477
Time (T) .0540

 

Table 4.15

The substituted estimates of the variance

components formulas

 

 

 

Estimate of the Final
Variance Components Values__
1. a§(e) = .1798 .1798 i
2. 0 (CTS) = .1903 - .1798/5 .0021
3. 62(RTS) = .3476 - .1798/5 .0036
4. 62(RCS) = .3348 - .1798/2 .0775
5. 62(RCT) = .1244 - .1798/30. .0000
6. 62(TS) = .4430 - .1798 - .1678/5(5) .0038
7. 62(03) = .4635 - .1798 - .1550/5(2) ..2118
8. (CT) = .2323 - .1798 - .0000 - .0105 /5(30) .0003
9. 62(R5) = .7641 - .1798 - .l678/5(2) .0417
10. 62(RT) = .2190 - .1798 - .1678/5(30) .0000
11. 62(RC) = 6.6689 - .1798 - .0000 - .1550/2(30) .1056
12. 62(S) = 44.7968 - .1798 — .4165 - .0954 - .1678/5(512) .8787
13. 62(T) = .0540 - .1798 - .0000 - .0954 - .l678/S(5)(30) .0000
14. 62(0) = .97.4470 - .1798 - 6.3895 - .0420 - 2.1182 - .2785
.0000 - .1550 - .0105/5(2)(30)
15. 62(8) = 16.3810 - .1798 - .0000 - .4165-.1678/5(2)(30) .0521

 

185

ities represent researchers using any one of those five sexual character-

istics as an indicant of sexual maturation (Table 4.16).

Table 4.16

The reliability formulas for the eight sources of error in this
4—way mixed model

 

 

Reliability Formulas ~
1. 0 WITHIN O‘(S)/OZ(S)+62(e)
2. p STABILITY OVER TIME 02(S)/62(S)+62(TS)+62(e)
3. p EQUIVALENCE ACROSS 62(S)/66(S)+62(CS)+62(8)
p SEXUAL CHARACTERISTICS
4. p STABILITY and EQUIVALENCE 62(S)/62(S)+62(TS)+62(CS)+02(e)
5. p RATERS 82(s>/62(s)+62(RS)+82(e)
6. p RATERS STABILITY OVER TIME 82(s)/62(S)+62(RS)+02(TS)+62(e)
7. p RATERS EQUIVALENCE ACROSS 62(5)/62(S)+62(RS)+62(CS)+62(e)
SEXUAL CHARACTERISTICS
8. p RATERS STABILITY and 62(s)/62(S)+62(RS)+62(CS)+62(TS)+
EQUIVALENCE 62(e)

 

By substituting the calculated estimates of the variance components
for their reliability formulas counterparts, the reliability coefficients

were calculated (Table 4.17).

Table 4.17

The substituted reliability formulas and their final value

 

Reliability, Substituted formulas Final Value

 

l.p WITHIN .8787/.8787+.1798 .830
2-0 STABILITY .8787/.8787+.0038+.1798 .827
3.9 EQUIVALENCE .8787/.8787+.21l8+.1798 .692
4.0 STABILITY and .8787/.8787+.0038+.2118+.l798 .690
EQUIVALENCE
5.p RATERS .8787/.8787+.0417+.1798 .799
6.p RATERS STABILITY .8787/.8787+.0417+.0038+.1798 .794
7.0 RATERS EQUIVALENCE .8787/.8787+.0417+.2118+.1798 .670
8.p RATERS STABILITY .8787/.8787+.0417+.0417+.0038+ .660
and EQUIVALENCE .2118+.1798

 

The p WITHIN indicated that for a fixed Rater, a fixed Sexual
Characteristic, and a fixed Time point the reliability of a rating was

.830. The p STABILITY OVER TIME indicated that for a fixed Rater and a

186

fixed Sexual Characteristic, the reliability across time was .827. That
finding suggests that raters can re-assess those five sexual character-
istics with little error. The p EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS
indicated that for a fixed Rater and a fixed Time point, the reliability
across those sexual characteristics was .692. The p STABILITY and
EQUIVALENCE indicated that for a fixed Rater, the reliability across time
and across sexual characteristics was .690 . The p RATERS indicated that
for a fixed Sexual Characteristic and a fixed time point, the reliability
across raters was .799. The p RATERS value of .799 demonstrated that
there was little error across raters. The p RATERS STABILITY indicated
that for a fixed Sexual Characteristic, the reliability across raters
and over time was .794. There was little error across raters and across
time. The O RATERS EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS indicated
that for a fixed time point, the reliability across raters and across
sexual characteristics was .670. The p RATERS STABILITY OVER TIME and
EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS indicated that the reliability
across raters, across sexual characteristics, and over time was .660.

Although the preceding reliabilities are interesting for under—
standing the nature of sexual maturation in a scientific sense, they do
not represent the practical situation. Researchers should use the SMIV
regression equation as their estimate of sexual maturation. Therefore,
the following reliabilities represent researchers using the average of
the five sexual characteristics as an indicant of sexual maturation.

To adjust for using the average score of the five sexual charact-
eristics as the indicant of sexual maturation, every 02(e) and 02(CS)
variable was divided by the number (n) of sexual characteristics (Tables

4.18 and 4.19).

187

Table 4.18

Adjusted reliability formulas

 

 

 

Reliability ' Formulas

1. p WITHIN 02(S)/62(S)+82(e)(1/n)

2. p STABILITY OVER TIME 62(S)/62(S)+62(TS)+62(6)(l/n)

3. p EQUIVALENCE ACROSS 62(S)/62(S)+62(CS)(l/n)+62(e)(1/n)
SEXUAL CHARACTERISTICS

4. o STABILITY and 62(S)/62(S)+62(TS)+62(CS)(l/n)+
EQUIVALENCE 62(e)(l/n)

5. p RATERS 62(S)/62(S)+62(RS)+62(e)(l/n)

6. p RATERS STABILITY OVER 62(5)/62(S)+62(RS)+62(RS)+62(a)(l/n)
TIME

7. p RATERS EQUIVALENCE ACROSS 02(5)/62(S)+02(Rs)+02(cs)(1/n)+
SEXUAL CHARACTERISTICS 62(e)(l.n)

8. p RATERS STABILITY and 02(S)/02(S)+62(RS)+62(TS)+62(CS)(l/n)+
EQUIVALENCE 62(e)(1/n)

Table 4.19

The adjusted substituted reliabilities formulas, and their
final values

 

 

Reliabilitv Substituted Formulas Final value
IT])wITHIH .8787/.8787+.1798(1/5) .961
2.,35TABILITY . .8787/.8787+.0038+.1798(1/5) .957
3. pEQUIVALSHCE .8787/.8787+.2118(1/5)+.1798(1/5) .918
4.;JSTABILITY and .8787/.8787+.0038+.2118(1/5)+.l798(1/5) .915

EQUIVALENCE -
5.;3RATERS .8787/.8787+.0417+.1798(1/5) .919
6.;3RATERS STABILITY .8787/.8787+.0417+.0038+.1798(1/5) .915
7.;3RATERS EQUIVALENCE.8787/.8787+.0417+.2118(l/S)+.1798(l/5) .880
8.;3RATERS STABILITY .8787/.8787+.0417+.0038+.2118(1/5) .876

and EQUIVALENCE +.1798(1/5)

The adjusted p WITHIN indicated that for a fixed Rater, a fixed
Sexual Characteristic, and a fixed Time point, the reliability of a rating
was .961. The adjusted p STABILITY OVER TIME indicated that for a fixed
Rater and a fixed Sexual Characteristic, the reliability across time was
.957. Those raters virtually re-assessed the sexual characteristics
without error when the analysis used the average score as the SMIV.

The adjusted p EQUIVALENCE ACROSS SEXUAL CHARACTERISTICS indicated

that for a fixed Rater and a fixed Time point, the reliability across

188

Sexual Characteristics was .918. The adjusted c>STABILITY AND EQUIVALENCE
indicated that for a fixed Rater, the reliability across Time and across
Sexual Characteristics was .915. The adjusted p RATERS indicated that

for a fixed Sexual Characteristic and a fixed Time point, the reliability.
across raters was .919. By using the average of the five sexual char-
acteristics as the indicant of the underlying trait of sexual maturation,
there was practically no error across raters. The adjusted p RATERS
STABILITY indicated that for a fixed Sexual Characteristic, the reliability
across raters and over time was .915. The adjusted p RATERS EQUIVALENCE
ACROSS SEXUAL CHARACTERISTICS indicated that for a fixed point, the relia-
bility across raters and across sexual characteristics was .880. The
adjusted p RATERS STABILITY OVER TIME and EQUILVANCE ACROSS SEXUAL CHAR
ACTERISTICS indicated that the reliability across raters, across sexual
characteristics, and over time was .876.

Question III asked whether or not physical education teachers could
reliably assess adolescent males' sexual maturation. The adjusted ‘DRATERS
reliability most directly answers that question. With a reliability
estimate of .919, those physical education teachers did reliably assess

the sexual maturation of the adolescent males in this sample.

Question IV. How does an estimate of skeletal age (ESA) that

 

encompasses chronological age, height, weight, and SMIV compare with

chronological age when Classifying adolescent males for motor proficiency

norms?

Before combining the 878 East Lansing subjects with the 700 Mason

subjects for this analysis, it seemed appropriate to determine whether or

'

not the two samples differed. Appendix L contains that discussion.

189

Because the review of literature demonstrated a void in the age
intervals that adolescent males reached the various stages in the
different sexual characteristics, Appendix M contains the proportion of
this sample's 1578 boys who reach the five stages of the five sexual
characteristics.

In order to compare ESA with chronological age (CA), it was
necessary to determine the ESA regression equation. Appendix H contains
that discussion. The BSA equation was ESA = 160.53 + 0.29(X(CA) -
159.27} + 0.33(X(H) - 63.01) + 0.23(X(W) - 108.30)+ 4.47(X(PH) - 2.95) +
4.26(X(P) - 3.02) - 0.83(X(A) - 1.86)+ 1.75(X(F) - 2.04) - 3.98(X(S) -
2.98). Table 4.20 shows the BSA and CA intervals that comprised the
annual increment groupings.

Table 4.20

The month ranges that comprised the annual increment groups

 

 

Group Month range

11.0 127 months to 138 months
12.0 139 months to 150 months
13.0 151 months to 162 months
14.0 163 months to 174 months
15.0 175 months to 186 months
16.0 187 months to 198 months
17.0 199 months to 210 months

 

Table 4.21 provides the BSA and CA within cell variances for the
five motor proficiency tests. While the CA classification procedure had

Table 4.21

The BSA and CA within cell variances for the five motor
proficiency tests

 

 

Sit-ups Pull-ups Standing ~ Vertical Shuttle
Long Jump Jump Run
ESA 800.28 10.33 93.82 8.66 0.87

CA 763.55 9.97 94.97 8.82 0.86

190

a lower within cell variance for sit-ups and pull-ups, the BSA classi-
fication procedure had a lower within cell variance for standing long
jump and vertical jump. And, the two classification procedures within
cell variances for shuttle run were equal.

Table 4.22 provides the subjects' distribution for the BSA and
CA groups. The BSA classification procedure dispersed the 1,578 subjects

Table 4.22

The subjects distribution for the BSA and CA groups

 

 

Sit-ups Pull-ups Standing Vertical Shuttle

Long Jump Jump Run

BSA 800.28 10.33 93.82 8.66 0.87
CA 763.55 9.97 94.97 8.82 0.86

 

more evenly through the age groups than the CA classification procedure.
Table 4.23 through 4.27 provide the mean value progressions for
the five motor proficiency tests in the BSA and CA groups.

Table 4.23

The mean value progressions for sit-ups in the BSA and CA groups
and the between-groups increments

 

 

 

 

Sit-ups
11.0 12.0 13.0 14.0 . 15.0 . 16.0 . 17.0
ESA 57.38 60.31 69.86 73.62 79.36 82.93 63.12
increments (2.93) (9.55) (3.76) (5.74) (3.57) (-19.81)
CA 51.59 54.81 67.45 76.13 85.25 79.94 91.20
increments (3.22) (12.64) (8.68) (9.12) (-5.31) (11.26)
Table 4.24

The mean value progressions for pull-ups in the BSA and CA groups
and the between-groups increments

 

 

 

Pull-ups
11.0 12.0 13.0 14.0 15.0 16.0 17.0
BSA 3.43 3.17 2.90 3.63 5.08 5.85 2.82
increments (-0.26) (-0.27) (0.73) (1.45) (0.77) (-3.03)
CA 2.82 2.47 2.84 3.29 3.97 4.01 2.51

increments (-0.35) (0.37) (0.45) (0.68) (0.04) (-l.50)

191

Table 4.25

The mean value progressions for standing long jump in the BSA
and CA groups and the between-groups increments

 

Standing Long Jump

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0 17.0
ESA 59.68 61.39 63.76 67.32 71.46 74.43 71.18
increments (1.71) (2.37) (3.56) (4.14) (2.97) (-2.25)
CA 57.44 59.86 64.52 68.19 72.21 73.13 80.60
increments (2.42) (4.66) (3.67) (4.02) (0.92) (7.47)
Table 4.26

The mean value progressions for vertical jump in the BSA and
CA groups and the between-groups increments

 

Vertical Jump

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0 17.0
ESA 12.24 12.89 13.67 14.81 15.86 16.49 - 15.65
increments (0.65) (2.28) (1.14) (1.05) (0.63) (-0.84)
CA 11.59 12.68 13.73 15.20 15.95 15.72 15.60
increments (1.09) (1.05) (1.47) (0.75) {-0.23) (-0.12)
Table 4.27 -

The mean value progressions for shuttle run in the BSA and
CA groups and the between-groups increments

 

Shuttle Run
11.0 12.0 13.0 14.0 15.0 16.0 17.0
BSA 11.32 11.15 10.95 10.70 10.53 10.56 11.06
increments (-0.17) (-0.20) (-0.25) (-0.17) (0.03) (0.50)
CA 11.46 11.33 10.07 10.64 10.62 10.54 10.04
increments (-0.13) (-0.44) (-0.25) {-0.02) (-0.08) (-0.50)

While for the 11.0, 12.0, 13.0, and 16.0 ages the ESA groups had
higher sit-ups mean values, for the 14.0, 15,0, and 17.0 ages the CA groups
had higher sit-up mean values. Whereas, the ESA groups increments advanced
positively until the final age group, the CA group increments stopped
advancing positively at the 16.0 age group.

For every age group, the BSA groups had higher pull-ups mean values.

Whereas the BSA groups advanced positively only between the 13.0 and 16.0

192

age groups, the CA groups advanced positively between the 12.0 and 16.0
age groups.

While for the 11.0, 12.0, and 16.0 ages the BSA groups had higher
standing long jump mean values, for the 13.0, 14.0, 15.0, and 17.0 ages
the CA groups had higher standing long jump mean values. Whereas the
BSA group increments advanced positively until the 17.0 age group, the
CA group advanced positively throughout the seven age groups.

While for the 11.0, 12.0, 16.0, and 17.0 ages the BSA groups had
higher vertical jump mean values, for the 13.0, 14.0, and 15.0 ages the
CA groups had higher vertical jump mean values. Whereas, the ESA group
increments advanCed positively until the final age group, the CA group
increments stopped advancing positively at the 16.0 age group.

While for the 13.0, 14.0, 16.0, and 17.0 ages the BSA groups had
higher shuttle run mean values, for the 11.0, 12.0, and 15.0 ages the CA
groups had higher Shuttle run mean values. Relative to shuttle run scores,
higher values indicated poorer scores. Whereas the ESA group increments
advanced negatively until the 16.0 age group, the CA group advanced
negatively throughout the seven age groups.

To answer question IV about whether or not a BSA classification pro-
cedure is superior to the CA classification procedure, the within-cell
variance analysis determined no clear-cut advantage for either classifi-
cation procedure. With the possible exception of the standing long jump
mean values, the mean values progression analyses also failed to demonstrate
a clear cut superiority for either the BSA or the CA grouping prOCedure.
However, Crampton, Dimoch, Greulich, and other researchers have shown
that while adolescent males are the same chronological age, they may be

of quite different biological ages. Therefore, the fact that the BSA

193

classification procedure dispersed the boys throughout the adolescent
age period demonstrates that it might be equating the boys on the basis
of biological age. Therefore, the BSA grouping might be demonstrated
that the differences in the motor proficiency abilities is due to higher
quality of effort or superior athletic competencies, and not biological
maturation differences.

While the preceding discussion completed the four questions posed
in this dissertation, the author pursued an additional classification
procedure that the results suggested. Even though the analysis is far
from complete, because that new procedure presented some intriguing
findings, Appendix 0 provides the concepts upon which the investigation

continues.

CHAPTER V

SUMMARY, CONCLUSIONS,

AND RECOMMENDATIONS

Summary

The purposes of this study were 1) to determine the proportion
of the variance in adolescent males' skeletal age explained by increments
in their chronological age, height, weight, and sexual maturation and
2) to determine how an estimate of skeletal age (BSA) compared with
chronological age (CA) when classifying adolescent males for motor
proficiency norms.

Increments in chronological age, height, and weight explained
45 percent, 20 percent, and 6 percent, respectively, of the variance in
skeletal age. The Sexual Maturation Indicator Value estimate (SMIV)
explained an additional 12 percent. In a regression equation estimate of
skeletal age (BSA), those four independent variables explained 83 percent
of the variance in the skeletal age of the adolescent males in this
sample.

By using the variances for the cells identically determined from
chronological age (CA) and BSA, neither index demonstrated a clear super-
iority. However, if researchers have been correct when they claim that
advanced biological maturation provides adolescent males with a certainty
of athletic superiority, then, because the BSA index is more likely to

eliminate that advantage, BSA may classify adolescent males for motor

194

195

proficiency norms better than CA.

Conclusions

 

The results suggest the following conclusions;

a) Without special training or competence, Boys Physical Education
teachers can reliably assess the sexual maturation of adolescent males.

b) The author's SMIV checklist provides reliable ratings for the
facial hair, axillary hair, pubic hair, penis, and scrotum and testes
development of adolescent males.

c) Because of the high degree of intercorrelation between penis
and scrotum and testes development, one of those ratings is not necessary
in the SMIV regression equation.

d) Because the reliability of the penis rating exceeded the scrotum
and testes rating, the SMIV regression equation that raters should use is;
SMIV = 126.47 + 1.74(X(F)) + 1.20(X(A)) + 2.75(X(PH)) + 6.68(X(P))

Where; (F) is the facial hair rating

(A) is the axillary hair rating
(PH) is the pubic hair rating, and
(P) is the penis rating.

e) Because 1) BSA classifies boys on the basis of biological matu-
ration and 2) BSA demonstrated within-cell variances equal to or slightly
better than chronological age, when classifying adolescent males for
motor proficiency norms, BSA is superior to chronological age.

f) The BSA regression equation that researchers should use is;

BSA = (-16.15) + .277(X(CA) + .212(X(H)) + .220(x(W)) + .594(X(SMIV))

Where; (CA) is chronological age in months

(H) is height in inches

(W) is weight in pounds, and

Ll "llo ‘LP, 4.

 

196

(SMIV) is the Sexual Maturation Indicator Value estimate in

months.

Recommendations _

 

Based on this investigation, the author recommends;

a) Replicate the first part of this study with a sample size
much larger than 30 subjects.

b) The high degree of intercorrelation between penis development
and scrotum and testes development indicates that the SMIV checklist
should combine those variables into a single genitalia variable.

c) As demonstrated in Appendix 0, workers should continue the
search for a classification index that eliminates the effect of biological
maturation on the outcome of athletic competitions for adolescent males.

d) Sponsors of athletic competitions for adolescent males should
re-examine their competitions in light of the effect of biological matu-
ration and either 1) support research for a classification system that
eliminates the effect of biological maturation, or 2) adjust the rules
of the competition such that they eliminate the effect of biological

maturation.

REFERENCES CITED

197

REFERENCES CITED

Acheson, Roy M. A method of assessing skeletal maturity from radiographs
J ANAT 88:498-508 1954.

American Association for Health, Physical Education and Recreation AAHPBR
Youth Fitness Test Manual Department of National Education
Association 1201 Sixteenth Street, N.W. Washington, D.C. 20036
1965.

Baldwin, Bird T. The growth of children from birth to maturity UNIV
IOWA STUDIES IN CHILD WELFARE 1(1):167-187 1921.

Baldwin, Bird T. and Lorle I. Stechler Mental growth curve of normal
and superior children UNIV IOWA STUDIES IN CHILD WELFARE 2(1):
1-61 1922.

Baldwin, Bird T. The determination of sex maturation in boys by a labor-
atory method J COMP PSY 8:39-43 1928.

Baldwin, Bird T. Laura M. Busby, and Helen V. Garside Anatomic growth of
children UNIV IOWA STUDIES IN CHILD WELFARE 4(1):1-88 1928.

Bardeen, C.R. The relation of ossification to physiological development
ARCH PHYS MED REHAB 2(6):1-8 1921.

Bayley, Nancy Skeletal X-rays as indicators of maturity J CONSULT PSY
4:69-73 1940.

Bayley, Nancy and Samuel R. Pinneau Tables for predicting adult height
from skeletal age: revised for use with the Greulich-Pyle hand
standards J PED 40:423-441 1952.

Bookwalter, Karl W. A critical evaluation of some of the existing means
of classifying boys for physical education RES QUART 10(3):ll9-
127 1939.

Bryan, A. Hughes and B.G. Greenberg Methodology in the study of physical
measurements of school children HUM BIOL 24:117-144 1952.

Buehl, Clara C. and S. Idell Pyle The use of age at first appearance of
three ossification centers in determining the skeletal status of
children J PED 21:335-342 1942.

Carter, Thomas M. Technique and devices used in radiographic study of
the wrist bones of children J ED PSY 17:237-247 1926.

198

Catchpole, H.R., W.W. Greulich, and R.T. Sollenberger Urinary excretion
of follicle stimulating hormone in young and adolescent boys
AM J PHYSIOL 123:32-33 1938.

Cattel, Psych Preliminary report on the measurement of ossification of
the hand and wrist HUM BIOL 6(3):454-471 1934.

Chang, K.S. Francis, Marjorie M.M.C. Lee, and S.J. Chan Sexual maturation
of Chinese boys in Hong Kong PED 34:804-811 1966.

Clarke, H. Harrison and Kay H. Peterson Contrast of maturational,
structural, and strength characteristics of athletes and non-
athletes 10 to 15 years of age RES QUART 32:163-176 1961.

Clarke, H. Harrison and James C.E. Harrison Differences in physical and
motor traits between boys of advanced, normal, and retarded
puberty RES QUART 33:13-25 1962.

Dimock, Hedley A research in adolescence CHILD DEV 6:177-195, 1935.

Dorfman, Ralph I., William Walter Greulich and Charles I Solomon The
excretion of androgenic and estrogenic substances in the urine
of children ENDO 21:741-743 1937.

Douglas, J.W.B. and J.M. Ross Age of puberty related to educational
ability, attainment, and school leaving age J CHILD PSY PSYCHIAT
5(4):185-196 1964.

Bdon, Gali Anovah: A fortran IV program to perform n-way analysis of
variance Occasional Paper 27 Office of Research Consultation
School for Advanced Studies College of Education Michigan State
University May 20, 1976.

Ellis, Richard W.B. Height and weight in relation to onset of puberty
in boys ARCH DIS CHILDH 21:181-189 1946.

Espenschade, Anna S. Motor performance in adolescence: including the
study of relationships with measures of physical growth and maturity
MONO SOC RES CHILD DEV 5(1):1-126 1940.

Espenschade, Anna S. Restudy of relationships between physical performances
of school children and age, height, and weight RES QUART 34:144-
153 1963.

Falkner, Frank Office measurement of physical growth PED CLIN N.A.
8(1):l3-18 1961.

Flory, Charles D. Sex differences in skeletal development CHILD DEV
6:205-212 1935.

Flory, Charles D. Predicting puberty CHILD DEV 6:1-6 1935.

199

Flory, Charles, D. Osseous development in the hand as an index of
skeletal development MONO SOC RES CHILD DEV 1(3):1-141 1936.

Freeman, Frank N. and Thomas M. Carter A new measure of the development
of the carpal bones and its relation to the physical and mental
development J ED PSYCH 15(5):257-270 1924.

Gallagher, J. Roswell and Carl C. Seltzer Maturity ratings and the
prediction of height of short l4-year-old boys PED 27:984-
989 1961.

Gates, Arthur I. The nature and educational significance of physical
status and of mental, physiological, social, and emotional
maturity J ED PSYCH 15:329-358 1924.‘

Greulich, William Walter, Ralph I Dorfman, Hubert R. Catchpole, Charles
I. Solomon, and Charles S. Culotta Somatic and endocrine studies
of puberal and adolescent boys MONO SOC RES CHILD DEV 7(3):l-84
1942.

Greulich, William W. The rational of assessing the deve10pmental status
of children from roentgenograms of the hand and wrist CHILD DEV
21:33-44 1950.

Gross, Elmer A. and Jerome A. Casciani The value of age, height, and
weight as a Classification device for secondary school students
in the seven AAHPBR Youth Fitness Tests RES QUART 33:51-58 1962.

Hellman, Milo Ossification of epiphysial cartilages in the hand AM J
PHYS ANTHRO 11:223-257 1928.

Hogben, Hamilton, J.A.H. Waterhouse, and Lancelot Hogben Studies in
puberty BRIT J SOC MED 2:29-42 1948.

Hunsicker, Paul and Guy G. Reiff AAHPBR Youth Fitness Test Manual 1976.

Hunt, Edward B., Grace Cooke, and J. Roswell Gallagher Somatotype and
sexual maturation in boys: a method of developmental analysis
HUM BIOL 30:73-91 1958.

Kelly, Harriet J. Anatomic Age and its relation to stature UNIV IOWA
STUDIES IN CHILD WELFARE 12(5):l-38 1937.

Krans, Hans and Rush P. Hirshland Muscular fitness and health JOPHER
24:17-19 1953.

Krans, Hans and Ruth P. Hirshland Minimum muscular fitness tests in
school children RES QUART 25:178-188 1954.

Kubitscheck, Paul E. Sexual development of boys with special reference
to the appearance of the secondary sexual characters and their
relationship to structural and personality types J NERV MBNT
DIS 76:425-451 1932.

200

Long, Eli and E.W. Caldwell Some investigations concerning the relation
between carpal ossification and physical and mental development
AM J DIS CHILD 1:113-138 1911.

Lowell, Frances and Herbert Woodrow Some data on anatomical age and its
relation to intelligence PED SEM 29:1-15 1922.

Marshall, Michael G. An investigation of the association between sexual
maturation and physical growth and motor proficiency in males
M.A. Degree Thesis Michigan State University 1967.

Marshall, W.A. and J.M. Tanner Variations in the pattern of pubertal
changes in boys ARCH DIS CHILDH 45:13-23 1970.

McBlhannon, Joseph C. The invalidity of the inspectional method of
ranking radiographs of the carpal area of the wrists J ED PSY
17:77—85 1926.

McSweeney, Maryellen and William H. Schmidt Quantal response techniques
for random predictor variables J ED STAT 2(4):257-288 1977.

Millman, Jason and Gene V. Glass Rules of thumb for writing the ANOVA
table J ED MBA 4(2):41-51 1967.

Meyzi, Olcay, Hulya Alp, Ayfer Yalcindag, Suleyman Yakacikli and Alper
Orphon Sexual maturation in Turkish boys ANN HUM BIOL 2(3):
251-259 1975.

Nicolson, Arline B. and Charles Hanley Indices of physiological maturity:
derivation and interrelationships CHILD DEV 24(1):3-38 1953.

Prescott, Daniel Alfred The determination of anatomical age in school
children and its relation to mental development HARV MONO IN
EDUCATION Series 1 Number 5 1923.

Pryor, J.W. Development of the bones of the hand: as shown by the X-ray
method BULL STATE COLL KENTUCKY 2(5):1-48 1905.

Pryor, J.W. Ossification of the epiphyses of the hand: X-ray method
BULL STATE COLL KENTUCKY 3(4):1-35 1906.

Pryor, J.W. The hereditary nature of variation in the ossification of
bones ANAT REC 1:84-88 1907.

Pryor, J.W. The chronology and order of ossification of the bones of
the human carpus BULL STATE UNIV KENTUCKY 1(7):1-26 1908.

Pryor, J.W. Some observations on the ossification of the bones of the
hand BULL UNIV KENTUCKY 8(11):1-67 1916.

Pryor, J.W. Differences in the time of development of centers of ossi-
fication in the male and female Skeleton ANAT REC 25:257-273 1923.

201

Pryor, J.W. Time of ossification of the bones of the hand of the male
and female and union of epiphyses with the diaphyses AM J PHYS
ANTHROP 8(4):401-410 1925.

Pryor, J.W. Difference in the ossification of the male and female
skeleton J ANAT 62:499-506 1928.

Pyle, S. Idell and Camille Menino Observations on estimating skeletal
age from the Todd and the Flory bone Atlases CHILD DEV 10:27-34
1939.

Pyle, S. Idell, Alice M. Waterhouse, and William Walter Greulich A
radiographic standard of reference for the growing hand and wrist
The Press of Case Western Reserve University Year Book Medical
Publishers, Inc. 35 East Wacker Drive Chicago 1971.

Rarick, Lawrence, Ionel Rapaport, and Vern Seefeldt Bone development
in Down's disease AM J DIS CHILD 107:7-13 1964.

Rarick, G. Lawrence and Frank L. Smoll Stability of growth in strength
and motor performance from childhood to adolescence HUM BIOL
39:295-306 1967.

Reynolds, Earle L. and Janet V. Wines Physical changes associated with
adolescence in boys AMER J DIS CHILD 82:529-546 1948.

Richey, Herman G. The relation of accelerated, normal, and retarded
puberty to the height and weight of school children MONO SOC RES
CHILD DEV 2(1):l-67 1937.

Rotch, Thomas Morgan Chronologic and anatomic age in early life JAMA
51:1197-1205 1908.

Rotch, Thomas Morgan A study of the development of the bones in childhood
by the Roentgen method, with the View of establishing a developmental
index for the grading of and the protection of early life ASSN AM
PHYSICIANS TRANS 24:603-624 1909.

Rotch, Thomas Morgan Roentgen ray methods applied to the grading of
early life AM PHYS ED REV 15:396-421 1910.

Rotch, Thomas Morgan and Harold Wellington Smith A study of the develop-
ment of the epiphyses of the hand and wrist for the purpose of
classifying the cadets at Annapolis ASSN AM PHYSICIANS TRANS
25:200-211 1910.

Sawtell, Ruth Otis Ossification and growth of children from one to eight
years of age AM J DIS CHILD 37:61-87 1929.

Sawtell, Ruth Otis Sex differences in the bone growth of young children
AM J PHYS ANTHROP 12:293-302 1929.

Scammon, Richard B. and Gordon H. Scott The technique of determining
irregular areas in morphological studies ANAT REC 35:269—277 1927.

202

Schiefley, Verda M. and William H. Schmidt Jeremy D. Finn's Multi-
variance, Univariate and Multivariate Analysis of Variance,
Covariance, and Regression Occasional Paper 22 Office of
Research Consultation School for Advanced Studies College of
Education Michigan State University October 1973.

Schonfeld, William A. and Gilbert W. Beebe Normal growth and variation
in the male genitalia from birth to maturity J UROL 48:759-
777 1942.

Schonfeld, William A Primary and secondary sexual characteristics
AM J DIS CHILD 65:535-549 1943.

Schonfeld, William A. Management of male pubescence JAMA 12l(3):177-
182 1943.

Seefeldt, Vernal D. A longitudinal study of physical growth and develop-
ment in Down's syndrome unpublished Doctoral dissertation
University of Wisconsin 1966. ‘

Smith, Harold W. Rotch method of roentgenographic age determination
US NAVAL MED BULL 7(1):1-20 1913.

Stevenson, Paul H. Age order of epiphyseal union in man AM J PHYS
ANTHRO 7(1):53-93 1924.

Tanner, J.M. Growth at adolescence Charles C. Thomas Springfield,
Illinois 1955.

APPENDICES

203

APPENDIX A

SEXUAL MATURATION INDICATOR VALUE (SMIV) CHECKLIST

204

SEXUAL MATURATION INDICATOR VALUE (SMIV) CHECKLIST

DATE

 

 

SCHOOL

 

 

Facial Hair (Ask the boys whether or not they shave.)

 

 

No facial hair

Unpigmented hair at the sideburns or the upper lip

Downy, lightly pigmented hair at the sideburns or the upper lip
Lightly pigmented, coarse hair at the sideburns or the upper lip
Terminal sideburns or mustache

AxillaryfiHair

 

 

No hair

Slight, unpigmented hair

Downy, lightly pigmented hair in one or both axilla

Small area (5 to 1 inch) of pigmented, coarse hair in both axilla
Coarse, curly, well-developed hair covers both axilla

 

 

Pubic Hair
No hair
Downy, unpigmented hair
Sparse, pigmented hair, note that it is straight
Small area (B to 1 inch) of curly, coarse pigmented hair
Curly, coarse, pigmented hair covers the pubic triangle
Penis Development

 

 

Early childhood size and configuration

Some slight enlargement evident

Increase in the shaft length, glans proportionately small
Proportional increase in length and diameter, but not to adult size
Adult size and configuration

Scrotum and Testes Development

 

SMIV

Early childhood size and configuration

Scrotum and testes visibly enlarged, one testes may_show lowered
One testes definitely has descended lower than the other testes

Scrotal sac has darkened, but the size is still less than adult

Adult size and configuration of the scrotal sac

formula: SMIV X + (Y - 4X)/10

Where X is the Pubic Hair rating and Y is the sum of the Facial Hair,
Axillary Hair, Penis Development, and Scrotum and Testes Development
ratings.

205

APPENDIX B

THE PROCEDURES THAT THE RADIOLOGISTS USED TO X-RAY THE HANDS
AND WRISTS OF THE SUBJECTS '

206

The Radiology staff at Michigan State University's Olin Health
Center used their general radiographic unit to X-ray both hands and
wrists of the 30 boys. The radiology assistant asked the boys to place
both hands on the X—ray plate, spread their fingers slightly, and gently
push their palms to the surface. With the General Radiographic Unit
42 inches above the X-ray plate, the MAX set at 60 milliroentgens and,
depending on the thickness of the youngster's hand, they set the milli-
voltage between 66 and 70, and exposed the boys' hands and wrists to the
X-ray cassette. To prevent fuzziness due to the boys' moving their
hands and to minimize the X-ray exposure time, the radiologist set the

exposure timer to its highest setting or less than one-half second.

207

APPENDIX C

THE PROCEDURES THAT PROFESSOR SEEFELDT USED TO ASSESS THE
HAND-WRIST X-RAYS AND HIS TEST-RETEST RELIABILITIES

208

While a graduate student at the University of Wisconsin, Professor
Vern Seefeldt assisted Lawrence Rarick and Ionel Rapaport in a study of
Down's disease (Rarick, Rapaport, and Seefeldt, 1964). Seefeldt assessed
the hand-wrist X-rays. To determine Seefeldt's ability to assess hand-
wrist X-rays Rarick and Rapaport required that he compare his readings
with experts. They found that Seefeldt showed no systematic bias.
Professor Seefeldt's test-retest correlations were .988 for 10 year old
children and .997 for 14 year old children. In his doctoral dissertation,
Professor Seefeldt's test-retest correlations were .978 for 9-10 year old
children and .990 for 13-14 year old children (Seefeldt, 1966). From
those experiences, Professor Seefeldt became a skilled assessor of hand-
wrist X-rays.

Professor Seefeldt uses the following methodology;

1) Search through the Greulich-Pyle male standards until you locate
the two standards between which the X-ray fits,

2) Following the order listed in Figure C-l, assess each of the
twenty-eight sites individually. Interpolate between the ages of the
two standards and assign a month value to each site.

3) Average the values you assign to the twenty-eight sites for your
skeletal age estimate.

Professor Seefeldt's skeletal age test-retest correlation for the
30 X-rays in this study was .96. Table C.l provides his test-retest

correlations for each of the twenty-eight sites.

209

Table C.1

Professor Seefeldt's test-retest correlations for the twenty-

eight sites

 

 

N Site Correlation
1 Distal radial epiphysis .933
2 Distal ulnar epiphysis .921
3 Capitate .889
4 Hamate .895
5 Triquentral .909
6 Lunate .921
7 Schaphoid .937
8 Trapezium .931
9 Trapezoid .942
10 First metacarpal epiphysis .929
11 Second metacarpal epiphysis .938
12 Third metacarpal epiphysis .940
13 Fourth metacarpal epiphysis .943
14 Fifth metacarpal epiphysis .948
15 Bpiphysis of the first proximal phalange .965
16 Bpiphysis of the second proximal phalange .946
17 Bpiphysis of the third proximal phalange .947
18 Bpiphysis of the fourth proximal phalange .945
19 Bpiphysis of the fifth proximal phalange .949
20 Bpiphysis of the second middle phalange .950
21 Bpiphysis of the third middle phalange .948
22 Bpiphysis of the fourth middle phalange .956
23 Bpiphysis of the fifth middle phalange .938
24 Bpiphysis of the first distal phalange .958
25 Bpiphysis of the second distal phalange .952
26 Bpiphysis of the third distal phalange .963
27 Bpiphysis of the fourth distal phalange .964
28 Bpiphysis of the fifth distal phalange .960

 

210

APPENDIX D

THE PROCEDURES THAT THE RADIOLOGISTS USED TO MEASURE THE
HEIGHT AND WEIGHT OF THE SUBJECTS

211

Height

Barefoot, the boys stood free on the Olin Health Center's sliding
anthropometer scale while a radiology assistant lowered the horizontal
bar to the midsagittal vertex of the head. Another radiology assistant
recorded the height to the nearest one-quarter inch.

Weight

Naked, except for their underwear, the boys stood on the Olin
Health Center's balance arm weight scale while a radiology assistant
determined their weight. Another radiology assistant recorded the

weight to the nearest one-half pound.

212

APPENDIX E

THE PROCEDURES THAT THE FOUR TEACHERS AND THE AUTHOR USED TO
ASSESS THE SEXUAL MATURATION OF THE 30 SUBJECTS WHO
REPRESENTED A RECTANGULAR DISTRIBUTION OF SEXUAL

MATURATION

213

Sexual Maturation

 

Using the SMIV checklist, the two East Lansing, Michigan and two
Mason, Michigan physical education teachers and the author assessed the
sexual maturation of the 30 boys twice within a thirty minute time period.
East Lansing Physical Education Director James M. Oestreich randomly
ordered the boys for each assessment. All five raters had some previous
experience with the SMIV checklist.

Naked, except for a towel wrapped around their waists, the boys
carried ten SMIV checklists. Each rater sat in his own well-lighted
assessment station. When the boys entered an assessment station, they
handed one checklist to the rater and stood on a tape line approximately
three feet from the rater. First, the raters asked the boys whether or
not they had shaved that day. Then, the raters asked the boys to hold
their towels in their right hand, raise both of their arms over their head,
and turn their head to the left and then to the right. While the boys
continued to look to the right, the raters completed the checklist. To

assess one boy took less than fifteen seconds.

214

APPENDIX F

THE PROCEDURES THAT THE TEACHERS USED TO MEASURE THE HEIGHT
AND WEIGHT OF THE SUBJECTS

215

Height

Barefoot, the East Lansing boys stood free on the physical
education department's sliding anthropometer scale while the physical
education teacher lowered the horizontal bar to the midsagittal vertex
of their heads. Assisting the teacher, a student recorded each boy's
height to the nearest one-quarter inch.

Barefoot, the Mason boys stood with their heels and backs against
a wall while the physical education teacher placed a level on the mid-
sagittal vertex of their heads. Assisting the teacher, a student recorded
each boy's height to the nearest one-quarter inch.

Weight

Naked, except for either their underwear or athletic supporters,
the East Lansing boys stood on the physical education department's balance
arm weight scale while the teacher determined their weight. Assisting
the teacher, a student recorded each boy's weight to the nearest one-half
pound.

Naked, except for either their underwear or athletic supporter,
the Mason boys stood on a calibrated bathroom scale while the teacher
determined their weight. Assisting the teacher, a student recorded each

boy's weight to the nearest one-half pound.

216

APPENDIX G

THE PROCEDURES THAT THE TEACHERS USED TO TEST THEIR STUDENTS ON
THE AAHPER'S MOTOR PROFICIENCY TESTS

217

Sit-ups

While in a prone position on the floor, the boys extended their
legs and abducted their feet to shoulder width. With a partner grasping
their ankles and holding their heels against the floor, the boys placed
their hands behind their neck and interlaced their fingers. Upon the
teacher's command of "Go," the boys sat up, touched an elbow to the
knee on the opposite side, and returned to the prone position. By alter-
nating their elbows and the knees, the boys repeated this movement to a
maximum of sit-ups. The boys could not rest between individual
sit-ups. If a boy rested too long between sit-ups or stopped entirely,
the teacher told the boy's partner to record the boy's sit-up score on

the class sit-up sheet.

Pull-ups

The East Lansing boys used a hand-over-hand traveling ladder for
their pullups. The Mason boys used a chinning bar for their pullups.
With their hands pronated (the backs of their hands toward their face)
and their feet off of the floor, the boys assumed a straight arm-straight
leg hanging position. Upon their teacher's or another student's command
of "Go," by flexing only the elbow joint, the boys elevated their body
until their chin cleared the handhold bar. Each time the boys' chin
cleared the bar, they received credit for one pullup. To begin another
pullup, the boys had to lower their body until they fully extended their
elbow joints. When the boys could no longer raise their chin above the
bar, they released the handhold bar and recorded their pullup score on

the class score sheet.

218

Standing Long Jump

 

«The physical education teacher placed perpendicular strips of
athletic tape on the gymnasium floor. A three foot strip of tape
functioned as the standing long jump starting line. A perpendicular strip
of tape extended ninety-six inches from the middle of the starting tape,
and it functioned as the measurement tape. Beginning thirty-six inches
from the starting tape, the teacher marked one inch intervals up to
ninety-six inches.

Standing with their toes to the starting tape and their feet
abducted to shoulder width, the boys straddled the measurement tape.
By flexing their knees and simultaneously swinging their arms backward
the boys began their jumping action. By simultaneously extending their
knees, plantar flexing their ankles, and throwing their arms forward and
upward, the boys jumped. Highly skilled boys learned to increase their
distance by extending their arms just before landing. The teacher deter-
mined the distance from the starting tape to their nearest heel upon
contact. They recorded the best score of three jumps.

Shuttle run

 

The physical education teacher placed parallel strips of athletic
tape thirty feet apart on the gymnasium floor.. One trip of tape acted as
the Starting Tape and the other strip of tape acted as the Block Tape.
Behind the Block Tape the teacher placed two 2x2x4 inch blocks of wood.
With the teacher or an experienced student operating the stopwatch, the
boys stood behind the starting line. Upon the timer's command of "Go,"
the boys sprinted to the Block Tape and grabbed one block, and sprinted
back to the Starting Tape. After placing the block behind the Starting

Tape, they sprinted back to the Block Tape and grabbed the second block,

219

then, by sprinting back past the Starting Tape, they completed the
shuttle run test. The timer determined their time to the nearest one-
tenth of a second. They recorded the best score of three runs.

Vertical Jump

 

Beginning at a height of five feet off the floor, the physical
education teacher placed a six foot strip of tape vertically on the
gymnasium wall with markings one inch apart. To determine their reaching
height, the boys stood flatfooted with one side of their body against the
wall. With the hand closest to the wall they reached as high as they
could. The number to which they reached became their reaching height
number. To determine their jumping height, the boys moistened the finger-
tips of the same hand, then by pliometrically extending their knees and
swinging their arms backward the boys began their jumping action. By
simultaneously throwing their arms upward, extending their knees, and
plantar flexing their ankles, the boys jumped as high as they could. At
the peak of their jump, the boys touched the measurement tape with the
moistened fingertips. The height where the fingertips touched the tape
became the boys' jumping height. By subtracting reaching height from
jumping height, the teacher or an experienced student assistant determined

the vertical jump score. They recorded the best score of three jumps.

220

APPENDIX H

THE PROCEDURES THAT THE TEACHERS USED TO ASSESS THE SEXUAL
MATURATION OF THEIR SUBJECTS

221

Using the SMIV checklist, physical education teachers from East
Lansing and Mason Junior High Schools assessed five primary and secondary
sex characteristics. In conjunction with their AAHPBR Youth Fitness
testing week, the physical education teachers daily asked seven or eight
boys to report to them after physical education class. In East Lansing,
the physical education office served as the station for the assessment
of sexual maturation. In Mason, the towel disposal room beside the
showers served as the station for the assessment of sexual maturation.
Both locations were well lighted.

Naked, except for a towel wrapped around them, the boys entered
the station and stood on a tape line approximately three feet from the
teachers. After writing the boy's name on the SMIV checklist, the teachers
asked the boys whether or not they shaved that day. Then, the teachers
asked the boys to hold the towel in their right hand, to raise both arms
over their head, and to turn their head to the left and then to the right.
While the boys continued to look to the right, the teachers completed the
checklist. To assess one boy and explain his SMIV score required less

than thirty seconds.

222

APPENDIX I

THE STEP-WISE REGRESSIONS FOR THE REORDERED CHRONOLOGICAL
AGE, HEIGHT, AND WEIGHT INDEPENDENT VARIABLES

223

Table I-l

After height, the percent of additional variance accounted for by
chronological age and the significance level

 

 

 

Percent of Additional Significance
Variance Level
Height 61.931
Chronological Age 2.856 P less than .1506
Table I-2

After chronological age, the percent of additional variance accounted
for by weight and the significance level

 

 

 

Percent of Additional Significance
Variance Level
Chronological Age 44.937
Weight 23.191 P less than .0002
Table I-3

After weight, the percent of additional variance accounted for by
chronological age and the significance level

 

 

 

Percent of Additional Significance
Variance Level
Weight 49.559
Chronological Age 18.569 P less than .0005
Table I-4

After height, the percent of additional variance accounted for by
weight and the significance level

 

 

 

Percent of Additional Significance
Variance Level
Height 61.931
Weight 3.978 P less than .0872
"I Table I-5

After weight, the percent of additional variance accounted for by
height and the significance level

 

 

Percent of Additional Significance ‘
Variance Level
Weight 49.559

Height 16.350 P less than .0013

224

Table I-6

After chronological age and weight, the percent of additional var-
iance accounted for by height and the significance level

 

 

Percent of Additional Significance
Variance Level
Chronological Age 44.937
Weight 23.191 P less than .0002
Height 2.625 P less than .1387

 

W

After accounting for their chronological age contribution, the
boys' height did have a significant additional contribution unique to it.
However, after accounting for the contribution of height, chronological
age did not have a significant additional contribution unique to it. The
reason for this is that chronological age is very highly correlated with
height, such that, after accounting for the common contribution of height
and chronological age, chronological age has nothing left to add.

Regardless ofthe order of chronological age and weight, both
variables had a significant contribution unique to them.

After accounting for height, weight did not have a significant
addition. The high intercorrelation between height and weight also
caused this finding. However, after accounting for weight, height did
have a significant additional contribution. Height has a contribution

that is unique to it beyond its common contribution with weight.

225

APPENDIX J

THE PROCEDURES FOR DETERMINING A SEXUAL MATURATION
INDICATOR VALUE(SMIV) REGRESSION EQUATION

226

Before determining a SMIV regression equation, we should examine

the sample correlation matrix for the five sexual characteristics (Table

 

 

J-l).
Table J-l
The correlation matrix for the five sexual characteristics

Facial Axillary Pubic Penis Scrotum

Hair Hair Hair Dev. Dev.
Facial Hair 1.000 1
Axillary Hair .849 1.000 . ‘
Pubic Hair .697 .670 1.000 g .
Penis Dev. .710 .671 .931 1.000
Scrotum Dev. .787 .731 .902 .954 1.000

 

 

Ideally, in a regression equation, the independent variables should
not be highly intercorrelated. However, because those variables relate
to some particular variable, in this case sexual maturation, they frequently
are highly intercorrelated. When two independent variables in a regression
equation are highly intercorrelated, sometimes the contribution of the
variable listed first devours all of the contribution of the variable
listed second. In this correlation matrix, the high correlation of penis
development and scrotum and testes development (.954) and pubic hair with
penis development (.931) and scrotum and testes development (.902) might
effect the SMIV regression equation.

To obtain a regression equation for estimating SMIV, the Multivariate
Regression Analysis estimated the dependent variable (skeletal age) from
the independent variables (the five sexual characteristics). Table J-2
contains the raw regression coefficients and mean values that the Multi-
variate Regression Analysis determined for the SMIV regression equation.

The mean deviated SMIV regression equation is:

SMIV = fi(SA) + 8(F)(X(F)- fi(F)) + 8(A)(X(A)- fi(A)) + 8(PH)(X(PH) - fi(PH)) *-

émmm- (1(8)) + 8(S)(X(S)-fl(S))

227

Table J-2

The raw regression coefficients and mean values that the
Multivariate Regression Analysis determined for
the SMIV regression equation

M

 

 

Raw Regression Mean
:: Coefficients (8) Values(fi) !
Facial Hair (F) 1.94 2.04
Axillary Hair (A) 1.21 1.86 (
Pubic Hair (PH) 2.77 2.95
Penis Dev. (P) 7.27 3.02
Scrotum Dev. (S) J -l.O6 2.98
Skeletal Age (SA) ---- g_160.53

 

The mean deviated SMIV regression equation is:
snxv = n(sA) + é<r>(x(r)-fi(r) + 8(A)(X(A))-G(A) + 8(PH)(x(PH)-fi(PH)) +

809) (x(m-fien + 8(S(X(S)-fi(s>)
To convert this mean deviated SMIV regression equation to non-mean deviated,
it is necessary to solve the following SMIV 8(0) equation;
SMIV 8(0) = fi(SA) - é(F)fi(F) —é(A)fi(A) - é(PH)fi(PH) - 8(P)fi(P) -

é(s)i1(s)

After substituting the B and U values from Table J-2 and solving the
equation, SMIV 8(0) equalled 127.35.

To demonstrate the effects of increasing the ratings of the five
sexual characteristics individually and to determine the common intervals
and range of the SMIV regression equation, Table J-3 substitutes the
ratings 1 through 5 for the sexual characteristics values. While a one-
level increase in facial hair, axillary hair, pubic hair, and penis
development represented monthly increases of 1.94, 1.21, 2.77, and 7.27,
respectively, a one-level increase in scrotum and testes development
represented a loss of 1.06 in the SMIV regression equation.

To have a decrease in SMIV when scrotum and testes development

increased demonstrated a problem in the regression equation. Why did

228

Table J-3

Substituting selected ratings for the five sexual characteristics
values into the SMIV regression equation

 

8(F) (X(F))+B(A)(X(A))+8(PH)(X(PH)+8(P)(x(P))+8(X)(X(S))+8 (0) = SMIV
1.94 (1) + 1.21 (1) + 2.77 (l) 7.27 (l) -l.06 (l) + 127.35=139.48

1.94 (2) + 1,21 (1) + 2.77 (1) 7.27 (1) -l.06 (l) + 127.35=141.42
1.94 (1) + 1.21 (2) + 2.77 (l) 7.27 (1) -l.06 (l) + 127.35=140.69
1.94 (1) + 1.21 (1) + 2.77 (2) 7.27 (l) -1.06 (l) + 127.35=142.24
1.94 (1) + 1.21 (1) + 2.77 (1) 7.27 (2) -l.06 (1) + 127.35=146.75
1.94 (l) + 1.21 (l) + 2.77 (l) 7.27 (1) -1.06 (2) + 127.35=138.42
1.94 (2) + 1.21 (2) + 2.77 (2) 7.27 (2) -1.06 (2) + 127.35=151.61
1.94 (3) + 1.21 (3) + 2.77 (3) 7.27 (3) -1.06 (3) + 127.35=163.74
1.94 (4) + 1.21 (4) + 2.77 (4) 7.27 (4) '-1.06 (4) + 127.35=175.873
‘1g94 (5) + 1.21 (5) + 2.77 (5) 7.27 (5) -1.06 (5) + 127.35=188.00

 

the beta weight for scrotum and testes development have a negative value?
High intercorrelation between independent variables can cause a negative
beta weight for one of the variables. Statisticians refer to this pheno-
menon as multi-colinearity. Multi-colinearity prevents the regression
analysis from determining the singular effects of the two highly corre-
lated independent variables.

Recall that the variables of scrotum and testes development and
penis development were highly correlated (.954). The procedure that
confirms the multi-colinearity problem is to eliminate one and then the
other of the variables from the list of independent variables and examine
the new beta weights. If the beta weights in those two lists are positive,
then the intercorrelation did cause the multi-colinearity problem.

Table J-4 contains the raw regression coefficients for the two
lists of independent variables. Because all of the raw regression co-
efficients are positive, the high intercorrelation between the penis
development variable and the scrotum and testes development definitely
caused the negative beta weight for the scrotum and testes development

variable.

229

Table J-4

Raw regression coefficients for the regression analyses minus
the penis development variable and minus the scrotum
and testes development variable

 

Minus Penis Dev. Minus Scrotum Dev.

 

Raw Regression Coefficients Raw Regression Coefficients
Facial Hair 1.74 0.88 -
Axillary Hair 1.20 1.26
Pubic Hair 2.75 5.46
Penis Dev. 6.68 ----
Scrotum Dev. ---- 6.05

 

The solution for multi-colinearity is to eliminate one of the
variables. Because the two variables were highly intercorrelated, they
contributed the same manifestation to the regression equation and to
eliminate one of the variables does not weaken the regression equation.
Therefore, the question became whether to eliminate the penis development
variable or the scrotum and testes development variable.

The mean-deviated minus scrotum and testes development (MSD)
regression equation is;

(MSD) SMIV = fi(SA) + 8(F)(X(F)-8(F)) + é(A)(X(A)-fi(A)) +

8(PH)(X()(PH)- fi(PH)) + 8(P)(x(P)-fi(P))
To convert this mean deviated (MSD) SMIV regression equation to a non-
mean deviated, it is necessary to solve the following (MSD) SMIV 8(0)
equation;
(MSD) snxv 8(0) = fi(SA) - é<r>fi<r> - 8(A)fi(A) - é<pH)fi(pH) - é<p>fi<p>
After substituting the Table J-4 8's and the Table J-2 8's and solving
the equation, (MSD) SMIV 8(0) equalled 126.47.

To demonstrate the effects of increasing the ratings of those four
sexual characteristics individually and to determine the common intervals
and range of the (MSD) SMIV regression equation, Table J-5 substitutes

the ratings 1 through 5 for the sexual characteristics.

230

Table J-S

Substituting selected ratings for the four sexual character-
istics values into the (MSD) SMIV regression equation.

 

(F) (X(F)) + (A)(X(A)) + (PH)(X(PH)) + (P)(X(P)) + 8(0) =(MSD) SMIV

 

1.74 (1) 1.20 (l) 2.75 (l) 6.68 (1) 126.47 = 138.84
1.74 (2) 1.20 (1) 2.75 (1) 6.68 (1) 126.47 = 140.58
1.74 (1) 1.20 (2) 2.75 (1) 6.68 (1) 126.47 = 140.04
1.74 (1) 1.20 (1) 2.75 (1) 6.68 (1) 126.47 = 141.59
1.74 (l) 1.20 (1) 2.75 (1) 6.68 (1) 126.47 = 145.42
1.74 (2) 1.20 (2) 2.75 (1) 6.68 (1) 126.47 = 151.21
1.74 (3) 1.20 (3) 2.75 (1) 6.68 (1) 126.47 = 163.58
1.74 (4) 1.20 (4) 2.75 (l) 6.68 (1) 126.47 = 175.95
1.74 (5) 1.20 (S) 2.75 (1) 6.68 (1) 126.47 = 188.32

 

 

A one level increase in facial hair, axillary hair, pubic hair,
and penis development represented monthly iHcreases of 1.74, 1.20, 2.75,
and 6.68, respectively in the (MSD) SMIV.

The mean deviated minus penis development (MPD) SMIV regression
equation is;

(MPD) SMIV = fi(SA) + 8(F)(X(F)-fi(F)) + é(A)(x(A))-G(A)) + é<PH)(X(PH)-
fi(pH» + 8(S)(x(s)fi(s))

To convert this mean deviated (MPD) SMIV regression equation to a non-

mean deviated, it is necessary to solve the following (MPD) SMIV 8(0)

equation:

(MPD) SMIV 8(0) = fi(SA)- 8(E)fi(r) - 8(A)fi(A) - é(pH)fi(pH) - 5(5) 8(5)

By substituting the Table J-4 (MPD) 8's and the Table J-2 fi's and solVing

the equation, the (MPD) SMIV 8(0) equalled 122.24.

To demonstrate the effects of increasing the ratings of those
four sexual characteristics individually and to determine the common
intervals and range of the (MPD) SMIV regression equation, Table J-6
substitutes the ratings 1 through 5 for the sexual characteristics.

A 1 level increase in facial hair, axillary hair, pubic hair, and

scrotum and testes development represented monthly increases of 0.88,

Table J-6

231

Substituting selected ratings for the four sexual character

istics values into the (MPD) SMIV regression equation

 

 

 

B(E)(x(E)) + 8(A)(X(A)) + B(PH)(X(PH) + 8(S)(X(S)) + 3(0) = (MPD)SMIV
0.88 (1) T 1.26 (1) + 5.46 (l) + 6.05 (l) + 122.24 = 135.59
0.88 (2) + 1.26 (1) + 5.46 (1) + 6.05 (1) + 122.24 = 136.77
0.88 (1) + 1.26 (2) + 5.46 (l) + 6.05 (l) + 122.24 = 137.15
0.88 (1) + 1.26 (1) + 5.46 (2) + 6.05 (1) + 122.24 = 141.35
0.88 (1) + 1.26 (1) + 5.46 (1) + 6.05 (2) + 122.24 = 141.94
0.88 (2) + 1.26 (2) + 5.46 (2) + 6.05 (2) + 122.24 = 149.54
0.88 (3) + 1.26 (3) + 5.46 (3) + 6.05 (3) + 122.24 = 163.19
0.88 (4) + 1.26 (4) + 5.46 (4) + 6.05 (4) + 122.24 = 176.34
0.88 (5) + 1.26 (5) + 5.46 (5) + 6.05 (5) + 122.24 = 190.49
1.26, 5.46, and 6.05, respectively in the (MPD) SMIV.
Table J-7 contains the SMIV, (MDS) SMIV, and (MPD) SMIV values.
Table J-7

Comparing the SMIV, (MSD) SMIV, and (MPD) SMIV values

 

Selected Values

 

 

(X(F)) (X(A)) (X(PH)) (X(F)) (X(S)) SMIV (MSD)SMIV (MPD)SMIV
(1) (1) (1) (1) (1) 139.48 138.84 135.89
(2) (1) (1) (1) (1) 141.42 140.58 136.77
(1) (2) (1) (1) (1) 140.69 140.04 137.15
(1) (1) (2) (1) (1) 142.24 141.59 141.35 '
(1) (1) (1) (2) (1) 146.75 145.42 ----
(1) (1) (1) (1) (2) 138.42 ---0 141.95 1
(2) (2) (2) (2) (2) 151.61 151.21 149.54 *
(3) (3) (3) (3) (3) 163.74 163.58 163.19
(4) (4) (4) (4) (4) 175.87 175.95 176.84
(5) (5) (5) (5) (5) 188.00 188.32 190.94

 

Comparing the (MSD) SMIV and (MPD) SMIV values failed to demonstrate

a Clear superiority for either regression equation. Therefore, the rater
reliability for measuring the penis development and scrotum and testes
variables will determine whether to use (MSD)SMIV or (MPD)SMIV. Nevertheless,
because the original SMIV regression equation contained a negative beta
weight for the scrotum and testes development did not prevent its use

when analyzing the additional contribution unique to SMIV.

232

APPENDIX K

THE STEP-WISE REGRESSIONS FOR THE REORDERED CHRONOLOGICAL
AGE, HEIGHT WEIGHT, AND SMIV INDEPENDENT VARIABLES

233

Table K-l

The step-wise regression for chronological age and SMIV

 

 

 

Percent of Additional Significance
Variance Level
Chronological Age 44.937
SMIV 30.031 P less than .0001
Table K-2

The step-wise regression for SMIV and chronological age

 

 

 

Percent of Additional Significance
Variance Level
SMIV 68.811
Chronological Age 6.158 P less than .0158
Table K-3

The step-wise regression for height and SMIV

 

 

 

Percent of Additional Significance
Variance Level
Height 61.934
SMIV 15.245 P less than .0003
Table K-4

The step-wise regression for SMIV and height

 

 

 

Percent of Additional Significance
Variance Level
SMIV 68.881
Height 8.377 P less than .0040
Table K-S

The step-wise regression for weight and SMIV

 

 

Percent of Additional Significance
Variance Level
Weight 49.559

SMIV 28.649 P less than .0001

234

Table K—6

The step-wise regression for SMIV and weight

 

 

Percent of Additional Significance
Variance Level
SMIV 68.811
Weight 9.397 P less than .0021

 

Regardless of the order of chronological age and SMIV, both
contributed a significant additional proportion of the variance. Regard-
less of the order of height and SMIV, both contributed a significant
additional proportion of the variance. And, regardless of the order of
weight and SMIV, both contributed a significant additional proportion of
the variance. The proportion of the variance in skeletal age when chrono-
logical age and SMIV were the independent variables, when height and SMIV
were the independent variables, and when weight and SMIV were the inde-

pendent variables was 75, 77, and 78 percent, respectively.

235

APPENDIX L

THE STATISTICAL ANALYSIS TO DETERMINE WHETHER OR NOT THE
878 EAST LANSING SUBJECTS DIFFERED FROM THE 700 MASON SUBJECTS

236

The Multivariate Analysis of Variance computer program analyzed
whether or not the two school's data were significantly different from
one another (Scheifley and Schmidt, 1973). In the Multivariate Analysis
of Variance analysis, the F-ratio for the multivariate test of equality
of the mean vectors indicates whether or not the grouping factors were
significantly different. Because this analysis is very sensitive towards
general F-ratic>significance,uthe univariate F-ratios indicates the
variables that caused the general F-ratio significance.

The factors upon which to group those data for the comparisons
were; two levels of a SCHOOL factor (East Lansing and Mason), four levels
of a GRADE factor (seventh Grade fall(7th F), seventh grade spring(7th S),
eighth grade fall(8th F), and eighth grade spring(8th SP)), and three
levels of a YEAR factor (year one(l), year two(2), and year three(3)).
Because East Lansing had sixth grade data and Mason had none, and Mason
had ninth grade data and East Lansing had none, the analysis compared only
the seventh and eighth grade subjects from both schools.

Because all of the data were cross-sectional, it was necessary to
determine whether or not the boys in the six test periods differed from
each other. Therefore, the first two comparisons contrasted the YEAR
factor with the GRADE factor for the East Lansing boys and the Mason boys
separately. Then, the third comparison contrasted the SCHOOL factor with

the GRADE factor across the YEAR factor, or a triple interaction analysis.

Contrast #1 (GRADE by YEAR for East Lansing only)

 

In a multi-factor design, to have the main effects and interactions
orthogonal to each other, the cell sizes must be equal or proportional.

Therefore, after determining the mean values for the cells containing

237

the total sample, forty subjects from each cell were randomly selected
as the analysis sanple. Then, the mean values in those analysis cells
were determined (Tables L-l through L-14). The mean values in the

analysis sample represent the mean values for the total sample.
Table L-l

East Lansing data's number of subjects in the total sample and
in the analysis sample

 

 

 

 

~. YEAR GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 48 -.40 i 68 - 4O 76 - 40 156 - 40
2 41 - 40 49 - 40 . 48 - 40 57 - 40
3 74 - 40 96 - 40 40 - 40 42 - 40
Table L-2

East Lansing data's chronological age mean values in the total
sample and in the analysis sample

 

GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 l49.69-150.01 155.79-155.6l 162.12-162.96 l66.83-l67.l9
2 149.45-149.41 155.76-155.74 163.15-163.08 l68.27-168.42
3 150.40-149.71 155.33-155.56 1631162163.16 167183-16718414_3

YEAR

Table L-3

East Lansing data's height mean values in the total sample and
in the analysis sample

 

 

 

YEAR_L, GRADE
7th Fall 7th Spring 8th Fall 8th Sprinqg_
1 61.27-61.18 16.35-6l.53 63.69-63.93 64.54-64.79
2 59.45-59.49 61.64-62.10 62.84-62.64 64.23-64.74

3 60.17-60.44 62.73-63.27 63.15-63.15 65.16-65.17

238

Table L-4

East Lansing data'sh019ht mean values in the total sample and

UEAR—
7th Fall

1 97.80-97.12
2 95.68-95.65
3 99.15-101.45

in the analysis sample

GRADE
7th Spring
108.40-109.85
105.98-106.76
110.54-111.20

8th Fall

114.11-118.82
111.50-110.14
111.39-111.39

Table L-5

' 8th Spring
122.95-120.43

127.07-127.41
123.09-123.40

East Lansing data's sit-ups mean values in the total sample and

in the analysis sample

 

 

YEAR 7th Fall

8th Spring

 

1 68.87-69.15
2 43.61-42.23
3 56.97-59.05

83.82-84.23
78.24-77.38

(

68.72-68.33

 

GRADE
7th Spring, 8th Fall
76.31-74.85 80.09-78.93
74.53-75.63 69.79-60.50
65.17-67.73 59.97-59.98
Table L-6

East Lansing data's pull-ups mean values in the total sample

and in the analysis sample

 

 

 

 

 

 

 

YEAR GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 3.42-3.18 2.98-2.20 3.55-3.50 3.46-3.88
2 2.41-2.30 3.00-2.93 3.33-3.30 2.68-2.53
3 3.01-2.87 2.79-2.98 3.63-3.63 4.05-4.05
Table L-7
East Lansing data's standing long jump mean values in the total
sample and in the analysis sample
YEAR -——;. , GRADE .
7th Fall 7th spring 8th Fall 8th Spring
1 64.60-65.35 66.91-65.75 67.31-66.50 70.58-70.45
60.39-60.15 67.63-67.53 64.73-65.10 70.98-71.60
3 60.84-61.35 67.81-67.70 64.30-64.30 72.79-72.88

 

239

Table L-8

East Lansing data's vertical jump mean values in the total sample

and in the analysis sample

 

 

YEAR
7th Fall

8th Spring

 

1 14.79-14.73
2 13.20-13.13
3 14.04-14.05

16.76-16.58
15.82-16.05
15.93-16.13

 

GRADE
7th Spring 8th Fall
14.50-13.93 15.66-15.58
14.80-14.85 14.19-14.13
13.69-13.73 14.06-14.60
Table L-9

East Lansing data's shuttle run mean values in the total sample

and in the analysis sample

 

 

 

 

 

 

 

 

 

 

 

GRADE
YEAR 7th Fall 7th Spring 8th Fall 8th Spring
1 10.93-10.94 10.30-10.41 10.52-10.44 10.14-10.16
2 12.10-12.13 10.65-10.62 11.14-11.19 10.12-10.22
3 10.76-10.75 10.62-10.67 10.84-10.84 10.25-10.24
Table L-10
East Lansing data's facial hair mean values in the total sample
and in the analysis sample
YEAR GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 1.31-1.30 1.66-1.55 2.45-2.55 2.25-2.13
2 1.83-1.80 1.88-1.90 2.12-2.08 2.68-2.58
3 1.22-1.10 1.97-2.10 1.90-1.90 2.31-2.35
Table L-ll
East Lansing data's axillary hair mean values in the total sample
and in the analysis sample
__ GRADE
YEAR 7th Fall 7th spring 8th Fall 8th Springfi ‘7—
1 1.08-1.08 1.74-1.63 2.29-2.40 2.55-2.43
2 1.24-1.25 1.63-1.68 1.94-1.88 2.74-2.65
3 1.27-1.33 1.86-1.85 2.42-2.43 2.67-2.68

 

 

240

Table L-12

East Lansing data's pubic hair mean values in the total sample
and in the analysis sample

 

 

 

 

'.YEAR.__. GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 1.83-1.88 2.84-2.90 3.17-3.30 3.38-3.45 ;
2 l.88-l.88 2.53-2.63 3.29-3.28 3.68-3.70
3 2.34-2.43 2.92-3 00 3.23-3.23 3.76-3.78
Table L-13

East Lansing data's penis development mean value in the total
sample and in the analysis sample

 

I

 

 

 

GRADE 4
YEAR 7th Fall 7th Spring 8th Fall 8th Spring -
1 2.08-2.08 2.79-2.80 3.04-3.03 3.13-3.20 l
2 2.17-2.15 2.51-2.63 3.02-3.03 3.58-3.63
3 2.57-2.50 2.91-3.00 3.43-3.43 3.81-3.80
Table L-14

East Lansing data's scrotum and testes development mean values
in the total sample and in the analysis sample

 

 

 

YEAR GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 2.25-2.28 2.85-2.83 3.16-3.20 3.26-3.23
2 2.10-2.10 2.69-2.75 3.08-3.05 3.67-3.70
3 2.12-2.08 2.54-2.48 2.78-2.78 3.74-3.73

 

Table L-15 provides the univariate F values and their associated
significance levels for the East Lansing GRADE by YEAR interaction analysis.

The general F-ratio for the East Lansing GRADE by YEAR interaction
was 3.0913. With a significance level less than .05, the analysis indi-
cated that East Lansing's four levels of the GRADE factor and three levels
of the YEAR factor interacted significantly. However, because the four
levels of the GRADE factor contained boys aged six months apart, the GRADE

by YEAR interaction was probable. Therefore, the univariate F's and their

241

Table L-lS

The univariate F values and their associated significance
levels for the East Lansing GRADE by YEAR

 

 

interaction

A_Variab1es Univariate F P less than
Chronol. Age .169 .9850
Height 1.419 .2057
Weight 1.132 .3426
Sit-ups 2.957 .0077
Pull-ups 1.267 .2714
Stand Lg Jp 1.764 .1048
Vertical Jp 1.879 .0827
Shuttle Run 7.793 .0001
Facial hair 5.310 .0001
Axillary hair .871 .5158
Pubic hair 1.139 .3385
Penis dev. .664 .6792

p_§crotum dev. 2.208 .0412

 

significance levels became the interesting statistics.

Because the univariate F analysis contained thirteen variables,
the probability level for each variable became .05/13 or .004. Figures
L.1 and L.2 graphically illustrate the interactions in shuttle run and
facial hair.

The second year test period 7th F boys having the very low mean
score, and all three year test period 7th Sp boys having higher scores
than the 8th F boys caused the GRADE by YEAR significant differences in
shuttle run. The third year test period 7th SP boys having a higher score
than the 8th F boys, and the first year test period P boys having a higher
score than the 8th Sp boys caused the GRADE by YEAR significant differences

in facial hair.

Contrast #2 (GRADE by YEAR for Mason only)

 

After determining the mean values for the cells containing the

total sample, twenty-five subjects from each cell were randomly selected

242
Figure L-l

The East Lansing data's shuttle run mean values GRADE by YEAR

 

 

 

 

interaction
10.24 10.“;
$0.11
' ”01*
10.4: ‘6‘“
10.6- “*“
ms - “
1043
10.8;
11.0 (0.99
11.24
“J?
IIJS
7th E 7th Sp 8tH F 8th'Sp '
Figure L-2
The East Lansing data's facial hair mean values GRADE by YEAR
interaction
2.6< 0
“‘5 1.58
2.44
4.35
2.2.
2.10 ‘
3.13
2.0
1.8 1.90 Mo
1.6
1.4
(.30
1.2
[-10

 

 

7th F 7th'Sp 8th F 8th'Sp

as the analysis sample.

were determined (Tables L-l6 through 29).

243

Then, the mean values for those analysis cells

Because the analysis sample

in the East Lansing cells contained forty subjects, by putting twenty-

five subjects in each Mason cell, the comparison of the East Lansing

and Mason data had proportional cells.

Because the author randomly

selected the analysis cell samples, again, the mean values in the analysis

sample represent the mean values in the total sample.

Table L-l6

Mason data’s number of subjects in the total sample and
in the analysis sample

 

 

 

 

 

.YEAR GRADE .
7th Fall 7th Spring 8th Fall 3th Spring

1 31-25 29-25 30-25 26-25

2 28-25 31-25 36-25 32-25

3 60-25 28-25 51-25 30-25

Table L-17
Mason data's chronological age mean values in the total
sample and in the analysis sample

YEAR GRADE

 

7th Fall
1 150.19-150.14
2 150.09-150.66
3 153.42-152.86

7th Springg
156.07-156.36
156.64-157.50
154.86-154.90

Table L-18

8th Fall
l65.18-164.68
162.53-161.22
162.17-163.04

3th Spring
172.60-173.02
167.77-167.12
168.10-168.04

Mason data's height mean values in the total sample and
in the analysis sample

 

7th Fall
1 60.21-60.10
2 60.42-60.48
3 60.53-61.05

GRADE

7th Spring
62.45-62.19
62.28-62.42
61.82-61.47

8th Fall
63.46-63.13
62.94-62.91
62.80-62.86

8th Spring
65.30-65.26
64.07-63.95
64.95-64.50

'.YEAR

244

Table L-19

Mason data's weight mean values in the total sample and

in the analysis sample

GRADE
7th Spring
109-31-110-12
116.95-119.72
108.71-105.12

5th Fall
126.85-123.84
110.62-113-34
116.39-112.48

7th Fall

1 95.40-95.32
2 101.07-100.84
3 103.30-108.43

Table L-20

#1

8th Spring 1

123.35-123.56 '
116.61-114.22 r

126.50-125.76

Mason data's sit-ups mean values in the total sample and

in the analysis sample

 

 

 

 

YEAR . G‘U mg .
7th Fall 7th Springg, th Fall 8th Spping
l 42.23-41.96 64.19-63.16 72.20-68.88 ~ -87.23-89.12
. 2- 35.96-33.48 49.84-49.36 -45.53-40.60 --- 78.78-86.04
3 p55927-55.40 71L86-71-72 69.20-73.80 76.57-75.60
Table L-21

Mason data's pull-ups mean values in the total sample and

in the analysis sample

 

 

 

 

 

 

 

 

7th Fall 7th Spring tn Fall 8th Spring,
1 2.77-2.88 3.93-3.84 3.40-3.00 6.35-6.36
2 3.75-3.60 2.68-2.36 3.28-3.00 5.22-5.44
3 2.93-2.28 3.04-3.32 2.76-2144 4-17-4-54
Table L-22
Mason data's standing long jump mean values in the total sample
and in the analysis sample
YEAR . G E .
7th Fall 7th spring 8th Fall 8th Spring
1 54.35-54.64 57.48-57-48 61.90-61.84 69.81-70.00
2 59.04-58.92 60.32-59-72 60-35-59.48 68.91-68.88
3 60.53-60.56 60.14-60.60 63.57-62.92 66.50-65.28

 

245
Table L-23

Mason data's vertical jump mean values in the total sample and

in the analysis sample

 

 

 

 

.YEAR GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 11.29-11.04 11.86-12.00 11.63-11.44 14.46-14.60
2 11.32-11.32 11.77-11.60 12.53-12.40 13.88-13.96
3 12.38-12.12 12.57-12.68 13.65-13.60 14.03-13.72
Table L-24

Mason data's shuttle run mean values in the total sample and
in the analysis sample

 

 

 

 

GRADE
YEAR
7th Fall 7th Spring 8th Fall 8th Spring
1 11.60-11.61 12.05-12.01 11.23-11.30 11.23-11.23
2 12.05-12.07 11.56-11.58 11.54-11.59 11.24-11.12
3 11.22-11.26 11.17-11.11 10.80-10.94 10.99-11.08
Table L-25

Mason data's facial hair mean values in the total sample and
in the analysis sample

 

 

 

 

YEAR ___ GRADE
7th Fall 7th Spring 8th Fall 8th Spring;
1 1.39-1.40 2.34-2.32 2.67-2.52 3.11-3.20
2 1.50-1.56 1.94-1.96 2.11-1.96 2.22-2.20
3 1.42-1.32 1.71-1.72 1.47-1.56 2.23-2,20
Table L-26

Mason data's axillary hair mean values in the total sample and
in the analysis sample

 

 

 

fl GRADE
YEAR 7th Fall 7th soring 8th Fall 8th Spring _g
1 1.52-1.40 2.17-2.24 2.93-2.84 2.85-2.92
2 1.53-1.60 2.13-2.16 1.97-2.00 2.16-2.12
3 1.67-1.56 1.93-1.88 l.88-l.84 2.33-2.40

 

246

Table L-27

Mason data's pubic hair mean values in the total sample and
in the analysis sample

 

 

 

 

YEAR ___7 fir GRADE
7th Fall 7th Spring 8th Fall 8th Spring
1 1.84-1.76 2.79-2.84 3.30-3.16 3.31-3.32
2 2.11-2.20 2.81-2.96 2.83-2.84 3.25-3.28
3 2.03-1.80 2.50-2.48 2.75-2.60 3.40-3.36
Table L-28

Mason data's penis development mean values in the total sample
and in the analysis sample

 

 

 

 

 

YEAR GRADE
_pp 7th Fall 7th Spring 8th Fall 8th Spring
1 2.26-2.24 2.41-2.40 3.33-3.20 3.23-3.24
2 2.14-2.20 2.61-2.68 2.56-2.44 3.28-3.24
3 2.00-1.80 2.25-2.24 2.71-2.68 3.10—3.00
Table L-29

Mason data's scrotum and testes development mean values in the
total sample and in the analysis sample

 

 

 

 

 

YEAR GRADE .
7th Fall 7th Spring 8th Fall 8th Spring
1 2.03.2.00 2.69-2.76 3.30-3.16 3.19-3.24
2 2.18-2.28 2.45-2.52 2.67-2.64 3.06-3.04
3 1.95-1.80 2.21-2.20 2.65-2.60 3.07-2.96
Table L-30

The univariate F values and their associated significance levels

for the Mason GRADE by YEAR interaction

 

 

Variables Univariate F P less than
Chronol. Age 3.076 .0063
Height .578 .7481
Weight 2.001 .0533
Sit-ups 5.222 .0001
Pull-ups 1.047 .3949
Stand Lg Jp 1.912 .0788
Vertical Jp 2.089 .0546
Shuttle run 2.159 .0471
Facial hair 1.492 .1818
Axillary hair 1.314 .2510
Pubic hair .700 .6500
Penis dev. 1.189 .3119
Scrotum dev. .726 .6287

 

 

247

The general F-ratio for the Mason GRADE by YEAR interaction was
2.3444. With a significance level less than .05, the analysis indicated
that Mason's four levels of the GRADE factor and three levels of the year
factor interacted significantly. However, because the four levels of the
GRADE factor contained boys aged six months apart, the GRADE by YEAR
interaction was probable. Therefore, the univariate F's and their signi-
ficance levels became the interesting statistic.

Only the significance level for sit-ups was below .004. Figure
L-3 graphically illustrates the interaction in sit-ups. The second year
test period 8th F boys having a lower score than the 7th Sp boys and the
third year test period 7th F boys and 8th Sp boys having mean scores that
were extremely high and low, respectively, caused the significant differ-

ences in sit-ups.

Figure L-3

The Mason data's sit-ups mean values GRADE by YEAR interaction

 

90 ‘ 3942
05.09
80 3
1':?.2:-----------------15'“ . 75.50
70 ‘ y
‘5'” 63.16 65.88
60 4
m
50 4 M.3;
1 *L96 (2
4O ‘10.»
30 < ’1‘“

 

 

7th F 7th Sp 8th 1? 8th Sp

248

Contrast #3 (SCHOOL by GRADE by YEAR)

 

The forty subjects for the East Lansing cells and the twenty-five

subjects for the Mason cells fulfilled the proportional cells requirement.

Tables L-30 through L-42 provide the East Lansing and Mason mean values

for their proportional cell samples.

Table L-31

East Lansing and Mason data's
in their proportional cell samples

chronological age mean values

 

 

 

YEAR 7th F ‘7th Sp TtBtfi F f:*8th Sp
EL M EL M EL M ’ EL M
1 . 150.01-150.15 155.61-156.36 l62.96-l64.68 l67.19-l73.02
2 149.41-150.66 155.74-157.50 163.08-161.22 l68.43-167.12
3 149.71-152.86 155.56-.54.90 163.16-163.04 l67.84-168.04

 

Table L-32

 

 

 

 

East Lansing and Mason data's-height mean values in their
proportional cell samples
Y 7th F 7th Sp 8th F 8th Sp
EAR EL M EL M EL M EL M
1 61-18-60.10 61.53-62.19 63.93-63.13 64.79-65.26
2 59.59-60.48 62.10-62.42 62.64-62.91 64.47-63.95
3 60.44-61.05 63.27-61.47 63.15-62.86 65.17-64.50
Table L-33

 

 

 

East Lansing and Mason data's weight mean values in their
prOportional cell samples
7th F 7th Sp 8th F 8th Sp
YEAR EL M EL M EL M EL M
1 97.12-95.32 109.85-110.12 118.82-123.84 120.43-123.56
2 95.65-100.84 106.76-119.72 110.14-113.34 127.41-114.22

3 101.45-108.54 111.20-105.12 111.39-112.48

123.40-124.76

249

Table L-34

East Lansing and Mason data's sit-ups mean values in their
proportional cell samples

 

 

 

 

7th F 7th Sp 8th F 8th Sp
YEAR E13 M EL M EL M EL M
l 69.15-41.96 74.85-63.16 78.93-68.88 84.22-89.12
2 42.23-33.48 75.63-49.36 70.50-40.60 75.38-86.04
3 59.05-65.40 67.73-71.72 59.78-73.80 68.33-75.60
Table L-35

East Lansing and Mason data's pull-ups mean values in their‘
proportional cell samples

 

8th Sp

 

 

 

 

 

 

 

 

 

 

 

W 7th I“ 7th Sp 8th F n
YEAR EL M EL M EL M EL M fl

7' 1 3.18-2.88 2.20-3.84 3.50-3.00 3.88-6.36
2 2.30-3.60 2.93-2.36 3.30-3.00 2.43-5.44 l

3 2.88-2.28 2.98-3.32 3.63-2.44 4.05-4.52

Table L-36
East Lansing and Mason data's standing long jump mean values
in their proportional cell samples

7th F 7th Sp 8th E 8th Sp ##-

YEAR 3L M EL M EL M EL M

1 65.35-56.65 65.75-57.48 66.50-61.84 70.45-70.00

2 60.15-58.92 67.53-59.72 65.10-59.48 71.60-68.88

3 61.38-60.56 67.70-60.60 64.30-62.92 72.88-65.28

Table L-37
East Lansing and Mason data's vertical jump mean values in their
proportional cell samples
7th F 7th Sp 8th F 8th Sp
YEAR EL M EL M EL M EL M

1 14.73-11.40 13.93-12.00 15.58-11.44 16.58-14.60

2 13.13-11.32 14.85-11.60 14.13-12.40 16.05-13.96

3 14.05-12.12 13-73-12.68 14.60-13.60 16.13-13.72

 

250

Table L-38

East Lansing and Mason data's shuttle run mean values in their
prOportional cell samples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

YEAR 7th F 7th Sp 8th F 8th Sp
EL M EL M EL M EL M
1 10.94-11.61 10.41-12.01 10.44-11.30 10.16-11.23
2 12.13-12.07 10.62-11.58 11.10-11.59 10.22-11.12
3 10.75-11.26 10.67-11.11 10.84-10.94 10.24-11.08
Table L-39
East Lansing and Mason data's facial hair mean values in their
proportional cell samples
jth F in Sp 8th 1“ 8th Sp .
YEAR EL M EL M EL M EL M 1
1 1.30-1.40 1.55-2.32 2.55-2.52 2.13-3.20 i
2 1.80-1.56 1.90-1.96 2.08-1.96 2.58-2.20 I
3 1.10-1.32 2.10-1.72 1.90-1.56 2.35-2.20
Table L-40
East Lansing and Mason data's axillary hair mean values in
their proportional cell samples
YEAR 76h F 7th Sp 8th F 8th Sp 5
EL M EL M EL M EL M 3
1 1.08-1.40 l.63-2.24 2.40-2.84 2.43-2.92
2 1.25-1.60 1.68-2.16 1.88-2.00 2.65-2.12
- 3 1.33-1.56 1.85-1.88 2.43-1.84 2.68-2.40
Table L-4l
East Lansing and Mason data's pubic hair mean values in their
proportional cell samples
7 7th F 7th Sp 8th F 8th Sp
YEAR EL M EL M EL M EL M
l 1.88-l.76 2.90-2.84 3.30-3.16 3.45-3.32 I
2 1.88-2.20 2.63-2.96 3.28-2.84 3.70-3.28 4

3 g,43-1.80 3.00-2.48 3.23-2.60 3.78-3.36

Table L-42

East Lansing and Mason data's penis development mean values in
their proportional cell samples

 

 

 

 

7th F 7th Sp 8th F 8th Sp ‘4
YEAR EL M EL M EL M EL M ;
I
l 2.08-2.24 2.80-2.40 3.03-3.20 3.20-3.24
2 2.15-2.20 ’2.63-2.68 3.03-2.44 3.63-3.24 7
3 2.50-1.80 3.00-2.24 3.43-2.68 3.80-3.00
Table L-43

East Lansing and Mason data's scrotum and testes development
mean values in their proportional cell samples

 

 

 

 

7th 8 7th Sp 8th F 8th Sp
YEAR EL M EL M EL M EL M J
1 2.28-2.00 2.83-2.76 3.20-3.16 3.23-3.24 '1
2 2.10-2.28 2.75-2.52 3.05-2.64 3.70-3.04
3 2.08-1.80 2.48-2.20 2.78-2.60 3.73-2.96

 

Table L-44 provides the univariate F values and their associated
significance levels for the East Lansing and Mason SCHOOL by GRADE by
YEAR triple interaction.

Table L-44

The univariate F values and their associated significance levels
for the East Lansing and Mason SCHOOL by GRADE by YEAR

 

 

interaction
Variables Univariate F P less than
Chronol. Age 2.397 .0267
Height 1.276 .2657
Weight 1.746 .1077
Sit-ups 3.140 .0048
Pull-ups 1.654 .1295
Stand Lg Jp 2.914 .0082
Vertical Jp 2.377 .0279
Shuttle Run 1.864 .0845
Facial hair 2.571 .0180
Axillary hair .897 .4963
Pubic hair .813 .5598
Penis dev. 1.037 .3999
Scrotum dev. 1.127 .3447

 

 

252

The general F-ratio for the SCHOOL by GRADE by YEAR triple inter-
action was 2.2593. With a significance level less than .05, the analysis
indicated the two SCHOOL levels across the four GRADE levels for the
three YEAR levels interacted significantly. However, because the four
levels of the GRADE factor contained boys aged six months apart, the
SCHOOL by GRADE by YEAR interaction was probable. Therefore, the univariate
F's and their significance levels became the interesting statistic.

Only the significance level for sit-ups was below .004. Figures

L-4, L-S, and L-6 graphically illustrate the interaction in sit-ups.

Figure L-4

The East Lansing and Mason data's sit-ups mean values
SCHOOL by GRADE interactions for the first year data collections

90 1 31.11
80 . 78.93 , 29.7.7.
EN“ MNSINQ
70 * I 75.35
63.15 60.08
60 ‘ 55.16
.p’
50 4 6"
4° ‘ an;

 

 

7th F 7th Sp 8th s- 8th Sp

253
Figure L-5

The East Lansing and Mason data's sit-ups mean values
SCHOOL by GRADE interactions for the second year data collections

 

 

 

 

90 1
80 4
‘flhto
13.12 ’3‘“ “‘

70 9
60 4 ‘

$4.65 $1.78
50 J
40 4

7th F 7th Sp 8th F 8th Sp

Figure L-6

The East Lansing and Mason data's sit-ups mean values
SCHOOL by GRADE interactions for the third year data collections

90 9

86.0?
80 3
77.38

70 1

501

40 ‘

 

30 ‘

 

 

7th F 7th Sp 8th F 8th Sp

254

The Mason 8th Sp boys having a mean value higher than the East
Lansing Sp boys may have contributed to the significant differences in
sit-ups for the first year's data collections. The East Lansing and
Mason 7th Sp boys having higher mean values than the 8th F boys contri-
buted to the significant differences in sit-ups for the second year's
collections. The East Lansing 7th Sp boys having an extremely high mean
value and the East Lansing 8th F boys having an extremely low mean value
contributed to the significant differences in sit-ups for the third year's
data collection.

Another analysis that also examines the differences between the
East Lansing and Mason subjects uses the least square estimates of the
SCHOOL contrast. By doubling the standard error for each variable and
adding and subtracting that value to and from its least square estimate,
the 95 percent confidence interval for each variable is determined (Table

L-45). If a confidence interval did contain zero, then the two school's

Table L-45

The least square estimates, standard errors, and confidence
intervals of the thirteen variables

 

 

Variable Least Square Twiee its Confidence Contained
Estimates Standard Error Interval Zero l
Chronol. Age -0.9033 2(0.4380) -1.78 to -0.02 No
Height 0.1533 2(0.2561) -0.37 to 0.67 Yes ‘
Weight -l.5204 2(1.7833) -5.08 to 2.04 Yes -
Sit-ups 5.7358 2(2.0489) 0.76 to 10.72 No
Pull-ups -0.4763 2(0.2115) -0.89 to -0.05 No
Stand Lg Jp 4.8629 2(0.6650) 3.52 to 6.20 No
Vertical Jp 2.2175 2(0.l936) 1.87 to 2.60 No
Shuttle run -0.6915 2(0.0650) -0.83 to -0.55 No
Facial hair -0.0496 2(0.0740) -0.18 to 0.10 Yes
Axillary hair -0.1425 2(0.0925) -0.32 to 0.04 Yes
Pubic hair 0.2354 2(0.0833) 0.08 to 0.40 No
Penis dev. 0.3242 2(0.0764) 0.16 to 0.48 No
Scrotum dev. 0.2479 2(0.0730) 0.11 to 0.39 No

 

255

data did not differ significantly. However, if a confidence interval
did not contain zero, then the two school's data did differ significantly.
If the data did differ between the two schools, then a judgement as to
whether or not the difference was of practical significance must be made.

Because the contrast subtracted the Mason values from the East
Lansing values, if a least square estimate was positive, then the East
Lansing subjects had the higher value. The East Lansing subjects had
higher values than the Mason subjects in height, sit-ups, standing long
jump, vertical jump, pubic hair, penis development, and scrotum and testes
development. Of those seven variables, the variables for which the 95
percent confidence intervals did not contain zero were sit-ups, standing
long jump, vertical jump, pubic hair, penis development, and scrotum and
testes development. Therefore, for sit-ups, standing long jump, vertical
jump, pubic hair, penis development, and scrotum and testes deve10pment,
the East Lansing subjects measured significantly higher than the Mason

The Mason subjects had higher values than the East Lansing subjects
in chronological age, weight, pull-ups, shuttle run, facial hair, and
axillary hair. Of those six variables, the variables for which the 95
percent confidence intervals did not contain zero were chronological age,
pull-ups, and shuttle run. Therefore, for chronological age, pull-ups,
and shuttle run, the Mason subjects measured significantly higher than the
East Lansing subjects. The East Lansing and Mason subjects did not measure
significantly different in height, weight, facial hair, and axillary hair.

Because the least square estimate analysis determined nine variables
that were statistically different between the East Lansing and Mason
contrast, the following suggestions about the practical significance of

those differences are provided:

256

1) Because classification procedures for motor proficiency norms
use yearly intervals, the .9 of a month advantage of the Mason subjects
is not of practical significance.

2) Because a 5.74 sit—up difference does not exceed the difference
between the mean values of the 7th F and 7th Sp boys or the 8th F and 8th
Sp boys, the 5.74 sit-ups advantage of the East Lansing subjects is not of
practical significance.

3) Because a .47 pull-ups difference exceeds the difference between
the mean values of 7th F and 8th F boys, the .47 pull-ups advantage of
the Mason subjects is of practical significance.

4) Because a 4.86 standing long jump difference exceeds the differ-
ence between the mean values of the 7th F and 7th Sp boys, the 4.86
standing long jump advantage of the East Lansing subjects is of practical
significance.

5) Because a 2.22 vertical jump difference exceeds the difference
between the mean values of the 7th F and 8th F boys, the 2.22 vertical
jump advantage of the East Lansing subjects is of practical significance.

6) Because a .69 shuttle run difference exceeds the difference
between the mean values of the 7th F and 8th F boys, the .69 advantage
of the East Lansing subjects is of practical significance.

7) Because a .24 pubic hair difference does not exceed the difference
between the mean values of the 7th F and 7th Sp boys or the 8th F and 8th
Sp boys, the .24 pubic hair advantage of the Mason subjects if not of
practical significance.

8) Because a .32 penis development difference does not exceed the
difference between the mean values of the 7th F and 7th Sp boys or the 8th

F and 8th Sp boys, the .32 penis development advantage of the Mason

PLEASE NOTE:

Page 257 is lacking in number
only. No text is missing.
Filmed as received.

UNIVERSITY MICROFILMS.

258

subjects is not of practical significance.

9) Because a .25 scrotum and testes difference does not exceed
the difference between the mean values of the 7th F and 7th Sp boys or
the 8th F and 8th Sp boys, the .25 scrotum and testes development advan-
tage of the Mason subjects is not of practical significance.

To summarize, the purpose of the three contrast analyses and the
least square estimates analysis was to examine the East Lansing and Mason
data to determine whether or not by combining those data, the statistical
analysis had a proper sample. The least square estimates analysis demon-
strated that the motor proficiency tests of pull-ups, standing long jump,
and shuttle run were of practical difference. However, the three-way
interaction analysis demonstrated that only the sit-ups difference was
statistically significant. Therefore, with some concern for the sit-ups
test, the author concluded that combining the East Lansing and Mason data

does provide a proper sample.

259

APPENDIX M

THE PROPORTION OF THIS SAMPLE'S 1578 BOYS WHO REACHED THE
FIVE STAGES OF THE FIVE SEXUAL CHARACTERISTICS

260

Table M-l

The observed proportions for facial hair

 

 

 

 

 

 

 

 

 

 

 

STAGES
AGES(mo.) N 1 2 3 4 5
132 15 100.0 0.0 0.0 0.0 0.0
138 45 71.1 26.7 2.2 0.0 0.0
144 85 74.1 20.0 5.9 0.0 0.0
150 234 59.4 30.8 7.7 1.7 0.4
156 283 46.6 36.0 12.4 3.9 1.1
162 284 37.7 29.9 21.5 9.5 1.4
168 248 28.2 27.4 25.8 14.1 4.4
174 159 15.1 21.4 27.7 28.9 _6.9
180 122 9.8 10.7 23.0 40.2 16.4
186 79 2.5 6.3 19.0 46.8 25.3
192 14 0.0 0.0 7.1 57.1 35.7
Table M-2
The observed proportions for axillary hair
AGES (mo. ) N STAGES
l 2 3 4 5
132 15 100.0 0.0 0.0 0.0 0.0
138 45 91.1 6.7 2.2 0.0 0.0
144 85 91.8 0.0 3.5 4.7 0.0
150 234 72.6 14.5 4.3 7.3 1.3
156 283 61.1 18.0 10.6 7.8 2.5
162 284 45.1 21.1 12.7 15.8 5.3
168 248 35.1 20.6 10.9 22.2 11.3
174 159 21.4 24.5 16.4 23.3 14.5
180 122 14.8 12.3 9.0 38.5 25.4
186 79 5.1 16.5 6.3 35.4 36.7
192 14 0.0 7.1 7.1 35.7 50.0
Table M-3
The observed proportions for pubic hair
AGES(mo.) N STAGES
l 2 3 4 5
132 15 80.0 13.3 6.7 0.0 0.0
138 45 60.0 20.0 17.8 2.2 0.0
144 85 58.8 24.7 7.1 9.4 0.0
150 234 32.9 30.3 17.5 17.9 1.3
156 283 20.1 30.7 26.9 17.7 4.6
162 284 5.6 24.3 28.2 32.7 9.2
168 248 9.3 14.5 22.6 35.5 18.1
174 159 4.4 7.5 30.8 35.2 22.0
180 122 2.5 6.6 17.2 37.7 36.1
186 79 2.5 1.3 8.9 36.7 50.6
192 14 7.1 0.0 0.0 35.7 57.1

 

261

Table M-4

The observed proportions for penis development

 

 

 

 

 

 

 

AGES(mo.) N STAGES
l 2 3 4 S
132 15 40.0 53.3 6.7 0.0 0.0
138 45 35.6 51.1 11.1 2.2 0.0
144 85 36.5 41.2 16.5 5.9 0.0
150 234 18.4 42.3 27.8 9.8 1.7
156 283 18.4 31.8 33.6 12.7 3.5
162 284 8 1 23.6 33.8 28.5 6.0
168 248 9.7 16.9 27.4 32.3 13.7
174 159 4.4 11.9 23.3 43.4 17.0
180 122 4.9 11.5 9.8 45.9 27.9
186 79 2.5 3.8 12.7 50.6 30.4
192 14 0 0 7.1 21.4 28.6 4;&2_
Table M-5
The observed prOportions for scrotum and testes deve10pment
H STAGES
AGLS (1210.) N 1 2 3 4 5
132 15 40.0 60.0 0.0 0.0 0.0
138 45 33.3 53.3 13.3 0.0 0.0
144 85 40.0 44.7 11.8 3.5 0.0
150 234 15.8 51.7 25.2 6.4 0.9
156 283 14.8 44.2 25.8 12.4 2.8
162 284 5.6 29.6 32.7 26.1 6.0
168 248 5.2 19.4 30.6 31.9 12.9
174 159 3.8 18.9 28.3 37.1 _11.9
180 22 0.8 17.2 17.2 45.1 19.7
186 79 2.5 5.1 17.7 48.1 p 26.6
192 14 7.1 7.1 28.6 35.7 21.4

 

Figures M-l through 5 graphically illustrate the proportion values

contained in Tables M-l through M-S.

Because the observed proportion figures for the five sexual charact-
eristics were somewhat irregular, especially at the low and high ages, the
author used the Quantal Response for Random Predictor Variables formula
to calculate the predicted proportion values contained in Table M-6 through
M-lO (Schmidt and McSweeney, 1977). Figures 6 through 10 graphically

illustrate those proportion values.

H I2 3&2? .~ ..~

\66

262

mam

OHN

vow

mmH

Nod

AnnuGOEv m0¢ AflUHUOAOZQmmU
mmd omH th mod Nod mma

I I D D _

mcofluuomoum Ham: Hmflomm Um>uwmno

H12 wusmfim

omH

gva

and

 

NMH

vow

flow
.00
70h
vom

 

SNOIIHOdOHd

80/

263

3528.5 mus. 98830225
3m 3m «on mg NS mm: 03 65 m3 SH 63 03 3: m3 «3

h I I I I I I I I I I I P .D I

2. .N

u.
«a
d.

:ofiuuoooum was: >Hgaafix< pm>hmmno

NI: mugmwh

vOH

10m

”om
.8
.2.
.8

 

SNOILHOdOHd

MIZ

9.21;...

C ~H

264

Amnugosv mot AdUHooqozommo
can can vow mmH mad omH owH vna mwa moa oma oma vva mma mad

I I I b I I I I I I I I I l

{ in...

 

cofiuuogoum new: UHQSQ pw>ummno

mu: mugmflm

.OH

.0N

"0m
.06
.on
vom

 

SNOILHOdOHd

265

mam

AmSHCOEV WU¢ AflUHuoqozommo
OHN VON mmH NmH omH owH VbH moa mmH and omH vgd mMH NMH

I I I I I I I I

£¢£VW

h abkls

ac};- "
51"Jae ‘

55“

mcofluuomoug ugwemon>wo meson pm>uwmno

Viz mugmfim

vOH

"om
.oo
.on
.00

 

SNOIIHOdOHd

\em

266

Amnugosv mum gHoOAozommo
man can vow mm: mm." 02 03 v: m3 «3 mma

I I I I I I I I I

an.
m I

5x 895 ..-’"

mCOHuHOQOHQ UCQEQOH®>®C mwumwrb GEM EDUOHOW U0>H®mn~0

ml: wusmflm

omH

I

£3 his

s‘me‘

VVH

mMH

.0m

.0m
.00
.0h
vow

 

SNOILEOdOHd

267

Table M-6

The predicted proportions for facial hair

 

 

 

 

 

 

 

AGES (mo .) STAGES
l 2 3 4 5
132 80.1 18.3 1.3 0.1 0.0
138 75.0 22.2 2.7 0.2 0.0
144 68.6 26.2 4.8 0.5 0.1
150 60.4 29.9 8.2 1.3 0.2
156 50.4 32.3 13.2 3.4 0.7
162 38.5 31.9 19.6 8.1 1.8
168 25.7 27.6 25.4 16.8 4.6
174 14.4 19.9 27.4 28.9 9.4
180 6.6 11.9 ' 24.4 41.2 15.9
186 2.6 6.0 18.6 49.8 23.0
192 0.9 2.7 12.3 54.0 29.7
198 0.3 1.2 8.0 54.7 35.8
204 0.1 0.5 4.8 53.1 41.4
210 0.0 0.2 2.9 50.2 46.7
216 0.0 0.0 1.7 46.4 51.8
Table M-7
The predicted proportions for axillary hair
STAGES
AGES(mo.) l 2 3 4 5
132 93.4 4.3 1.5 0.7 0.1
138 89.5 6.4 2.5 1.4 0.1
144 83.7 9.3 4.0 2.7 0.4
150 74.3 13.0 6.0 5.2 1.4
156 62.8 16.9 8.5 9.3 2.3
162 48.3 20.2 11.1 15.3 5.1
168 33.2 21.5 12.9 22.4 10.0
174 20.1 20.2 13.2 29.1 17.3
180 10.9 16.9 12.0 33.5 26.7
186 4.3 12.8 9.9 35.6 37.1
192 2.4 8.9 7.6 33.4 47.7
198 1.0 5.9 5.4 30.2 57.6
204 0.3 3.7 3.7 26.0 66.3
210 0.1 2.2 2.4 21.6 73.6
216 0 0 1.3 1.6 17.4 79.6,

 

268

Table M-8

The predicted proportions for pubic hair

 

 

 

 

 

 

 

STAGES
132 66.6 24.2 6.8 2.3 0.1
138 57.1 28.1 10.0 4.5 0.2
144 45.5 30.6 14.9 9.1 0.8
150 31.4 32.5 20.2 14.2 1.7
156 21.9 27.5 24.7 21.8 4.0
162 12.9 22.3 26.7 29.7 8.4
168 6.1 15.6 24.6 35.9 17.8
174 3.1 9.9 22.5 38.7 25.8
180 1.5 5.7 17.5 37.5 37.7
186 0.5 3.0 9.5 32.9 54.0
192 0.2 1.4 8.0 26.7 63.7
198 0.1 0.6 4.9 20.3 74.1
204 0.0 0.3 2.8 14.7 82.2
210 0.0 0.1 1.6 10.3 88.0
216 0.0 0.0 0.9 7.0 92;;__
Table M-9
The predicted proportions for penis development
STAGES

AGES(mo.) 1 2 3 4 5
132 35.9 47.4 15.6 0.8 0.1
138 31.8 45.9 19.6 2.5 0.3
144 27.3 44.9 23.8 3.9 0.7
150 22.4 39.4 27.8 8.7 1.7
156 17.4 33.6 30.5 15.0 3.6
162 12.3 26.2 30.8 23.8 6.9
168 7.9 18.5 28.5 33.3 11.9
174 4.5 11.6 23.1 42.8 18.0
180 2.3 6.3 17.2 48.9 25.4
186 1.1 3.4 11.0 52.1 32.4
192 0.5 1.7 7.3 51.4 39.1
198 0.2 0.8 4.5 49.3 45.2
204 0.1 0.4 2.8 45.7 50.9
210 0.0 0.2 1.6 42.1 56.2
216 0.0 0.1 0.9 437.9 61&;_

 

269

Table M-10

The predicted proportions for scrotum and testes development

 

 

 

AGES(mo.) 1 2 STAELS 4 5
132 39.6 50.5 8.8 1.1 0.1
138 32.9 52.1 12.6 2.2 0.3
144 26.1 51.6 17.3 4.2 0.7
150 19.7 48.4 22.5 7.8 1.5
156 13.8 42.4 27.3 13.5 3.1
162 8.9 33.8 30.3 21.2 4.8
168 5.2 24.4 30.8 29.8 9.7
174 2.7 16.0 27.7 39.0 14.7
180 0.9 11.2 23.0 45.9 20.3
186 0.6 5.3 17.7 50.3 26.2
192 0.3 2.8 12.9 52.1 31.9
198 0.1 1.4 9.2 52.0 37.3
204 0.0 0.7 6.2 50.4 42.6
210 0.0 0.3 4.1 47.9 47.5
216 0.0 0.2 2.7 44.7 52.2

 

A comparison of the similarity of the observed proportion values
with the predicted proportion models showed that the models fit the
observed proportion model. Therefore, by examining Tables M-6 through
M-10 and Figures M-6 through M-lO, workers can determine the proportion
of this 1,578 adolescent male sample that the raters rated as being in each
of the five sexual characteristic's stage at 6 month intervals from 132
months to 216 months. For example, Table M-8 and Figure M-8 show that at
age 162 months (135 years), the raters rated; 12.9 percent of the 1,578
boys not yet started their pubic hair deve10pment, 22.3 percent as Stage 2
pubic hair, 26.7 percent as stage 3, 29.7 percent as stage 4, and 8.4

percent as having completed their pubic hair development.

©‘E Q.~::.~.Q I \.\Q

270

AmSuCOEV w0< A<UHOOAOZOMEU

mam OHN vow mmH NmH me own vhH me NOH mmH omH vva

I I I I P I I I I P

In.
a.
I ..
‘\
.... \fiflvn.
0) a .
on fo- 00!“ As
a u
.. 4..
xx...

wcofluuomoum Ham: Hafiomm Uwuofipwum

wlz ougmflm

:0”

"on
.00
.on
.00

 

SNOIIHOdOHd

N.12 9.2:. ._...~

271

mam

0

Cam

P

vCN

cod

AnnuCOEv $04 A<UHOOAOzom=U
Ned can can Qua mod mod

I I I

wgofluuomoum uwmm xumHHflx< pwuoflpwum

biz wusofim

mmH

0mm

vva

QMH

NMH

 

ION

.0m
.00
.05
vom

 

SNOILHOdOHd

$12 2 .5: .7.

272

mam

OHN

VON

Am£uGOEV WU¢ AdUHUOAozommU

mmH Naa owa

.l//

f

omH

 

«5H mod mod and end va QMH

D I I I I F I

mcoHuuomoum uwgm oflngm pmuowpwum

mu: 9:62

row

"on
.00
.oh
.om

 

SNOIIHOdOHd

273

on

oaw.

vow

I

mma

I

me

Amsugoev mug AdUHuoqozommo

omH

omH qua wwH on omH omH vva

I I I I I I I

 

mcofluuomoum uCOEmoaw>wo magma pwuoflwwum

mlz wusmfim

mMH

NMH

I:

 

SNOIIHOdOHd

O~\Z 3133.4..— Ow

274

on

OHN

Amnugoav mug quHOoqozcmmO

vow mma me oma omH vhH mmd me omH omH

I I P I I I I I I

It;

 

 

MCOMDHOQOME ungQoHo>oQ mmumme pgm Edu0uom pauoaUmum

can: musgflm

vva

.P

mma

NMH

10H

VON

.8
.8
.2
rom

 

SNOIIHOdOHd

275

APPENDIX N

THE REGRESSION EQUATION FOR THE ESTIMATE OF SKELETAL AGE (ESA)

276

Before determining the regression equation for the estimate of
skeletal age (ESA) that incorporates chronological age, height, weight,
and SMIV, the correlation matrix for those four variables should be
examined (Table N-l). The two anthropomorphic measures of height and

Table N-l

The sample chronological age, height, weight, and SMIV
correlation matrix

 

 

Chronological Age Height Weight SMIV
Chronological Age 1.000
Height .698 1.000
Weight .388 .718 1.000
SMIV .560 .699 .537 1.000

 

weight correlated the highest (.718), while chronological age and weight
correlated the lowest (.388). SMIV and height (.699) and chronological

age and height (.698) correlated equally, but SMIV and chronological age
correlated at only .560.

To obtain the regression equation for estimating skeletal age from
chronological age, height, weight, and SMIV, the Multivariate Regression
Analysis predicted the dependent variable (skeletal age) from those four
independent variables. Table N-2 contains the raw regression coefficients
and mean values that the Multivariate Regression Analysis determined for
the ESA.

Table N-2
The raw regression coefficients and mean values that the Multi-

variate Regression analysis determined for the
ESA regression equation

 

 

Raw Regression Mean

Coefficients (8) Values (u)
Chronological Age (CA) .277 159.27
Height (H) .212 63.01
Weight (W) .220 108.30
SMIV .594 160.57

ESA --- 160.53

ESA

277

The mean deviated (ESA) regression equation is:

MSA) + 8(CA)(X(CA)-fi<cm + émxm-fimn + 300(me - fi<w>> +

8(SMIv)(X(SMIV) - fi(SMIV))

To convert this mean deviated regression equation to non-mean deviated,

it is necessary to solve the following ESA 5(0) equation;

ESA 8(0)

After substituting the Table N-2 5 and u values and solving this equation,

ESA 5(0) equalled -16.15.

{3(sm - é<cmfi<0m - 'émifimi - é<w>fi<w1 - é(smv>fi<smxv)

To demonstrate the effects of different chronological age, height,

weight and SMIV values on ESA, Table N-3 substitutes 144 and 168 months

for chronological age, 60 and 72 inches for height, 100 and 150 pounds

for weight, and 151.61 and 175.87 months for SMIV.

Table N-3

The 151.61

and 175.87

Substituting selected chronological age, height, weight, and
SMIV scores into ESA

 

 

5(CA)(X(CA))+ 5(H)(X(H))+ 3(W)(X(W)) + B(SMIV)(X(SMIV))+ 8(0) = ESA

7277—1144) + .212 (60) + .220 (100) + .594 (151.6l)+(-l6.15)= 148.51
.277 (168) + .212 (60) + .220 (100) + .594 (151.611+(-16,15)= 155,16
.277 (144) + .212 (72) + .220 (100) + .594 (151.61)+(-16,15)= 151.05
.277 (168) + .212 (72) + .220 (100) + .594 (151.61)+(-16.15)= 157.71‘
.277 (144) + .212 (60) + .220 (150) + .594 (151.61)+(-16,15)= 159,51
.277 (168) + .212 (60) + .220 (150) + .594 (151.61)+(-l6.15)== 166.16
.277 (144) + .212 (72) + .220 (150) + .594 (151.61)+(-16.15)= 162.06
.277 (168) + .212 (72) + .220 (150) + .594 (l51.61)+(-16.15)= 168.71
.277 (144) + .212 (60) + .220 (100) + .594 (178.03)+(-16.15)= 162.92
.277 (168) + .212 (60) + .220 (100) + .594 (175.87)+(-l6.15)= 169.57
.277 (144) + .212 (72) + .220 (100) + .594 (175.87)+(-16.15)=‘165.42
.277 (168) + .212 (72) + .220 (100) + .594 (175.87)-(-16.15)= 172.12
.277 (144) + .212 (60) + .220 (150) + .594 (175.87)+(-16.15)= 173.92
.277 (168) + .212 (60) + .220 (150) + .594 (175.87)+(-l6.15)= 180.57
.277 (144) + .212 (72) + .220 (150) + .594 (175.87)+(-16.15)==176.47f
.277 (168) + .212 (72) + .220 (150) + .594 (175.87):(-16.15)«=183.12‘

 

months for SMIV represented the values when all five sexual characteris-

tic ratings equalled 2 and 4, respectively.

A 1 month increase in chrono-

logical age increased ESA .277 of a month, a 1 inch increase in height

278

increased ESA .212 of a month, a 1 pound increase in weight increased ESA
.220 of a month, and a 1 month increase in SMIV increased ESA .594 of a
month.

As another way of examining the impact of those four independent
variables on the dependent variable, provided that the range for chrono-
logical age is 60 months, the range for height is 20 inches, the range
for weight is 100 pounds, and the range for SMIV is 50 months, by multi-
plying the beta weights for those variables times their ranges, the result-
ing value demonstrates the maximum impact potential for each variable.
Table N-4 contains the beta weights, the ranges, and the maximum impact
potentials for those four independent variables.

Table N-4

The beta weights, ranges, and maximum impact potentials for
the four independent variables in the ESA regression

 

 

equation
Variables Beta Weights x Ranges = Max. Impact Potential
Chronological Age .277 x 60 = 16.20 ESA months
Height .212 x 20 = 4.24 ESA months
Weight .220 x 100 = 22.00 ESA months
SMIV .594 x 50 = 29.70 ESA months

 

SMIV and weight provide the highest impact potentials, with chrono-
logical age's impact about one-half of SMIV and height's impact very small.
For chronological age to have the same impact as SMIV, the adolescent
period would have to last almost 9 years. For height to have the same
impact as SMIV, the boys would have to grow over 115 feet during adoles-
cence. And, for weight to have the same impact as SMIV, the boys would
have to gain 135 pounds during adolescence.

To determine the precision that the analysis established the raw
regression coefficients for ESA, you examine whether or not the confidence

interval for each variable contains zero. A confidence interval equals

279

the raw regression coefficient plus and minus twice its standard error.
Table N-S contains the raw regression coefficients and standard errors
for the four variables in ESA.

Table N-S

The raw regression coefficients and standard errors for
chronological age, height, weight, and SMIV

 

 

Raw Regression Coefficients Standard Errors
Chronological Age .277 .142
Height .212 .604
Weight .220 .088
SMIV .594 .142

 

Table N-6 contains the confidence intervals for those four
variables.
Table N-6

The confidence intervals for chronological age, height, weight

 

 

and SMIV
Raw Regression + twice its _ Confidence
Coefficient - Standard Error - Interval
Chronological Age .277 + and — 2(.l42) = (-.007 to .561)
Height .212 + and - 2(.604) = (-.996 to 1.42)
Weight .220 + and - 2(.088) = (.044 to .396)
SMIV .594 + and - 2(.l42) = (.310 to .878)

 

The confidence intervals for chronological age (-.007 to .561) and height
(-.999 and 1.420) both included zero. Therefore, for this sample the
Multivariate Regression analysis did not establish the raw regression
coefficients for chronological age and height very precisely.

Observed data contains measurement error. And, if one of the
independent variables in a regression equation contains more measurement
error than the other independent variables, the analysis' estimated raw
regression coefficients can incorrectly attentuate the raw regression
coefficient of the variable with the higher measurement error. Consequent-

ly, the interpretations of the impact of the individual independent var-

280

iables can be misleading. Regression analyses on observed data cannot
account for measurement error. In this analysis the chronological age,
height, and weight independent variables should have far less measurement
error in them than the SMIV independent variable. The solution is to use
a Structural Multivariate Regression analysis to remodel the analysis.
Structural Regression analyses take into account the measurement_error
contained in the SMIV variable (Lisrel, 1977).

The latent raw regression coefficients indicate whether or not
the measurement error in SMIV distorted the ESA raw regression coeffi-
cients. While the Structural Multivariate Regression Analysis developed
its own SMIV regression equation, it required that one of the sexual
characteristic's beta weights equal 1.000. Therefore, the pubic hair
characteristic beta weight was set equal to 1.000.

Table N-7 contains the parallel beta weights. While from the

Table N-7

The observed and structural ESA beta weights for chronological
age, height, weight, and SMIV

 

Chronological Age Height Weight SMIV
Observed Beta Weights .277 .212 .220 .594
Structural Beta Weights .258 .120 .240 7.308
observed analysis to the structural analysis, chronological age's beta
weight decreased slighly (.277 t0 .258) and weight's beta weight increased
slightly (.220 to .240), height's beta weight decreased dramatically
(.212 to .120). Because the SMIV scale for the observed analysis ranged
from 1 to 5, those beta weights cannot be compared directly.
To compare those beta weights, a table similar to Table N-3 had

to be developed for the structural values. From the discussion of the

impact of the independent variables on the structural ESA, the relation—

281

ship between the observed and structural SMIV becomes apparent.

To obtain the structural ESA regression equation, the Structural
Multivariate Regression Analysis predicted the dependent variable
(skeletal age) from the three independent variables of chronological age,
height, and weight and the five sexual characteristics. However, whereas
the analysis of the observed data provided beta weights for those eight
variables individually, the structural analysis determined beta weights
for chronological age, height, weight, and SMIV. Table N-8 contains the
latent regression coefficients from the structural analysis and the mean
values from the analysis on the observed data.

Table N-8

The latent regression coefficients from the structural analysis
and the mean values from the analysis of the observed data

L

 

 

Latent Regression Mean

Coefficients (8) Values (E)
Chronological Age .258 159.27
Height .120 63.01
Weight .240 108.30
SMIV 7.308 cannot use
Skeletal Age ----- 160.53

 

Because the structural SMIV was in a different scale than the SMIV
from the observed data, the structural ESA regression equation cannot use
the observed Q(SMIV) value. Solving the following equation provided the
fi(SMIV) value for the structural ESA regression equation.

Structural (SMIV = {(Af A)-1} {(A’fi)}

Where A_is the vector of the structural weight estimates for the five
sexual characteristics and fi_is the vector of the observed mean value
estimates for the five sexual characteristics. Table N-9 contains the
A vector and the a vector. After substituting the Table N-9 values and

solving the equation, the structural fi(SMIV) equalled 3.03.

282

 

Table N-9
The A vector and the fi vector
A vector = ( 0.689 0.719 1.000 1.053 0.769)
p vector = ( 2.04 1.86 2.95 3.02 2.98)

 

The mean deviated structural (S)ESA regression equation is;

(S)ESA = fi(SA) +8 (0A)(X(CA)-fi(cai+é(ni(x<n>-fi(n)+é(W)(X(w1-fi(w11+
8(SMIV)(X(SMIV)-fi(SMIV))

To convert this mean deviated equation to a non-mean deviated, it is

necessary to solve the following equation;

(S)ESA 8(0) = fi(SA)-B(CA)fi(CA)-é(n)fi(H)-é(W)fi(w>-é(smxv)fi(smxv)

After substituting the Table N—8 values and the calculated structural

(SMIV) value and solving the equation, the ESA 8(0) equalled 63.764.

To demonstrate the effects of increments of chronological age,
height, weight and SMIV values on the (S)ESA values, Table N-lO substi—
ltutes 144 and 168 pounds for weight, and 2.00 and 4.00~for SMIV.

Table N-lO

Substituting selected chronological age, height, weight, and
SMIV scores into the (S)ESA regression equation

 

 

370A)(X(0A))+é(8)(X(H))+é(w1(X(w11+8(smxv)(X(SMIv))+ 870) = ESA

.258 (144) + .120 (60) + .240 (100) + 7.308 (2.00)4—63.764 = 146.73
.258 (168) + .120 (60) + .240 (100) + 7.308 (2.00)H-63.764 = 152.92
.258 (144) + .120 (72) + .240 (100) + 7.308 (2.00)H-63.764 = 148.17
.258 (168) + .120 (72) + .240 (100) + 7.308 (2.00)*’63.764 = 154.36
.258 (144) + .120 (60) + .240 (150) + 7.308 (2.00)4'63.764 = 158.73
.258 (168) + .120 (60) + .240 (150) + 7.308 (2.00)*'63.764 = 164.92
.258 (144) + .120 (72) + .240 (150) + 7.308 (2.00)*'63.764 = 160.17
.258 (168) + .120 (72) + .240 (150) + 7.308 (2.00)4-63.764 = 166.36
.258 (144) + .120 (60) + .240 (100) + 7.308 (4.00)4-63.764 = 161.35
.258 (168) + .120 (60) + ,240 (100) + 7.308 (4.00)4-63.764 = 167.54
.258 (144) + .120 (72) + .240 (100) + 7.308 (4.00)4-63.764 = 162.79
.258 (168) + .120 (72) + .240 (100) + 7.308 (4.00)4.63.764 = 168.98
.258 (144) + .120 (60) + .240 (150) + 7.308 (4.00)4.63.764 = 173.35
.258 (168) + .120 (60) + .240 (150) + 7.308 (4.00)+63.764 = 179.54
.258 (144) + .120 (72) + .240 (150) + 7.308 (4.00)+63.764 = 174.79
.258 (168) + .120 (72) + .240 (150) + 7.308 (4.00)g:63.764 == 180.28

 

283

As another way of examining the impact of those four independent
variables on the dependent variable, provided that the range for chrono-
logical age is 60 months, the range for height is 20 inches, the range
for weight is 100 pounds, and the range for SMIV is 4 levels, by multi-
plying the beta weights for those variables times their ranges, the
resulting value demonstrates the maximum impact potential for each variable.
Table N-ll contains the beta weights, ranges, and maximum impact poten-
tial for the four independent variables.

Table N-ll

The beta weights, ranges, and maximum impact potentials for the
four independent variables in the (S)ESA regression

 

 

equation.
Variables (S)ESA Beta Weights x Ranges = Max. Impact Poten.
Chronological Age .258 x 60 = 15.48 (S)ESA months
Height .120 x 20 = 2.40 (S)ESA months
Weight .240 x 100 = 24.00 (S)ESA months
SMIV 7.308 x 4 = 29.33 (S)ESA months

 

SMIV and weight provide the highest impact potentials, with chrono-
logical age's impact about one-half of SMIV and the impact of height is
minimal. For chronological age to have the same impact as SMIV, the
adolescent period would have to last almost 95 years. For height to have
the same impact as SMIV, the boys would have to grow over 20 feet during
adolescence. And, for weight to have the same impact as SMIV, the boys
would have to gain almost 122 pounds during adolescence.

Table N-12 contains the ESA maximum impact potentials from the
observed (0) and structural ESA maximum impact potentials. With the
possible exception of height, the Structural Multivariate Regression

analysis redistributed the beta weights very slightly.

284

Table N-12

The (0)ESA and (S)ESA maximum

impact potentials

 

Variables (0)ESA Maximum Impact

Potentials

Chronological Age 16.20 (0)ESA months

Height 4.24 (0)ESA months
weight 22.00 (0)ESA months
SMIV 29.70 (0)ESA months

(S)ESA Maximum Impact

Potentials
(S)ESA months
(S)ESA months
(S)ESA months
(S)ESA months

15.48

2.40
24.00
29.23

 

Because the structural analysis eliminated the SMIV measurement

error, the structural sample correlation matrix could be different from

the observed sample correlation matrix.
matrix was its sample variance matrix.
analysis' PHI matrix.

Table N-l3

The structural analysis' PHI

Table N-13 contains the structural

The structural analysis' PHI matrix

 

 

Chronological Age Height Weight SMIV
Chronological Age 174.713
Height 40.267 19.056
Weight 111.556 68.243 474.051
SMIV 8.453 3.372 12.161 1.160

 

Dividing the off-diagonal variances by the product of the square

roots of the diagonal variances provided the structural sample correlation

matrix.

Table N-14

Table N-14 contains the structural sample correlation matrix.

The structural sample correlation matrix

 

 

Chronological Age Height Weight SMIV
Chronological Age 1.000
Height .698 1.000
Weight .388 718 1.000
SMIV .594 (.560) 717 (699) - .519 (.537) 1.000

 

In parenthesis beside the SMIV correlations are the comparable observed

SMIV correlations.

While from the observed analysis to the structural

analysis the SMIV-chronological age correlation increased (.560 to .594)

285

and the SMIV-height correlation increased (.699 to .717), the SMIV-weight
correlation decreased (.537 to .519).

The structural analysis provided another statistical procedure
for estimating the reliability of the five sexual characteristic ratings.
Because the structural analysis used the average of the five rater's two
assessments for each sexual characteristic as that characteristic's
observed value, the following reliabilities represent the reliability of
those averaged ratings.

The following variance (02) equation provides the values that
substitute into the structural reliability equations.

02(X(i)) = A2(i)02(€) +02(e(i))
where the A'x are the structural analysis' estimates of the sexual char-
acteristic beta weights, the 02(€)is the structural analysis' estimate
of the SMIV variance, and the 02(e)'s are the structural analysis' esti-
mates of the sexual characteristics measurement error and appear as the
diagonal elements of the theta delta matrix. Table N-lS contains the
02(X(i)) equations for the five sexual characteristics and their solutions.

Table N-15

The 02(X(i)) equations for the five sexual characteristics
and their solutions

 

 

mi) 02(5) 02min 02min
Facial Hair .689 X 1.160 + .390 = .940
Axillary Hair .719 X 1.160 + .538 = 1.138
Pubic Hair 1.000 X 1.160 + .159 = 1.319
Penis Dev. 1.000 X 1.160 + .058 = 1.344
Scrotum Dev. .769 X 1.160 + .036 = .722

 

The following reliability (0) equation provides the structural
reliability values for the five sexual characteristics;
9(x(i)) = 12(1) 02(g)+ 02(x(i))

Table N-16 contains the p(x(i)) equations for the five sexual character-

286

Table N-l6

The o(x(i)) equations for the

five sexual characteristics
and their solutions

 

 

A271) x 02(5) c“(x(i)) = o(x(i))
Facial Hair .689 X 1.160 .940 = .586
Axillary Hair .719 X 1.160 1.138 = .527
Pubic Hair 1.000 X 1.160 1.319 = .879
Penis Dev. 1.053 X 1.160 1.344 = .957
Scrotum Dev. .769 X 1.160 .722 .950

 

istics and their solutions.

The averaged penis deve10pment ratings

provided the highest reliabilities (.957), and the averaged axillary

hair ratings provided the lowest reliability (.527).

The sexual char-

acteristics that provided the second, third, and fourth highest reliabil-

ities were scrotum and testes development (.950), pubic hair (.879), and

facial hair (.586).

287

APPENDIX 0

AN ADDITIONAL CLASSIFICATION PROCEDURE
THAT THE RESULTS SUGGESTED

288

The tremendous difference in the distribution of the subjects
by ESA when compared to CA suggested another way of classifying the
boys. By crossing CA with BSA, I determined the prOportion of the boys
in this sample who displayed accelerated, standard, and delayed biological

maturation (Table 0-1). Those subjects whose CA was higher than ESA, I

Table 0-1

The proportion of the boys in this sample in each cell
when CA is crossed with BSA

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

11.0 l7(43.6) 45(20.7) 10(l.8) DELAYED
12.0 20(51.3) 104(47.9) 140(25.1) 31(6.5) ' 4(1.7)

13.0 2(5.1) 50(23.0) 215(38.S) 109(22.9) 14(5.9)

14.0 15(6.9) 150(26.9) 173(36.3) 47(19.9) 4(8.5)
15.0 3(1.4) 38(6.8) 136(28.6) 110(46.6) 16(34.0)
16.0 4(o.2) 24(5.0) 53(72.5) 21(24.7)
17.0 ACCELERATED 1(o.2) 3(0.6) 7(3.0) 5(10.6)

 

labeled delayed—maturers. Those subjects whose CA equalled ESA, I labeled
standard-maturers. And, those subjects whose CA was lower than ESA, I

labeled accelerated-maturers.

In Table 0-2, I provided the percentage of accelerated, standard,
and delayed-maturers.
Table 0-2

The percentage of the accelerated, standard, and delayed-maturers
across the six CA intervals

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 56.4 31.4 34.6 34.3 25.9 12.8
Standard 43.6 47.9 38.5 36.3 46.6 44.7

Delayed ---- 20.7 26.9 29.4 27.5 42.5

289

Because the 12.0 through the 15.0 age range contains the greatest
variability in adolescent male biological maturation, I will consider
those ages of prime importance in the following analyses. The percent of
standard-maturers decreased from 48% to 36% from 12.0 to 14.0 but returned
to 42% at 15.0. The percent of the accelerated-maturers increased from
31% at 12.0 to a high of 35% at 13.0 and decreased to 26% at 15.0. The
percent of delayed-maturers increased from 21% at 12.0 to 29% at 14.0 and
settled back to 28% at 15.0.

To verify this new classification procedure, I needed to determine
the variances for the motor proficiency tests (Table 0-3).

Table 0-3

The variances for the motor proficiency tests when classifying
by CA, BSA, and the accelerated, standard, and delayed

 

 

classifications
Sit-ups Pull—ups Standing Vertical Shuttle
' Long Jump Jump Run
CA 763.55 9.97 94.97 8.82 0.86
ESA 800.28 10.33 93.82 8.66 0.87
Acclerated 739.72 9.84 103.94 10.30 0.91
Standard 757.98 9.78 89.25 7.32 0.78
Delayed 802.07 9.27 71.49 6.78 0.81

 

For the sit-ups test, the accelerated and standard categories
both displayed slightly lower variances than either CA or BSA. The
accelerated-maturers had more homogeneity than either the standard or
delayed-maturers. The reason for this finding might be that more of the
accelerated-maturers attained the maximum number of sit-ups. If that
hypothesis is correct, then the accelerated-maturers will be more homo-
geneous than the standard or delayed-maturers during the early adolescent
years.

For the pull-ups test, the accelerated, standard, and delayed cate-

gories all had lower variances than either CA or BSA. The delayed-maturers

290

had more homogeneity than the other categories. The reason for that
finding might be that the delayed-maturers cannot do very many pull-ups,
therefore, they score similarly. Conversely, some accelerated-maturers
can do many pull-ups, hence, they display less homogeneity. If that
hypothesis is correct, then the variances for the delayed-maturers should
increase across CA.

For the standing long jump test, the standard and delayed cate-
gories both displayed lower variances than either CA or BSA. The delayed
category had a 24% decrease in variance from the ESA variance. The
delayed-maturers again had a lower variance than either of the other two
categories. The reason for this finding might be that a few highly-
skilled and motivated accelerated-maturers score far better than their
peers. If this hypothesis is correct, then the variances for the
accelerated-maturers will increase across CA.

The results of the vertical jump test parallel the standing long
jump results.

For the shuttle-run test, the standard and delayed categories both
displayed slightly lower variances than either CA or BSA. The standard-
maturers had slightly more homogeneity than the delayed-maturers and 14%
more homogeneity than the accelerated-maturers. This finding is a puzzle.
Apparently, biological maturation does not influence the shuttle run tests
in the same manner as it effects the other tests. Because the shuttle run
test measures agility rather than strength or power, the hypothesis might
be that biological maturation does not effect agility.

In Tables 0-4 through 0-8, I have presented the variances for the
five motor proficiency tests across the six CA intervals. The asterisk

indicates that the variance was lower than either the CA or ESA variances.

291

Table 0-4

The sit-ups variances across the six CA intervals for the
accelerated, standard, and delayed-maturers

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

Accelerated 7.14* 6.00* 7.25* 13.55 15.47 30.20

Standard 9.10* 5.19* 8.80* 8.86* 15.89 4.38*

Delayed ---- 6.59* 7.57* 9.49* 13.55 6.48*
Table 0-5

The pull-ups variances across the six CA intervals for the
accelerated, standard, and delayed-maturers

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

Accelerated 7.14* 6.00* 7.25* 13.55 15.47 30.20

Standard 9.10* 5.19* 8.80* 8.86* 15.89 4.38*

Delayed ---- 6.59* 7.57* 9.49* 13.55 6.48*
Table 0-6

The standing long jump variances across the six CA intervals
for the accelerated, standard, and

delayed maturers

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 95.64 76.97* 89.12* 125.44 157.41 288.50
Standard 20.03* 65.66* 72.01* 105.54 115.91 173.93
Delayed ----- 61.22* 67.42* 70.19* 88.05* 92.83*
Table 0—7

The vertical jump variances across the six CA intervals for
the accelerated, standard, and delayed-maturers

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 7.92* 8.19* 8.80* 12.55 12.25 17.30
Standard 4.01* 4.88* 6.85* 11.09 8.52* 9.53
Delayed ---- 5.89* 5.29* 6.83* 11.70 7.47

 

292

Table 0-8

The shuttle run variances across the six CA intervals for
the accelerated, standard, and delayed-maturers

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 1.32 1.11 0.90 0.82* 1.01 0.90
Standard 0.50* 0.73* 0.83* 0.90 0.63* 0.55*
Delayed ---- 0.83* 0.93 0.77* 0.72* 0.69*

 

For the sit-ups test, I hypothesized that the accelerated-maturers
would have lower variances during the early years than the standard and
delayed-maturers. Not only does that hypothesis stand up, but at 14.0,
15.0, and 16.0, the variances for all of the categories decreased drama-
tically. The variance for the 15.0 standard-maturers reduced 41% from the
CA variance.

For the pull-ups test, I hypothesized that the delayed-maturers
had lower variances because they could not do many pull-ups. Therefore,
as the CA intervals increased, the variances for the delayed-maturers
should increase. From 12.0 to 15.0, the variance for the delayed-maturers
increased from 6.59 to 13.55. The pull-ups results provides no pattern.
At 12.0 and 14.0, the standard-maturers have the lowest variances. At
13.0, the accelerated-maturers have the lowest variance. And, at 15.0,
the delayed-maturers have the lowest variance. However, we should note
that with the one exception of the accelerated-maturers at 14.0, all of
the categories for the 12.0, 13.0, and 14.0 CA levels have lower variances
than either CA or ESA. And, the variance for the 12.0 standard-maturers
is 38% less than the CA variance.

For the standing long jump test, I hypothesized that the highly-
skilled and motivated accelerated-maturers created the higher variances

for the accelerated-maturers category. Therefore, as the CA intervals

293

increased, the variances for the accelerated—maturers should increase.
From 12.0 to 15.0, the variance for the accelerated-maturers increased
from 76.97 to 157.41. There are two very clear patterns in the standing
long jump results. (1) As CA increases, the variances for all three
categories increase. (2) At every CA level, the accelerated-maturers
show the least homogeneity and the delayed-maturers Show the most homo-
geneity. Also, we should note that with the two exceptions of the
accelerated and standard-maturers at 14.0, all of the categories for the
12.0, 13.0, and 14.0 CA levels have lower variances than either CA or ESA.

For the vertical jump test, I hypothesized the same situation as
for the standing long jump test. And indeed, from 12.0 to 15.0, the
variances for the accelerated-maturers increased from 8.19 to 12.25 with
a high of 12.55 at 14.0. Unlike the standing long jump finding, however,
every category did not increase as CA increased. The accelerated and
standard-maturers increased from 12.0 to 14.0, but they both decreased
from 14.0 to 15.0. Also, unlike the standing long jump findings, the
relationship between the three categories did not remain constant across
the CA intervals. At 12.0 and 15.0 the standard—maturers had the most
homogeneity. At 13.0 and 14.0, the delayed-maturers had the most homo-
geneity. Also, we should note that with the three exceptions of the~
accelerated-maturers at 13.0 and 14.0 and the standard-maturers at 14.0,
all of the categories for the 12.0, 13.0, and 14.0 CA levels have lower
variances than either CA or BSA.

The lower variances for the 12.0, 13.0, and 14.0 CA levels for
pull-ups, standing long jump, and vertical jump indicate that this method
classifies adolescent males during the ages of the greatest variability

in their biological maturation better than CA or BSA, and, some of the

294

reductions in variance are considerable. For example, the 38% reduction
for standard-maturers at 12.0 for pull-ups, the 35% reduction for delayed—
maturers at 12.0 for standing long jump, and the 44% reduction for
standard-maturers at 12.0 for vertical jump.

For the shuttle run test, the results were puzzling. The variances
of the accelerated-maturers decreased from 12.0 to 14.0 and increased
from 14.0 to 15.0. The variances of the standard-maturers increased
from 12.0 to 14.0 and decreased from 14.0 to 15.0. The variances of the
delayed-maturers increased from 12.0 to 13.0 and decreased from 13.0 to
15.0. The standard-maturers had the most homogeneity at 12.0, 13.0, and
15.0 with the delayed-maturers the most homogeneous at 14.0. Three of
the four variances from 12.0 to 15.0 were lower than either CA or ESA
for both the standard and delayed-maturers. Only the accelerated-maturers
at 14.0 had a variance lower than either CA or ESA. Those results confuse
an already puzzling situation. The shuttle run test appears uneffected
by biological maturation.

At this point in analyzing the data, I realized that the analysis
was not what I wanted. True, I wanted to classify the boys at each CA
level on the basis of the relationship between their CA and BSA. However,
crossing CA and BSA intervals did not provide that. Recall that the month
range for the 13.0 year old is 151 to 163 months. If a boy's CA was 151
months and his BSA was 150 months, then the analysis classified him a
delayed-maturers. Obviously, the analysis mis-categorized him. Further,
if a boy's CA was 151 months and his BSA was 163 months, then the analysis
classified him a standard-maturer . A twelve month difference in CA and
BSA should not be a standard-maturer. Therefore, I re-programmed the

analysis to subtract ESA from CA. If the difference was -6.00 or more,

295

then I had the analysis classify the boy as an accelerated-maturer. If
the difference were between -6.00 and +6.00, then the analysis classified
the boy as a standard-maturer. And, if the difference were +6.00 or more,
then the analysis classified the boy as a delayed-maturer.

Except that the analysis classified the boys by this more precise
procedure, Tables 0-9 through 0-13 parallel Tables 0-4 through 0-8.

Table 0-9
The sit-ups variances across the six CA intervals for the more

precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age
11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 924.62 814.40 814.40 694.72* $00.21* 925.36
Standard 976.76 679.39* 870.26 885.60 480.51* 408.31*
Delayed ------ 1115.61 869.84 709.35* 629.13* 887.32

Table 0-10
The pull-ups variances across the six CA intervals for the more

precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 8.20* 4.97* 6.79* 12.61 13.11 19.88
Standard 6.44* 5.91* 8.25* 8.99* 15.08 14.52
Delayed ~--- 7.66* 7.85* 8.39* 17.32 9.18*

 

Table 0-11

The standing long jump variances across the six CA intervals for
the more precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 118.65 78.16* 83.19* 98.70 91.48* 264.97
Standard 35.13* 66.43* 72.36* 113.21 131.87 96.19

Delayed ------ 56.14* 68.58* 70.73* 95.68 93.60*

296

Table 0-12

The vertical jump variances across the six CA intervals for
the more precise accelerated, standard, and
delayed-maturers classifications

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

Accelerated 14.16 6.89* 8.02* 10.84 10.02 12.73

Standard 3.51* 5.16* 7.25* 10.90 8.81 7.62*

Delayed ----- 6.39* 5.35* 6.37* 10.22 10.80
Table 0-13

The shuttle run variances across the six CA intervals for the
more precise accelerated, standard, and
delayed-maturers classifications

 

Chronological Age

 

 

v 11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 1.87 1.32 0.84* 0.73* 0.75* 0.76*
Standard 0.59* 0.67* 0.90 0.94 0.74* 0.73*
Delayed ---- 1.00 0.88 0.74* 0.64* 0.89

 

A comparison of the tables revealed the following:

(1) For the sit-ups test, the more precise procedure found more
homogeneous classifications for the accelerated-maturers at 13.0, 14.0,
and 15.0, the standard-maturers at 12.0, and the delayed-maturers at 13.0
and 15.0. To the list of variances that were lower than either CA or ESA,
the new procedure added the standard-maturers at 12.0.

(2) For the pull-ups test, the more precise procedure found more
homogeneous classifications for the accelerated-maturers at 12,0, 13.0,
14.0, and 15.0, the standard-maturers at 13.0, 14.0, and 15.0, and the
delayed-maturers at 14.0. The new procedure had the same groups in its
list of variances that were lower than either CA or ESA. Additionally,
compared to the CA variances the reduction in variance of the accelerated-
maturers at 12.0 was 50%.

(3) For the standing long jump test, the more precise procedure

found more homogeneous classifications for the accelerated-maturers at

297

at 12.0, 13.0, 14.0, and 15.0, the standard-maturers at 15.0, and the
delayed—maturers at 12.0. To the list of variances that were lower than
either CA or ESA, the new procedure added the accelerated-maturers at
15.0, but lost the delayed-maturers at 15.0. Although the variances for
all of the accelerated-maturers groups decreased, their relationships to
the standard and delayed-maturers did not change.

(4) For the vertical jump test, the more precise procedure found
more homogeneous classifications for the accelerated-maturers at 12.0,
13.0, 14.0, and 15.0, the standard-maturers at 14.0, and the delayed-
maturers at 14.0 and 15.0. To the list of variances that were lower than
either CA or ESA, the new procedure added the acceleratedematurers at
13.0, but lost the standard-maturers at 15.0.

(5) For the shuttle run test, the more precise procedure found
more homogeneous classifications for the accelerated-maturers at 13.0,
14.0, and 15.0, the standard-maturers at 12.0, and the delayed—maturers
at 13.0, 14.0, and 15.0. To the list of variances that were either lower
than either CA or ESA, the new procedure added the accelerated-maturers
at 13.0 and the delayed-maturers at 12.0.

Before examining the mean values for the new classifications, we
should examine the distribution of the 1578 subjects (Table 0-14).

Table 0-14
The distribution of the 1578* subjects across the six CA intervals

by the more precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated ll 37 166 178 72 12
Standard 17 96 231 170 108 30
Delayed 0 47 156 148 67 16

 

*Sixteen subjects were in the 17.0 CA interval

298

If physical education teachers construct their classes to meet
the needs of the standard-maturers, then they will not meet the needs of
58% of the boys at the 13.0 CA level.

Tables 0—15 through 0-19 provide the mean values for the five
motor proficiency tests across the CA intervals.

Table 0-15
The sit-ups mean values across the six CA intervals for the

more precise accelerated, standard, and delayed—
maturers classifications

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

Accelerated 41.27 54.35 69.71 76.85 85.11 65.58

Standard 55.53 50.98 65.19 72.04 86.46 84.97

Delayed ----- 54.32 64.46 77.42 84.51 82.88
Table 0-16

The pull—ups mean values across the six CA intervals for the
more precise accelerated, standard, and delayed—
maturers classifications

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0

Accelerated 2.00 1.76 2.92 4.16 6.36 4.33

Standard 2.24 2.59 2.95 3.51 5.92 7.97

Delayed ---- 3.34 3.46 3.25 4.99 6.13
Table 0-17

The standing long jump mean values across the six CA intervals
for the more precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 55.36 59.81 66.04 71.78 76.31 69.33
Standard 57.59 59.03 63.49 66.72 71.90 76.53

Delayed ----- 59.83 62.13 64.64 69.16 69.56

299

Table 0-18

The vertical jump mean values across the six CA intervals for
the more precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 10.82 12.95 14.30 16.31 17.42 15.00
Standard 11.47 12.42 13.48 15.03 15.94 16.37
Delayed ----- 12.70 12.85 13.78 14.55 14.44
.1.
Table 0-10 x

The shuttle run mean values across the six CA intervals for the
more precise accelerated, standard, and delayed-
maturers classifications

 

Chronological Age

 

 

11.0 12.0 13.0 14.0 15.0 16.0
Accelerated 11.45 11.51 10.81 10.39 10.46 10.91
Standard 11.55 11.33 10.97 10.70 10.55 10.34
Delayed ----— 11.31 11.04 10.88 10.76 10.96

 

For the sit-ups test, while within each category the mean values
increase across CA, within each CA interval the pattern varies. The
accelerated—maturers have the highest mean values at 12.0 and 13.0. At
14.0, the delayed-maturers have a slight advantage over the accelerated-
maturers for the highest mean value. And, the standard-maturers have a
1.3 sit-up advantage over the accelerated—maturers who lead the delayed-
maturers by .6 situps at 15.0. Apparently, biological maturation does not
effect the sit-ups test.

For the pull-ups test, the results display a paradoxical pattern.
From 12.0 to 13.0, one pattern appears and from 14.0 to 15.0 the opposite
pattern appears. Between the 13.0 and 14.0 CA intervals, the pattern
reversed. At 12.0 and 13.0, the delayed-maturers have the highest mean
values and the accelerated—maturers the lowest. However, at 14.0 and 15.0,
the accelerated-maturers have the highest mean values and the delayed-

maturers the lowest. The accelerated and standard-maturers both show

300

mean value increases from 12.0 through 15.0. But, the mean values for
the delayed-maturers are essentially equal at the 12.0, 13.0, and 14.0
CA levels followed by an increase at 15.0.

For the standing long jump test, within each category the mean
values increased across CA. However, the mean values for the three
categories were essentially equal at the 12.0 level. Thereafter, the
accelerated-maturers displayed a clear advantage. Biological maturation
definitely effected the results of the standing long jump test.

For the vertical jump test, the pattern followed the standing
long jump results. And, the shuttle run test followed the standing long
jump and vertical jump results. Whereas the shuttle run variances defied
interpretation, its mean values demonstrated that biological maturation

effects the shuttle run test.

ICHIGQN STQTE UN V.

I LIBRARIES
H I” [HIUIWIHIIIIWH
9 7623

312 310070