LIERARY
Mlchigan Eitate
1 University

 

 

 

PLACE It RETURN BOX to move this checkout from your record.
To AVOID FINES ntum on or bdon date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Afflnnatlvo ActionlEqud Opportunity Institution
M1

 

 

 

BESLLEXCE OE I
RESPOXV
.&\' N'ESUGA"

 

THE INFLUENCE OF RACIAL IDENTITY AND CULTURAL VALUES ON
RESPONSES TO BIODATA EMPLOYMENT ITEMS:
AN INVESTIGATION OF DIFFERENTIAL ITEM FUNCTIONING

By

David John Whitney

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1 995

 

I}: KILLENCE OF
RESPONSE

 

 

Tm prtscm sud} pm?

‘17 .3.“ _ 1" .

I. ..g:oup incm .ws 1
?*~ A 4'. .
tart 3:3; dcmcg'ashlc

at: mm! will): prefer

mat {1961) we met

153m for cultural vai-
Iii-5,1": '
m. emplomcm itc'
r I ‘ .
Jae malted ugh: \\

3:25: ‘ '
melpams to ex:

ding 50cm
AM,

in;

7"...
="-‘..“.‘-">"a '
'5“‘phLlC
I"ace Th
\ .‘
t1.”

ABSTRACT
THE INFLUENCE OF RACIAL IDENTITY AND CULTURAL VALUES ON
RESPONSES TO BIODATA EMPLOYMENT ITEMS:
AN INVESTIGATION OF DIFFERENTIAL ITEM FUNCTIONING
By

David John Whitney

The present study proposed and examined a conceptual model of the influences of
racial subgroup differences in responses to biodata employment items. The model
suggested that demographic race influences racial identity, which in turn was expected to
lead to cultural value preferences. The cultural value typology of Kluckhohn and
Strodtbeck (1961) was used as a theoretical framework of cultural values. Differences in
preferences for cultural values were hypothesized to lead to differential item functioning
on biodata employment items.

One hundred eight White and ninety-nine Black college students were used as
research participants to examine the hypotheses of the study. Each participant was
administered a test battery consisting of 107 biodata items, along with scales assessing
random responding, social desirability, cultural values, cognitive ability, racial identity,
and demographic information.

Results indicate little support for the expectation that racial identity classification
would explain additional variance in the adoption of cultural values over that explained
by demographic race. The strength of racial identity was found to impact the adoption of

several cultural values, however. Mixed results were found for expectations of Black-

 

T-A . “Ta. ‘0‘
‘12 “Tim” “53:1 \\

5...: ;. 15:1: relatix‘c cc;

but» “1

rim cor‘istcm with i

122...? wines and :3;

5122131". of the bit ‘3'..-

m: mum's. Pair;

 

$23215) Ibc race of in:
rimmed by boz‘n cu”;
:rzzon :lm a won in?

:‘21‘ It 1‘. A
- “mes van success

W“ 9--
‘fi-i mZCmCes

White preference differences for cultural values. Although Blacks and Whites clearly
differed in their relative endorsements of various cultural values, these differences were
not always consistent with theory. Some evidence was found supporting the expectation
that cultural values and race would impact responses to biodata employment items.
Overall, 21% of the biodata items were influenced by differences in cultural values
between respondents. Thirty-three percent of the responses to biodata items were
influenced by the race of the respondent. Although responses to only 9 items were found
to be influenced by both cultural values and race, evidence was found supporting the
contention that a priori inspection of biodata employment items for the presence of

cultural values can successfully predict items that would be susceptible to racial subgroup

response differences.

thcughl feel 1 has.
tr: 1‘» most inx‘zflua'bic _

‘2: at bat. Many their.

 

7-5311 --1 strongly believe

2mm ur'rh my fellow "i

:5? 3° mugh 1hi8 proce
3.321%“. I ,

"“.- C11 thrlmd. an.
‘; 301': like 63er in the

2:": .
'F‘m‘agam fix:

I
‘Wi Show {00%

11‘: a. .

in. .
“5:3".

«.035. COTE] Gable S

2:2
‘k .
Show! “u
V3]
1331.],

ACKNOWLEDGMENTS

Although I feel I have learned a great deal over the last 5 years at MSU, I strongly
believe my most invaluable lessons have come from the individuals I have shared my
time with here. Many thanks are extended to my fellow graduate students (both past and
present) - I strongly believe my graduate education has relied heavily upon my
interaction with my fellow “partners in crime”. Not to mention the social support needed
to get me through this process! A special thanks to those in my cohort- Dennis Devine,
Stan Gully, Jen Hedlund, and Eleanor Smith. I’ll never forget our trips to Denny’s late at
night (more like early in the morning), the tumult we all felt our first year, the occasional
camping extravaganza, the celebrations at Chi Chi’s, the long gripe sessions which did a
world of good, snow football games, lunches at the Break Place (chili, anyone?), singing
nonsense out of tune (okay, maybe that’s more me than any of you), the shared
expectations, Coral Gable’s Happy Hour, and of course, the many hours of just plain
spending time together. You’ve all influenced my life immeasurably - I sure as hell
hope it’s for the better!

My deepest thanks are extended to Rick DeShon and John Hollenbeck for your
encouragement and criticisms throughout my dissertation process. I truly appreciate your
help on this project! I have been extremely fortunate during my tenure at MSU to have

not only 1 mentor, but 2. I was lucky enough to have them both serve on my dissertation

iv

 

" A 4 1.3 CL
lr"~fi' She Linus“
ashram: I can possz.

flel
‘ ' 'pi'oqrrfiqou‘
‘THQ‘QN 34.01.qu
fl.

  
  
  

3:75:33); however. I \I 2.
sites Tasks for hcipir. g
'25: 2:: Nth wonderfal ins:

This dissertation vs 0'

533: and ovcrmghr '31:;

£72332 It

6.

committee. Mike Lindell and Neal Schmitt have provided me with more during my 5
years here than I can possibly describe appropriately in this limited context. I need to
thank you both for supporting me with TA’S and RA’S throughout my 5 years. More
importantly, however, I want to thank you for your constant patience and belief in my
abilities. Thanks for helping to clarify my many, many, many confusions along the way.
You are both wonderful instructors in NO psychology as well as life.

This dissertation would never have been completed without Neal’s financial
support and overnight turnaround on revisions. I don’t know how you do it, but I do
appreciate it.

I want to thank A.J. Alejano for demonstrating that a dissertation could be
completed from start to finish in a single year. Thanks so much more, though, for your
constant confidence and encouragement. E-mail is a cool thing.

From the very beginning of this dissertation to its final completion, no one has
provided as much constant emotional support and understanding as Regina O’Connell.
Thanks for listening to me like no else ever has.

Finally, I want to acknowledge the greatest teachers in my life, my parents -
Carmine and Marion Caccamo, and John and Mary Whitney. This accomplishment is

yours as much as mine.

1570f TABLES

SET OF FIGURES .

 

 

 

 

FRIED'L’CIION
Dismal Item Fun:
Differential liar.
Mercurial Item
SUD ,tttit't Empioxme
Real" Identit} Iheon
Racial Idenut} I
Ctttumi Influences ,
TORard a Def“ ,-
\aluc OTICUIAUC
mill VS. Soda] Cl
Nuts and I
II! PM! Stu-dt
H L:flpothesis 1
HE’Uthcsis 2
H‘s’tes 3
Himthcsis 4
E‘I‘Dthesis 5

‘ N'theis6

urn"
anD
Staph

TABLE OF CONTENTS

Page
LIST OF TABLES . . . . . . . . . . . . . . . . . ix
LIST OF FIGURES . . . . . . . . . . . . . . . . . xi

INTRODUCTION . . . . . . . . . . 1

Differential Item Functioning arid Cognitive Ability . . . . . . . 5

Diflerenfial Item Functioning Methodology. . . . . . . . 7

Diflemnfial Item Functioning: Findings . . . . . . . . . 13
Subjective Employment Tests . . . . . . . . . . . . . 23
Racial Identity Theory . . . . . . 27

Racial Identity Theory. Toward a Meaningful Construct . . . . 33
Cultural Influences . . . . . . . . . . . . . 35

Toward a Definition of Culture . . . . . . . . . . . 37

Value Orientations . . . . . . . 38
Racial vs. Social Class Difiemnws In Value Orientations. . . . . . 53
Cultural Values and Item Responding. . . . . . . . . . . 56
The Present Study . . . . . . . . . . . . . . . 58
Hypothesis 1 . . . . . . . . . . . . . . . . . 59
Hypothesis 2 . . . . . . . . . . . . . . . . . 6O
Hypothesis 3 . . . . . . . . . . . . . . . . . 61
Hypothesis 4 . . . . . . . . . . . . . . . . . 62
Hypothesis 5 . . . . . . . . . . . . . . . . . 64
Hypothesis 6 . . . . . . . . . . . . . . . . . 65

Sample . . . . . . . . . . . . . . . . . 66
Measures Administered. . . . . . . . . . . . . . . 66

My"
IA

“LI-II! '

Demptvc Imiomai;
Dmomhaily Dev.
Race and 333mm;
CufimI Value Patter?II
Cain-rat Value Differ:
[airman ARE.
Hierarchical Re:
Calm Vaiues 8; Bi-
Rear Subpom C;
REIEI Sili‘gmups 33‘:
R3221 SHHETOUPS. Ex;

 

Blot‘ihta Res;
Rm SUIETOLI’QS. Ac:
Biotiaia Res;

RIC The Real SIOQ”
3H mt‘aRace: “Inch

 

 

Bi’SSION. . . -
Hymmcsis 1: Race 5L
IImthesis 2: Race V!
I};titiei53: Racial:

Twical Ame
Reheat: Cultura
II}}mheis S: Cumin
Hypothesis 6: Racial

, VaIucs.tooT'1
Etpfetaton' Analvsis‘
H‘Iot'aior) Anaji'sis
Hpntatiom and Con"
Racial Identity \

RESULTS .
Descriptive Information.

Demographically Determined Race and RaCial Identity Classification.

Race and Background Information. . . .

Cultural Value Pattems and Racial Subgroup .

Cultural Value Differences Between Racial Subgroups
Univariate Analyses . . . .
Hierarchical Regression Analyses

Cultural Values & Biodata Responses . .

Racial Subgroups, Cultural Values, and Biodata Responses

Racial Subgroups and Biodata Responses . .

Racial Subgroups, Expected Cultural Value Influences, &

Biodata Responses .

Racial Subgroups, Actual Cultural Valrie Influences, &

- Biodata Responses . . . . .

Race: The Real Story?. . .

DIF and Race: Which Racial Subgroup was Hurt More? .

DISCUSSION. . .
Hypothesis 1: Race & Racial Identity. . .
Hypothesis 2. Race vs. Racial Identity, and Cultural Values .
Hypothesis 3: Racial Subgroups and Cultural Values of the
Typical American.

Hypothesis 4: Cultural Value Differences between Racial Subgroups.

Hypothesis 5: Cultural Value Differences and Biodata Responses

Hypothesis 6. Racial Subgroups and Biodata Responses (& Cultural

Values, too!). . .

Exploratory Analysis 1: Variance In Race. . .
Exploratory Analysis 2: DIF and Overall Test Score .
Implications and Conclusions . . . .

Racial Identity .

Racial Subgroups and Cultural Values . .

Subgroup Differences In Biodata Responding

Recommendations. . . . . .
Study Limitations .

Racial Identity. .

Racial Subgroups and Cultural Values . .

Racial Subgroups and Biodata Responding .
Future Directions . . . . . .

APPENDIX A .

APPENDIX B .

vii

 

74
74
84
86
89
94
95
95
11 1
121
122

125

137
139
147

153
153
154

155
155
157

157
159
159
160
160
161
163
167
168
168
169
170
172

175

222

25031 E .

LIST OF REEREXCES

APPENDIX C .

APPENDIX D.

APPENDIX E .

LIST OF REFERENCES .

viii

226

233

255

290

7:4: I: Thermally-tier

 

.‘el; Orerali Meats. S'
aes; Principal Comm
left: Variable Intereo:

.IIS; Correlation of R
Cu'iturai V:

e26: Resu'tts of t-tesz:

.v.
38a,
5

... f; Items of Mes:

tiz-Iei
I RCSLIIQ OIH
‘12: S
W‘ of
:3

Table 1:

Table 2:

Table 3:

Table 4:

Table 5:

Table 6:

Table 7:

Table 8:

Table 9:

Table 10:

Table 11:

Table 12:

Table 13:

Table 14:

Table 15:

LIST OF TABLES

Theoretically-derived Dominant Cultural Value Orientation Patterns .

Overall Means, Standard Deviations, and Reliability Estimates
Principal Components Analysis of Racial Identity Strength
Variable Intercorrelations

Correlation of Race and Racial Identity Classification with
CulturalValues

Results of t-tests Examining Race and Background Information .
Results of t-test Comparisons of Cultural Value Alternatives - Blacks
Results of t-test Comparisons of Cultural Value Alternatives - Whites
Cultural Value Patterns of Blacks and Whites .

Results of t-test Comparisons of Cultural Value Differences by Race

Results of Hierarchical Regression Analyses with Cultural Values
Regressed on SES, Race, and Racial Identity Strength

Summary of Racial Group Preferences for Cultmal Values .

Results of the test for significant difi‘erences in Chi-square for Loglinear
Analyses examining Cultural Value‘Option Choice Interaction .

Degree to which Cultural Value Orientations Influenced Biodata
Response Option Selection, as Predicted . . .

Results of the test for significant difference in Chi-square for Loglinear
Analyses examining Racial Subgroup*0ption Choice Interaction

50

75

79

81

85

87

91

92

93

96

98

108

113

119

124

Tel: 16; Number of Bio.
Consimz

IE: IT. Biodata Items III
Rota} Stir.

3:318: Results orog:
Chariot}.

 

.3: 19' Result: of Ion

b

v y -
Imam: '.
3

III; Results of 102:5
I'r'oanicit)‘.

The]: Racial Group En

 

Deletion of .

Table . 16.

Table 17:

Table 18:

Table 19:

Table 20:

Table 21:

Table 22:

Number of Biodata Items Exhibiting DIF by Racial Subgroup
Consistent with Theoretical Expectations.

Biodata Items Exhibiting Both a Cultural Value‘Option Choice and
Racial Subgroup*Option Choice Interaction . . .

Results of Logistic Regressions with Race Regressed on SES, Sex,
Urbanicity, Cognitive Ability, and Cultural Values .

Results of Logistic Regressions with Race Regressed on SES, Sex,
Urbanicity, Cognitive Ability, and Biodata Items .

Results of Logistic Regressions with Race Regressed on SES, Sex,

Urbanicity, Cognitive Ability, Cultural Values, and Biodata Items .

Racial Group Encumbered by Flagged Biodata Items

T-tests of Overall Biodata Test Score by Racial Subgroup Following
Deletion of DIF Items. . . . . . .

135

138

141

143

145

149

151

 

Fgre I: A Conceptual J
ngz; Interaction of R-
7;: 3: Interaction of RI.
@342 Inhe'action of R3.

7.5.15: Endorsement of

 

 

LIST OF FIGURES

Figure 1: A Conceptual Model of the Research .

Figure 2: Interaction of Racial Identity Strength & Race when DV is Mastery .
Figure 3: Interaction of Racial Identity Strength & Race when DV is Linea!
Figure 4: Interaction of Racial Identity Strength & Race when DV is Individual

Figure 5: Endorsement of Cultural Values by Race .

 

104

105

106

110

INTRODUCTION

The search for an unbiased estimate of an individual's performance abilities has
long been a mainstay of Industrial/Organizational psychologists. Ethical and legal
concerns related to test fairness continue to demand the thoughtful consideration of
practitioners and researchers alike. Most of the research in this area has considered
differential item performance of disparate racial groups on cognitive ability tests (e.g.,
Butler-Omololu, Doster, & Lahey, 1984; F reedle & Kostin, 1990; Rudner, Getson, &
Knight, 1980; Schmitt, 1988; Schmitt & Dorans, 1990; Stricker, 1982). This literature
has produced many suggestions for fair cognitive ability test construction, but few
insights into the causes of differential performance on particular test items for selected
subgroups. Yet selection test batteries rarely consist solely of cognitive ability tests.
Despite the paucity of knowledge concerning factors influencing differential responding
on cognitive ability test items, still less is known about the influences of differential
responding on subjective test items such as personality, biodata, and situational judgment
items. The current study examines difl'erential item functioning of items fiom tests with
subjective responses. Unlike many of the investigations conducted on cognitive ability
tests, the present study combines past empirical findings with theoretical speculation to
offer several a priori hypotheses concerning characteristics of biodata employment items

that are likely to lead to differential item functioning.

2

Figure 1 presents a model of the proposed research. Initially, a brief review of the
item bias literature examining cognitive ability tests is pursued in order to lay the ground-
work pertaining to the current status of the field. Theoretical offerings into the causes of
differential responding will then be considered. Specifically, the current study proposes
to investigate the oft-suggested, but seldom researched, cultural influences on item
response inclinations. The proposed study argues that before considering cultural
influences, researchers must initially focus their concern on accurately classifying
individuals in terms of their cultural heritage. Racial identity is offered as a
psychological construct that would be expected to account for greater variance in cultural
identification than traditionally utilized physical characteristics such as skin color. The
current study proposes that identification of an individual's racial identity should help
explain an individual's preference for certain cultural values. The concept of racial
identity is explored, and several criticisms of current conceptualizations of the construct
are offered. Although the construct of culture also has suffered from conceptual
ambiguity, the work of F. Kluckhohn and Strodtbeck (1961) has been useful in
establishing several cultural values which are hypothesized to vary across peoples and
cultures (e.g., Beutler & Bergen, 1991; Carter, 1990; Chapman, 1981; Inclan, 1985).
These cultural values are expected to be consistent with the philosophies and culture
associated with a particular racial group. These same values, when reflected in options to
biodata employment tests, are proposed to result in differential response patterns for

people with disparate racial identities.

 

5030.3». 05 .0 .0303. 3:300:00 (

up 0539.”.-

 

 

mEm:
EoE>anm
Smegm

:0 “:0

k
I

 

 

 

 

 

 

mo:_m> >552
Doug—:62 .

 

 

 

 

 

 

 

 

5.88: on. a .28: 3.58:8 <
3 959“.

one:

 

 

 

A not: Is Intuit:

agent pQ‘ChOIO'glCa:

:Z-tterm race (eg. D:

   
 
 

soon for this common
esc'hologisu and Ar.
airy (fish. 1995). G
airtigator use and d
has: appropriately de:
cm°~’I‘~mahzatiorls of I
iti'ition of sitar one
171967) definition of St

‘EJUUpIharis social

4

A note is required to explain the terminology used to identify racial subgroups.
Recent psychological literature has been rife with arguments and opposing definitions of
the term race (e.g., Dole, 1995; Jensen, 1995; Rushton, 1995; Yee, 1994). A major
reason for this controversy is attributable to the fact that the use of the term “race” by
psychologists and Americans in general does not correspond to any particular biological
entity (Fish, 1995). Given the disparate conceptions of race, how should a well-meaning
investigator use and define race for research purposes? The answer to the question is
most appropriately determined on a case by case basis. However, given the diverse
conceptualizations of race, it is incumbent upon the researcher to provide a clear
definition of what one means by the term. The present study borrows Van den Berghe’s
(1967) definition of social race as its conception of race. According to this view, race is
“a group that is socially defined but on the basis of physical criteria” (cf. Lindell & Perry,
1992, p. 140).

For the purposes of this paper racial identity will be defined in terms of Janet
Helms’ definition. According to Helms (1990), racial identity refers to the "belief
systems that evolve in reaction to perceived differential racial-group membership" (p. 4).

The terms Asians, Blacks, Hispanics, and Whites are used in the present study to identify
racial subgroups formed on the basis of a single-item self-report measure. Alternatively,
the terms Afi'ican-Americans, Asian-Americans, Hispanic-Americans, and White-
Americans are used to denote subgroups formed on the basis of the assessment of an

individual's self-reported ascribed and reference group racial identity.

es. item. it is first a;
Differential Item In
In reference to s

7151 “its: the

 

 

 

multiply. dint

lust “hat If}:

We may ther

units. and any:
Em m mt“ early 0';-
heng to the Arm}
Its: , Med:

'1 I131: the III:

Clmif}. a him

of that Claim

memofs

uanSesn (p ‘
Hi.- s
oush Such Charg.

tel

WWW“ the

131: On
PTor
Med gro

5
Before a formal examination of the cultural influences on subjective employment
test items, it is first necessary to examine the meanings of test and item bias.
Differential Item Functioning and Cognitive Ability
In reference to standardized intelligence tests, Thorndike (1927) lamented:
"Just what they measure is not known; how far it is proper to add, subtract,
multiply, divide, and compute ratios with the measures obtained is not known;
just what the measures obtained signify concerning intellect is not known.
We may refer to these defects in order as ambiguity in content, arbitrariness in
units, and ambiguity in significance"(p. 28).
Even in these early days of ability testing, Thorndike was not alone in his criticism.
Reacting to the Army Alpha cognitive ability testing program, Walter Lippmann
(1922) stated:
"I hate the irnpudence of the claim that in fifty minutes you can judge and
classify a human being's predestined fitness for life. I hate the pretentiousness
of that claim, I hate the abuse of scientific method which it involves. I hate
the sense of superiority which it creates, and the sense of inferiority which it
imposes" (p. 146).
Although such charges may be levelled against all testing procedures in general,
criticisms such as these have spurred research interest into the differential impact of
tests on protected groups.
Two parallel but separate lines of research have examined differential

performance of selected subgroups (generally Blacks and Hispanics vs. Whites, or

melts. entries) I

shite. These hum:

 

 

 

tilewglltile ahtltt} ::
statues on the trite: .
Isiah-oat of test bias
Cole. I973 Ledsirlks.
mimic to C leary.

“A ICSt is biase

 

3 cTiteriori for

hm che cr

6

males vs. females). On the one hand researchers have examined bias in the test as a
whole. These investigations, for example, examine whether Blacks score lower on
the cognitive ability test (the predictor) than would be indicated by their relative
standings on the criterion measure the test was selected to predict. Although several
definitions of test bias have been introduced (e.g., Darlington, 1971; Thorndike, 1971;
Cole, 1972), Ledvinka (1979) states that the commonly accepted definition is
attributable to Cleary. Cleary (1968) states that:

"A test is biased for members of a subgroup population if, in the prediction of

a criterion for which the test was designed, consistent non-zero errors of

prediction are made for members of the subgroup. In other words, the test is

biased if the criterion score predicted fi-om the common regression line is

consistently too high or too low for members of the subgroup" (p. 115).

A second research approach has been the investigation of possible item bias.

The presumption of item bias investigations has been that there are some identifiable
cultural influences contained in the items which lead to differential responding by
members of difiemnt cultural subgroups. Once this content has been identified, the
biased items would presumably be removed fiom the test in an effort toward
increasing test fairness. Perhaps the earliest of such investigations was conducted by
Bells, Davis, Havighurst, Herrick, & Taylor (1951) who attempted to determine what
type of test items were most advantageous to children of high socioeconomic status

(SES), as well as those items on which children of low SES performed well. Only

 

3-3;; the last deco-CL:
- morass: ITAIS‘S

Angeli I I 993) ;
sent: ': conflict den
geera‘l‘ts accepted 5:.

er...) able (or profit

 

mobilities of ensue
="‘¢lfill€0\.LSl)‘ pros-r.
ltrlll (I981). for CI
group more than ano
some kind uhich ext
353811113] perform};
“33H." “’6 meanrt
iferenrtal item fir;
itiliitiorl of item b
[hieremjd Item F

“got? (199.

I?
K

em -. ,
4101mm:

7
within the last decade, however, have such item bias investigations become
commonplace (Tatsuoka, Linn, Tatsuoka, & Yamamoto, 1988).

Angoff (1993) indicates that as use of the phrase item bias flourished, a
semantic conflict developed between statistical and social uses of the term. The
generally accepted statistical definition of bias has argued that an item is biased if
equally able (or proficient) individuals, fi'om different groups, do not have equal
probabilities of answering the item correctly (Angoff, 1993). Other authors
simultaneously provided a social component to this definition. Shepard, Camilli, and
Averill (1981), for example, append the phrase "a kind of invalidity that harms one
group more than another." This second definition obviously requires a judgment of
some kind which extends beyond the straightforward statistical illustration of
differential performance. This dual use of the term bias led to confusion over what
exactly was meant by item bias. In order to assuage some of this confusion, the term
dzfl'erential item fimctiom‘ng (or, DIF) emerged to refer solely to the statistical
definition of item bias.

DWrential Item Functioning Methodology

Angoff (1993) provides an overview of the development of modern differential
item functioning methods, part of which is summarized here. Despite early efforts by
Cardall and Coffman (1964) and Cleary and Hilton (1968) to employ the use of an
analysis of variance methodology to investigate difl‘erential performance of test items,
these methods failed to catch the attention of subsequent researchers. Based in part

on Thurstone's (1925) absolute scaling techniques, Angoff (1972) offered a method

marred to a dcita-p

l
o

 

requires the calcula:
meg-um: conversi; I
|
denies are then pic-t
ms. fie plotting I
mm “tether bi;

“WW 3 pamcula;

~54»: _ .
"~sw0mard 11116”. L

 

 

 

an}; to consider the
to a faillmc to dttect t1
Faber items which .
tint-aces in their di

The method vi:
33:0le for differem

1 '1"
RA l {1'10ch, was 01-4.

AM 4.
wrung [0 “m re

M
L“ Q“ be Wading-

 

8
referred to as delta-plot or transformed item-difficulty (TID). The T‘ID method
requires the calculation of item p—values for each subgroup of interest, and for the
subsequent conversion of each p-value to a normal deviate. These pairs of normal
deviates are then plotted on a bivariate graph, with the two groups represented on the
axes. Afier plotting each of the pairs of normal deviates, the resulting ellipse
indicates whether bias is a problem. An item by group interaction is detected
whenever a particular item falls some distance from the plot of points. Despite its
straightforward interpretation and simple logic, this method has been criticized for
failing to consider the discriminating power of each item on the test, which may lead
to a failure to detect truly biased items (e.g., Cole and Moss, 1989; Shepard, 1981).
Further, items which are classified as biased may only appear to be so, due to
differences in their discrimination.

The method which has become the standard in the detection of DIF readily
accounts for difl‘erences in discrimination. This method, the item response theory
(IRT') model, was originally explicated by Lord (1952) and Lord and Novick (1968).
According to item response theory, the performance of an examinee on a specific test
item can be predicted by a set of latent trait variables. Further, the relationship
between the set of traits underlying item performance and the test performance of
examinees can be described by an item characteristic curve (ICC) (Hambleton,
Swaminathan, & Rogers, 1991). The ICC is an S-shaped curve of the proportion of
individuals at the same ability level who answer the given item correctly. Assuming

the test items are unidimensional and the item measures the same ability, an identical

 

am wouii be tugs.
coursing he sang.
, I
give firm mm e\:_~
An ICC is a m;
PW“ of succcsl
333951311125 of the :
MESH: curve c;
model mum rel-mi

mm Inc"1‘15 Sine:

The one New: 1};

 

What!

Pit.“

(1:-

hi Comm.
5 the difficulty
ism: number 0

1.5ka 10

32 large; the Value of

 

 

9
curve would be expected to be found for all samples, regardless of the subgroups
composing the samples. If two groups produce different item response curves for a
given item, there exists evidence that the above assumptions have been violated.

An ICC is a mathematical function which considers the relationship between the
probability of success on an item, the ability measured by the test, and the
characteristics of the item (Hambleton, Swaminathan, & Rogers, 1991). The item
characteristic curve can be composed of one to three parameters. The one parameter
model (ofien referred to as the Rasch model) is the most simplistic of the item
response models since it considers only a single parameter: the difficulty of the item.

The one parameter IRT model can be represented mathematically as:

efl-bt)
1 + em-bl)

 

Pz(9) =

where,

P,(9) is the probability that a randomly chosen examinee with ability 9 will answer
item i correctly.

b,- is the difficulty parameter of item i

n is the number of test items, and

e is the natural logarithm, 2.718

The larger the value of b, , the greater the difficulty of the item. This model is of

limited usefulness in the investigation of DIF because it assumes that all items are

    

mi mm:
mini com‘t‘é'i
has: Shepard I 1'“
Rszb one-patient
mid to TTD 1p; '.
The two para
i230 includes a dis;
meter mode‘i. w‘:
meter. denoted ;
2:1in scale. Thus.
berm indixiduals
Hmbiflon. Sm:
mm C, This pa:
“w inner the item
dimming the item
imam ”3901156 0;:

Whammy as;

 

10
equally discriminating (similar to the TID model above) and that items cannot be
answered correctly by guessing - an unlikely assumption in multiple choice tests.
Indeed, Shepard (1981) found a correlation of .99 between the DIF indices of TID and
Rasch one-parameter model. Since both models are similarly derived, the criticisms
provided to TID apply directly to the Rasch model as well.

The two parameter model is more satisfactory than the one parameter because
it also includes a discrimination index, but the most complete IRT model is the three
parameter model, which also adds an index for guessing. The discrimination
parameter, denoted an is proportional to the slope of the ICC at the point b, on the
ability scale. Thus, items with steeper slopes are more useful for drawing distinctions
between individuals at disparate ability levels. The final parameter is referred to by
Hambleton, Swaminathan, & Rogers (1991) as the pseudo-chance parameter, and is
denoted c, This parameter represents the probability that examinees of low ability
will answer the item correctly. This value is often lower than the chance probability
of guessing the item correct, as many examinees of low ability will be attracted to
incorrect response options. The three parameter IRT model can be represented

mathematically as:

“‘1”

ubl H‘

P5191

isthe probabzl

 

 

 

in :' corml)’.

1.

is the dlSCfimj:
is the difficult}
151139 pseudo-c
is a scaling fa:
I0 the normal 0

is the “antral To.

 

 

11

eDaMG-lu)

p.(e)=c.+(1-CJW

where,

Pi(9) is the probability that a randomly chosen examinee with ability 9 will answer

item i correctly.

a, is the discrimination parameter of item i

b, is the difficulty parameter of item i

c,- is the pseudo-chance parameter

D is a scaling factor introduced to make the logistic function as close as possible
to the normal ogive function

11 is the number of test items, and

e is the natural logarithm, 2.718

Another technique capable of identifying DIF was originally proposed by
Holland and Thayer (1988), and is based upon the work of Mantel and Haenszel
(1959). The Mantel-Haenszel technique requires the formation of a 2x2 table
developed by examining the frequencies of the correct vs. incorrect item responses,
and the two subgroups the researcher wishes to compare, matched on ability. The

Mantel-Haenszel index is calculated for each ability level. This index is the ratio of

353mm 0.:
ategones a co:
fact-e: by which
excee; the odd
Addie-mil): ti
hint about the
ism-ates of (ii!
item dimming
{1936) 9L1; €651
03.539 Watch}

3031 the ]

TIM; d1 gm

12
odds that one subgroup answered the item correctly to the odds that the second
subgroup answered the item correctly. Thus, when there is no difference in subgroup
performance on the item, the Mantel-Haenszel index is 1. Across all matched
categories a common odds ratio (am) is then produced, which determines the average
factor by which the odds that a member of one group responds correctly to the item
exceed the odds that a member of the second group responds correctly to the item.
Additionally, this common odds ratio is sometimes then transformed to center the
index about the value of zero. Although the Mantel-Haenszel technique can provide
estimates of differential item difficulty, it does not provide estimates of differential
item discrimination (as IRT does). Therefore, Angoff (1993) and Dorans and Kulick
(1986) suggest that the use of Mantel-Haenszel techniques be accompanied by a plot
of the Mantel-Haenszel indices by score interval.

Both the Mantel-Haenszel and the IRT approaches to the detection of DIF
examine difi'erential responses to the correct option of the item. Examining only the
correct choices for DIF would initially make sense, since it is only these options
which directly affect test scores. However, recent work has begun to examine each of
the distractors of an item (e.g., Dorans, Schmitt, & Bleistein, 1992; Green, Crone, &
Folk, 1989; Veale & Foreman, 1983). The reasoning behind this comprehensive
differential item functioning (CDIF), or alternatively, differential distractor
functioning (DDF), is that if two subgroups respond differentially to a distractor, this
could lead more members of one subgroup to get the item incorrect. Investigation of

differential distractor functioning has the further advantage of increasing the

twist b} a
9:)? will incr
comsed oft
named DD
table of optic:
categories It
item choice hj
titles are 3:
”mm is the
“htther tests,
3mm)“ ma
8; F 01k 1 l 989'
Option Choice
tn the main ef
Difefemjaz 1i

REScam]

H1iml‘mck’logie:

13

probability of finding difi‘erential response patterns over DIF. This is true simply
because, by examining each of the distractors instead of just the single correct item,
DDF will increase the number of options examined fourfold in a test with items
composed of four distractors each. Green, Crone, & Folk (1989), for example,
examined DDF by employing a methodology which used a three-way contingency
table of option chosen by subgroup by ability level. A main effect for option choice
corresponds to the differences in popularity of the various options. The interaction of
item choice by ability levels identifies the extent to which test-takers of differing
abilities are attracted to the various options of the item. The interaction of greatest
interest is the interaction between option choice and subgroup, which assesses
whether test-takers of different subgroups choose the various options more or less
frequently than would be expected based upon their ability estimates. Green, Crone,
& Folk (1989) indicate that differential distractor functioning is found only when the
option choice by ability interaction explains additional variance beyond that attributed
to the main efi‘ects or other interactions.
Difierential Item Functioning: Findings

Research examining difl'erential responding generally employs one of the
methodologies described above. Examinees are initially grouped based on some
characteristic of interest to the researcher such as race or gender. Comparisons are
made between groups by controlling for ability, as determined by the total test score.
These studies, then, are hoping to uncover the relative strengths and weaknesses of

one group over another, much like ipsative tests. This information regarding the

Writing is due to

 

|
memes SAT and k
speeific influences 11
mm conflicting
ed“ '
hangs of the resea
re mined below.

Researchers h.”
Don-'15. Schmitt. &
1"" '
rte-0). Differentia‘
of :w ‘ ‘ -

o 21th mate]
serious ofatest '1
mt, _

generally re

muted cotmterpa:

(190*
‘ ~"s for “
CXBmplz‘

”Waits failed

14
influences on differential responding could then be utilized by future test makers.
Since DIF researchers often operate under the presumption that differential
responding is due to some sort of cultural influence, much of the current DIF research
examines SAT and GRE verbal items, which would presumably contain more culture-
specific influences than mathematical items. Although much of this research has
produced conflicting findings, there are some trends in this research. The major
findings of the research examining the factors which influence differential responding
are outlined below.

Researchers have consistently found evidence of differential speededness (e.g.,
Dorans, Schmitt, & Bleistein, 1992; Schmitt & Bleistein, 1987; Schmitt & Dorans,
1990). Differential speededness refers to divergent response rates between members
of two ability-matched groups to items appearing at the end of separately timed
sections of a test. The research indicates that when differential speededness occurs,
Whites generally reach the final items on the test more often than their ability-
matched counterparts from other ethnic groups. Dorans, Schmitt, and Bleistein
(1992), for example, compared the rate at which Asians, Blacks, Whites, and
Hispanics failed to reach items on two verbal sections of an SAT. Although no
difi‘erential speededness was found between Whites and Asians, both Blacks and
Hispanics at a given ability level tended to answer fewer items than Whites. Dorans
et a1. (1992) suggest that these findings may be due in part to differential strategy use

by diverse races. They speculate that Whites may more readily skip over difficult

 

hm: mm L

themsmm:

   
  
 

11135 312531.91 1.
Miami M H“;
new one Oi so
he???” and A,
m, High-JCS “'0“;
m and in 5.33
“gm 1m difTerenT
med: the P650”:
M a single spc'ling
Pit-panic M03333
Manon (such as a
ho; ifism‘c M01
hpohescs. Attbougl
he Watts, this fine
meiomed lihites 0
its effect was found 01
mi United US: of t

3%011 of hypothe

 

15
items than members of other ethnic groups, allowing White examinees to reach more
of the items in a timed section of the exam.

In an attempt to uncover sources of differential responding, Schmitt (1988)
hypothesized that Hispanics and Whites would respond differentially to verbal items
containing one of several characteristics: (1) cognates (2) false cognates (3)
homographs, and (4) content of special interest. Specifically, Schmitt (1988) argued
that Hispanics would outperform a group of Whites matched on ability when the
correct response for the SAT verbal item contained a word with a common root in
English and in Spanish. However, the presentation of an item containing a word
which has different meanings in the two languages (a false cognate) was expected to
impede the performance of Hispanics. English words which have multiple meanings
but a single spelling (homographs -- ex. bark) were similarly expected to impede
Hispanic performance. Finally, verbal passages which were relevant to the Hispanic
population (such as a passage about Mexican-American women) were expected to
favor Hispanic performance. Limited support was found for several of these
hypotheses. Although Hispanics did tend to outperform Whites in items containing
true cognates, this finding was not consistent across all items. Hispanics
outperformed Whites on reading passages of special interest to Hispanics. However,
this effect was found only when the passages were specifically targeted toward this
group. Limited use of false cognates and homographs in the test disallowed

examination of hypotheses involving those variables. Schmitt (1988) argues that

 

fim research soul

n in thick 916 h)?

 

I
Schmitt and D
.er‘n' items These :

l' U
:I

shed ETS Ste.

m

Asses to Mites. lie

 

possibic interest no ';
itneen content of in
nnogaphs on DIP s
sens. key and dim;
Knits cleariy indie
1"! facial minority gr
tflhtes. Finally. 8

r“53°39‘11” on item

16
future research would benefit from the construction of an experimental form of the
test in which the hypothesized factors could be controlled.

Schmitt and Dorans (1990) continued the investigation of the DIF of SAT
verbal items. These researchers synthesized the major findings of several
unpublished ETS studies that compared the responses of Hispanics, Blacks and
Asians to Whites. Items were coded by the authors as being of special interest, of
possible interest, no particular interest. Results indicated a positive relationship
between content of interest and DIF for Hispanics and Blacks. The effect of
homographs on DIF was investigated by coding the presence of homographs in all
stems, key and distractors, and then correlating this with the standardized DIF values.

Results clearly indicate that items which contain homographs are harder for each of
the racial minority groups examined than these items were for the sample composed
of Whites. Finally, Schmitt and Dorans examined the possible influence of vertical
relationships on item response patterns. Vertical relationships were defined as those
items in which a word or words in the stem could be associated with a word in the
key or any distractor which is independent of the analogical relationship between the

two words in the stem. As an example, Schmitt and Dorans (1990) provide the

following:
PHYSICIAN: PATIENT :: (A) nurse: hospital
(B) guard: warden (C) informer: agent
(D) attorney: lawyer (E) accountant: client

 

niafiors‘rip bet“?

 
  
   
 

regions'nip. since L”

Schmitt and D
ions in a stem an:
Hispanics Blacks
vertical relationship
only HTS-panics were
Comet option. all :1
min] relations'ni;

Schmitt and I
539.3185 Pertaining
Psalm: between
ofksiam with POC
Tom for 1103-113“.
COX-fining a large

k

engiish 35 their be

17

The correct answer is option (E) since both option (E) and the stem represent a
relationship between subject and recipient. However, option (A) represents a vertical
relationship, since both the stem and option (A) are associated with medical terms.

Schmitt and Dorans (1990) hypothesized that vertical relationships between
words in a stem and words in a non-key distractor would cause negative DIF for
Hispanics, Blacks, and Asians. However, positive DIF was expected when the
vertical relationship was found between the stem and the keyed option. Although
only Hispanics were found to benefit fiom vertical relationships between the stem and
correct option, all three racial minority groups performed more poorly on items with
vertical relationships between stem and incorrect options than did White test takers.

Schmitt and Dorans (1990) further examined the ETS research for specific
findings pertaining to ethnic groups. Differential item functioning seemed most
prevalent between Asian Americans and Whites only when the sample was composed
of Asians with poor English language skills. Whereas negative DIF was commonly
found for non-native English speaking Asians on SAT verbal items and math items
containing a large degree of verbal material, very little DIF was found for Asians with
English as their best language. In relation to Hispanic populations, Schmitt and
Dorans (1990) uncovered no new evidence of differential responses beyond those of
Schmitt (1988) discussed above. Research examining the DIF of Blacks on the SAT
has found that the greatest amount of differential responding occurs on analogy items

(Rogers & Kulick, 1987; Scheuneman, 1978). The Schmitt and Dorans (1990) review

 

ofETS researCh inc;
With SAT verbal ite:
;- - I

The effect or it.

on he SAT and Gra;

Freedle and Kostin I. Z

 

Blacks performed be:
non-ms) than their
min items uhich w
Mining item diffze
m ihC GRE. Freedle
{'5’th hFTDOthesis th.‘
his Freedle. Kost

mm Procedm “

18
of ETS research indicates that Black examinees often experienced greater difficulty
with SAT verbal items containing vertical relationships and homographs.

The effect of item difficulty on the differential item functioning of verbal items
on the SAT and Graduate Record Examination (GRE) for Blacks was investigated by
Freedle and Kostin (1990). These researchers surprisingly found that on both tests,
Blacks performed better on harder verbal items with little context (analogies and
antonyms) than their ability-matched White counterparts, and performed worse on
easier items which were relatively context fi'ee. Very little DIF was found when
examining item difficulty for sentence completion and reading comprehension items
on the GRE. F reedle and Kostin (1990) used one unpublished ETS study as support
for the hypothesis that Blacks and Whites use different strategies to answer difficult
items. Freedle, Kostin, and Schwartz (1987) instructed students to use a verbal
protocol procedure when presented with SAT-type analogies. Blacks were found
more often than Whites to employ the use of a strategy referred to as indirect
induction - that is, Blacks were more likely to employ a reasoning method which
suggested a partial grasp of the word meanings in the analogy stem, which can then
subsequently be used to limit the available options fi'orn which the ultimate choice is
made. Further analyses by Freedle, Kostin, and Schwartz (1987) demonstrated that
use of an indirect induction strategy could increase the chances of accurately
responding to a difficult item by as much as 80 percent. Alternative explanations
provided by Freedle and Kostin (1990) concerning their finding that Blacks were

more likely to respond correctly to more difficult GRE verbal items than their ability-

mined Whit: COL.
int effects

hither supp;
5123122165 is proud;
prior high students
tin runs 1: oi'tovtt'
item fraction suhtr.
its» subtract 2 tr:-
‘hm the test ta};
“iii? the same (it:
that two item I}:

maid favor ta:

19
matched White counterparts include violations of unidirnensionality assumptions and
floor efi‘ects.

Further support for the possibility that DIF may result from disparate solution
strategies is provided by Tatsuoka, Linn, Tatsuoka, and Yamamoto (1988). Over 500
junior high students were instructed to solve fraction subtraction problems in one of
two ways. Following instruction, students in both groups were administered a 40-
item fraction subtraction exam. This test was composed of six separate types of items
(e.g., subtract 2 fractions with the same denominator; mixed number subtraction
where the test taker needs to find a common denominator; mixed number subtraction
with the same denominator; etc.). Of these six types of items, it was hypothesized
that two item types would favor neither instruction method, and two types of items
would favor each of the two instruction methods. Initial results demonstrated no DIF
between the two instructional methods. Speculating that the students in the two
groups may actually be utilizing a method they were not specifically taught, further
analyses examined DIF resulting from actual solution methodology employed by the
student. These analyses found that strategy used differentially affected which items
the student responded to correctly. The authors argue that further investigation of
DIF should examine some cognitive process variables, rather than examine DIF based
upon simple demographic variables which would be expected to have largely
unknown relationships to more relevant instructional and cognitive variables.

Scheuneman and Gerritz (1990) employed the use of Mantel-Haenszel

techniques to examine DIF in reading comprehension items. However, these

members not on?

 

m 1,0 between IT.

(stream; Page"?

I
These potentially I'
cirfieristics of ti:
indrc’ion deduction
i3; metre of the c

phases; short lists 0

the ‘90 (e.gu predie;

 

Mantel-Haenszel in:
wmblllcd TO the pm
more than 2596 of IT“

the race Compansom

I1

20

researchers not only examined differential responding between Blacks and Whites,
but also between men and women. These researchers carefully identified components
of reading passages from nearly 300 items from six forms each of the SAT and GRE.

These potentially DIF-inducing components included: (1) passage content, (2)
characteristics of the question posed in the stem (e.g., transformation of information;
induction/deduction; passage-based inferences; rhetorical inferences; or, reasoning),
(3) structure of the options (e.g., complete sentences and stand-alone phrases; simple
phrases; short lists of 1-4 words; or, numbers), and finally, (4) semantic context of
the text (e.g., predicates; arguments; or, modifiers). Following the calculation of
Mantel-Haenszel indices, regression was utilized to examine which variables
contributed to the prediction of DIF. Content of the passages was found to explain
more than 25% of the variance in the Mantel-Haenszel values for both the gender and
the race comparisons. For race, analyses revealed inconsistent results across the two
tests. On the SAT, narrative or scientific passage content favored Whites. When the
stem demanded the use of rhetorical inferences or reasoning ability, Whites again
performed better than a matched group of Black examinees. Finally, White
examinees were slightly more likely to respond correctly to SAT reading
comprehension items when the options were in the form of a sentence. Blacks
performed better on items with word Options and passages of greater length. The
analysis of the GRE items revealed only limited overlap with the SAT findings.
White examinees once again performed better on passages with scientific content.

Other GRE item characteristics which favored White performance required the use of

 

micnnsi silent...

tort or number. 8 .

   
  
 
  

hurhldes content t
easily more con:

innit: mil-CS. \M'

mfiormation of in
ms “uh 5cm
km and (

“an“ ofan ite

mm Tues

21
transformed information or rhetorical inferences, and options composed of a single
word or number. Blacks performed better on longer passages and those with
humanities content than did Whites. The analyses examining sex differences were
generally more consistent across the two test types. Females performed better on
longer passages, with narrative or humanities content, in which the stem required
transformation of information or rhetorical inferences. Males performed better on
passages with science topics, with options in sentence form, and items in which the
stem required some passage-based inferences or reasoning. Based on these findings,
Scheuneman and Gerritz (1990) argue that in order to understand the observed
difficulty of an item, it is necessary to consider the various personal attributes of the
examinees. These researchers point out that in examining DIF between any two
groups, the researcher is contrasting a multitude of background characteristics that are
likely to have resulted in the development of somewhat different patterns of abilities
and personal attributes. Pointing to the gender differences in passage content, for
example, Scheuneman and Gerritz (1990) argue that these findings are consistent with
the fact that women have traditionally enrolled in more literature and Humanities
courses, whereas men have taken more science courses. Thus, there exists a strong
possibility that cultural difl‘erences in prior learning, experience, and interest patterns
may exert sufficient force to cause DIF. A possible alternative explanation is that
students choose those courses in which they have innately superior abilities. Higher
success rates would then be attributable to innate abilities. Scheuneman and Gerritz

(1990) argue that despite the possible multiple influences on DIF, variations in the

'. ”Wm: cxpe. .1

$533021
val"

Tue R5333:
2i 3 533075 Willi
rrlsde specific 2‘:
Tris of 306655?
”web has sag;
dramatic to dif
However. both it
an ambiguous.
iferential respc
Mthtts (e.g.,
Tameka. 5: Ya:
“:33 reSearcher:
15C 0f den-10m;

mum? n

22

background experiences of individuals, and subsequent DIF, warrants further
investigation.
Summary

The research examining differential responding has provided some indication as
to the factors which influence disparity in item response patterns. These factors
include specific characteristics of the item stem, of the options used, as well as of the
type of processing or inferences the examinee is expected to perform. Further,
research has suggested that some of these differences in propensity to respond may be
attributable to difl‘erences in solution strategies used by the groups of interest.
However, both within and across studies the findings of DIF are often inconsistent
and ambiguous. The lack of clear-cut and consistent findings in the causes of
differential responding to cognitive ability items has been lamented by many
researchers (e.g., Cole, 1981; Scheuneman & Gerritz, 1990; Tatsuoka, Linn,
Tatsuoka, & Yamamoto, 1988). Tatsuoka et a1. (1988) argue convincingly that all too
often researchers examining difiemnfial responding have relied too heavily upon the
use of demographics (e.g., self-reported race or sex) as the determinant of group
membership. This emphasis on research examining racial and gender characteristics,
however, is no doubt a reflection of societal concerns for fairness in testing.

Although there is an underlying belief that background experiences may
influence members of different races to respond differentially, it is the cognitive
strategies actually employed by the examinees which are of greater theoretical interest

to psychologists. Scheuneman and Gerritz (1990) provide evidence that the use of

races. firher in. e
5:23} different .
tie: the solmio
arms the re}:
mi St‘bsequent n
Snbl'et'tive [hip

The Va
mes: Tales of
35a ingjcaj Star
“215‘ Scorcg

Shim df’mr

2magisiLc‘fi‘dft

~ A
Ett-
‘u .

23

different solution strategies may lead to differential item response. Despite evidence
provided by Freedle and Kostin (1990) and Dorans, Schmitt, & Bleistein (1992)
suggesting that different solution strategies may be employed by members of different
races, further investigation is necessary to examine why members of different races
employ different solution strategies. Specifically, little is known about the way in
which the solution strategies employed by examinees are influenced by their cultural
backgrormd. Perhaps then, the greatest benefit would be provided by research which
examines the relationships between demographics, cognitive-psychological variables,
and subsequent response patterns on assessment tests.
Subjective Employment Tests

The vast majority of research examining DIF has scrutinized the differential
response rates of selected groups on cognitive performance. Cognitive ability tests
are a logical starting point for the examination of item bias. Each year, thousands of
students’ scores on such standardized cognitive ability tests as the SAT and GRE are
used to determine their fate for future educational opportunities. Cognitive ability
testing is used for much more than just aiding the college admission process,
however. Cognitive ability tests are one of the most popular employment selection
tests in use today. The popularity of cognitive ability tests is attested to by the fact
that there exist a multitude of meta-analyses examining the validity of these tests
across a variety of employment contexts (e.g. Schmidt, Pearlman & Hunter, 1980;
Schmidt, Htmter, & Caplan, 1981; Schmitt, Gooding, Noe, & Kirsch, 1984). These

studies have demonstrated that the popularity of cognitive ability is well-deserved,

rd: iii-'1” on Vii;
ground .45 wit}
2in £05155“ch a:
xenon of supen'isc
sofew would dou't‘t
unpason to other
teroziio' tests cog
Eaves across job:

Despite the we
911303: Option avai
”9.3533211 aptitude. 1
altiOb Mom-lane
”shined Thus

L’Wni . -

24
with population validity coefficients across all jobs of .54 with a training success
criterion,and .45 with a job proficiency criterion (Hunter & Hunter, 1984). Using
more conservative analytic assumptions for the correction of restriction in range,
Hartigan and Wigdor (1989) report validity coefficients in the .2 to .4 range with a
criterion of supervisory ratings for the General Aptitude Test Battery (GATB). Even
so, few would doubt the impressive validities of cognitive ability tests. In
comparison to other commonly used predictors such as interviews, job samples, and
personality tests, cognitive ability tests have often been shown to have equal or better
validities across jobs (Hunter & Hunter, 1984).

Despite the well-deserved popularity of cognitive ability testing, these tests are
only one option available to human resource managers in their examination of
applicant aptitude. Assuming, for example, that a cognitive ability test correlates .4
with job performance, over 84% of the variance in job performance remains
unexplained. Thus, additional selection tests are used to help account for the
remaining variance in job performance. Interviews, work samples, assessment
centers, biodata, situational judgment inventories, and personality tests are all
possible methods of validly assessing an applicant's potential to the organization.
Whereas the former three test types in the above list require assessment of individuals
or small groups, the latter three are commonly used paper and pencil tests which can
be administered to large groups with relative ease. Although each of these tests
require a subjective response on the part of the candidate, can be administered in a

multiple choice format, and are assigned "correct" option keys, there are subtle

irate: among Sues
semis personal h
ties are ksigned to
lift: the indixidual .
argument tAllpon
a: he basis of critic:
5- 3'3 156d to asset
53232011 or dam“
Write to Fleis‘r.
5531‘ {116131331}-
mm hm fGuns;
198:" lmpfessive
mm? 311d sin

333m & Carter

25
differences among these tests. Biodata scales are composed of those items which
concern the personal backgrounds and life experiences of applicants. Personality
scales are designed to assess the distinct organization of characteristics which help
define the individual and determine the person's pattern of interaction with the
environment (Allport, 1961). Finally, situational judgment items are often developed
on the basis of critical incidents and the behavioral dimensions required on the job,
and are used to assess how an applicant would act when presented with a specific
situation or dilemma. Impressive validities have been offered for each of these tests.
According to Fleishman (1988), biodata scales often have validities in the low .403.
Several meta-analyses examining the validity of biodata against supervisory rating
criteria have found validities above .35 (Hunter & Hunter, 1984; Reilly & Chao,
1982). Impressive validity estimates have also been found in some studies for
personality and situational judgment tests (e.g., Barrick & Mount, 1991; Motowidlo,
Dunnette & Carter, 1990).

Some scientists have challenged the validities of these tests based on their
potential for fakability. In particular, research has examined the degree to which
applicants are likely to fake responses on biodata inventories. Mumford and Owens
(1987) point out that the research has demonstrated that although self-report bias may
produce response shifts, this bias is not sufficient to change the ordering of
individuals. Further, these authors argue that social desirability and acquiescence
seem to have only minimal impact upon biodata measures. The susceptibility of these

scales to outright faking has also been thoroughly examined. There seems to be some

  
   
  
 
 

31,5 hat although :
is $1335 vfinen a?

Depite ’he pottul;
1" 7;: been no em:
remit examining note
raising the differen'
tgte ability tests. E
it trainee to subieeti
renal ‘Wkgmtmd. o
=133-33mm can he said -
tub or rational?
“its 3" dfiltfininec
at some criterion of it

me. ~‘
'1)- Inthc fatior

“Oral dlmcnsion

26

consensus that although it is possible to fake responses to these tests, applicants rarely
do so, especially when instructed that such deception can be detected (Cascio, 1975).

Despite the popularity of personality, biodata, and situational judgment tests,
there has been no examination of the potential for item bias in these tests. The lack of
research examining potential DIF on these subjective tests is unfortunate, especially
considering the differences between these tests and the more heavily investigated
cognitive ability tests. Biodata, personality and situational judgment scales require
the examinee to subjectively select an option for each item based on their preferences,
personal background, or inclinations. That is, there is no objective standard by which
one option can be said to be "better" than another. Rather, the correct option is either
empirically or rationally keyed. In the case of empirical keying of an item, the correct
options are determined based upon their ability to discriminate the examinees based
on some criterion of interest (such as supervisory ratings, promotion rate, turnover
rate, etc). In the rational scoring approach, items are developed and keyed to assess
behavioral dimensions believed to be essential to adequate job performance. In either
case, any particular item may have a single correct option, multiple correct options, or
be weighted such that some items are viewed as "more correct" than others. These
characteristics of personality, biodata, and situational judgment items are clearly
different fiom cognitive ability tests, in which each item is designed to have only one
correct option. Further, it is clear from the definitions of these types of tests that
personal characteristics of the individual examinee are more likely to have an impact

on item responses to subjective tests than on cognitive ability tests. To the extent that

Eggs of disparai
{3933555 L0 [.3656 If

All mmmIlOl
mic-meat tests a: .1
regz'ii'g item bias is
rich iniitiduals dif:
fastening requires Ll
refs-able on which ir
Cf'ar and sex as the 1
:31 332131 identity mi;

Won of differ:

Rich] Identity Th e0 r
RWher-s “15L

 

27
members of disparate races or genders are subjected to differing experiences,
responses to these test items are likely to reflect these differences.

An examination of the potential for differential item functioning in subjective
employment tests appears long overdue. However, the testing of a priori hypotheses
regarding item bias is highly dependent upon accurate assessment of the ways in
which individuals differ from one another. That is, the search for differential item
functioning requires the selection of theoretically meaningful and psychometrically
defensible on which individuals may be differentiated. In contrast to the routine use
of race and sex as the measures of individual variation, the following section pr0poses
that racial identity might serve as a more useful classification variable for the
examination of difl‘erential item functioning.

Racial Identity Theory

Researchers wishing to examine differences between certain groups of people
must at some point choose a criterion on which to categorize those individuals.
Thoughtful consideration of the proper criterion has strongly been urged whenever
research is conducted (Astin, 1964; Nagle, 1953; Stuit & Wilson, 1946). In the field
of testing, researchers have relied almost exclusively upon self-reported and/or
observable demographic features as the criterion for dividing individuals into groups
of interest. Thus, individuals are classified based upon their skin color, sex, socio-
economic status, or educational attainment. The choices of race and sex, in particular,
are undoubtedly influenced by legal and social concerns in testing. Ever since the

passage of the 1964 Civil Rights Act which prohibited hiring decisions based on race,

rim test 5::
amaze of bit
Lettuce subseq-

Helms (19%
TICK: for other 1:
rfitgmeig' \n'
agitation ar
353mm: demc

ngfgt' .- .
"£34113 111 Sign

e1 fall‘s not
w for Y.
I;

h “ere Open

28
sex, religion, or national origin, these easily reported demographic variables have
become the primary focus of researchers examining test and item bias. However, the
blind use of such "observable" criteria may be unwarranted. As Helms (1992)
accurately points out, researchers examining differential test performance between
Black and Whites rarely consider the particular biological characteristics presumed to
influence test scores. Rather, race is used as a proxy for some undetermined
composite of biological, cultural and educational differences, which are presumed to
influence subsequent test scores.

Helms (1992) argues that the use of such observable criteria as skin color as a
proxy for other presumed criteria disregards the fact that there exists considerable
heterogeneity within racial characteristics. Indeed, she persuasively argues that many
complications arise in determining subgroup assignment based on self-report or
observable demographic categories due to the prevalence of interracial heritage and
variability in skin colors and cultural identifications. Yee (1983) utilizes personal and
historical examples to demonstrate how demographic race classifications can be
misused. As an Asian-American Yee checked "colored" when registering to vote in
his new Southern township in 1964. The public registrant examined the voter
registration, quickly erased Yee's response, and substituted white. Heatedly, the man
said, "Y‘all's not colored, don't you know that? Y'all's white!". This reaction was
unexpected for Yee and his wife, who had recently moved from California, where
Asians were openly discriminated against. Further confusion over the status of Asian

Americans can be demonstrated by an 1854 Supreme Court of California case which

l
355?“ uhf'hfl the te‘

 

arrie: eonsiction fc
r;....}\ founded by a
here there mistakenly 1
Ta '53» W965 Of pee
m 35 113533115 For

‘.

 

attuned Even Lodz
the classified as 812.;
trays a 'scientific' c
may in Sonth Afri
Billiidals as Whites

Tim. “a ”as all ot‘

LC t
“‘5.
Stif.

29
decided whether the testimony of an Asian American was sufficiently valid to uphold
a murder conviction for a white defendant. The Court reasoned that this country was
originally founded by Columbus on his voyage to the West Indies, and the natives
here were mistakenly termed Indians. Rationalizing that the world was composed of
only three types of peoples - Whites, Blacks, and Indians - the Court classified
Asians as Indians. Further, since Indians had no legal rights, the testimony of the
Asian (who was now classified as Indian) was thrown out, and the conviction
overturned. Even today in America, individuals need only have 1/16 Afiican ancestry
to be classified as Black (McRoy & Zurcher, 1983). Further evidence that race is not
always a "scientific" classification is found in the classifications used until very
recently in South Africa. Here, many inconsistencies abound as to classification of
individuals as Whites, Blacks and Coloreds. Japanese, for example, were considered
White, whereas all other Asians were classified as Colored. Finally, the memorable
scene in the film Emafimpa in which the (unbeknownst) Jewish protagonist is
used to demonstrate the fine features exemplary of Nordic youth, bears witness to the
fact that physical features are unreliable sources of scientific inquiry. These
argtunents suggest that, despite established research methodology, over-reliance upon
self-report or observable demographic characteristics may be too simplistic and
inaccurate as a criterion for classifying individuals for DIF analyses. These
arguments echo the urgings of Tatsuoka et a1. (1988), who recommended the
classification of individuals based upon some theoretically reasonable criterion (such

as cognitive strategies) rather than simple demographics. Further, a similar argument

 

was made '0} Novick a
I

gar; memiership are
should be used as the 'r
The Challenge. tit
as would allow in J
It erbrace either bio}..-
553511321115 of race V\
ET 3321;]; [ht amoL
with SchCdon Leg Peri

Which have .
“muted 1

 

mm\c])mi d
the t .
TTwin:dafi‘t’a'ences

t; -
‘~’~‘~ net

DIF- Altgmm
”Tahoio '.
Elsal characre

30
was made by Novick and Ellis (1977), who argued that classifications based on ethnic
group membership are merely surrogates for some educational disadvantages which
should be used as the basis of group formation.

The challenge, then, would be to develop a theoretically meaningful criterion
which would allow individuals to be classified into psychometrically defensible
subgroups. In abandoning demographic variables, researchers would be encouraged
to embrace either biological or psychological variables. Possible biological
determinants of race would require the use of measurable biological characteristics.
For example, the amount of melanin in the skin might be measured and correlated
with selection test performance, although Helms (1988) indicates that the few studies
which have attempted to assess the relationship between cognitive performance and
objectively measured race-related biological characteristics have failed to find any
significant difl‘erences. It is perhaps important to note that such studies examined test
bias, not DIF. Alternatively, the relationship between item responding and
psychological characteristics of individuals could be investigated. Psychological
characteristics seem a likely choice as a criterion for DIF considering the prevailing
underlying assumption of some as yet unspecified cultural influence on differential
response patterns. These psychological variables may possibly mediate the
relationship between demographic variables and test performance. Psychological
variables have the further advantage of being readily measurable, given reliable and

valid scales, and willing subjects.

mg DH: and 165
we. as fairness ”“05
are. are to 511'??let ‘
tea: the societal CODC
rambles) chosen as LT
atria; with the demo
.10.th in allouing :

To examine the r
femur): Identity refer.
mil‘lion that he or 5?
rule: serve as a measu
meal mammal

"We.” -
malt: chmfication b”
C.

Could the Psycho

 

 

31

The major problem with the use of psychological variables as the basis for sub-
group determination for DIF analyses is that these sub-groups may fail to be useful
from a societal standpoint. That is, much of the research energy invested in
examining DIF and test bias has been conducted solely to satisfy societal concerns
with test fairness across specific racial or gender groups. Any psychological variables
which are to supplant the use of demographics, therefore, must at least to some degree
retain the societal concern with race and sex issues. Ideally, the psychological
variable(s) chosen as the criterion for forming sub-groups would provide considerable
overlap with the demographic classifications, but provide greater accuracy and
flexibility in allowing for individual differences.

To examine the psychological impact of race, researchers could examine racial
identity. Identity refers to a sense of group or collective belonging based on one's
perception that he or she shares commonalities with a particular group. Identity, then,
could serve as a measurable psychological attribute which would provide greater
theoretical understanding of the underlying cognitive processes of an individual than
would classification based on simple demographics.

Could the psychological variable of racial identity provide researchers with
additional information beyond categorization based on mere demographic
charactertistics? If so, some individuals of a particular demographic classification
(e.g., Asian, Black, Hispanic, or White) would be expected to adopt a racial identity
inconsistent with their demographic race. A White individual, for example, might

adopt an African-American racial identity. Is this possible? Some racial minorities

 

«3,4,5;- group. In a re:
for Blacks and was
:‘eg- Gait 8; C larle
tr} yang Blacks are

be possible for White

American racial idem.

 

For. Wayne. lndiana ‘
a to a complete uph
called “Wiggers' (a c

Med 3 series of h;

32

living in America may be expected to adopt the racial identity of the dominant White
racial group. In a recent study assessing the preference of Black preschool children
for Blacks and Whites, Spencer (1984) found evidence consistent with earlier studies
(e.g., Clark & Clark, 1939; Williams & Roberson, 1967) which demonstrated that
very young Blacks actually preferred Whites to members of their own race. Would it
be possible for Whites to similarly identify racially with minority groups? Recently,
there has been some press detailing White individuals who have assumed African
American racial identities. A recent Esquire article profiled several White youth from
Fort Wayne, Indiana whose apparent identification with African American culture had
led to a complete upheaval in this small midwest town (Carroll, 1994). These so-
called "Wiggers" (a contraction of the racially motivated slur "white nigger") had
received a series of harassments fiom townspeople including death and bomb threats,
prompting investigations by the FBI and denunciations by the NAACP. Obviously,
then, both majority and minority group members can potentially assume racial
identities which are not consistent with societal expectations based on demographic
characteristics.

The current study proposes that subgroup classification for the investigation of
DIF should be based on the psychologically-oriented racial identity of an individual,
rather than the individual's biologically determined racial characteristics. Following
an investigation of the meaningfulness of the construct of racial identity, it will be
proposed that racial identity influences one's exposure to, and subsequent

internalimtion of, cultural values.

3:3; Identity: Tom:

The theoretical
mail}- championed
liiflielms. 1984; P;
unity theory (RlTl 1
hernhe'slup. Specific
E3 waned different:
tria' identity is 8 p5};
mmidp- An atter
officntity: PUSOnal,
Emu: mnflity be
P311111 to faCial belief
mid“! adopts a Pa

aeration According

 

33
Racial Identity: Toward a Meaningfid Construct
The theoretical development of the construct of racial identity has been
primarily championed by Janet Helms and her colleagues (e.g., Carter & Helms,
1987; Helms, 1984; Parharn & Helms,1981). According to Helms (1990), racial
identity theory (RIT) is concerned with the psychological implications of racial-group
membership. Specifically, RIT examines the "belief systems that evolve in reaction
to perceived differential racial-group membership" (p. 4). By definition, therefore,
racial identity is a psychological variable originating from one's racial-group
membership. An attempt is made by racial identity theorists to consider three forms
of identity: personal, reference group, and ascribed. Personal identity refers to
generic personality beliefs about oneself. These characteristics need not necessarily
pertain to racial beliefs. Reference group identity refers to the extent to which the
individual adopts a particular racial group to guide one's thoughts, feelings, and
behaviors. According to Helms (1990), one's reference-group identity is reflected by
examining one's organizational memberships, ideologies, and entertainment choices.
Ascribed identity refers to the individual's stated racial affiliation and commitment.
These components of the racial identity construct are reminiscent of Gross, Mason, &
McEachem's (1958) view of achieved, ascribed, and assumed roles.
Unforttmately, the racial identity theories which are present in the literature
today are not particularly useful for classifying individuals into separate racial
identity groups, Rather, these scales were primarily developed to assist clinical

psychologists better understand the racial issues presented by their clients. Separate

30:;Eetei} dis:
5:: Blacks and '
Prim racial- i
Zii'lff Black D:
In order
mm researcl'
are ascribed ide
‘T‘T on dflelo;
all dCVClOpm
Ere—a] Classific
T0 test the

37's .
T can e1Oped \

34
measurement scales are administered to individuals - presumably based upon their
demographically-determined race - in order to determine the stage of racial
development they have achieved. Nigrescence theory attempts to explain the racial
identity development of Blacks, while Whites are thought to progress through a
completely distinct series of racial development stages. These current stage theories
for Blacks and Whites have numerous conceptual and methodological shortcomings.
Further, racial identity theories are conspicuously absent for individuals who are
neither Black nor White.

In order to utilize racial identity as a criterion for subgroup formation, the
present research requires the assessment of racial identity in terms of reference group
and ascribed identity. Although Helms and her colleagues have worked extremely
hard on developing scales useful for assessing stages of White- and Afi-ican-American
racial development for clinical purposes, I believe these scales are less useful for
general classification purposes of an individual of any race into general categories.

To test the hypotheses proposed in this research, a new scale of racial identity
was developed which (1) allows the classification of individuals into one of several
possible racial identification categories (i.e., Afi'ican-American, Asian-American,
Hispanic-American, White-American, etc.), and (2) directly assesses both reference
group and ascribed components of racial identity to provide some mm of the
strength of the individual's racial identification. This two component scale is
provided in Appendix A. This scale was created by the present author in order to

classify individuals in terms of psychological racial identity, instead of traditional

1.3.3.1021 i
be sensitive to
nitdaal’s pa
r131 identir}
approach “35
media: the re
”muses It i
“72:1:th no“
Cultural rum
\lUCh of

ateomficaL I

35

classification according to observable physical features of the individual. As opposed
to traditional demographic classification strategies, the proposed approach to racial
classification is likely to be influenced not only by observable characteristics, but also
be sensitive to individual differences in cultural and education backgrounds. Since an
individual's past experiences are thought to influence not only the development of
racial identity but also one's adoption of a specific culture, this new classification
approach was an attempt to provide a psychological variable which was thought to
mediate the relationship between race and suspected cultural influences on item
responses. It is the investigation of possible cultural influences on item responding to
which we now turn.
Cultural Influences

Much of the work examining differential response patterns has been
atheoretical. Indeed, thisresearch usually assumes one of two forms. The first
approach is a purely statistical approach which typically compares the use of two or
more techniques for the identification of DIF. The second approach is more
concerned with identifying factors influencing DIF . However, even the second group
of researchers have not been overly concerned with theory. That is, these researchers
typically obtain a database of SAT or GRE items and statistically examine the
differences in item responding between selected groups. Once these items have been
identified, items are scrutinized in the hope of determining similarities across flagged
items. The assumption underlying most of the work on differential item responding is

that various groups respond difi‘erentially due to disparate cultural experiences and

Schemcman a
ramr'ible fo:
'resrirs cones:
lazing cape

may be linked

36
influences. Freedle and Kostin (1990), for example, attribute their findings of
differential response patterns between Blacks and Whites on cognitive ability verbal
items to cultural differences in "nuances of meaning to fi'equently used words, which
presumably reflects varying everyday experiences among cultures" (p. 341).
Scheuneman and Gerritz (1990) provide further speculation that life experiences are
responsible for DIF across racial groups. Scheuneman and Gerritz (1990) state,
"results concerning the content of reading passages suggest those differences in prior
learning, experience, and interest patterns between . . . Black and White examinees
may be linked with DIF and deserve further study." Why has such subsequent
investigation not been conducted? Perhaps the primary reason is the method typically
employed in this research - analysis of archival data. As Schmitt (1988) indicates,
there is a need to develop experimental forms of exams which incorporate
hypothesized factors which might influence DIF. Further, investigators must specify
what they mean by cultural experiences. All too often culture is used as an
amorphous term which seems equally applicable to life experiences, racial heritage,
and ethnic background. In contrast, what is needed is specific theoretical
specification of which cultural influences might influence DIF, and then experimental
investigation of whether these cultural difl‘erences are in fact associated with
differential item responding. In order to accomplish this goal, researchers must begin
with a clear specification of what they mean by cultural influences. The present study
hypothesizes that difi'erences in cultural values across individuals of difl'erent racial

identity will influence item responding. It is argued that individuals of difi‘erent racial

1331111135 W111 bf C
ma palm“:
,Tarara'a drfinu‘

The word '
c’miciefi by K
WES 02121211) t
was elayed 21'.
ion 15: of Ihé
111 m of the
modded by th
Midi Greel
a115111511“! dei

Qfa Specific p

37

identities will be differentially exposed to and therefore motivated to adopt, different
cultural preferences. These differences are further expected to influence the responses
an individual provides to subjective employment test items.
Toward a definition of culture

The word "culture" has itself experienced a lengthy development which is
chronicled by Kroeber and C. Kluckhohn (1952). The present day meaning of culture
was originally established in 1871, though its first inclusion in an English dictionary
was delayed at least 50 more years. Use of the word "culture" seems to have evolved
fiom use of the word "civilization". Even today, most dictionaries tend to define each
in terms of the other. For example, one alternative entry for the word "culture"
provided by the 1992 edition of the Random House Webster's Collegiate Dictionary
defines culture as "a particular form or stage of civilization, as that of a nation or
period: Greek culture". This same dictionary offers the following among the
alternative definitions for the word "civilization": "any type of culture, society, etc.,
of a specific place, time, or group: Greek civilization". Kroeber and C. Kluckhohn
(1952) point out that this complete overlap between the two words was not as
common in other languages such as German and French. Even here, however, there
was considerable confusion as to what was meant by culture. Perhaps only when
considered from a scientific viewpoint, therefore, does culture become more clearly
defined. Based on an assemblage of 164 scientific uses of the term "culture", Kroeber
and C. Kluckhohn (1952) developed 7 general categories of definitions: descriptive,

historical, normative, psychological, structural, genetic, and miscellaneous. Each

33.;gorx comers
. . 1 - f" :1
322130115 0 ~L-

savvy!" ng’qhfi
“ha-bol- A_‘ksi\

. 3 I - ‘
Remy: (13:1:

38
category contains lists of definitions which employ a common theme. Descriptive
definitions of culture provide broad definitions with emphasis on the specification of
content. Historical definitions emphasize a common social heritage or tradition.
Normative definitions emphasize rules, ideals, values, and behavior. Psychological
definitions emphasize adjustment, learning, and habit. Structural definitions
emphasize the patterning or organization of society. Finally, genetic definitions stress
culture as a product or artifact of society.
What conclusions can be drawn regarding the meaning of culture? Kroeber and

C. Kluckhohn (1952) provide a comparatively succinct yet specific definition.
According to these authors, culture is composed of "patterns of and for behavior
acquired and transmitted by symbols, constituting the distinctive achievements of
human groups, including their embodiments in artifacts; the essential core of culture
consists of traditional (i.e., historically derived and selected) ideas and especially their
attached values" (p. 35) (emphasis added).

Value orientations

If one is to examine the cultural differences between people, therefore, perhaps

it would be wise to begin with an investigation of disparities in values across cultures.
Detailed examinmion of differences in values across cultures has been conducted by
F. Kluckhohn and Strodtbeck (1961). Although proposed over thirty years ago, F.
Kluckhohn and Strodtbeck's (1961) typology of cultural value orientations remains
the premier classification of cultural values. This typology, therefore, would seem the

ideal framework for determining difi‘erences in values between cultures. Indeed,

car-1:13:15 hart mo:

Swizbcck'sllgél)

coasting reg.- BC

1935')

u

mils: problems. F -

 

rm roughly the 58
I: the solxing of 3 p1
minted in number
ofpossible solutiom
fa dinough all sol;
soiudom are preferr;

sermons according

 

marinas are def;

374 ~4
- W165. resulting 1

awashable elemi
i; directive elem?“I
13w.»

Juan and thou
{p.41 The three eler
I} A -

ha\’i0rs
‘3‘
w. One S
World\1m .

39
clinicians have recommended the increased usage of the F. Kluckhohn and
Strodtbeck's (1961) framework in order to provide therapeutic improvements to
counseling (e.g., Beutler & Bergan, 1991; Carter, 1990; Chapman, 1981;1nclan,
1985).

Kluckhohn and Strodtbeck (1961) argue that all societies encounter highly
similar problems. Further, for many of these problems human societies have often
found roughly the same answers. However, while there do exist alternative solutions
to the solving of a particular human problem, these alternative solutions are not
unlimited in number. That is, human problems must be solved within a limited range
of possible solutions. Most importantly, F. Kluckhohn and Strodtbeck (1961) argue
that although all solution alternatives are present at all times in all societies, certain
solutions are preferred for any given society over other solutions. These preferred
solutions, according to these theorists, can be termed value orientations. Value
orientations are defined as, "complex but definitely patterned (rank-ordered)
principles, resulting from the transactional interplay of three analytically
distinguishable elements of the evaluative process - the cognitive, the affective, and
the directive elements - which give order and direction to the ever-flowing stream of
human acts and thoughts as these relate to the solution of 'common human' problems"
(p. 4). The three elements are further clarified by the theory of value orientations.
The direction function is thought to provide the selection or choice between
alternative behaviors; the cognitive ftmction is conceived as providing a fiamework

for one's worldview; and the My; function is concerned with the amount of

metro-ml 6359' or ‘

 

armed to prefer on;
05313331101111 and
“"3231 soiutions. C'
mgr-123d possible 5
thermal ordering ..
pearl: solutions to .
What are the pr.
£3132 societies? F , I!
373316 “ll-115 basic ;
Via IS the relation 01

 

3“!“ .Q

4O
emotional energy or commitment that is available. Although a particular culture is
assumed to prefer one value orientation over others, it is an essential part of the theory
of F. Kluckhohn and Strodtbeck (1961) that each culture is influenced by all of the
optional solutions. Cultures vary in value orientations, therefore, not in the
recognized possible solutions, nor in the choosing of a single solution, but in the
preferential ordering of solutions. That is, cultures vary in the degree to which the
possible solutions to common human problems are preferentially emphasized.

What are the problems which are assumed to be universally encountered by all
human societies? F. Kluckhohn and Strodtbeck (1961) argue that humans commonly
grapple with 5 basic problems: (1) What is the character of innate human nature? (2)
What is the relation of the individual to nature (and supernature)? (3) What is the
temporal focus of human life? (4) What is the modality of human activity?, and (5)
What is the modality of a person's relationship to others? In order, these problems
may be summarized as concerns with orientations toward (1) human nature, (2)
person-nature, (3) time, (4) activity, and (5) relations.

The first basic value orientation refers to questioning of .the innate goodness of
human natm'e. Is humanity basically Evil? Is it basically Good? Or is it some
mixture of Good and Evil? This value orientation is further divided along lines of
flexibility. That is, each of the three possible categories of Evil, Good, or Mixed
(sometimes referred to as Neutral) are further subdivided based on the flexibility of
the human nature. The resulting six possible human nature orientations are: (a) Evil

and unalterable (b) Evil and perfectible (c) Good and unalterable ((1) Good and

333251: (c) an 113'.

O

 

marathon-5:1“. I;
353151: alternari rve-s
Session Thus. or
“an wandered

The second ha
and name. F. lilac},-

Faun 0f how the i

 

 

 

 

Scbjugazion-to-X arm
in essence. followers
peson's role is simpi
Same) “ho refuse
littered a Subjugazic
HihNamm, Person
37mm mm P

353:0. . .
Inez T111515

‘15 '
Ne opmslle c
as-
.5 mt natural f0
1‘3».

41
corruptible (e) an immutable mixture of Good and Evil, or (t) a mixture subject to
change. Although Kluckhohn and Strodtbeck (1961) originally proposed these six
possible alternatives to human nature, subsequent work has disregarded the flexibility
dimension. Thus, only the three general alternatives of Good, Evil, and Mixed are
typically considered.

The second basic value orientation represents the relationship between person
and nature. F. Kluckhohn and Strodtbeck (1962) see three possible solutions to the
problem of how the individual "fits into" the natural world. The first option is
Subjugation-to-Nature. Fatalism is the most dramatic consequence of this position.
In essence, followers of this option feel that, when it comes to Nature or God, the
person's role is simply to accept the inevitable. The elderly patient (or, Christian
Scientist) who refuses further medical treatment is an example of a person who has
adopted a Subjugation-to-Nature value orientation. A second position is Harmony-
with-Nature. Persons adopting this value orientation believe that there is no
separation between person, nature, or supernature. Rather, one is merely an extension
of the other. This is the orientation attributed to many American Indian tribes. The
final possible alternative for the Person-Nature orientation is Mastery-over-Nature.
This is the opposite end of the continuum of Subjugation-to—Nature. The basic belief
hereisthatnaturalforcesofallkindsaretobeovercome andputtotheuseof
humans. Industrialized societies adopt this perspective as they view nature as
something to overcome, or be used as a mere resource for achieving human needs and

desires.

 

Tn: value 0T-
mit 13 hOUI'S. min

305: invariant in Ie

   

mez'a’enst which 1
15'. time orientate.“
531151311 tradition

maid pnmanly cm;

 

:71: Past or consequc
has: cultures in “hi
Iii-11cm and Preset

The founh V311
1mm 111061: of 581

1““ mm Possibl-

42

The value orientation of time has less to do with concern how one divides time
units via hours, minutes, or days, and more to do with one's conception of what is
most important in terms of the Past, Present, or Future. That is, of these three time
orientations, which is most influential to the individual? Societies which possess a
Past time orientation tend to emphasize the maintenance, or restoration, of long-
established traditions. A culture which is characterized by a Present time orientation
would primarily emphasize short-term concerns, without regard to the traditions of
the Past or consequences for the Futrn'e. Finally, a Future orientation is possessed by
those cultures in which people anticipate future events to such an extent that Past
traditions and Present concerns are de-emphasized.

The fourth value orientation - Activity - is concerned with the culture's
primary mode of self-expression, and how others typically judge that expression.
Again, three possible solutions are identified which are best viewed as part of a
continuum: Being, Being-in-Becoming, and Doing. Being represents the least degree
1 of societally imposed restriction upon behavioral self-expression. Rather, Being
emphasizes spontaneous expression of human personality. Impulses and desires drive
activity to a greater degree than goal-oriented behavior. Being-in-Becoming
orientation emphasizes activity which is directed toward the goal of development of
all aspects of the selfas an integrated whole. As in the Being orientation, the Being-
in-Becoming orientation is more concerned with what the human being is, than what
the human can accomplish. A crude example of an individual with a Being-in-

Becoming orientation might be the philosophy student at a liberal arts school, who is

3‘35 by a desire 11:1

 

101:th By ca
“3:2". results in ac;

£12321 observer. "._
3111 the perceived Vt
cm W“ thing:

The final orien

 

Ili‘fi‘iwort of value.
1‘11 3" "W hsro
societies “1111 an In;
@1111de m deft:
3am AlthOug}
whom-‘1 in the pu
11121111131“! fo
tits: mm mlecr
a: DUI 330mm
Comm Ofientar

11m indmdual- l:

of“ pan
again take We

43
driven by a desire to determine his role in the world, not by the desire to achieve top-
notch grades. By contrast, the Doing orientation is primarily concerned with activity
which results in accomplishments which can be assessed by the standards of some
external observer. Thus, according to the Doing orientation of activity, self-worth
and the perceived worth of others is based largely upon the individual's actions on
other persons, things, or situations.

The final orientation in the F. Kluckhohn and Strodtbeck (1962) theoretical
framework of values examines the individual's relationships with others. Once more
there are three hypothesized orientations: Individualistic, Communal, and Lineal. In
societies with an Individualistic orientation, each individual's relationships to other
individuals are defined according to goals (and roles) which are autonomous in
structure. Although the individual in an individualistic society is given a great deal of
autonomy in the pursuit of his or her goals, the individual is still expected to maintain
a healthy respect for the goals of other individuals. Thus, students working on an in-
class group project are expected to be cooperative with their fellow classmates, but
are not discouraged from having their own goals of achieving the highest grade. A
Communal orientation emphasizes the goals and welfare of the group over the goals
of the individual. In these societies, the individual is recognized as a human only as
he is a part of a particular social order. In societies with a Lineal orientation, group
goals again take precedence over the individual. Here, however, there are additional

emphases on maintaining continuity of the group over time, and preserving the

44
"ordered positional succession." This orientation is observed wherever kinship ties
are the basis for societal structure.

In summary, the cultural values typology of F. Kluckhohn and Strodtbeck
(1961), consists of five value orientations, each with three alternatives. Each culture
is thought to rank each one of the possible alternatives. The ranking patterns which
are most influential in a particular culture are termed the dominant patterns. The less
influential rankings of alternatives are referred to as variants. Thus, this theory
clearly recognizes individual variation in the ranking of value orientation alternatives.

When comparing across cultures, it is possible to compare the dominant patterns
across cultures, or, alternatively, only the highest ranked alternative to each value.
The dominant value patterns are expressed such that the most preferred alternative of
the 3 alternatives is expressed first, followed by the second most influential
alternative, and finally the third. Thus, if the dominant pattern of alternatives for a
particular society is a over 1: over 9, this would be expressed as a > b > c. Further, if a
particular culture has no preference between two alternatives b and c, the dominant
valuepatternwouldbe expressedasa> b=c.

F. Kluckhohn and Strodtbeck (1962) ofi‘er some speculation as to the highest
ranked alternative to each of the value orientation patterns possessed by the "typical"
American (read, White European American). These theorists think that many
Americans view basic human nature as Evil, but perfectible, based in large part on the
strong Puritan influence of early Americans. However, they further argue that a

growing number of Americans hold a Mixed perspective of human nature. In contrast

The dominant val ue

 

nerefore would be

   
   
 

1mm thought
Perhaps nowhere is

he aged that me.

 

telosion -— serve
Tris saggests the do
me > chrmorrr' w:
”Emmi. accordin

31:3 regard the Pa

45
to those who would view that control and discipline are necessary to achieve
goodness, the latter view would be more understanding of momentary indulgences.
The dominant value pattern for the Human Nature orientation of European Americans
therefore would be: Mixed > Evil > Good. Not surprisingly, the man-nature
orientation thought to be adopted by the typical American is Mastery-over-Nature.
Perhaps nowhere is this as self-evident as our reliance on technology. Indeed, some
have argued that major technological failures - such as the space shuttle Challenger
explosion - serve as mere reminders that humans will never truly control nature.
This suggests the dominant value pattern for the Person-Nature orientation: Mastery-
over > Harmony with > Subjugation to. The time orientation of the "typical"
American, according to F. Kluckhohn and Strodtbeck (1961), is Future. Americans
often regard the Past with a cynical view, are discontented with their Present, and
look expectantly toward a better tomorrow. The dominant value pattern of European
Americans for the Time dimension, therefore, would be: Future > Present > Past.
The activity orientation of most Americans is strongly one of Doing. That is,
Americans judge their own self-worth as well as that of others primarily upon their
accomplishments, and their potential for accomplishments. A typical question a
father may ask a daughter about her new beau, for example, is "What does he do for a
living?". Clearly, then, the "typical" American is concerned with Doing over the
emphasis on spontaneous expression associated with it Being orientation. Thus, the
dominant value pattern can be represented as: Doing > Being-in Becoming > Being.

Finally, Americans are typically assumed to pru'sue an Individualistic relational

 

l ' '- pl
memos 1111ng

   

eerily interested

..::2e. and 50:13; r
._, l 1.3.;ng ”I; > ,
1).. "hung“.sm

1:- KlllCthOhn a

 

3321115 b\ examinil
1m States. Two 1
it were primarily l
alligator-c1 the fiftl

“33.21111 “1115 one

46
orientation. This generalization is based upon the perception that Americans are
generally interested less in maintaining the goals of some groups, as they are in
pursuing their own autonomous goals, which typically include increased wages,
prestige, and social mobility. The dominant value pattern consistent with this logic
is: Individualism > Communal > Lineal.

F. Kluckhohn and Strodtbeck (1961) provided initial evidence for their typology
of values by examining the value orientations of five communities in the southwest
United States. Two of these communities were American Indian (N avaho and Zuni),
two were primarily English-speaking commrmities (a Mormon village and a farming
village), and the fifth community was composed of Mexican-Americans. Differences
in 4 of the 5 value orientations were demonstrated between cultm'es. In particular, the
dominant value orientations expressed by the Mexican-American group were the most
divergent when compared with those of the other groups. As would be expected, the
least difi'erentiation between value orientations was between the Mormon and the
farming communities. In comparison to the Mexican- and Native-American
commrmities, these groups preferred the Individual relations orientation, the Doing
activity orientation, the Mastery-over-Nature man-nature orientation, and the Future
alternative of the time orientation.

Application of the value orientation model of F. Kluckhohn and Strodtbeck
(1961) has been repeatedly suggested for use in clinical psychology (e.g., Beutler &
Bergan, 1991; Carter, 1990; Chapman, 1981; Inclan, 1985). Citing research that has

shown that similarity in values between therapist and client aids the therapy process,

 

Beads and Berg?
ifg'flcwglh bastd 0’

.‘e'

mx'idecre' 311111 '
532.333 (1981 l 5".

rte: tnderstood \\
to an illustration C
pathologists vs. R
r.3 proponents c
idiom: a Present
relational orientati.

32d Harm. Thus

itt'elopment of th

47

Beutler and Bergan (1991), for example, have proposed that therapists adopt a
framework based on the work of F. Kluckhohn and Strodtbeck (1961) that would
provide credibility to therapy involving the examination of personal values. Further,
Chapman (1981) suggests that major theoretical frameworks in psychology can be
better understood when the value-orientation implicit in the fiamework is explicated.
As an illustration, Chapman outlines the dominant value orientations for Gestalt
psychologists vs. Radical Behaviorists. Based on the analysis of the writings of the
major proponents of these theories, Chapman (1981) asserts that Gestalt psychologists
advocate a Present time orientation, a Mixed human nature orientation, an Individual
relational orientation, a Becoming activity orientation, and Harmony between man
and nature. Thus, Gestalt psychologists stress the importance of aiding the
development of the individual's growth potential by focussing on the individual's
current concerns. In contrast, Radical Behaviorists stress the active processes of past
reinforcement contingencies on the individual. Therefore, Chapman (1981) suggests
that the dominant value orientations of proponents of Behaviorist theory are: a Past
time orientation, Mixed human nature orientation, Communal relational orientation, a
Doing activity orientation, and a Mastery-over-Nature (i.e., control) relationship
between man and nature.

The cultural values of Blacks have received very limited empirical
investigation (Carter, 1990). However, several authors have outlined the traditional
philosophies of African societies, fiom which at least some of the traditional values

orientations of African Americans might be derived. Specifically, both McFadden

  

a; Gbekobou (19:
God a the origins:
searched by God
m 3 W501] 15 V1:
“my “inch is 1:.
“151m 0311' becat‘.
ofa cemmlmn). A
the :3:-335mm. .1 a:
Nobles, 1980-]. The

Social relatiOm are

 

 

 

E1 individual. inclt‘

48
and Gbekobou (1984) and Nobles (1980) painstakingly outline philosophies in
traditional Afiican societies. According to these theorists, Africans traditionally view
God as the originator and sustainer of mankind. The natural world is viewed as
controlled by God. Man is seen as a union of both biological and spiritual life. As
such, a person is viewed as a combination of both good and bad elements. Unity is a
concept which is intimately familiar to Africans. The individual person is viewed as
existing only because others exist. That is, to be a human is equated with being part
of a commrmity. According to Mbiti (1970), an African's self-concept is revealed in
the expression, "1 am because we are; and because we are, therefore, I am" (cf.
Nobles, 1980). The emphasis in African philosophy is clearly upon the community.
Social relations are characterized by spontaneity. Conversations can spring up with
any individual, including those one has never previously met. Traditional African
society further sees no distinction between one's beliefs and one's actions. Actions
are viewed as indicants of what one believes.

Afiican societies place great emphasis on the preservation of traditional life-
styles. The oral tradition, in which beliefs and acts are a part of the community,
ensured this preservation. Time is very elastic to Afiicans, though major concern is
placeduponthepastandthepresent. Althoughallofmanis viewedaspartofa
community, kinship determines the strongest ties. According to Nobles (1980),
kinship, along with strong beliefs in the survival of the tribe, combined to form one of
the strongest cohesive devices in traditional Afiican life. McFadden and Gbekobou

(1984) concur, citing the family as the most basic unit of life. In African society,

heel-tire. children .
|

 

mitt-132.1 has the :

These philoso-
I

ctirre of F. Klucia‘
5331a large degre:
rare. Nobles (198
a possessing a "m:
dominant African p;
:51]. The Strong

13133101153111? between;

 

Gods ultimate pow
are Person enema:
if ”151101531113 hem
to > Mas-ref}, Ol‘eh

“mag in the l—‘l’es<

#311011. PQW = f

33110.13] Aftican Sc

4'
or

ag.
n

1111115“ on kinship
Xian

49
therefore, children are considered an invaluable possession. The more children an
individual has, the more glory is brought to that individual.

These philosophies clearly suggest the dominant value orientations of Afiican
culture of F. Kluckhohn and Strodtbeck's (1961) typology of values. Although there
is not a large degree of evidence concerning the Afiican philosophy of basic human
nature, Nobles (1980) points out that traditional Afiican groups generally viewed man
as possessing a "twin" nature of both good and bad elements. This suggests the
dominant Afiican pattern for the Human Nature orientation would be: Mixed > Good
= Evil. The strong emphasis on unity suggest that Afiicans would strongly view the
relationship between Person and Nature as one of harmony. However, beliefs of
God's ultimate power over the universe suggests that Afiicans also value a Nature
over Person orientation. It seems the dominant Afiican value orientation pattern for
the relationship between Person and Nature would be: Harmony with > Subjugation
to > Mastery over. The importance of traditional life, combined with the emphasis
on living in the present, suggest the following dominant pattern for the Afiican's Time
orientation: Past = Present > Future. In consideration of the spontaneity present in
traditional African social relationships, it is suggested that dominant Activity pattern
for Afiicans would be: Being > Being -in-Becoming > Doing. Finally, the strong
emphases on kinship and community clearly suggest the following dominant African
pattern for the Relations orientation: Lineal > Communal > Individualistic. Table 1
provides a summary of the theoretical value orientations for individuals of

Afrocentric and White Eurocentric cultures.

ry.... .53-,33

 

In
Mr

letters

 

 

50

Table l:
Theoretically-derived Dominant Value Orientation Patterns

1! l D . . ! fi' ! . 1M . E .
Human Nature Mixed>Good-Evil Mixed>Evil>Good
Person-Nature Harmony>Subj>Mastery Mastery>Harmony>Subj
Time Past-PresenDFuture Fume>Present>Past
Activity Being>Becoming>Doing Doing>Beeoming>Being

Relations Lineal>Comrnunal>lndiv Indiv>Cornmunal>Lineal

 

 

Th: question
the dominant Vain]
L1985 '5: argues that
bug in America.
Clfitmr must have ':

mession of the u

it'mi'ioral elemenq

 

mmfim 0f slave;
well 35 Pftscn'ed a

aimion Follow

51

The question which arises at this point is whether Blacks might possibly retain
the dominant value orientation patterns associated with African philosophy. Nobles
(1980) argues that in order for African philosophy to maintain an influence on Blacks
living in America, two prerequisites must occur. First, the carriers of the original
culture must have been isolated from the new culture. Secondly, the behavioral
expression of the traditional culture must not openly conflict with the cultural-
behavioral elements of the new culture. Nobles (1980) argues that the holding of
Black slaves in captivity worked to maintain traditional African philosophies. The
institution of slavery simultaneously provided isolation fi-om the American culture, as
well as preserved the oral traditions of African philosophy by discouraging even basic
education. Following the Civil War, racial segregation continued to encourage the
maintenance of African philosophies. Further, inner city ghettoes may be seen as
continuing isolation even today. However, the degree to which African philosophies
are maintained today in the late 20th centtu'y is questionable, especially in
consideration of television's impact on society, as well as the advent of improved
educational equality between the races.

An initial investigation of the difl‘erenees between the value orientations of
Blacks and Whites was conducted by Carter (1990). In a comparison of 300 Black
with 500 White college students, Carter (1990) found many similarities between these
groups in terms of the most important alternative chosen by each group for each value
orientation. Both groups had the highest mean response to the Mixed alternative of

Human Nature, Harmony with Nature for the Person-Nature orientation, Future

maximum for Tim
fixative for file
Backs and Whites
wine orientations.
mad White A...
mung “inch “as i
arm-m “m fom
mm b." bOIh gmt
flamed dominant .'
FORE imcncans v
majors than dii

animation 3.1an .

Rmdm' BlaCks

 

52
orientation for Time, Doing alternative for the Activity dimension, and Individualistic
alternative for the Social Relations orientation. This research indicates that both
Blacks and Whites generally share the most typical alternative to each of the basic
value orientations. Further, these typical responses seem to be more in line with the
expected White American values than with African philosophies. Indeed, the only
finding which was inconsistent with the expected dominant White American
alternative was found for the Nature orientation. Harmony with Nature was typically
chosen by both groups over Mastery over Natm'e, which is more consistent with the
expected dominant alternative for Blacks. This evidence would seem to suggest that
young Americans, whether Black or White, are much more alike in their value
orientations than different. However, significant differences in preferences for value
orientation alternatives were detected when the entire dominant value patterns were
considered. Blacks were more likely than Whites to endorse the (a) Evil alternative
of the Human Nature orientation, (b) the Subjugation to Nature alternative of the
orientation examining the relationship between Person and Nature, (c) the Past
oriented Time orientation, (d) the Being-in-Becoming activity orientation, and (e) the
Lineality alternative of the Social Relations orientation. Of these, the Subjugation to
Nature, Past oriented, and Lineality alternatives are each preferences which are
consistent with African philosophy. Carter (1990) argues that the Black preference
for the Evil alternative of the Human Nature orientation may result from personal
experiences with racial discrimination. Consistent with expectations, Whites in the

Carter study expressed a significantly greater preference for the Individualistic

 

mamas; t
sari} the s

| ,
.. at!”
"385

L ‘H

011

(F2

amicable
33303? of t
mm: of g
a533115 plat
501m COR \‘2
mom mm
3mm“.
gums i5 116:
Mill vs, S-

The pr.
mz’LF-UOIB
ssoo
1% applied

is?! -
Hi In Class

53
alternative of the Social Relations orientations. However, contrary to expectations,
Whites reported a significantly greater Communal Social Relations orientation, and
expressed a greater preference for the Being alternative of the Activity orientation.

As an initial comparison of the value orientations of Black and White college
students, the study by Carter (1990) reveals some interesting findings. Most
noticeably, the value orientation alternatives selected most often by both groups were
exactly the same. Further, there was a great deal of similarity in the preferential
ordering of the value orientation patterns for the two groups. However, there were
considerable differences in the mean responses to the alternatives. Although the
majority of these mean difi‘erences were either consistent with theory or explainable
in terms of group experience, several of these difimnces were unexpected. Although
it seems plausible at this point to say that Blacks and Whites share a preference for
some core values, there does seem to be an influence of Afro-centric and Euro—centric
values, respectively, upon the comparative preferences for various value orientation
alternatives. However, fm'ther examination of the value orientations of these two
groups is necessary before drawing firm conclusions.

Racial vs. Social Class Diflerenees in Value Orientations

The preceding paragraphs implicitly argue that the observed difl‘erences in value
orientations for members of different races are associated with disparate cultural
backgrounds (i.e., Afro-centric, Brno-centric, Hispanic, etc.). Similar argmnents have
been applied to provide explanation for the findings that Black students often do less

well in classroom achievement than their White counterparts. Kaplan (1963), for

54

example, points out that "the term culturally deprived is now viewed with some
disfavor by certain sociologists and educators, more accurate terms they submit are
culturally different or economically deprived. They maintain that children from
impoverished families are not culturally deprived in the sense that they are
cultureless. Rather, their cultures and heritages differ from those cultivated by the
middle class schools they attend" (cf. Atolagbe, 1980). An alternative explanation for
differential values of members of different races is posed by those who argue that
social class and economics are primarily responsible for these findings.

In an attempt to investigate the validity of these two competing explanations for
the differences in expressed values, Malpass and Symonds (1974) administered a
questionnaire assessing values relevant to work, personal characteristics, physical
living conditions, creativity, and religiousness to high school students of difi‘erent
race and social class. These researchers found that social class differences were best
able to accormt for differences in 4 of the 5 types of values examined. Only values
toward religiousness were better explained by race. However, these findings can be
questioned based on the limited sample utilized by Malpass and Symonds (1974). Of
the 142 high school student subjects, only 27 were Black. All of these 27 were
labelled as low social class.

Solorzano (1992) oflered a more serious challenge to the argument that

differences in academic achievement should be attributed to differences in cultural
values. Solorzano (1992) compared the college aspirations of over 16,000 White 8th

graders with the college aspirations of nearly 3,000 Black 8th graders. Contrary to

55
suggestions that Blacks place low value on educational achievement, findings .
revealed that, after controlling for SES, Black eighth graders possessed higher
educational aspirations than their White counterparts. Further, when the parents of
many of these students were asked about their college aspiration for their children,
Black parents reported higher levels of aspirations for their children than did White
parents, after controlling for social class. Solorzano (1992) uses these findings to
argue that differences in cultural values should not be used to explain differences in
academic achievement aspirations. Rather, he suggests that researchers closely
examine the influence of social class upon observed racial differences.
These arguments are in direct contrast to the reasoning put forth by Nobles

(1980). In his attempt to develop a Black Psychology, Nobles argues

"Black Americans derive their most fundamental self-definition fi'om several

cultural and philosophical premises which we share with most West African

'tribes'. . . Black Psychology is something more than the psychology of so-

called underprivileged peoples, more than the experience of living in ghettoes

or of having been forced into the dehumanizing conditions of slavery. . . Its

unique status is derived from the positive features of basic African philosophy

which dictate the values, customs, attitudes and behavior of Africans in Africa

and the New World" (p. 18).
The argument that cultural values are determined by something more than mere social
class status is bolstered by the findings of Chandler (1971). Chandler’s (1971)
investigation of Mexican Americans revealed that even though increased educational

attainment was associated with the increased adoption of modern values, the

dominant value orientation held by this sample remained very traditional.

1’5 asked It

..
Tfl"

QPE'QFYDQ

56

Clearly, two competing arguments have been presented which claim to be
determinants of the adoption of value orientation patterns. It seems clear that at this
point neither possibility can be entirely ruled out. Further research examining the
dual influences of social class and cultural influences on the development of value
orientations appears warranted. Although the present study is not primarily
concerned with the influence of social class on the expression of cultural values,
analyses will be conducted to control for the effects of SES on cultural. The
following section proposes that an individual's cultural values are sufficiently strong
so as to influence item responding on subjective employment tests.
Cultural Values and Item Responding

Job applicants completing a subjective employment test such as a biodata test
are asked to choose which of several multiple choice options they feel is most
indicative of their own past actions, most similar to their own feelings, or most likely
their course of action, given a specified situation. In comparison to objective
cognitive ability-type tests in which the examinee relies on their skills and
knowledges to produce a correct response, examinees completing a subjective test
must depend upon their own feelings, propensities, and dispositions to provide a
response. Could cultural values influence responses to these types of subjective
employment tests? If the prolific value theorist Milton Rokeach is correct, the answer
would seem to be a definite "yes".

According to Rokeach (1973), values are enduring because during childhood

they are instilled in us in an absolute, all-or-none manner. Children are instructed that

 

a 8’33 ‘2
$111 15 65

{19‘ -' 3*?"

‘J’

xiii. f8;
[ff-$5511 tn:
straits (h:
53323111 of V
Once
affair-c. a
Rokeach. m
comiStcnt \
its by pref
S’l‘poned ‘t
Giver,

M81} 1;

57
a certain value is always desirable. Therefore, we are instructed as children to believe
that it is desirable to be not just a little honest, but to be completely honest. Rokeach
(1973) argues that as the child matures, independently held values often come into
conflict, requiring a preferential ordering of these values. For example, conflict
between the values of honesty and obedience to parents may arise when a parent
instructs the child to lie. This process gradually produces a hierarchically ordered
system of values, in which each value is ordered in priority to every other value.

Once formed, values have a strong motivational component as well as cognitive,
affective, and behavioral dimensions (Rokeach, 197 6). Values, according to
Rokeach, are determinants of both attitudes and behaviors. This definition is
consistent with Allport's (1961) assertion that "a value is a belief upon which a man
acts by preference"(p. 454). The idea that values determine preferences is further
supported by C. Kluckhohn's (1952) idea that values determine "oughtness".

Given the impact of values on preferences, attitudes, and behavior, it seems
extremely likely that cultural values will exert an influence on the responses
examinees provide to biodata employment tests. As stated above, biodata instruments
assess the personal backgrounds and life experiences of the individual. Additionally,
biodata items often assess the respondent’s personal preferences. Applicants
administered a biodata inventory are asked to answer each item by choosing from
several multiple choice response options. Given the subjective nature of these tests,
respondents must rely on their preferences, attitudes, and past behaviors for

responding. These are exactly the components that Rokeach argues are influenced by

V2335- Tm
m"; be con:
The Preset.

Deep:
rafiorm' of
far the dete
hardening
his is $15:
of a prion i

3%"; -
stilt} 1:515

58

values. Therefore, to the extent that there are cultural value differences between
people of different racial identity, it would be expected that biodata item responding
will be consistent with the values associated with a particular culture.
The Present Study

Despite a voluminous literature examining the possibility of item bias, the
majority of the existing literature is focussed on differentiating the various techniques
for the detection of such bias, rather than exploring factors leading to differential item
ftmctioning. The literature which does seek to uncover the factors leading to item
bias is susceptible to criticism in that it all too often is conducted without the benefit
of a priori hypotheses, nearly universally limits the examination of bias to cognitive
ability tests, and uses a criterion of physical characteristics as the basis for
determination of subgroup membership. Further, any significant findings revealed by
this research is often attributed to some global "cultural" or experiential difl‘erences
between the subgroups examined. The proposed study will rectify these
shortcomings of the past by:
(1) examining differential item functioning for biodata employment tests;
(2) defining subgroup membership in terms of perceived racial identity;
(3) specifying the cultural values thought to influence item responding; and
(4) testing specific hypotheses.

Figure 1 represents the overall model of differential item functioning which is
examined in the proposed study. Although cultural disparity has routinely been

offered as an explanation for observed differences in item bias investigations (e.g.,

influence rte
mural val:
amine the

' rmrfit

suggestion t

l ' ‘ .
an .....
t1.) tit-ctr. 0’:

.1M _
“magma
The
a?

59

Dorans, Schmitt, & Bleistein, I992; Freedle & Kostin, 1990; Scheuneman & Gerritz,
1990), very little research has examined the actual dimensions of culture which
influence item responding. In providing an initial investigation of the influence of
cultural values on the response patterns of examinees, the present study proposes to
examine the utility of racial identity, rather than observable racial characteristics, as
the grouping variable. The initial proposition of the model offers the rather intuitive
suggestion that racial identity will be related to, and largely determined by, one's
physical, observable racial characteristics.

Hypothesis 1:

Perceived racial identity will be moderately related to a single item measure of

self-reported race.
Although this hypothesis argues that racial identity will be determined in part by the
traditional physically observable racial categories, it is not believed that racial identity
is completely determined by observed racial characteristics. Due to individual
differences in cultural, educational, and socioeconomic experiences, as well as
heterogeneity in the physical expression of biologically determined racial
characteristics, perceived racial identity is proposed to provide a more accurate
classification of the race of an individual than physical characteristics.

The second hypothesis examines the relationship between physically
determined race, racial identity, and cultural values. According to Helms (1990),
racial identity refers to an individual's perception that (s)he shares a common racial

heritage with a particular racial group. Whereas biological markers of race (e.g., skin

if”
Hui .-
4»

60

color) enable us to categorize individuals on mere physical characteristics, racial
identity allows categorization based upon actual psychological characteristics of the
individual. These psychological characteristics would be expected to exhibit a
stronger relationship with other psychological variables than would variables formed
on the basis of physical characteristics. Therefore, it is expected the development of
specific cultural values will be more strongly related to racial identity than physically
determined race.

Hypothesis 2:

Perceptions of racial identity will be more strongly related to cultural values

than a single item self-reported race measure.

F. Kluckhohn and Strodtbeck's (1961) typology of values suggests that peoples
of differing racial identities would be expected to exhibit differences in their
dominant value orientation patterns. According to this logic, the dominant value
orientation pattern associated with a particular cultural heritage (i.e., Eurocentric,
Afi'ocentric, etc.), would be largely maintained in American society, even if the
individual was born and raised in the United States. The individual's exposure to the
value system associated with their racial identity, which would presumably be
provided by parents, grandparents, and others of the same racial identity, would
outweigh the influence of exposure to the typically diversified American culture in
the formation of cultural values. In the opinion of the present author, this argument is
absurd. In this day of mass communication and public education, Americans who

have been born and raised in the United States are likely to be very similar in their

61
dominant cultural value patterns. Although groups such as the Amish provide
evidence that mass communication cannot be equated with mass influence, it is
expected that the dominant cultural value patterns of African- and White-Americans
will be highly similar.
Hypothesis 3:
The dominant cultural value pattern for both of the racial identity groups
examined in this study will be consistent with the proposed dominant value

patterns of the typical White American. Specifically,

Human Nature: Mixed > Evil > Good

Person-Nature: Mastery > Harmony > Subjugation
Time: Future > Present > Past

Activity: Doing > Being in Becoming > Being
Relations: Individualistic > Communal > Lineal

This hypothesis is consistent with Atolagbe's (1980) conclusion that African and
White Americans have values which are more alike than different. Initial evidence
for hypothesis 3 is further provided by Carter (1990). As discussed above, the
dominant value pattern expressed by Blacks and Whites in Carter’s sample of college
students was not exactly consistent with prior theoretical speculation of the typical
White cultural value pattern. Most strikingly different from the above hypothesized
patterns was the emphasis on the Harmony alternative of the Person-Nature value
orientation. However, rather than base the hypotheses regarding the dominant value
patterns of "typical" Americans on the result of a single study (e.g., Carter, 1990),
these hypotheses are based upon the preponderance of theory (e.g., Beutler & Bergan,

1991; F. Kluckhohn & Strodtbeck, I961).

to suggest 1
Rater. it i:
fitteiapm
mdt‘rscme
Pfiference
0381131301.
1), The cc
m't'efore.
m be Cor
Therefore
mm n

In or
weiopm

mum.

62

Although the dominant value patterns exhibited by Americans of different racial
identity are not expected to differ, it is by no means the intention of the present author
to suggest that cultural value orientations do not discriminate between these groups.
Rather, it is expected that racial identification will have a more subtle influence in the
development of value orientations such that subgroups will differ in their preferential
endorsement of the various value orientation alternatives. 'Ihese differential
preferences are expected to be consistent with the proposed dominant value
orientation patterns associated with a particular racio-ethnic group's culture (see Table
l). The concept of racial identity is relatively novel in psychological research.
Therefore, the influence of racial identity strength on the adoption of cultural values
will be considered exploratory.
Therefore, this hypothesis argues that racial identity classification and racial identity
strength may interact to determine the possession of cultural values.

In order to dispel the rival hypothesis that social class determines the
development of cultural values, social class will be statistically controlled across
groups.

Hypothesis 4:
After controlling for social class, the relative endorsements of each value orientation

alternative will be predicted by racial identity classification. Specifically, it is

expectedthat:

4A. African-Americans will place greater emphasis on the Good
alternative of the Human Nature orientation than their White-
American counterparts. '

 

4L
4t

41

4f

4f

Carter's (19'
81th and Whit:
(19901 found sex-e

Wilma Mos

ileit to endorse t

 

63
4B. African-Americans will place greater emphasis on the
Subjugation to Nature alternative of the Person-Nature
orientation than their White-American counterparts.
4C. African-Americans will place greater emphasis on the
Harmony with Nature alternative of the Person-Nature
orientation than their White- American counterparts.

4D. African-Americans will place greater emphasis on the Past
alternative of the Time orientation than their White-American

counterparts.
4E. African-Americans will place greater emphasis on the Being
alternative of the Activity orientation than their White-
American counterparts.
4F. African-Americans will place greater emphasis on both the
Lineal and Communal alternatives of the Relations orientation
than their White-American counterparts.

Carter's (1990) examination of the difi‘erences between the value orientations of
Blacks and Whites again lends some initial support to hypothesis 4. However, Carter
(1990) found several differences which were not consistent with prior theoretical
speculation. Most notably, White-Americans were found to be significantly more
likely to endorse the value orientation indicating Communal relationships than were
Blacks. Once again, however, the above hypothesis has been formed on the basis of
theory that has not yet received an adequate test from the results of a single study.

The primary aim of the present study is to examine the relationship between
cultm'al values and difi‘erential item firnctioning on subjective employment tests. The
remaining hypotheses, therefore, concentrate on the relationship between culture and
item responding. Hypothesis 5 examines the relationship between differences in
endorsement of the various value orientation alternatives, and responses to biodata

test items containing options which have been a priori classified as reflecting a certain

value orientation.

 

Bipolht‘
|
That wil
mmme
oriented."
taint on:

sewn

5A. B.

 

Hypothesis 5:

There will be a significant relationship between endorsement of biodata item

options a priori classified as reflecting some alternative of a cultural value

orientation, and endorsement of the corresponding alternative of the cultural

value orientation scale. Specifically, the following hypotheses will be

examined:

5A.

5B.

SC.

SD.

5E.

5F.

56.

Biodata item options classified as reflecting people's good human
nature will be more likely to be endorsed by individuals who have high
scores on the Good scale of the Human Nature value orientation than
those with low scores on this scale.

Biodata item options classified as revealing human's destined fate will
be more likely to be endorsed by individuals who have high scores on
the Subjugation to Nature scale of the Person-Nature value orientation
than those with low scores on this scale.

Item options classified as revealing people's oneness with nature will
be more likely to be endorsed by individuals who have high scores on
the Harmony with Native scale of the Person-Nature value orientation
than those with low scores on this scale.

Item options which emphasize the past are more likely to be endorsed
by individuals who have high scores on the Past scale of the Time
value orientation than those with low scores on this scale.

Item options which emphasize the present are more likely to be
endorsed by individuals who have high scores on the Present scale of
the Time value orientation than those with low scores on this scale.
Item options which are classified as indicative of emphasizing
spontaneous behaviors will be more likely to be endorsed by
individuals who have high scores on the Being scale of the Activity
value orientation than those with low scores on this scale.

Item options which are classified as emphasizing the importance of the
group over the individual will be more likely to be endorsed by those
individuals who have high scores on the Lineal or Communal scales of
the Relations value orientation than those with low scores on these
scales.

Hypothesis 6 finally examines the relationship between cultural values

and actual differential item fimctioning on subjective employment tests.

Hypotbt.
Ween r.
biodata :

“inch Clif

——-_'_

6t

68.

60.

F171”?

61'.

.3???
.—

6F.

F?

f!

for

 

65

Hypothesis 6:

When racial identity is used as the grouping factor, DIF will be detected in

biodata items when an option in the item contains a cultural value alternative

which differs in preferential rating across subgroups. Specifically,

6A.

6B.

6C.

6D.

6E.

6F.

Biodata items which contain a rationally keyed option which reflects
the basic good nature of man will be differentially less difficult for
African-Americans.

Biodata items which contain a rationally keyed option which reflects
sentiments of human subjugation to nature will be differentially less
difficult for African Americans.

Biodata items which contain a rationally keyed option which reflects
sentiments of human harmony with nature will be differentially less
dificult for African Americans.

Biodata items which contain a rationally keyed option which
emphasizes the past or present will be differentially less difficult for
Biodata items which contain a rationally keyed option which
emphasizes spontaneity will be differentially less difficult for Afiican
Americans.

Biodata items which contain a rationally keyed option which
emphasizes concerns of the group will be differentially less difficult
for African Americans.

Although each of these hypotheses are worded so as to indicate that African-

Americans would benefit from the inclusion of biodata items reflecting cultural

values, it is important to note that African-Americans would be expected to

experience diflerentially more difficulty than White-Americans when a non-keyed

biodata option reflects one of the above cultural values.

sheets it
shite:
{initiate

muster

1103113110” (

METHOD

Sample

Two hundred sixteen college students at a large Midwestern university served as
subjects for this study. Participants were initially recruited through a psychology
subject pool, and received credt toward fulfillment of come requirements in
exchange for their participation. Additional Black participants were recruited by
announcements posted on campus. These participants were paid $10 each for
participation in the 2 hour study.

Following data collection, the data was screened for out-of-bounds and random
responding. Based on this inspection, 9 subjects were discarded from the analyses.
Further, the analyses examined only those subjects reporting a Black or White
demographic classification. The resulting sample size for analysis was 108 Whites,
and 99 Blacks. Thirty-six percent of the sample was male. The median age of the
sample was 20.18.

I l l . . 1

Subjects completed a test battery consisting of a demographic information sheet,
a tw0ocomponent racial identity scale, a cultural values scale, a cognitive ability scale,
random response and social desirability scales, and numerous biographical

information (biodata) items. Each of these scales is presented in Appendix A.

66

DEMON

Par:

racia ide:
item ratia

and tserl':

67
Demographic Information
Participants recorded demographic information pertaining to the individual's

racial classification, age, sex, parental education, and annual household income
during childhood.

Racial Identity

Racial identity was assessed by two scales created by the present author -

racial identity classification and racial identity strength (see Appendix A). The 15
item racial identity classification scale was intended to assess both reference group
and ascribed identity. The classification items inquired about the racial/ethnic group
the individual spends the majority of his/her time with, feels most comfortable with,
admires most, etc. The individual's perceived racial identity was determined by
Whatever racial/ethnic group the individual endorsed most frequently. The racial
identity strength scale used in the present study was initially composed of 20 items
assessing the strength of an individual's racial identity. Following factor analysis, this
30316 was reducted to 8 items (see Results). The racial identity strength items
”(mired the individual to indicate on a 7 point Likert-type scale the degree to which
race is a conscious and important part of his/her identity. The mean response to these
items indicated the strength of one's racial identity.

Cultural Values

A modified version of the Intercultural Values Inventory (ICV) (Kohls, Carter,
8‘ Helms, 1984) was used to assess cultural value orientations. This scale is

composed of 150 items and was created to assess the 5 cultural value orientations and

at; anemi-
Each value i
into sum=
For each ite:
rite scale a
a: would rai
Eetel of end
mm;

Carter
from .50 to
air 9 of .
Biodata Em]

The pr
Witt to ider

“‘73 Pmior

68
their alternative solutions originally proposed by Kluckhohn and Strodtbeck (1961).
Each value orientation is assessed by 30 items. These thirty items are further divided
into subscales such that each value orientation alternative is assessed by 10 items.
For each item, respondents will indicate their level of agreement on a 7 point Likert-
type scale as to whether the item reflects a value by which they either live their lives
or would raise their children . The subscales were scored by computing the mean
level of endorsement of all items composing the scale. Thus, for each of the 5 value
orientations, the individual received 3 subscale scores.

Carter and Helms (1990) report internal consistency reliability estimates ranging
from .50 to .79 for the 15 subscales of the ICV, with a mean internal consistency
estimate of .66.

Biodata Employment Test Items

The present author searched pre-existing item banks of biographical data tests in
order to identify employment items which reflected the value orientations and
alternatives proposed by Kluckhohn and Strodtbeck (1961). Only item banks which
were previously used in selection testing, or were intended for selection purposes,
were examined. Items which were identified by the present author as assessing one or
more of the value orientation alternatives were selected for possible inclusion in the
present study. Items were initially selected if either the stem, or one or more options
reflected a value orientation alternative.

Two additional raters were then asked to read each of the biodata items and

indicate the value orientation alternative or alternatives to which it was related. These

mfg W61? 2
m 53:11} 0}
mg W35 C
maiden!" 5
$553" mid
stem or an i
my. one of
llis 1370695
Hm Na:

1
It

‘1
an“

. and
tier-notes
its present
alternative
that the it
Two i
5% item
BA dimer
and asked u
«it the job.
ibis to C
Maintain a l
Rita's Were

69
raters were assigned the additional task of rating their degree of confidence that the
item stem or option reflected the particular value orientation they had assigned. This
rating was conducted on a 4 point Likert-type scale ranging from 1 "Not at all
confident" to 4 "Very confident". Biodata items were selected for inclusion in the
present study only if (a) two of the three raters (including the author) agreed the item
stem or an item option reflected a particular value orientation alternative, and (b) at
least one of the raters assigned a confidence rating of "3" or higher to their rating.
This process resulted in the inclusion of 107 biodata items. The value orientations of
Human Nature, Person-Nature, Time, Activity, and Relations were reflected in 5, ll,
13, 42, and 36, of the items respectively. Several items were judged to reflect
alternatives from more than one value orientation. Each of the biodata items used in
the present study is found in Appendix A, along with a listing of the value orientation
altemative(s) judged to be reflected in each item, and the name of the scale fi'om
which the item was obtained.

Two independent raters were asked to develop a scoring key for each of the
biodata items based on the job of police officer. Each rater was provided with seven
KSA dimensions thought to represent job-relevant KSAs for the job of police officer,
and asked to develop a scoring key for each item which reflected the important KSAs
for the job. These KSA dimensions were the Ability to Organize, Plan, & Prioritize,
Ability to Communicate Orally, Ability to Relate Effectively with Others, Ability to
Maintain a Positive Image, Ability to Attend to Detail, and Initiative and Motivation.

Raters were also given the opportunity to omit a scoring key for a particular item if

it in “as
moment's.
[ls-repanci
user. A to:
separate biod
limo-tom res;
Poor m
tirelessly to
minimal!)
“Ere dimbu
ail} 311W:
PmdPaflts r
from all anal
m‘th "Sin

70
the item was felt to be completely unrelated to the job of police officer. These raters
recommended the identical scoring key for 85/107, or 79%, if the items.
Discrepancies between the two independent scorers were resolved by the present
author. A total biodata test score was developed by summing across each of the
separate biodata items.
Random response and Social Desirability scales

Poor motivation levels could potentially lead subjects to respond randomly or
carelessly to the items composing the test battery. In order to detect individuals of
exceptionally low motivation, random response items developed by Hough (1990)
were distributed throughout the battery. This scale is composed of 7 items which are
easily answered correctly if the individual takes the time to read the item thoroughly.
Participants with incorrect responses to two or more of these items were excluded
from all analyses. Low internal consistency reliability estimates found in past
research using this scale (e.g., Hough, 1990) are attributable to the low response
variability to the items composing the scale.

Additionally, two social desirability scales were administered to subjects based
on the work of Paulhus (1988). These scales are intended to assess both positively
biased honest answers, as well as deliberate faking. Paulhus (1988) found test-retest
reliability estimates of.65 and .69 after a 5 week period, as well as acceptable internal
consistency reliability estimates (.68 to .86). It should be noted that the inclusion of

the social desirability scales was for exploratory purposes only.

Cognac Abi.‘

General
its. ales we
Pest-sci Tes
retails {.88
tide variety c
l‘onderiic in -

“4...,

L'90“ ar

35011110 Corn;

and “mien:

7]

Cognitive Ability

General cognitive ability was assessed by use of the Wonderlic Personnel Test.
Test takers were given 12 minutes to complete the 50 items composing the Wonderlic
Personnel Test. The Wonderlic has demonstrated acceptable internal consistency
reliability (.88 to .94) and criterion-related validity estimates (rxy = .22 to .53) across a
wide variety of jobs (Wonderlic Personnel Test, Inc., 1992). The inclusion of the
Wonderlic in the present test battery was primarily for exploratory purposes.
Procedure

Upon arrival for study participation, subjects were informed that they were
about to complete employment-type items designed to assess attitudes, preferences,
and experiences. Subjects were informed that many of the items they were
completing were very representative of actual employment items that they would
likely encormter if applying for a job. Participants were told that they could pretend
that they were completing the tests as part of an application process for a position in a
local law enforcement agency. In order to motivate subjects to respond accurately,
subjects were informed:

”Following the completion of all data collection, the 6 individuals with the highest
test scores will each receive $25 cash. However, included in the test battery are items
designed to detect lying and faking. Subjects who are detected as either lying or
faking will not be eligible for one of the $25 awar ."

Subjects were initially administered the Wonderlic Personnel Test. Following

the 12 minute testing, test forms were collected, and the remaining measures were

tight-med. Pei

littoral mes;

 

The data V
(minor: and
berm mitt"

Although
honoring ha'
iii and Mann
Muse and l
these Procedtr
tossed “c.
[111%me l
55min it
Mpoged m
2553' dichou
mm Optic]
theme \‘aryi
flattening I
hired.

RegenL

amine p01:

72
distributed. Participants were given as much time as required to complete the

additional measures.

Analyses

The data were initially scanned for out of bound responses and outliers.
Correlation and hierarchical regression were used to investigate the relationships
between racial variables and cultural values.

Although several procedures for the identification of differential item
frmctioning have gained scientific acceptance, the two most popular have been the
IRT and Mantel-Haenszel procedures. However, traditional uses of both the item
response and Mantel-Haenszel procedures require dichotomously scored items. Thus,
these procedures require the use of a scoring key in which some responses are
considered "correct" and others are considered "incorrec ". Use of these approaches is
unwarranted in the present investigation. Unlike many previous investigations of
difl’erential item fimctioning in which researchers have analyzed pre-existing data sets
composed primarily of objective items, the present study utilizes items which are not
easily dichotomously scored. Although some biodata items may have only one
correct option, many biodata items are scored in such a manner that different options
receive varying degree of credit. A method of examining differential item
ftmctioning that was capable of analyzing polychotomous responses was therefore
desired.

Recently, several authors have suggested using item bias procedures which

examine polychotomous responses (e. g., Dorans, Schmitt, & Bleistein, 1992; Green,

Crone & Folk.

 

tog-knee anal }
chain. and gr:
option choice i

Htpothes;
thionship bet

option selected.

 

tines - once to
Wes the in
is noted above
tittn “a judge.-
910x315 corn;
Mining the 1
his decision V
“M Value 2
imam-'3 jUdl'.

titanium3pc

73
Crone & Folk, 1989). Green, Crone, and Folk (1989), for example, suggest use of
log-linear analysis examining the main effects and interactions of ability, option
choice, and group membership. An item is considered to exhibit DIF whenever the
option choice by group interaction is necessary to explain the model.

Hypotheses 5 and 6 were pursued using loglinear analysis to examine the
relationship between racial identity group membership, cultural value alternative, and
option selected. For each of these hypotheses, loglinear analyses were repeated 107
times -- once for each of the biodata items included in the study. Hypothesis 5
examines the influence of cultural values on responses to biodata employment items.
As noted above, each biodata item was selected for inclusion in the study only if the
item was judged to reflect a particular cultural value orientation. Although 3
subscales compose each cultural value orientation, Hypothesis 5 was tested by
examining the influence of only a single cultural value alternative per biodata item.
This decision was made in response to concerns with family-wise error rates. The
cultural value alternative examined for a particular item was the cultural value
alternative judged most closely associated with the item, based on the direction of the

stem and response options.

“5. ”‘l'r'ita in" 'I

 

This int:
at Oliilt Site;
The me:
5&5 variables '.

mim' mine 5;

“$10.78“?

 

RESULTS

EHIE.

This introductory section is intended to provide descriptive information regarding
each of the scales used in the present study.

The means, standard deviations, and coefficient alpha reliability estimates of all
study variables are presented in Table 2. The alpha reliability estimates of the fifteen
cultural value scales composed of 10 items each were irritally estimated to be in the range
of .54 to .78, with a mean = .66. These estimates were consistent with reliabilities
reported in previous research (e.g., Carter, 1984; Carter & Helms, 1987). Using ten items
for each of the cultural value scales, Carter (1984) reported alphas ranging from .37 to
.73, with a mean reliability of .52. Carter and Helms (1987) also used the full ten items
in each of the 15 scales, and reported alphas ranging from .50 to .76, with a mean of .65.
However, in an attempt to increase coefficient alpha reliability estimates for these scales
in the present study, principal components analysis was used to determine the underlying
factor structure of each scale. These results were closely scrutinized to determine
whether the resulting factors were consistent with the theoretical construct the items were
intended to assess. Next, reliabilities were calculated for those items composing factors
which were consistent with the intended construct. Item-total correlations were
inspected. If this empirical analysis indicated that dropping a particular item would
increase scale reliability, the item was once again inspected for theoretical consistency
with the intended construct. If the item was determined to be theoretically consistent
with the intended construct, the item was retained. If the item was judged to be only
tangentially related to the intended construct, it was dropped, and the reliability was re-
calculated. This procedure resulted in the deletion of an average of 2.7 items from each

74

 

Ext an 87:.

Van'ablg
beef Tw‘qu

Rica. 13:an
At!

Sui

Father Etta:
Moths Educ:
hm:
Ut‘mitin'

"k, I“! ll'.‘

Eni

Muted

Good
Wetter
Hmony

“my

Past

Present

Far:

2:“
rte-m.

Do . Bee

1%

Cost.
WWW

Overall Means, Standard Deviations, and Reliability Estimates

 

75

Table 2:

 

 

Black and White Samples Combined

Variable lanterns Mean SD or
EmsLInfonnsticn

Racial Identity Strength 8 3.76 .73 .87
Age ' 1 20.18 2.14 n/a
Sex8 1 .36 .48 n/a
Father Education 1 4.26 1.25 n/a
Mother Education 1 4.31 1.08 Na
Income 1 3.83 1.53 n/a
Urbanicity 1 3.05 .84 n/a
Evil 8 2.87 .57 .73
Mixed 7 3.85 .42 .59
Good 7 3.79 .45 .64
Subjugation 8 3.15 .63 .78
Harmony 9 3.64 .42 .79
Mastery 8 2.83 .51 .67
Past 6 2.48 .54 .66
Present 10 3.55 .48 .79
Future 9 3.60 .45 .72
Being 7 3.66 .43 .69
Being-in-Becoming 6 3.57 .46 .64
Doing 7 3.84 .44 .73
Lineal 8 2.70 .51 .69
Collateral 6 3.10 .45 .55
Individual 4 4.12 .49 .73
Cognitive Ability 50 23.39 5.96 .72
Biodata Test Score 1 14 88.45 14.16 .82
Random Response Scale 7 .27 .45 n/a
Self-deception 17 7.20 2.67 .58
Impression Management 20 6.34 3.56 .80

 

llcoded0=-=f'emale,lI-male

 

Black Sample

t"H

Is r
has”: IO‘s-Ir

its 13:23:} .
Age
Su‘

in: Educanc
More: Bataan
hm:

76

 

 

 

Table 2 (cont’d)
Black Sample Only
Yarnblc M M SD
Emufllnfonnstim
Racial Identity Strength 8 4.22 .56
Age 1 20.13 1.70
Sex8 1 .23 .42
Father Education 1 3.87 1.23
Mother Education 1 4.11 1 11
Income 1 3.35 1 47
Urbanicity l 3.45 70
W '
Evil 8 2.97 .60
Mixed 7 4.00 .35
Good 7 3.80 .48
Subjugation 8 3.42 .55
Harmony 9 3.67 .41
Mastery 8 2.90 .49
Past 6 2.56 .57
Present 10 3.59 .48
Future 9 3.71 .49
Being 7 3.64 .43
Being-in-Becoming 6 3.59 .46
Doing 7 3.94 .42
Lineal 8 2.85 .48
Collateral 6 3.14 .47
Individual 4 4.14 .49
Minimum
Cognitive Ability 50 20.80 5.00
Biodata Test Score 114 90.77 14.08
Random Response Scale 7 .32 .47
Self-deception 17 7.87 2.53
Impression Management 20 6.90 3.35

 

acodedo-It‘emale,1'I-male

Waite Sampzl

1"“

hi?“ l"?!
in. menu-r;
5:1

Sn‘

Fame: Edna:
Home: Educ:
Income

 

 

77

 

Table 2 (cont’d)
White Sample Only
laudable Lumen: MeaL SD n
EmaLInfonnation
Racial Identity Strength 8 3.34 .60 .77
Age 1 20.19 2.49 n/a
Sexa 1 .47 .50 n/a
Father Education 1 4.62 1.16 Na
Mother Education 1 4.50 1.03 Na
Income 1 4.27 1.46 Na
Urbanicity 1 2.67 .77 No
Evil 8 2.78 .52 .72
Mixed 7 3.71 .42 .60
Good 7 3.78 .41 .59
Subjugation 8 2.90 .60 .80
Harmony 9 3.60 .42 .82
Mastery 8 2.77 .52 .70
Past 6 2.41 .50 .69
Present 10 3.51 .49 .82
Future 9 3.50 .38 .59
Being 7 3.69 .43 .65
Being-in-Beeoming 6 3.54 .46 .69
Doing 7 3.75 .44 .72
Lineal 8 2.56 .49 .73
Collateral 6 3.06 .43 .53
Individual 4 4.11 .49 .75
Cognitive Ability 50 25.76 5.79 .63
Biodata Test Score 122 86.33 13.97 .81
Random Response Scale 7 .22 .42 nla
Self-deception 17 6.58 2.67 .
Impression Management 20 5.82 3.69

 

'codedfifemale, 1 -male

0111* 15 cu
roamed ir
. 15 win
reiiabilin 6
1113311} 01:

The
any; Hy c
minis w:
mggmedi
Title 3. T
labels and
mgh as
10 “hich r.
We of s:
W] groL
”1336 11m
Giant's m
macaw 01'
Mrs. A
£3510? hes
I“ item:
.633ij
3211213125
km Stale

78

of the 15 cultural value sub-scales. The mean of each of the remaining items was
computed in order to form a scale score. The items composing the final scale for each of
the 15 cultural value alternatives are found in Appendix A. The revised coefficient alpha
reliability estimates for the fifteen cultural value scales reported in Table 2 improved
slightly over the original estimates (range: .55 to .78, mean = .69).

The twenty items created by the present author to assess racial identity strength
originally exhibited an internal consistency reliability of a = .83. Principal components
analysis was used to examine the underlying factor structure of these items. This analysis
suggested four interpretable factors. These factors, with item loadings, are presented in
Table 3. The items composing these distinct factors were examined to determine factor
labels and interpretation. The first factor appears to assess the construct of racial identity
strength as originally conceived in this study. Namely, the first factor assesses the degree
to which racial identity is perceived to be an important component of the individual's
sense of self. The second factor seems to reflect an extreme preference for one's own
racial group, or perhaps reflects a behavioral manifestation of racial identity strength.
These items reflect the degree to which the individual avoids other racial groups in favor
of one's own group. The third factor assesses personal meditation and reflection on the
concept of race. The fourth factor appears less interpretable than the previous three
factors. Although factors 2 and 3 represent potentially interesting constructs, the first
factor best represents the construct of racial identity as intended in the present study.

Two items were removed from this first factor following rational inspection of each item:
"Generally I don't give much thought to racial issues", and "I believe we should de-
emphasize racial differences”. The internal consistency reliability of the remaining 8-
item scale was .87. This factor was used in all further analyses as the measure of racial
identity strength.

79

Table 3:

Principal Components Analysis of Racial Identity Strength

 

Factor

Item 1 2 3 4
1 am strongly committed to my racial group .79 .32 -.18 .09

membership.
Matters of race are very important to me. .70 .14 .22 -.03
My race is an important part of who I am. .69 .37 -.05 .08
I am proud to be a member of my race. .66 .23 -.14 -.06
I think one's racial identity is an integral part of

one's personality. .62 .23 .12 .19
Generally, I don't give much thought to racial

issues. (R) .62 -.10 .46 .04
I rarely consider what race I am. (R) .62 .06 -. 12 .15
I strongly identify with members of my own race. .60 .44 -.19 .23
I wish I were a member of some other race. (R) .51 .27 .05 .43
I believe we should de-emphasize racial

differences. (R) .48 -.35 -.38 .28
If given the opportunity, I'd spend the vast

majority of my time with members of my

own racial group. .27 .70 -.27 .16
I tend to socialize with members of my own

racial group. . 13 .70 .00 .28
l consciously seek out members of my own race

for social activities. .32 .67 -.18 -.07
I generally feel the most "attached” to other

members of my own race. .31 .51 -.06 .41
I've never given much thought to the racial

experiences of others. (R) .16 -.25 .78 -.06
I've never really given much thought to what

it would be like to be a member of some other

race. (R) -.07 -.06 .73 -.07
I generally value die opinions of individuals

of my own racial identity more than I value

the opinions of individuals of other racial

groups. .29 .45 -.46 -.05
As a youth, I walirrhnarily exposed to individuals -.01 -.03 -.06 .74

of my own rac' group.
Some people think I spend too much time acting like

people of some race other than my own. (R) .11 .16 .02 .71
I believe there are more positive characteristics

associatedwithotherracesthanmy own. (R) .10 .28 -.15 .35

 

11 is
1852‘}. "’ SI
0

1 V .
$153321 3:
T-

L
32839138
“ '1.
Willi 1.:
51...“;1‘:

Name 1

80

It is also important to note the low coefficient alpha for the seven item random
response scale. This estimate is consistent with previous findings (e.g., Hough, 1990),
and can be attributed to the low variability in responses to the scale.

The intercorrelations among the primary variables included in the study are
presented in Table 4. Of particular importance are the intercorrelations of the 3 cultural
value alternatives which are intended to assess 1 of the 5 cultural value orientations. For
example, the Good, Mixed, and Evil subscales are all intended to assess the Human
Nature orientation. These subscales could be viewed as representing a continuum of
human nature. If these separate subscales were found to correlate highly with one
another, it would be proper to combine the subscales (with reverse coding if needed) into
a single composite scale. Inspection of Table 4 reveals that the 15 scales assessing
cultural values have only low to moderate intercorrelations. Indeed, examining the
intercorrelations for 2 scales at the opposite poles of a cultrn'al value orientation, the
absolute value of these correlation are all in the range of .1 to .34. This evidence of
discriminant validity argues against combining any of the 3 scales assessing one of the 5
cultural value orientations into composite scales. Rather, the 15 scales assessing cultural
values were examined separately for each of the analyses described below.

Intercorrelations across cultural value subscales assessing distinct cultural value
orientations are also low to moderate. However, the intercorrelation between the Future
and Doing subscales appears particularly strong (r = .61). This seems sensical since the
Future subscale assesses the degree to which the individual plans for future events, while
the Doing subscale assesses the degree to which the individual is accomplishment-
oriented. Accomplishment often requires significant planning abilities.

Several additional correlations are also noteworthy. A surprisingly high
correlation was observed between cognitive ability and the Subjugation to Nature

subscale (r = -.47). It appears that individuals with lower cognitive ability are more likely

IEO-uinfithULOc-un O-GIP- I>
u' O-AHIIF

81

v—

we.
8.-
no.
no.
2.
N_.
«N.
3.
2.
2.
2.
3.
mm.
Nv.
mm.
2.
o.—

n—

S.-
N.
N.
N— .-
_ n.
a _ .
1.
8.
5m.
n _ .
N.
N.
N.— .
S .-
8.
N.
no.

a.—

N—

no.
3.
2.
3.-
3.

S.
2.
an.
3.

3.
_N.
2.

.N.
N.
on.
2.

=

8.
8.-
no.
N _.
n _ .-

:—

N.-
3.
N_. -
a _ .-
3.-
8.-
_ _.-
cm.-
_ _.-
3.

_ _.-
.N.-
N.-
2 .-
8.-
$.-
3.-
N.-
«N.-

N.-
2. -
No.

S.

_ _.-
N _ .-
3.-
no. -

2.

2 .-
3 .-
2 .-
5.-
. _.-
S .-
N.-
N.-
_ _.-
n _ .-
n _ .-
8.-

8.-

cN.-
3.-
no.

8.

8.
8. I
ma.
3.
N_.

N- .-
No.-
6N.-
2 .-
3 .-
3. -

o. _

guess—Eeflofi— czar!»

a. sea.

2 .-
ri
NN. -

uv.
n _ .
vN.

mm.-

.3253
5.889.:
§a08§8
58.5.

.5 seem
.3232.
.2223
.85..

9.59

we_£ouom-:.-m:._om
mean

23:...
.5an
“man—
.0282.

ace-Ea: .

acmfiuafizm
e80
.852

=>m

E5. 2538
2:8:—
538:3 .8an
3.333— 85¢"—

a
wm<
5225 £52 33..

.946 been: 38:
.83.

6N
.NN
.NN
.oN
.mN
.N
.N
.NN
._N
.3
.m—
.2
.2
.2

~Nnvvi~or~oo

2.3:er V 0-: sh.

82

c.—

«N

N_.
c.—

N

5N

a. .

3.-

me.-
c.—

3

8. no.” 8.

an. 8 no.

on. 8. 3.

3r 8: _ _.-
c. No. 5 ~ .

N N 3

8.

N_.
8.-
S.
2.-
9..
c.—

NN

8.2.5. 9 oz...

8.
on.
on.
9N.-
av.
«N.
N.
N.

N

g

3.-
2.
N_.

am.
_ N.
No.-
N.
am.

a—

N.
2.
2.-
on.
n_.
on.
«N.
3.
ON.
_ _.
6..

cc.-
8.-
. _.
8.-
2 .
on.
N.
3 .
N.
N.
Na.
n _ .
o. _

h—

a

0—2: a _ 6.2.5 I o 33 a
56:25. 58.5.. I . 585:2. 823 n c :38 u

 

8.
Na.
8.
no.

2.-

Ease:
5:30.95
8:88.23
See—am

a: see...
.3222.
.22....8

.oN
.NN
.NN
.oN
.mN
.N
.N

3:5 .NN

58
3588.538
2.3.

88:...

5805

.3.

E282

._N
.3
.2
.m—
.2
.2
.m—

IO fill:
SUITES

‘1'}?
.tfism
U

were l
a; at:

stall

L)"
(b

83

to endorse the notion that humans are at the whim of Nature. Another surprising
correlation was found between overall scores on the biodata test, and the Impression
Management subscale assessing socially desirable responding (r = .53). Individuals who
were likely to manage self-impressions were also likely to score highly on the biodata test
(as evidenced by the police officer scoring key). This finding brings into question exactly
what biodata instruments measure - actual beliefs and experiences, or the individual’s
perceptions of the “correct” answer.

Evidence that the sample used in the present study is not unrepresentative of the
larger population is found in the intercorrelations of dummy-coded race with the SES
variables. Whites were more likely to have experienced greater household income during
childhood, and their parents were more likely to receive more education than the parents
of Blacks. Examination of the intercorrelation of cognitive ability with dummy-coded
race provides further evidence that the sample used in this study is not considerably
different from the general population - consistent with Wonderlic database information
(W onderlic Personnel Test, Inc., 1992), Whites performed about one standard deviation
better on the Wonderlic than Blacks (r = -.42). Cognitive ability scores do not appear to
be strongly related to overall score on the biodata test (r = -.05).

Descriptive Information Summary

Although the coefficient alphas of some cultural value scales were quite low,
factor analysis and rational deletion of items was used to improve scale internal
consistency estimates considerably. The coefficient alpha reliability estimates of each of
the scales reported in Table 2 appears sulficient for use in scientific inquiry.

Inspection of Table 4 provides evidence that the scales were measured
appropriately. Evidence of discriminant validity across subscales of the cultural value
orientations argues against the formation of composite scales. Rather, the sample used in

first-28552111

seer}. ten.

 

wiggle file:
as found I
= .93 '1. Fur
considered
Valle Ame
an 41mm
considered
African An
Missing
{0’3 made

he: 1'an'a't

84

the present study reflects societal findings that Blacks, relative to Whites in American

society, tend to have parents with less education and earnings.

 

Hypothesis 1 asserts that racial identity classification will be moderately related to
a single item measure of self-reported race. Table 4 reveals a strikingly high correlation
was found between dummy coded race, and dummy coded racial identity classification (r
= .93). Further inspection of the data revealed that of the 108 participants indicating they
considered themselves demographically as White, 106 were classified as possessing a
White American racial identity. Only 2 of these individuals were classified as possessing
an African American racial identity. Similarly, of the 99 participants indicating they
considered themselves demographically as Black, 97 were classified as possessing an
African American racial identity. Only 2 of these individuals were classified as
possessing an White American racial identity. In contrast to Hypothesis 1’s expectations
for a moderate relationship between demographic race and racial identity classification,
the 2 variables appear nearly identical.

Hypothesis 2 asserts that racial identity classification would explain more
variance in cultural values than would demographic race. In order to determine whether
this hypothesis was supported, the significance of differences in correlations between
each of the variables with each of the fifteen cultural value scales were examined using
the procedure recommended by Downie and Heath (1970). The results of this
examination are presented in Table 5. This table illustrates that there are no significant
differences in the correlations between demographically—determined race and racial
identity classification with any of the cultural value scales, with the exception of a single

pair of correlations (Being), both of which were near zero. This evidence indicates that,

 

 

Exil

. lei
Good
Subjugauc
Hmrny
Mates
Past
Perm
Fran
Bang
Beng-in-I
501mg
11a
Collars-ta”
bdlvliua

r\
altar:
“‘1- €le

85

Table 5:
Correlation of Race and Racial Identity Classification
with Cultural Values
1r- . ' -' I'ruurur (:5
Evil .17
Mixed .35
Good .02
Subjugation .42
Harmony .08
Mastery 12
Past 14
Present 08
Future .24
Being -.06
Being-in-Becoming .05
Doing .21
Lineal .29
Collateral .09
Individual .02

 

'0
I.

"' indicates a significantly different correlation between measures of demographic status

and cultural value measm'e (p < .05)

 

3:20.22 1

..

J '.
tame.
realm

lz-rvzrmzz

w

psych:

Id

f

n1

86

at least for this sample, assessing psychological racial identity provides little additional
information over a single item demographic response. Therefore, there was no empirical
support for Hypothesis 2. Given the traditional concern in both psychological assessment
and selection hiring to consider self-reported race, all fiirther analyses in the present study
requiring grouping by race were based upon the single item measure of self-reported race.
Summary

A higher than expected correlation was found between demographic race and
psychological racial identity classification (r = .93). Due to this finding, the expectation
that an individual's adoption of cultural values would be more strongly related to
classification based on their psychological racial identity rather than mere demographic
characteristics did not receive empirical verification. Neither Hypothesis 1 nor 2 received
complete support. Given the customary use of self-reported race for the racial
classification of individuals in psychological research, all further analyses requiring racial
subgroups used a single item measure of self-reported race as the grouping variable.

Waugh

Differences in background information across the two racial subgroups were
examined using independent-samples t-tests. The means and standard deviations for both
Black and White racial subgroups for each of the following t-tests are reported in Table 6.

A listing of the actual items and response options used to assess background information
is available in Appendix A.

Although no difference was detected in the average ages of these two groups
(t(185)= .04, p>.05), White participants were less likely to be female than were Black
participants (t(204) = 3.72, p<.01). Whites scored significantly more highly than Blacks

on each of the socioeconomic variables. Whites were more likely to be raised with a

Variable

 

Age

Sex‘

Householu
Peter’s E
Mother's l
Urbain:
Cogmive
Cog. Abi.‘
01ml B
Self-Dec-
“Picnic
Random ]

Racial Id-

macaw
racial SI

87

Table 6:
Results of t-tests Examining Race and Background Information

 

 

Black White

Mable? Mean SD Mean SD

Age 20.18 1.70 20.19. 2.49
Sex“ .23 .42 .47 .50
Household Income“ 3.35 1.47 4.27 1.46
Father's Education“ 3.87 1.23 4.62 1.16
Mother's Education“ 4.11 1.11 4.50 1.03
Urbanicity‘ 3.45 .70 2.67 ’ .77
Cognitive Ability“ 20.80 5.00 25.76 5.79
Cog. Ability adjusted for SES" 20.88 5.00 25.75 5.79
Overall Biodata Score" 90.77 14.09 86.33 13.97
Self-Deception“ 7.89 2.52 6.58 2.67
Impression Management“ 6.90 g 3.35 5.82 3.69
Random Response Scale“ .44 .65 .25 .48
Racial Identity Strength" 4.22 .56 3.34 .60

 

"' indicates significant difference between
racial subgroups, p<.05

higher f2
and l
lines 1
545.0%
cluster
pinup;
Birch
educatic

and not

growing

“trifle

M raci
lien afi
W165 1
35.32, p

Scared fl

impress
{(305) =
lllites a
“1'3931 =
alies.

mom F

110m [Dee

88

higher family income, (t(202) = 4.49, p<.01). Whereas Blacks on average reported their
annual household income during childhood as just above the $35,000 - $44,999 range,
Whites on average reported their childhood household income to be just above the
$45,000 to $59,999 range. The fathers of White participants were on average more
educated than their Black counterparts (t(220) = 4.51, p<.01), and the mothers of White
participants were generally more educated than Black participants, (t(200) = 2.61, p<.05).
Blacks generally reported fathers with education less than "some college" and mother's
education to be just above "some college", Whites generally reported that their fathers
and mothers had received "some college". Blacks were more likely to have been raised in
urban areas than Whites (t(204) = -7.59, p<.01. Whereas Whites on average reported
growing up in towns of somewhere between 2,000 and 100,000 people, Blacks on
average reported growing up in cities composed of between 10,000 and 100,000 people.

A significant difference was also detected in the cognitive ability scores for the
two racial groups, t(203) = 6.61, p <.01, with Whites outperforming Blacks on this test.
Even after statistically controlling for the effects of parental education and income,
Whites were found to score significantly higher on the Wonderlic (AR2 = .14, F(1,202) =
35.32, p < .01). In contrast, Blacks outperformed Whites on the biodata test rationally
scored for the job of police officer, (t(203) = -2.27, p<.05).

Blacks also scored significantly more highly on both the self-deception and
impression management subscales assessing social desirability (t(205) = -3.55, p<.01 and
t(205) = -2.19, p<.05, respectively). Prior to deletion of cases who responded carelessly,
Whites and Blacks differed significantly in responses to the random response scale
(t(193) = -2.44, p<.05), indicating that Blacks were more careless in their responding than
Whites. This difference evaporated following the discarding of cases with more than two
random responses. Finally, the importance of racial identity was found to be significantly
more meaningful for Blacks than for Whites (t(205) = -10.87, p<.01).

Sam-m

B
pitcipa
Blacks o:
inn urb
Won-ial:
we alsc
manager

bemom

l “1.1.31 ‘

students 1
the 3 win
t‘llll‘alcn
mm

"31831111

[5901? 0ft

89

Summary
Blacks and Whites were found to differ on a number of variables. Black

participants tended to be female more often than White participants. Whites outscored
Blacks on each of the socioeconomic variables, and were more likely to have been raised
in an urban environment. While Whites had a higher average score on the 50 item
Wonderlic Personnel Test, Blacks outscored Whites on the overall biodata test. Blacks
were also found to score more highly on both the self-deception and impression
management subscales of the social desirability measure. Finally, Blacks were found to
be more concerned with racial identity than Whites.

WW

Carter (1990) found the cultural value patterns of Black and White college
students to be highly similar. Hypothesis 3 similarly posits the preferential ordering of
the 3 cultural value alternatives for a particular cultural value orientation will be
equivalent for both Blacks and Whites. Note that examination of this hypothesis requires
confirming the null hypothesis. Therefore, this hypothesis is considered exploratory, and
is meant to illustrate the similarity in the patterning of cultural values across the two
racial subgroups. The hypothesized cultural value patterns in Hypothesis 3 are based on
theory of cultural values. These theory-based hypothesized patterns differ fiom some of
Carter's (1990) empirical findings. This hypothesis is therefore further useful as an
examination of theoretical cultural value patterns vs. previous empirical findings.

In order to assess the cultm'al value orientation patterns for a particular racial
subgroup, paired samples t-tests were conducted between each of the 3 cultural value
alternative scales assessing a single cultural value orientation. For example, in order to

examine Blacks' preference for the Good, Mixed, and Evil alternatives of the Human

Nan:
those
exari
Hume
ofienz

which

90

Nature orientation, several steps were followed. First, data records were selected only for
those individuals who designated themselves as Black. Second, paired samples t-tests
examined the preferential endorsement of the Good, Mixed, and Evil subscales of the
Human Nature orientation. Finally, a cultural value pattern for the Human Nature
orientation was formed by placing a greater than sign (e.g., >) between any 2 scales
which difl‘ered significantly, and an equal sign (e.g., =) between those scales which did
not differ significantly. This analysis was repeated for the remaining cultural value
orientations of Person-Nature, Time, Activity, and Relations. Each of these analyses was
then repeated, selecting only those individuals who reported a demographic classification
of White.

Table 7 presents the means, standard deviations, and significant differences across
the 5 cultural value orientations for Blacks. Table 8 does the same for Whites. The
cultural value patterns across the 5 value orientations of Human Nature, Person-Nature,
Time, Activity, & Relations for both Blacks and Whites are presented in Table 9.
Inspection of Table 9 reveals a replication of Carter's (1990) findings that Blacks and
Whites share nearly identical cultural value patterns. This expectation is consistent with
hypothesis 3. However, the cultural value patterns presented in Table 9 are more
consistent with Carter's (1990) findings than with the theoretical expectations reflected in
hypothesis 3. Specifically, based on theoretical grounds, it was posited that the typical
American's view of Human Nature would place greater emphasis on the Evil alternative
than the Good alternative. As seen in Table 9, although both Blacks and Whites agreed
that the Mixed orientation was very important, both groups placed greater emphasis on
the Good alternative than the Evil alternative. The Person-Nature orientation of both
racial subgroups also failed to support theoretical speculation. Although the Mastery
alternative was expected to be most highly valued by typical Americans, both racial
subgroups had a strong preference for the Harmony with Nature alternative, followed by

91
Table 7:

Results of t-test Comparisons of Cultural Value Alternatives - Blacks

 

 

 

 

 

 

 

 

 

 

Humanflamreflriematipm Mean SD
Evil 2.97 .60
Mixed 4.00 .35
Good 3.80 .48
Evil vs. Mixed: t(98) = -16.14, p <.01

Evil vs. Good: t(98) = -9.45, p < .01

Mixed vs. Good: t(98) = 3.74, p <.01
Won Mean SD
Subjugation 3.42 .55
Harmony 3.67 .41
Mastery 2.90 .49
Subjug. vs. Harmony: t(98) = -4.30, p <.01

Subjug. vs. Mastery t(98) = 8.20, p<.01

Harmony vs. Mastery: t(98) = 12.28, p <.01
W Mean SD
Past 2.56 .57
Present 3.59 .48
Future 3.71 .49
Pastvs. Present: t(98)=-15.10,p<.01

Past vs. Future: t(98) = -15.65, p<.01

Present vs. Future: t(98) = -2.06, p<.05

E . . D . . l I SD
Being 3.64 .44
Being-in-Becoming 3.59 .46
Doing 3.94 .42
Being vs. Being-in-Bec.: t(98) = 1.13, p>.05

Being vs. Doing: t(98) = -6.10, p<.01
Being-in-Bec. vs. Doing: t(98) = -6.02, p<.01

_ ' ' n Mean AL

Lineal 2.85 .48
Collateral 3.14 .47
Individual 4.14 .49
Lineal vs. Collat: t(98) = -5.64, p<.01

Lineal vs. Individ.: t(98) = -17.44, p<.01

Collar. vs. Individ.:

t(98) - 45.04, p<.01

 

SSH

IPPF

P 5

‘LEE[

tLEE

unEQIL Ch

IL L at»

 

Results of t-tests Comparisons Between Cultural Value Alternatives - Whites

ll ll {1' 'n

92
Table 8:

 

 

Evil
Mixed
Good

Evil vs. Mixed:
Evil vs. Good:
Mixed vs. Good:

WWII

 

 

Subjugation

Harmony

Mastery

Subjug. vs. Harmony:
Subjug. vs. Mastery
Harmony vs. Mastery:

I' D' .

 

 

Past
Present
Future

Past vs. Present:
Past vs. Future:
Present vs. Future:

!"D"

 

Being
Being-in-Becoming
Doing

Being vs. Being-in-Bec.:
Being vs. Doing:
Being-in-Bec. vs. Doing:

81' 0..

 

 

Lineal
Collateral
Individual

Lineal vs. Collat:
Lineal vs. Individ.:
Collat. vs. Individ.:

 

Jean SD
2.78 .52
3.71 .42
3.78 .41
t(107) = -19.88, p<.0]
1007) = -14.11, p< .01
t(107) = -1.41, p>.05
Mean SD
2.90 .60
3.60 .42
2.77 .52
1007) = -9.87, p <.01
1007) = 2.16, p>.05
t(107) = 12.52, p <.01
MEL SD
2.41 .50
3.51 .49
3.50 .38
t(107) = -17.37, p <.01
t(107) = 48.53, p<.01
t(107) = .17, p>.05
Jean SD
3.69 .43
3.54 .46
3.75 .44
t(107) = 3.57, p<.01
«107) = 429, p>.05
t(107) = 4.01, p<.01
Mean 8L
2.56 .49
3.06 .43
4.11 .50
t(98) = -10.93, p<.01
t(98) = -20.67, p<.01
t(98) = 45.46, p<.Ol

 

Vain:
Hana

P6501

93

Table 9:
Cultural Value Patterns for Blacks and Whites

 

 

Xalueflrientatien Blacks Whites

Human Nature: Mixed>Good>EVil Mixed=Good>Evil '
Person — Nature: Harmony>Subjug.>Mastery Harmony>Subjug.>Mastery
Time: Present=Future>Past Present=Future>Past
Activity: Doing>Being=Being-in-Bec. Being=Doing>Being-in-Bec.
Relations: 1ndividual>Collat.>Lineal lndividual>Collat.>Lineal
macs

N = 99 Blacks

> denotes signigicant difference (p<.05) between mean scale scores
= denotes no significant difference (p>.05) between mean scale scores

my

_ 5?

KT)

94

the Subjugation to Nature viewpoint. Neither group endorsed the Mastery alternative as
highly as these other alternatives. Both racial subgroups shared the same patterning for
alternatives to the Time orientation. Once again, however, the pattern produced failed to
perfectly support theory. Despite the expectation that the typical American would
endorse the Future alternative more highly than the Past or Present alternatives,
endorsement of the Futm'e alternative did not differ significantly from endorsement of the
Present alternative. The patterns for the Activity orientation were the most interesting.
Blacks and Whites did not share the same patterning of cultural value alternatives.
Although as expected both groups placed a large degree of emphasis on the Doing
alternative , the White group surprisingly placed an equally great emphasis on the Being
alternative. The patterns for the Relation orientation were consistent with expectations
based on theory. Both groups preferred the Individual alternative over both the Collateral
and Lineal alternatives. Hypothesis 3, therefore, received only partial support.
Summary

As proposed in hypothesis 3, Blacks and Whites were found to share highly
similar cultural value patterns. However, the cultural value patterns produced by these
groups were in some cases not consistent with theory. Taken together, these findings
replicate similar findings by Carter (1990).

C1 1111 HE: E 3.15]

Hypothesis 4 argues that racial subgroups will differentially endorse various
culttual value alternatives. In order to test this hypothesis, univariate statistics were
initially attempted. These findings are discussed below. Hypothesis 4 was then further
investigated using hierarchical regression in order to statistically control for variables

which might provide alternative explanations for the univariate results.

" HEIEE‘

 

[711310

ma
repon
to en:
orient

slat
Writ:
Time
these
Alibi

Xauu

95

Univariate Analyses

Racial subgroup differences in the preferences for particular cultural values were
initially examined using t-tests. The means, standard deviations, and t-values are
reported in Table 10. Consistent with hypothesis 4, Blacks were more likely than Whites
to endorse the cultural value alternatives of Subjugation to Nature from the Person-Nature
orientation, the Past alternative of the Time orientation, and the Lineal alternative of the
Relations orientation. Contrary to the hypothesis, however, Blacks also outscored
Whites on the Evil alternative of the Human Nature orientation, Future alternative of the
Time orientation, and Doing alternative of the Activity orientation. It was thought that
these three cultural value alternatives would be endorsed more strongly by Whites.
Although Blacks were also more likely to endorse the Mixed alternative of the Human
Nature orientation, no a priori expectations existed for this cultural value alternative.

Given the possible influence a large sample size may have on the determination of
statistical significance when using t-tests, hypothesis 4 was further investigated using
Cohen's (1977) definition of a small effect size to determine statistical significance.
Specifically, any mean difference between the 2 racial sub-groups on a particular cultural
value that exceeded .2 standard deviations was judged to be significant. Using this
criterion with subgroup N's equal to 100, a difl‘erence of .2 standard deviations or larger
would occur with a probability approximately equal to or less than .02. This procedure
identified the same 7 significant differences in endorsement of cultural values between
Blacks and Whites as previously identified by the t-tests. Thus, sample size did not
unduly influence the determination of subgroup differences.
Hierarchical Regression Analyses

Although Blacks endorsed certain cultural values more strongly than Whites,
there are several possible explanations for this finding other than cultural value
differences. First, it is possible that this finding merely reflects difl‘erences attributable to

96

Table 10:

Results of Heat Comparisons of Cultural Value Differences by Race

 

Black Black White White

M Mean SD Mean SD t-test

Human Nature

Evil 2.97 .60 2.78 .52 t(194) = -2.51, p<.05
Mixed 4.00 .35 3.71 .42 t(203) = -5.35, p<.01
Good 3.80 .48 3.78 .41 t(192) = -.31, p>.05
Person-Nature

Subjugation 3.42 .55 2.90 .60 t(205) = -6.60, p<.01
Harmony 3.67 .41 3.60 .42 t(204) = -1.11, p>.05
Mastery 2.90 .49 2.77 .52 t(205) = -1.77, p>.05
Time

Past 2.56 .57 2.41 .50 t(196) = -2.08, p<.05
Present 3.59 .48 3.51 .49 t(204) = -l .14, p>.05
Future 3.71 .49 3.50 .38 t(l85) = -3.48, p<.01
Activity

Being 3.64 .44 3.69 .43 t(203) = .79, p>.05
Being-in—Becoming 3.59 .46 3.54 .46 t(204) = -.71, p>.05
Doing 3.94 .42 3.75 .44 t(204) = -3.09, p<.01
Relations

Lineal 2.85 .48 2.56 .49 t(204) = -4.43, p<.01
Collateral 3.14 .47 3.06 .43 t(l99) = -l .32, p>.05
Individual 4.14 .49 4.11 .50 t(204) = -.33, p>.05

50:10
mg
cm?
used.
011 iii!

:fien

um

50ch

97

socioeconomic status between the two racial groups. Second, it is also possible that the
strength of an individual's sense of racial identity may further explain the adoption of
certain cultural values. To investigate these possibilities, hierarchical regression was
used. Specifically, a score on one of the 15 cultural value scales was initially regressed
on income and parental education. This step was intended to statistically control for the
effects of socioeconomic status. In the second step, the cultural value alternative was
regressed on dummy coded race. A significant change in R2 at this point indicates that
race explains additional variance in the cultural value alternative over that explained by
socioeconomic status. In a third step, strength of racial identity entered the equation. A
significant change in R2 at this step indicates that the degree of importance an individual
places on racial identity influences endorsement of cultural values beyond race and SES.
In the final step, the cultural value alternative was regressed on the interaction of race by
racial identity strength. A significant change in R2 at this final step indicates that the
influence of racial identity on endorsement of a cultural value is dependent upon the
individual's race. Table 11 presents the results of these regressions.

Close inspection of Table 1] reveals that when the cultural values are regressed
on race following the addition of socie-economic variables, race explains unique variance
for 6 of the cultural value alternatives. This finding replicates the above t-tests in which
racial differences were detected in the preference for the Evil, Mixed, Subjugation, Past,
Future, and Lineal cultural value alternatives. As discussed above, only the findings that
Blacks were more likely to support the Subjugation, Past, and Lineal alternatives supports
Hypothesis 4. However, unlike the previous analysis, the Doing alternative failed to
explain a significant portion of tmique variance, following the addition of socioeconomic
variables.

When entered into the regression equation following the addition of race and the

socioeconomic variables, the strength of an individual's racial identity added significantly

98

Table 11:
Results of Hierarchical Regression Analyses with Cultural Values
Regressed on SES, Race, and Racial Identity Strength

 

 

 

 

DV is Evil

mm: E SE B 32 A32
STEP 1:

Income .06 .03 .03 .03
Mother Educ. -.01 .05

Father Educ. -.08 .04

STEP 2:

Dummy-coded Race .22 .08 .06’ .03‘
STEP 3:

Racial Identity Str. .06 .07 .07“ .00
STEP 4:

Race’RIStr. Int. -.18 .14 .08“ .01
DV is Mixed

W 1: sm 32 A32
STEP 1:

Income -.05 .02 .04‘ .04‘
Mother Educ. .02 .03

Father Educ. -.03 .03

STEP 2:

Dummy-coded Race .26 .06 .13“ .09"
STEP 3:

Racial Identity Str. .16 .05 .18‘ .04‘
STEP 4:

Race‘RIStr. Int. .02 .09 A . 18" .00
DVflGmfl

man: a sea 112 A32
STEP 1:

Income -.04 .02 .03 .03
Mother Educ. .02 .04

Fadrer Educ. -.04 .03

STEP 2:

Dummy-coded Race -.04 .07 .03 .00
STEP 3:

Racial Identity Str. .05 .05 .04 .00
STEP 4:

Race‘RIStr. 1m. .12 .11 .04 .01

DH

STEE
inter
Mats
Fart:

99

 

 

 

 

 

 

 

 

 

Table l l (cont'd):

DV is Subjugation

3 JL a2 A32
STEP 1:
Income -.00 .03 .03 .03
Mother Educ. .01 .05
Father Educ. -.09 .05
STEP 2:
Dummy-coded Race .52 .09 .18“ .15‘
STEP 3:
Racial Identity Str. .19 .07 .21‘ .03“
STEP 4:
Race’RIStr. Int -.21 .14 .22‘ .01
DV is Harmony 2

a Jan 32 As
STEP 1:
Income -.04 .02 .02 .02
Mother Educ. .02 .03
Father Educ. -.01 .03
STEP 2:
Dummy-coded Race .03 .06 .02 .00
STEP 3:
Racial Identity Str. .03 .05 .02 .00
STEP 4:
Race‘RIStr. Int. .03 .10 .02 .00
DV is Mastery

J sea 112 A32

STEP 1:
Income .00 .03 .01 .01
Mother Educ. .01 .04
Father Educ. -.04 .04
STEP 2:
Dummy-coded Race .1 1 .08 .02 .01
STEP 3:
Racial Identity Str. .12 .06 .04 .02"
STEP 4:
Race‘RIStr. Int -.37 .12 .08‘ .04"

 

‘ indicates p<.03

100

Table 11 (cont'd):

 

 

 

 

 

 

DV is Past 2
W B JILL J: an
STEP 1:

Income .00 .03 .01 .01
Mother Educ. -.01 .04

Father Educ. -.04 .04

STEP 2:

Dummy-coded Race .14 .08 .02 .02‘
STEP 3:

Racial Identity Str. .09 .06 .04 .01
STEP 4:

Race‘RIStr. Int. -.01 .13 .04 .00
DV is Present 2
Mariam: L sea 112 Au
STEP 1:

Income .02 .03 .03 .03
Mother Educ. .08 .04

Father Educ. -.08 .04

STEP 2:

Dummy-coded Race .07 .07 .03 .00
STEP 3:

Racial Identity Str. .13 .06 .06“ .02‘
STEP 4:

Race’RIStr. Int. -.08 .12 .06 .00
DV is Future

mm: J sea 32 A32
STEP 1:

Income -.02 .02 .04 .04
Mother Educ. .05 .04

Father Educ. -.07 .03

STEP 2:

Dtunmy-coded Race .18 .06 .07“ .04‘
STEP 3:

Racial Identity Str. .01 .05 .07" .00
STEP 4:

Race‘RIStr. Int. -.03 .11 .07" .00

 

‘ indicates p<WS

101

 

 

 

 

 

 

 

 

Table 11 (cont'd):

DV is Being 2 2
Brink]: J Sui R AR
STEP 1:
Income -.03 .02 .03 .03
Mother Educ. .02 .03
Father Educ. -.03 .03
STEP 2:
Dummy-coded Race -.11 .06 .04 .01
STEP 3:
Racial Identity Str. .06 .05 .05 .01
STEP 4:
Race‘RIStr. Int. .13 .24 .06 .01
DV is Being-in-Becoming 2 2

SE3 8 AB
STEP 1:
Income -.00 .03 .01 .01
Mother Educ. .03 .04
Father Educ. -.05 .03
STEP 2:
Dummy-coded Race .02 .07 .01 .00
STEP 3:
Racial Identity Str. .10 .06 .03 .02
STEP 4:
Race'RIStr. Int. .17 .1 1 .04 .01
DV is Doing
2mm: m 32 A32
STEP 1:
Income -.04 .02 .07’ .07“
Mother Educ. .04 .03
Father Educ. -.08 .03
STEP 2:
Dtunmy-coded Race .12 .06 .09“ .02
STEP 3:
Racial Identity Str. .06 .05 .09‘ .01
STEP 4:
Race‘RIStr. Int. .01 .10 .09‘ .00

 

‘ indicates p<.05

 

 

STEP 3:
PM] Id

YEP 4-
PW.

102

 

 

 

 

 

 

Table l l (cont'd):
DV is Lineal 2 2
Elijah]: L SE B J AB
STEP 1:
Income .03 .03 .08" .08‘
Mother Educ. -.02 .04
Father Educ. —. 12 .04
STEP 2:
Dummy-coded Race .26 .07 .14“ .06‘
STEP 3:
Racial Identity Str. .08 .06 .15“ .01
STEP 4:
Race’RIStr. Int. -.31 .1 1 .18‘ .03“
DV is Collateral
unable B SE B 32 4532
STEP 1:
Income .03 .03 .02 .02
Mother Educ. .03 .04
Father Educ. -.07 .03
STEP 2:
Dummy-coded Race .08 .07 .03 .01
STEP 3:
Racial Identity Str. .05 .05 .03 .00
STEP 4:
Race'RIStr. Int. -. 12 .1 1 .04 .01
BY is Individual
m a jar; 32 A31
STEP 1:
Income .01 .03 .01 .01
Mother Educ. -.03 .04
Father Educ. —.03 .04
STEP 2:
Dummy-coded Race .00 .07 .01 .00
STEP 3:
Racial Identity Str. .15 .06 .04 .03‘
STEP 4:
Race‘RIStr. Int. .33 .12 .08‘ .04“

 

" indicates p<.05

 

to the prei
indixidua‘
account t
additional
adoption 1
identity st
for the cul
presented
F1,-
Spetific c1
mean res;
”PP-011mg
their endo
higher in 1
00111101 Na
high in rat
millions.
much mor
W113 cor
mates '1
flmfive
txpetted tc
WM, 1
time
he Wm 11;

103

to the prediction of the cultural values of Mixed, Subjugation, Mastery, Present, and
Individual. That is, even after taking the efl‘ects of socioeconomic status and race into
account, the degree to which race was important to the individual helped to provide
additional information concerning the adoption of these cultural values. Finally, the
adoption of three cultural values was best explained by examination of the race by racial
identity strength interaction. This interaction added a significant degree of predictability
for the cultural values of Mastery, Lineal, and Individual. These three interactions are
presented in Figures 2, 3, and 4, respectively.

Figures 2, 3, and 4 were constructed by plotting the average endorsement of a
specific cultural value for both racial groups one standard deviation above and below the
mean response on racial identity strength. Examination of Figure 2 yields evidence
supporting hypothesis 4. As racial identity strength increases, Blacks tend to decrease
their endorsement of the Mastery cultural value alternative. In contrast, Whites who are
higher in racial identity are more likely to endorse the notion that humans are meant to
control Nature. Figure 3 fails to provide support for the hypothesis that Blacks who are
high in racial identity are more likely to endorse an authoritative, Lineal view on human
relations. Although this figure suggests that Blacks who are iow in racial identity are
much more likely than a similar group of Whites to endorse this alternative, the two racial
groups converge in their endorsement of the Lineal alternative as perceived racial identity
increases in importance. The race by racial identity strength interaction for the Individual
alternative also fails to support the hypothesis (see Figure 4). Although Blacks were
expected to be less likely to endorse the Individual alternative as racial identity strength
increased, this group was actually more likely to endorse the Individual alternative as the
importance of racial identity increased. The endorsement of Individualism by Whites, on
the other hand, was uninfluenced by the degree of importance of racial identity strength.

«N GLn-Au‘rfl

104

If!

 

 

3d

and

 

tea: a 2. =2... 8:. a £9.85 5...: 33— a 5.39.2..—
a 35E ..

 

nun.“

 

tumu fibula“.-

'-

105

 

up.»

If:

 

.85.. a E as... 8:. e 59.25 3.83 is. 2 gases...
in 9...»...— ..

 

"V Quack-uh

qr“

If!

 

4 .6

 

 

1:23—3— 3 >6 5:3 8n”—

 

e 533% £2.93 33. a 5:85.:
s 955 a

 

sessé

107

While it does seem that Blacks and Whites differ on their preferential
endorsement of several cultural values, these preferences are complex and in some cases
inconsistent with theory. As hypothesized, Blacks are more likely to adopt the cultural
values of a Subjugation to Nature and the Past Time orientation. Also consistent with
hypothesis 4 is the finding that Blacks who are high in racial identity are less likely than a
similarly highly racially identified group of Whites to endorse the Mastery over nature
orientation. However, contrary to the hypothesis, Blacks were more likely than Whites to
endorse an Evil view of basic human nature and a future time orientation. Further, in
contrast to expectations, Blacks who felt their racial identity was an important part of
their lives were less likely than those with low racial identity to endorse a Lineal
perspective of human relations. Rather, Blacks high in racial identity strength were more
likely to endorse an Individual view of human relations. Table 12 provides a summary of
these findings. Hypothesis 4 clearly received only mixed support from the data.

Is it possible that the observed cultural value differences merely reflect differences
in socially desirable responding between racial subgroups? This question is raised by the
finding that Blacks were more likely than Whites to respond in a socially desirable
manner. In order to test the possibility that Blacks endorsed cultural values more
frequently as a result of the desire to respond in a socially acceptable manner, a final set
of hierarchical regressions were performed. In these regressions, a cultural value
alternative was initially regressed on SES, then the 2 social desirability scales, followed
by dummy coded race, then racial identity strength, and finally, the interaction of race
and racial identity strength. These regressions were performed using each of the 15
cultural value alternatives as dependent variables. The interpretation of these regressions
was identical to that discussed above. The desire to respond in a socially acceptable
manner did not influence the relationship between race and endorsement of cultural

values.

lliret
Good
Subju
Harm
Maste
Pat‘
Presen
Flltun

Being

Domg

Lineal

108

Table 12:
Summary of Racial Group
Preferences for Cultural Values

[:1 Illl G E n. II'CI 11:1

Evil Blacks
Mixed Blacks
Good No difference in endorsement
Subjugation" Blacks
Harmony No difference in endorsement
Mastery" Whites with high racial identity
Past" Blacks
Present No difference in endorsement
Future Blacks
Being No difference in endorsement
Being-in-Becoming No difference in endorsement
Doing No difference in endorsement,
after controlling for SES
Lineal Blacks low in racial identity
strength; convergence high in
racial identity str.
Collateral No difference in endorsement
Individual Whites low in racial identity strength;

Blacks very high in racial identity strength

 

* indicates the finding is consistent with theory-based expectations of racial subgroup
differences in cultural values

 

hmmi

Nah
saei
mdorsc
hunt
hunt

Maia
Bin:
immm
amen
hasm
diam
mwmn
wmmmx
“idles.

lithe

109

Summary

Convincing evidence was found that cultural value differences exist between
Blacks and Whites. Figure 5 plots the mean level of endorsement of the cultural value
scales for these racial subgroups. Although Blacks and Whites do differ in their relative
endorsement of the cultural value alternatives, equally persuasive evidence indicates that
the cultural value differences between these racial groups are not consistent with
theoretical expectations.

Using univariate tests, several cultural value differences were detected between
Black and White racial subgroups. In particular, Blacks were more likely to endorse the
Evil and Mixed alternatives of the Human Nature orientation, Subjugation to nature
alternative of the Person-Nature orientation, Past and Future alternatives of the Time
orientation, and the Doing alternative of the Activity orientation. Although several of
these subgroup differences were consistent with Hypothesis 4, the findings that Blacks
differentially preferred the Evil, Future and Doing alternatives were surprising. It is
important to note that each of the race differences in cultural values as detected by the
univariate procedures was due to greater endorsement on the part of Blacks compared to
Whites. In order to control for several possible alternative explanations for the findings,
the effects of SES and racial identity strength were examined with hierarchical regression.
After performing this analysis, it was still found that Blacks generally had a differential
preference for the cultural value alternatives of Evil, Mixed, Subjugation to nature, Past
and Future. This analysis also revealed that Blacks low in racial identity strength
differentially preferred the Lineal alternative of the Relations orientation more than
Whites. Finally, the Individual alternative of the Relations orientation was endorsed
more fi'equently by Whites low, and Blacks high, in racial identity strength. These
findings strongly suggest that although the racial subgroups examined in this study differ

in their endorsement of cultural values, these differences are not consistent with theory.

110

 

 

megs; 1.1
9.05m IOI

 

 

13091419"!

baa-urbane

1919121100
‘* uomfinlqns

. 50198

‘ ”mm-I

‘ 10091:!

" 188:!
MIN

‘ AuouueH

laeun
" 611100

’ 13009

" pexm
1103

 

 

qt-
GP

 

\( ....\/>l\

 

 

cone 3 eo=_a> .8330 he ueeEeEOeem
"m 952“. .

: m6

; m...

.- mN

m6

-- rev

 

dliifl
items

' 1
lillili
6

mil
d1'St‘lbl

”his:

111

The ultimate goal of the present study was to determine whether cultural value
differences between racial subgroups might influence responses to biodata employment
items. The next section considers whether culttual value differences between individuals

might lead to disparate selection of biodata response options.

CulmrallaluesfliodataResmnses

Do cultural values influence responses to biodata employment items? In order to
determine whether responses to commonly used biodata items were consistent with the
cultural values of the respondent (e.g., hypothesis 5), loglinear analyses were conducted.
Specifically, the polychotomous loglinear approach proposed by Green, Crone, & Folk
(1989) was used to determine whether cultural values influenced responding to any of the
107 biodata items used in the ctu'rent study. Initially, each cultural value alternative was
trichotomized to represent low, medium, and high levels of endorsement of that cultm'al
value alternative. Cultural value alternatives were trichotomized by dividing the
distribution of all endorsements into thirds. Those individuals in the lowest one-third of
the distribution of endorsement responses to a value orientation alternative were assigned
to the low group. Those in the middle 1/3 were assigned to the average endorsement
group for that particular value orientation. Finally, those scoring in the upper third of the
distribution of responses for a particular value orientation alternative were considered to
represent a high level of endorsement.

Recall that each biodata item was originally selected for inclusion in this study if
one or more option was judged by independent raters to reflect a cultural value
alternative. In order to test hypothesis 5, loglinear analysis was conducted by forming a
3-way contingency table as follows. For each biodata item, the frequency of each option
choice (e.g., a,b,c, or d) was computed for each racial group (i.e., Black or White) at each

112

level of the specified cultural value alternative (i.e., low, medium, or high). The loglinear
analysis was composed of three main effects (racial group, cultural value alternative, and
option choice), three 2-way interactions (racial group*cultural value, racial group*option
choice, and cultural value*option choice), and one 3-way interaction (racial
group*cultural value*option choice). The influence of cultural values on option choice
was examined by comparing two models. The first model was composed of all main
effects and all 2-way interactions, and the second model was composed of all main effects
and all 2-way interactions except the cultural value‘option choice interaction. If the
difference in chi-squares between the two models was significant, this provided evidence
that the level of endorsement of the cultural value alternative influenced option response
choice. Table 13 presents a complete listing of the biodata items which were found to be
influenced by cultural values. Specifically, this table presents the biodata item number,
the difference in chi-squares and df between the two loglinear models, the significance
level of this difference, and the culttu'al value alternative examined. A list of each of
these biodata items is found in Appendix B. Using a criterion of alpha equal to .05,
responses to twenty-two of the 107 biodata items included in this study were influenced
by endorsement of cultural values. Thus, only 21% of the biodata responses could be
attributed to cultmal value differences between respondents. This result is surprising,
considering that the biodata items were chosen for inclusion in the present study only if 2
out of 3 independent raters agreed that one or more response options for a particular
biodata item reflected one of the cultural value alternatives suggested by Kluckhohn and
Strodtbeck (1961).

Although loglinear analyses demonstrated that cultural values did influence option
selection for some of the biodata items included in the study, mere examination of the
loglinear models fails to provide information regarding whether the raters' a priori

classification of a particular option as reflecting a cultural value alternative was in fact the

113

Table 13:

Results of the test for significant difference in Chi-squares for
Loglinear Analyses examining Cultural value*0ption Choice Interaction

 

Item # A x2 A (If Cultural value
Alternative

2 13.38“ 6 Good

3 13 .47" 6 Evil

6 20.18" 6 Doing

1 1 1940* 8 Evil

15 23 .49" 8 Future

21 17.00" 4 Future

25 17.80"“ 8 Past

26 12.87“ 6 Doing

28 28.91 ** 4 Lineal

33 19.34" 6 Doing

47 51.37" 8 Lineal

59 19.39" 8 Doing

63 13.34“ 6 Doing

64 27.73 " 10 Being-in-Becoming

68 23.50" 7 Doing

71 17.06“ 8 Future

73 29.51 " 10 Doing

75 14.15" 4 Lineal

77 51.1 I" 10 Evil

87 33.86" 4 Present

96 16.46” 2 Future

1 10 15.44’ 6 Collateral

 

* indicates significant difi'erence between log-linear models, p < .05
“ indicates significant difference between log-linear models, p < .01

__ ____ —-.-...r.-—--—-

opioz
their
is '(
tidal

cider

Nair:
inner

810de

114

option differentially preferred by respondents who strongly endorsed the cultural value
alternative. As an example, if the biodata option 3 was a priori selected as most reflecting
the "Good" alternative of the Human Nature orientation, did individuals who strongly
endorsed the Good alternative prefer option 3 more than individuals who less strongly
endorsed the Good alternative? In order to determine whether cultural values influenced
the selection of biodata options as expected, the cultural value*option interaction for each
of the 22 flagged biodata items was graphed. Appendix D presents these graphs. These
graphs were formed by graphing the proportion of respondents from each level of
endorsement of the cultural value (i.e., low, medium, of high) that selected each biodata
option for a particular item. A brief description of each of these interactions follows:
Biodata item 2:

As expected, those who strongly endorsed the Good alternative of the Human
Nature orientation were more likely to report that their experience with people has
demonstrated that there exists a lot of good in all people.

Biodata item 3:

Consistent with expectations, individuals who strongly endorsed the Evil
alternative of the Human Nature orientation were more likely than those individuals who
less strongly endorsed the Evil alternative to indicate that exams should be closely
proctored because few students are completely honest.

Biodata item 6:

As hypothesized, those who strongly endorsed the Doing alternative of the
Activity orientation were more likely to indicate that they plan their weekends either in
general terms or in detail, in comparison to those who less strongly endorsed the Doing

alternative.

115

Biodata item 11:

Consistent with the hypothesis, participants who strongly endorsed the Evil
alternative of the Human Nature orientation were more likely to strongly agree that they
tend to be cynical and skeptical of other's intentions.

Biodata item 15:

As expected, respondents who strongly endorsed the Future alternative of the
Time orientation were more likely to keep a schedule written down weeks in advance
than were those who less strongly endorsed the Future alternative.

Biodata item 21 :

In support of the hypothesis, individuals who strongly endorsed the Future
alternative of the Time orientation were more likely to do a lot more planning before
starting each day than those who did not strongly endorse the Future alternative.

Biodata item 25:

Contrary to expectation, participants who strongly endorsed the Past alternative of
the Time orientation were less likely to report feeling thoroughly interested by museum
visits than those who less strongly endorsed this alternative.

Biodata item 26:

Consistent with expectations, those who strongly endorsed the Doing alternative
of the Activity orientation were more likely to report more home of their day is spent in
constructive work than those who less strongly endorsed the Doing alternative.

Biodata item 28:

As expected, in contrast to those who weakly endorse the Lineal alternative of the

Relations orientation, those who strongly endorse the Lineal alternative are more likely to

assert that people should have greater respect for authority.

 

116

Biodata item 33:
Consistent with the hypothesis, respondents who strongly endorsed the Doing
alternative of the Activity orientation were more likely than others to report that hard

work makes them feel good.

Biodata item 4 7:

As expected, individuals who strongly endorsed the Lineal alternative of the
Relations orientation were more likely than those who weakly endorsed the Lineal
alternative to strongly agree that we should look to our religious authorities for decisions
on moral issues.

Biodata item 5 9:

Consistent with expectations, participants who strongly endorsed the Doing
alternative of the Activity orientation were more likely than those who weakly endorsed
this alternative to be bothered most by people who lack initiative.

Biodata item 63:

As expected, respondents who strongly endorsed the Doing alternative of the
Activity orientation were more likely than those who weakly endorsed this alternative to
be considerably or highly disturbed when they leave something unfinished.

Biodata item 64:

Consistent with expectations, individuals who strongly endorsed the Being-in-
Becoming alternative of the Activity orientation were more likely to very often choose
classes, projects, or assignments simply to learn something new than were individuals
who less strongly endorsed the Being-in-Becoming alternative.

Biodata item 68:

117

As expected, participants who strongly endorsed the Doing alternative of the
Activity orientation were much more likely to strongly agree that they work hard to
accomplish their goals than were individuals who less strongly endorsed this alternative.
Biodata item 71:

Contrary to expectation, individuals who strongly endorsed the Future alternative
of the Time orientation were no more likely than others to recognize that they do not have
a required piece of equipment or part prior to actually needing it.

Biodata item 73 :

As expected, those who strongly endorsed the Doing alternative of the Activity
orientation were more likely than those who less strongly endorsed the Doing alternative
to assert that they are a productive person who always gets the job done.

Biodata item 75 :

Consistent with the hypothesis, individuals who strongly endorsed the Lineal
alternative of the Relations orientation were more likely than those who weakly endorsed
this alternative to respect people in positions of authority.

Biodata item 77:

Consistent with expectations, participants who strongly endorsed the Evil
alternative of the Human Nature alternative were more likely than those who less strongly
endorsed this alternative to strongly agree that people will take advantage of you if you
let them.

Biodata item 87:

As expected, those who strongly endorsed the Present alternative of the Time

orientation were more likely to report they believe in living for the moment than were

individuals who weakly endorsed the Present alternative.

Brod

alter
pin

8:00

into

being

all:

Clpti

Value
the 0
their
Prise]

Orient

118

Biodata item 96:

Consistent with expectation, participants who strongly endorsed the Future
alternative of the Time orientation were more likely than others to use a date book or
planner.

Biodata item I 1 0:

The interpretation of this graph supports the notion that those who strongly

endorse the Collateral alternative of the Relations orientation are less likely to prefer

being along than are those who weakly endorse this alternative.

Twenty of the twenty-two biodata items that demonstrated a significant cultural
value‘option choice interaction received empirical verification consistent with a priori
expectations. That is, raters' a priori classification of a particular option for a biodata
item was found to be preferred by individuals who strongly endorsed the targeted cultural
value in 20 of the 22 flagged biodata items. Of these twenty items, 1 item was related to
the Good alternative, and 3 items were related to the Evil alternative of the Human Nature
orientation. One biodata item which was a priori classified as reflecting both the Past and
Present alternatives, and 3 items classified as reflecting the Future alternative of the Time
orientation received empirical verification. The choice of response option was influenced
as expected by the Being-in-Becoming alternative for 1 biodata item, and by the Doing
alternative of the Activity orientation for 7 biodata items. Raters' classification of biodata
items as reflecting the Relations orientation alternatives of Collateral and Lineal received
empirical verification 1 and 3 times respectively.

Table 14 collapses the information regarding the degree to which each of the 15
cultural value alternatives influenced biodata response option selection into the 5 cultural
value orientations. Although only 5 biodata items were a priori judged to reflect 1 of the

3 Human Nature alternatives, responses to 4 of these items were formd to be influenced

119

Table 14:
Degree to which Cultural Value Orientations Influenced
Biodata Response Option Selection, as Predicted

 

 

Proportion of Percentage of

Items Items

Receiving Receiving
Cultural Value Orientation Empirical Empirical
Washed SW SW
Human Nature 05 80%
Good 1/1 100%
Mixed 0/0 -
Evil 3/4 75%
Person-Nature 0/10 0%
Subjugation 0/3 0%
Harmony 0/1 0%
Mastery 0/6 0%
Time 45/15 40%
Past 1/1 100%
Present 1/2 50%
Future 4/ 12 33%
Activity 8/41 20%
Being 0/8 0%
Being-in-Becoming 1/10 10%
Doing 7/23 30%
Relations 4/36 11%
Lineal 3/7 43%
Collateral 1/18 6%
Individual 0/11 0%

Total 20/107 19%

120

by the individual's endorsement of this value orientation. In other words, one's view of
basic human nature is likely to influence responses to biodata items which include
response options which refer to the Evil or Good intentions of individuals. However, the
fact that only 5 biodata items were a priori classified as reflecting a Human Nature
orientation indicates that few biodata items in the several biodata test banks used for this
study actually reflect this orientation. None of the 10 items classified a priori as
reflecting one of the Person-Nature orientation alternatives were found to be influenced
by the degree to which the individual endorsed this cultural value orientation. For the
biodata items a priori classified as reflecting an alternative of the Time orientation,
responses to over one-third (6/15) of these items were found to be influenced by the
individual's endorsement of this cultural value orientation as expected. The vast majority
(72%) of the biodata items included in the present study were thought to reflect either the
Activity (11 = 41) or Relations (11 = 46) cultural value orientations. Biodata items which
were a priori classified as reflecting either an Activity or Relations orientation were
influenced by endorsement of these cultural values for 8 and 4 of the items respectively.
(Alternatively, 20 and 11 percent of the time, respectively.) Overall, of the 107 biodata
items that raters agreed reflected cultural values, responses to 20 biodata items were
actually shown to be influenced by cultural values as expected. Consistent with
hypothesis 5, these findings suggest that cultural values do have some impact on
responses to biodata items. However, some cultural values seem more likely than others
to influence biodata responses.
Summary

Are biodata responses contingent upon the cultural values of the respondent? The
analyses discussed above strongly suggest that the individual's cultural values do
influence responses to biodata items. Responses to biodata items which assess the
individual's perception of human nature and time orientations are especially likely to be

121

influenced by the cultural values of the individual. Fortunately, perhaps, these biodata
items were relatively rare in the several biodata databanks used in this study. Biodata
items which were classified as reflecting the culttual value orientations of Activity and
Relations were much more prevalent in the several databases examined, but responses to
these items were less influenced by the cultural values of the respondent. Even so, 20 of
the 107 biodata items examined in this study were found to be influenced by the cultural
values of the individual as predicted. This is far greater than the approximately 5 items
which would be expected by chance alone.

3.15] CI]!!! ”3.! B

Do cultural value differences between racial subgroups influence biodata option
selection? Hypothesis 6 could be tested by comparing the fit of the saturated model
discussed above composed of the cultural value‘racial subgroup‘option interaction, with
a second loglinear model which is composed of all two-way interactions, but excludes the
three-way interaction. If the fit of the second model was significantly worse than the
model containing the 3-way interaction, we would have evidence that culttu'al value
differences between the racial subgroups influenced option choice for that particular
biodata item. The problem with such an analysis is that ignores the influence of
participants' overall test score. Thus, any findings regarding response option selection
difi'erences across racial subgroups may merely be reflecting differences in overall ability
on the biodata test. In order to control for possible performance differences across racial
subgroups, it would be necessary to include a fourth main effect - performance level.
This could be a trichotomized variable indicating low, medium, and high levels of ability
on the overall biodata test. The loglinear analyses could then proceed by comparing a
model composed of all 3-way interactions (performance level‘racial subgroup‘option

122

choice; performance level*cultural value‘option choice; performance level‘racial
subgroup‘cultural value; and, racial subgroup‘cultural value‘option choice) with a
second model composed of all three-way interactions except the racial subgroup*cultural
value‘option choice interaction. A significant difference between the fit of the two
models would support the notion that cultural value differences across racial subgroups
influenced option choice for that particular biodata item. Unfortunately, such an analysis
would require a very large sample size in order to ensure that each cell in the contingency
table possessed a sufficient number of observations. The sample size used in the present
study precludes such an analysis. Therefore, in order to examine whether cultural value
difi‘erences across racial subgroups influenced option choice, differences in preferences
for particular biodata response options were initially examined across racial subgroups,
after controlling for overall biodata test score. Following this analysis, items exhibiting
differential item functioning were interpreted in light of the a priori expectations
regarding cultural values differences. These analyses were conducted in order to
determine whehter the cultural value correlates of biodata responses were identical to the
race correlates of biodata responses. If responses to identical biodata items were
influenced by both the adoption of cultural values and race, we could conclude that
cultural value difiemnces between racial subgroups is likely to produce DIF for biodata

employment items.

B'lSl IE! 13

Loglinear analyses were conducted to examine the possibility of differential
responding to the biodata items by Blacks and Whites. The approach suggested by
Green, Crone, and Folk (1989) was again used to examine differential responding across
each option composing a particular biodata item. In order to determine whether racial

123

subgroup influenced option choice, the loglinear analyses conducted in this section were
composed of 3 main effects: biodata Option choice, racial subgroup, and performance
group. Performance groupings were created by trichotomizing the overall score achieved
on the biodata test into low, medium, and high performing groups. Recall that overall
test score was formed by initially providing a rational key (based on the job of police
officer) for each item. Overall test score was produced by summing individual item
scores across the 107 biodata items and 7 non-random response items. The effect of
including the seven non-random response items was in essence equal to the addition of a
constant to the overall biodata test score, since cases with more than 1 non-random
response were deleted fiom all analyses.

Loglinear analysis was conducted by forming a 3-way contingency table as
follows. For each biodata item, the frequency of each option choice (e.g., a,b,c, or d) was
computed for each racial group (e.g., Black or White) at each performance level (e.g.,
low, medium, or high). The loglinear analysis was composed of three main effects (racial
group, performance level, and option choice), three 2-way interactions (racial
group*performance level, racial group‘option choice, and performance level‘option
choice), and one 3-way interaction (racial group‘performance Icvel‘option choice). The
influence of racial subgroup on option choice was examined by comparing the model
composed of all main effects and all 2-way interactions, with a second model composed
of all main effects and all 2-way interactions except the racial subgroup‘option choice
interaction. If the difference in chi-squares between the two models was significant, this
provided evidence that preferences for biodata response options were dependent upon
one's racial subgroup.

Table 15 presents a complete listing of the biodata items which demonstrated
differential responding across racial subgroups. Specifically, this table presents the

biodata item number, the difference in chi-squares and df between the two loglinear

124

Table 15:
Results of the test for significant difference in Chi-squares for
Loglinear Analyses examining Racial Subgroup*Option Choice Interaction

 

MILL—All Adi Cnltttnmmeoritntttinn
2 10.1 8" 3 Human Nature

3 11.89" 3 Human Nature +
6 18.83" 3 Activity +

8 26.22" 4 Relations

10 31.16" 4 Human Nature
11 9.76“ 4 Human Nature
12 14.03” 4 Person-Nature +
18 14.56" 2 Time

20 10.74" 1 Relations

23 14.88" 2 Time

33 10.82" 2 Activity +

40 9.76“ 4 Activity +

42 10.53“ 4 Activity +

44 15.74“ 4 Activity +

47 40.83" 4 Relations +

53 8.79“ 3 Activity

57 9.52" 2 Activity

58 1 1.67* 5 Activity

69 11.42“ 4 Relations +

72 12.07" 4 Activity

74 18.49" 3 Relations +

77 36.61" 4 Human Nature
83 15.56" 4 Relations +

87 14.25" 2 Time

89 33.41 ** 4 Relations +

94 15.41 * 7 Relations +

95 6.77“ 2 Relations +

99 14.77“ 4 Relations

100 9.45“ 3 Relations +

103 20.23" 3 Activity

104 9.23“ 3 Activity +

106 6.16“ 2 Person-Nature +
110 8.31“ 3 Relations

112 6.22“ 1 Relations +

1 14 8.96" 3 Relations

 

* indicates significant difference between loglinear models, p<.05
” indicates significant difierence between loglinear models, p<.01
+ indicates flagged item is consistent with theory

125

models, and the significance level of this difference. Additionally, this table provides the
cultural value orientation the item was a priori expected to reflect. Appendix C presents a
complete list of each of these biodata items. Overall, 35 out of 107, or 32.7% of the
biodata items exhibited differential responding across racial subgroups. Blacks and
Whites apparently were attracted to different response options for a surprisingly large
percentage of the biodata items.

mu .h .0.) .‘0 tr: . ' Ii '1.‘ 5. = M“? (“l-1011'

'.

In order to determine which options of a particular item were differentially
preferred by one of the racial groups, each of the 35 significant racial subgroup‘option
choice interactions were graphed. These graphs were formed by graphing the proportion
of respondents from each racial group that selected each biodata option for a particular
item. These 35 graphs are presented in Appendix E. A brief description of the
interpretation of each graph follows. In addition to the interpretation of the racial
subgroup*option choice interaction, brief comment is provided as to whether the finding
was consistent with a priori expectations regarding cultural value differences.

Biodata item 2:

Whites have a slight preference for options which reflect the concept that, in their
own experience, most people have a lot of good, or some good. Blacks are more likely
thanWhites to prefer the option indicating that it has been their experience that people are
about as good as they have to be. This finding is contradictory to the a priori theoretical
expectation that Blacks would be more likely to endorse options reflecting the idea that
human nature is basically " good". Interestingly, this finding is consistent with Carter’s
(1990) empirical finding.

126

Biodata item 3:

Whites differentially prefer the option reflecting the sentiment that exams should
be closely proctored in order to carefully watch the few "bad apples". Blacks are more
likely than Whites to endorse the honor system for test examination. This finding is
consistent with the a priori hypothesis that Whites would be more likely to endorse
options reflecting the idea that human nature is basically "evil".

Biodata item 6:

Whites are more likely than Blacks to endorse the option indicating that weekends
are planned in general terms. Blacks are more likely than Whites to plan few or no
weekend activities, and simply do what feels right at the time. This finding is consistent
with a priori expectations that Blacks would view activity as more spontaneous, and
Whites would be more "doing" oriented.

Biodata item 8:

Although both groups most frequently stated that the most important thing to
them is family and fiiends, Blacks preferred the "religion" option more than Whites. This
is consistent with the expectation that Blacks would be more likely to endorse societal
authority (i.e., the Lineal alternative of the Relations orientation).

Biodata item 10:

Whites preferred options 2 & 4 as the ways they would discipline a child. These
options reflected sentiments referring to encouragement of good behaviors, and
attempting to reason with a child. Blacks were more likely than Whites to prefer options
5 & 1. Option 5 reflects physical punishment, while option 1 refers to withdrawal of
pleasure. This finding contrasts with the expectation that Whites would be more likely
than Blacks to believe that basic human nature is "evil", and therefore requires

ptmishment

127

Biodata item 1 1:

When asked if they tend to be cynical and skeptical of other's intentions, Blacks
expressed greater preference than Whites for the "strongly agree" option. Whites, more
strongly than Blacks, preferred the disagree option to this item. This finding contrasts
with the expectation that Whites would be more likely than Blacks to believe that the
basic human nature is "evil".

Biodata item 12:

Blacks differentially preferred to try to reschedule an important meeting if faced
with treacherous road conditions, whereas Whites preferred to leave early for the
appointment, and use extra driving precautions. This finding was consistent with
expectations for the Mastery alternative of the Person-Nature orientation, that Whites
would be more likely than Blacks to feels they could control nature.

Biodata item 18:

Blacks diflermfially prefer to plan every detail while on vacation, while Whites
differentially preferred to make only general plans. This finding was inconsistent with
the expectation that Whites would be more detail and planning oriented than Blacks (i.e.,
the Future alternative of the Time orientation).

Biodata item 20:

Whites, difiemnfiafly more than Blacks, preferred to be in a dance-band than a
soloist in an opera. In contrast to expectations regarding the Collateral alternative of the
Relations orientation, Blacks were more likely to express the desire to be singled out,
rather than a part of a group.

Biodata item 23:

Concerning present activities, Blacks were more likely than Whites to endorse the
option indicating they make rather precise and detailed plans. Whites differentially
preferred broad, general plans. This finding was inconsistent with the expectation that

128

Whites would be more planning oriented than Blacks (i.e., the Future alternative of the
Time orientation).
Biodata item 33:

Whites differentially preferred the option indicating that hard work makes them
feel good. Blacks differentially preferred the option that they don't mind hard work now
and then. Consistent with expectation that Whites would be more Doing oriented, Whites
were more likely to report that hard work makes them feel good.

Biodata item 40:

Whites were more likely than Blacks to indicate that they sometimes feel upset at
the end of the workday because they had not accomplished enough. Blacks were more
likely than Whites to say they never feel upset because they had not accomplished
enough. Consistent with the a priori expectation that Whites would be more Doing
oriented, Whites were more likely to report being upset at the end of the day for not
accomplishing enough.

Biodata item 42:

Whites are more likely than Blacks to admit "sometimes" making up work to
keep from being bored. Blacks are more likely than Whites to indicate they "seldom" do
this practice. Once again, this finding is consistent with the expectation that Whites
would be more accomplishment, or Doing, oriented.

Biodata item 44 :

Whites are more likely than Blacks to indicate "sometimes" feeling guilty when
they take time off to relax. Blacks are more likely than Whites to prefer the option
indicating they never feel guilty when taking time off to relax. This finding is also
consistent with the a priori expectation that Whites would be more Doing oriented.

129

Biodata item 4 7:

Blacks are much more likely than Whites to strongly agree or agree that we
Should look to our religious authorities for decisions on moral issues. Consistent with the
expectation that Blacks would be Lineally oriented on the Relations orientation, religious
authorities are seen as more acceptable sources of guidance by Blacks than Whites.
Biodata item 53:

Whites were more likely than Blacks to keep on reading when they encounter a
word they are unfamiliar with, while Blacks diflerenfially preferred to immediately
consult a dictionary. No a priori expectations regarding racial subgroup differences were
made for this item, which was classified as reflecting the Being-in-Becoming alternative
of the Activity orientation.

Biodata item 5 7:

Whites differentially preferred lectures, while Blacks differentially enjoyed
discussion classes in college. No a priori expectations regarding racial subgroup
differences were made for this item, which was classified as reflecting the Being-in-
Becoming alternative of the Activity orientation.

Biodata item 58: e

The interpretation of this item is that Blacks differentially prefer to stop only at
planned points along a vacation trip in a car. This finding did not support the expectation
that Whites would be more likely to be accomplishment oriented.

Biodata item 69:

Whites were more likely than Blacks to have no involvement in civic
organizations, while Blacks were more involved, up to 3 memberships. This finding
supports the a priori expectation that Blacks would be more community oriented than

Whites (i.e., Collateral alternative of the Relations orientation).

130

Biodata item 72:

Relative to Blacks, Whites expressed greater interest in speculating on the nature
of the universe or the human condition. No a priori expectations regarding racial
subgroup differences were made for this item, which was classified as reflecting the
Being-in-Becoming alternative of the Activity orientation.

Biodata item 74:

Blacks were less likely to report being disrespectful to high school teachers than
Whites. Consistent with the expectation that Blacks would be more likely than Whites to
accept authority (i.e., the Lineal alternative of the Relations orientation), Blacks reported
less disrespect for their high school teachers than Whites. '

Biodata item 77:

Blacks were much more likely to strongly agree that most people will take
advantage of you if you allow them to. Whites differentially preferred the option
indicating they were neither in agreement nor disagreement with this proposition. In
contrast to expectations, Blacks were more likely than Whites to indicate that people are
basically "evil".

Biodata item 83:

Whites were more likely than Blacks to indicate that they sometimes put in a lot
of extra efl‘ort to buy independence, while Blacks were more likely to indicate that they
have never worked harder to buy independence. Consistent with expectations regarding
the Individual alternative of the Relations orientation, Whites reported greater willingness
to sacrifice in order to gain independence than Blacks.

Biodata item 87:

Whites differentially preferred living for the moment, while Blacks were more

likely than Whites to believe in working for the future. In contrast to expectations

131

regarding the Future alternative of the Time orientation, Whites were not more likely than
Blacks to plan for the future.
Biodata item 89:

Blacks were more likely than Whites to report that community/voltmteer work is
important to them. Consistent with a priori predictions concerning the Collateral
alternative, Blacks reported more community involvement than Whites.

Biodata item 94:

Blacks were more likely than Whites to report they often have observed people
put friendships and personal goals aside to try to achieve a greater good for the larger
group. Whites differentially preferred the "sometimes" option. Consistent with
expectations regarding the Collateral alternative, Blacks were more attracted to the option
indicating instances of self-sacrifice for the good of the group.

Biodata item 95:

Whites differentially preferred the option which indicates they have seldom hurt a
friend to benefit themselves, while Blacks differentially preferred the option indicating
they have never done this. As expected, Whites were more likely than Blacks to choose
the response option consistent with the Individual alternative of the Relations orientation.
Biodata item 99:

Blacks were differentially more likely to report they have "often" been accused by
others of doing something solely because it Was in their best interest. Whites were more
likely than Blacks to report they have ”seldom" been so accused. Contrary to
expectations for the Collateral alternative of the Relations orientation, Blacks were more
likely to report having been accused of doing something in their own self-interest.
Biodata item [00:

Consistent with expectations regarding the Collateral alternative, Blacks reported

more involvement in community groups than Whites.

132

Biodata item 103:

As a child when something was broke in an effort to see "what makes it tick",
Blacks report (more often than Whites) that their parents became angry and punished
them. Whites are more likely than Blacks to report that their parents tried to explain that
it was wrong, or tried to help them find the answers they were seeking. No a priori
expectations regarding racial subgroup differences were made for this item, which was
classified as reflecting the Being-in-Becoming alternative of the Activity orientation.
Biodata item 104:

Whites are more likely than Blacks to be bothered somewhat by an unfinished
chore. This finding is consistent with expectations regarding the Doing alternative, that
Whites would be more accomplishment oriented than Blacks.

Biodata item 106:

Whites were more sure than Blacks that what happens in the long run is mostly
under their control. Consistent with expectations regarding the Mastery over Nature
alternative, Whites reported feeling in greater control of their destiny than Blacks.

Biodata item I I 0:

In contrast to expectations regarding the Collateral alternative, Blacks preferred
being alone more than Whites.
Biodata item 1 12:

Blacks are more likely than Whites to have won a commendation for civic
involvement since high school. Consistent with expectations regarding the Collateral
alternative, Blacks were more likely to receive public recognition for community

involvement.

133

Biodata item 1 14:

If given an important project at work that needed more attention, Whites indicate
they would minimally attend to other projects due to the realization that all work time
could not be devoted to the more important project. Blacks are more like to respond that
they would divide their time equally among projects. In contrast to expectations
regarding the Collateral alternative, Blacks did not prefer the option of requesting help

from co-workers when given an important project.

Referring once more to Table 15, the last column represents the cultural value
orientation to which each of the above 35 items was classified a priori. Notice that a star
(*) is located next to the cultural value orientation of some items. A star has been placed
in the row only if (1) the biodata item was included in the present study because it was
classified as reflecting a particular cultural value orientation, (2) the item actually did
exhibit differential responding by racial subgroup, and, (3) the interpretation of this
interaction was consistent with a priori expectations, as portrayed in Table 1. For
example, the row for biodata item 3 in Table 15 is starred because the interpretation of
the graph of the racial subgroup‘option choice interaction is consistent with the a priori
expectation that Whites, more than Blacks, would view basic human nature as "evil".
Examination of Table 15 reveals that the a priori expectations regarding the influence of
cultural values were empirically supported in 18 of the 35 items which exhibited
differential responding by race. Of the total set of biodata items, 18/107, or 17%, were
accurately predicted to exhibit differential item fimctioning across racial subgroups due to
expected cultural value differences. Thus, the proposition that cultural values could be a
priori identified which would lead to differential responding across racial groups received
partial supportw

134

Table 16 reports the number of biodata items a priori classified as reflecting each
cultural value orientation, the number and percentage of these items exhibiting
differential item responding across racial subgroups, and the number and percentage of
items which demonstrated DIF consistent with theoretical expectations. Five biodata
items were included in this study which were thought to reflect the cultural value of
Human Nature. Each of these five items exhibited differential responding across racial
subgroups. However, only 1 of these 5 items were interpreted as originally expected.
Contrary to expectations, Blacks did not view human nature as more " good" than Whites.

Indeed, for those biodata items thought to reflect the cultural value of human nature,
Blacks were generally more likely than Whites to endorse a negative, cynical view of
human nature.

Of the eleven biodata items included in the study which were classified as
reflecting a Person-Nature orientation, only two items (20%) were flagged for differential
responding across racial groups. Both items supported the expectation that Whites would
adopt a greater sense of mastery over nature than Blacks.

Fifteen biodata items were originally classified as reflecting a Time orientation.
Of these items, 3 exhibited differential responding across racial subgroups. The
interpretation of the racial subgroup‘option interaction for each of these items was
inconsistent with a priori expectations, however. Contrary to expectations, Blacks, not
Whites, tended to choose biodata options that depicted greater concern with future events.

F orty-one biodata items were originally classified as reflecting an Activity
orientation. Of these 41, eleven were found to exhibit differential responding across
racial subgroups. Four of these eleven items were a priori classified as assessing the
”being-in—becoming" alternative of the Activity orientation (biodata items 53, 57, 72, and
103). No a priori expectations were made regarding which racial subgroup, if any, would

135

Table 16:

Number of Biodata Items Exhibiting DIF by

Racial Subgroup Consistent with Theoretical Expectations

Cultural Value

.I‘I .791:

 

 

Human Nature
Person - Nature
Time

Activity

Relations

Total

# of DIF % of DIF
# of Items Items # %
Biodata by Racial by Racial Consistent Consistent
‘u‘ 0'4—‘I—o lqult ”I I'vu‘ .1 I“.

5 5 100% 1 20%

10 2 20% 2 20%

1 5 3 20% O 0%

41 11 27% 6 15%

36 14 39% 9 25%

107 35 33% 18 17%

136

prefer this alternative. Thus, only seven of the Activity items which exhibited differential
responding across race were a priori classified as expected to differ across race. The
interpretation of 6 of these 7 items was consistent with theoretical expectations. The
differential response patterns indicate that Whites do seem to have a greater "doing"
orientation than Blacks - they prefer biodata options referring to liking hard work, being
bothered by an unfinished chore, and feeling guilty when not working.

Finally, 36 of the biodata items were originally classified as reflecting the
Relations orientation. Of these 36 items, 39% (i.e., 14 items) exhibited differential
responding across racial subgroups. Nine of the 14 items were consistent with theoretical
expectations. Responses to the biodata items indicate that Blacks are more likely to
endorse options reflecting respect for elders and religion - consistent with the Lineal
alternative of the Relations orientation. Also consistent with expectations, Blacks report
more civic involvement than Whites, indicating a possible communal affiliation.
However, Blacks were also more likely to desire to be alone, and were more likely than
Whites to have been accused of behaving in their own self-interest.

Summary

Are racial subgroup differences in responding to biodata employment items
consistent with expected cultural value differences between these groups? The findings
of the present study strongly suggest that biodata items assessing Human Nature and
Relations orientations are likely to lead to difi'erential responding across racial subgroups.

Subgroup diflemnces in item responding were found for 100% of the biodata items a
priori classified as reflecting one’s view of basic human nature, and 39% of the items
classifed a priori as reflecting one’s view of interpersonal relations. A sizeable
percentage of items reflecting the other cultural value orientations also exhibited
difiemnfial item functioning. In particular, 20% of the biodata items a priori classified as

137

reflecting the Activity orientation were found to lead to differential responding across
racial subgroups.

However, the differential preferences for options on these biodata items were
often not consistent with theory. Although nearly 1/3 of the biodata items included in
this study demonstrated DIF, only 17% of the items were consistent with theory. This
finding suggests the need for the revision of theory concerning the cultural value
differences between racial subgroups. Theoretical revisions regarding how racial

subgroups view basic human nature are particularly needed.

tau t :. 1,: tr: : ' ti ‘I' 5: 3010:..zi'1ml.‘

I.
at

If racial subgroup difl°erences in item responding were directly attributable to
cultural value differences, we would expect a significant cultural value*option choice
interaction for each of the items exhibiting a significant race*option choice interaction.

In order to examine this possibility, the biodata items which exhibited a cultural
value‘option choice interaction were compared with the biodata items which were found
to have a significant racial subgroup‘option choice interaction. Responses to 9 of the 107
biodata items were found to be influenced by both cultlnal values and racial subgroup.
These items, along with the cultural values assessed, are reported in Table 17. Clearly,
this analysis confirms the finding that biodata items which inquire about an individual's
view of basic human nature are likely to lead to differential response patterns across
racial subgroups. Of the 5 items included in this study which were classified as reflecting
the Human Nature orientation, 4 items were found to exhibit both significant cultural
value’option choice and racial subgroup‘option choice interactions. Few of the items
classified as reflecting any of the other 4 cultural value orientations were likely to

138

Table 17:
Biodata Items Exhibiting Both a Cultural Value*Option Choice
and Racial Subgroup*Option Choice Interaction

 

 

 

 

Cultural Value Cultural Value
Bindataltem # Altemtive Drientatinn
2 Good Human Nature
3 Evil Human Nature
6 Doing Activity
1 1 Evil Human Nature
33 Doing Activity
47 Lineal Relations
77 Evil Human Nature
87 Present Time

1 10 Collateral Relations

139

experience both significant cultural value*option choice and racial subgroup‘option
choice interactions.
Summary

The previous analyses examining differential response options for each racial
subgroup examined the degree to which a priori theory-based classifications for cultural
value differences actually influenced responses to biodata items. This section considered
direct empirical evidence of cultural value and racial subgroup influences on option
choice selection. This analysis found that subgroup differences in response selection to
only 9 of the 107 (8%) biodata items could be attributed directly to cultural value
differences, as measured by the 15 cultural value alternative scales. Given an alpha level
of .05 for determination of statistical significance, we would expect approximately 5 out
of evey 100 items to achieve statistical significance, assuming the items were
independent. However, the assumption of independence is certainly violated in the
present study. This indicates that family wise error rates are somewhat higher than 5/ 100.
Considering this fact, the finding that 9 items were found to flagged in both cultln'al

values and race analyses is not particularly impressive.

W

The above analyses provide convincing evidence that Blacks and Whites differ in
responses to identifiable types of biodata items. This finding remained even following
statistically controlling for the effects of socioeconomic status and social desirability.
However, as reported above, researchers have bemoaned the construct validity of race.
Dole (1995), for example, has argued

140

“I propose, therefore, that applied and scientific psychologists drop race because it
is a hopelessly ambiguous and politicized descriptor of alleged human
subspecies.”(p. 40).

Although applied Industrial/Organizational psychologists may be less concerned with
what race is and more concerned with maintaining compliance with legal conceptions of
race, theoretical I/O psychologists must maintain greater concern for whether any
findings regarding race are possibly attributable to other factors. The present study
attempted such an analysis. Specifically, the proposed mediating relationship between
race, cultural values, and biodata responses was examined by three series of analyses
utilizing hierarchical logistic regression. These analyses were conducted in order to
examine whether cultm'al values and biodata responses might explain additional variance
in race, even afier controlling for several demographic and ability variables. The control
variables included SES, sex, a measure of the size of the city the individual grew up in
(urbanicity), and cognitive ability.

In the first stage, race was initially regressed on SES, and then sex, cognitive
ability, urbanicity. After controlling for these variables, race was regressed onto 1 of the
15 cultural values. If the cultural value explained additional variance in race, the
improvement in model fit was significant. This process was repeated for each of the 15
cultural value subscales. The results of each of these analyses are reported in Table 18.
Five cultlnal values explained additional variance in race, after controlling for cognitive
performance and each of the demographic variables. These cultural value alternatives
were Mixed, Subjugation to Nature, Future, Doing, and Lineal. Thus, variance in race
remained after controlling for each of the cognitive performance and demographic
variables. Some of this variance was explained by differences in cultural values.

A second analysis was performed which again regressed race onto each of the
SES, sex, cognitive ability, and urbanicity variables in the initial 2 steps. In the third

141

E .32: 5 308323: on. he .33 8582»: 2: 8.8%.: :38?an—

 

au. 2. 3. ma. 3. 2. 3. an. 2. S. «e. 3. G. S. 3. Eon—29......—
og Qt- and- 3N- 2: as and- m... 2... 2... a. 3.“- mm. 36- Gr .28an
8. me. an; 8. 3.- Sr mm.— mm. 3. 3. mm. 3.. Na. 3.. 8. 33322.5
3. .53—38.....—
R4 2825
z .- be? 3:38
$2. xom
a... £2535
S. 30826.3...—
M‘N 38.50
«at 2:8:—
8. eafioaem 35o:
3..- 828:3 35a...

[WPWP‘II
1319131101)
WWI
Sumo
ocg-m-Sureg
Smog
strung
1m
159d
WIN
Known};
uonaflnfqns
13°09
WIN

11421

85— 83355 .528an

as?» .2530 es. 553. 35.56 53:35 .5 .QO
.3 60.8591 33— .53 3880..qu 35!: me 8.38”—

£— 035—.

142

step, however, one of the 107 biodata items entered the equation. If the response to the
biodata item explained additional variance in race over that accounted for by each of the
control variables, the improvement in model fit was significant at this stage. This
analysis was repeated for each of the 107 biodata items. Significant results are reported
in Table 19. Responses to 21 of the 107 (20%) biodata items explained additional
variance in race, following controlling for the cognitive performance and demographic
variables.

In the final stage, race was once again initially regressed onto the demographic
and cognitive performance variables. In the third step, the 5 cultural values which
explained additional variance in race in stage I entered the equation. Finally, race was
regressed onto the biodata items which indicated improvement in model fit in stage 2. If
the demographic, cognitive performance, and cultural value variables completely
mediated the relationship between race and biodata responses, the improvement in model
fit would be non-significant at this stage. Alternatively, if the demographic, cognitive
ability, and cultural value variables only partially mediated the relationship between race
and biodata item responses, the improvement in model fit at this step would achieve
significance. The results of this analysis are reported in Table 20. Examination of the
table reveals that the relationship between race and 6 of the biodata items was completely
mediated - items 3, 45, 47, 52, 55, and 77. This relationship between biodata responses
and race found to be only partially mediated by demographic, cognitive ability, and
cultural value variables for 15 of the biodata items.

Summary

The present study found that responses to 20% of the biodata items explained
additional variance in race after controlling for the effects of cognitive ability, and the
demographic variables of sex, urbanicity, and SES. This 20% is considerably less than
the approximately 1/3 of the biodata items exhibiting differential item functioning when

143

.2 .622: 3 220805522 22 .3 .26. 3:323»? 22 8.8%.: 2:02.329:—

3. no. me. 3. 3. 3. me. we. 3. 3. 3. 3. 3.

Q .~
3 .-
an:
cc.—
and"
MVN
an:

8.
av...

$2: 33m «35 :3:. 32m 32m 33m 233 33m 22m 23m 205 me:—

255 3.3: :5. $5.? 2.2an since: .5: gm
.3 238E3— ooamgzi 33.8.33— 33»?— ue 8.33—
“an 035—.

 

32:393.:—

Eoz 5:205
305329:—

384.60

3:? 2:58
xom

Ease:

32:25.33—

2826

2:8:—
:e28:2m 3:82
828:3 3.2:"—

85. ”o_::_.a> 2.92:9un

144

.2 .922: 3 2.08323: 22 2o .33 8585:»: 22 «.3822: 2:953an—

3. no. 3. me.

333 82m 303 505

:3:—en! a— 03:.

3. S.

2.05 2.95

3.

2.05

3.

R .~
3 ..
mm:
cc.—
and"
n‘ .N
«N:
be.
cor

$05

8.2

 

25803.58—

Eu: 833m

Eon—«>93...—

3825

3:2 3:38
sec.

22.52::

Eon—38...:—

33:50
2:85
=o23:2m 3.202
5.332....— 8.2a..—

Hozata> .52.—2.00

145

83m

2.

320

E .32: 5 2.383955 05 no .26. 8585:»? 05 «88:2: .582,an—

3. me. an. we. 3. ma. 3. 3. 3. 3.

«man 32m 33m 33m 32m nnoa 83m 32m «05 co:—

3.8— 3333 a... .853, .3330 55.2 2525.5 5.35.3.5
£5 .85 .3 638.33— 83— .53 2.88933— 8325 .3 3.33—
“ea 033—.

3.

S.

3.2-
2..
3..
3.
n:
3.—

3.

R <
3 ..
nn..
3..—
S.

3N
«N:
8.

cc:
n05

85.

 

3952533—
30: 3233

303963...—

38»on
.354
~58
283m

swam—.35

852

«5338...:—

33.830
35.2 2:38
5a

£253.:

30325.5...—

28:60
2:85
cow—8:3.— .8502
£338 25.

a.£§> 2.02.2.3

146

E .32: 5 .583an 05 mo .32 8:85am? 05 8.865 .553an—

mc. we. 3. me. 3. 3.

335 395 393 52m 2.03 2.05

6.2.3 2 2.3

3. me.

3.

3.2..
ms.
8..
we.
2.—
_N.—

R 4
E ..
an...
3.—

3.

m‘ .N
am:
5.
av...
2.2m sea

 

30525.3...—
Eo= 823.:

Eon—03.2.3—

E8300
32.3
~58
23:...

stamina

v0.32

2.05953...—

2325

E5 2:38
x3.

£25:

«33263:.—

253.6
2.32:
338:3 .3502
cofiuzum 35¢..—

82 "2.3.; .5235

147

the demographic and cognitive ability variables are not statistically controlled. One-third
of the cultural value subscales provided additional variance in race after the addition of
the control variables.

These analyses suggest a mixed interpretation regarding race. On the one hand,
race was significantly predicted by far fewer biodata responses after statistically
controlling for the effects of SES, urbanicity, sex, and cognitive ability, than were found
when utilizing DIF analyses which did not include such statistical controls. On the other
hand, despite arguments that race is an ambiguous construct which merely serves as a
proxy for other measurable variables, the analyses suggest that some differences in
cultural values and biodata item responses can be attributed to race even after controlling
for cognitive ability, sex, urbanicity and SES.

Contrary to the conceptual model presented in Figure 1, little evidence was found
indicating that the relationship between biodata responses and race is completely
mediated by cultural values (and the control variables). Indeed, this relationship was
found for only 6% of the biodata items in the study. An additional 14% of the biodata
items were found to be partially mediated by cultural values and the control variables.

The 35 items exhibiting differential item ftmctioning across racial subgroups were
also examined to determine whether this DIF would be expected to produce lower overall
test scores for Whites or Blacks. Specifically, the scoring key developed for the biodata
items was examined for each of the 35 flagged items. If an item was keyed such that the
differentially preferred option for Whites received the greatest score, the item was
considered to differentially encumber Blacks. If an item was keyed such that the
difi'erentially preferred option for Blacks received the greatest score, the item was

148

considered to differentially hinder Whites. If the source of the differential item
functioning for a particular item was an option or options that were not keyed as correct,
the item would not produce positive or negative effects for either group. It is important to
note that this analysis is meant to serve as an illustration of DIF for one possible
occupation - police officers. Because the rational keying of biodata items may differ
depending upon the occupation examined and the particular employee characteristics
desired, the results reported below are meant to serve as an example of how DIF could
affect the overall test score of a particular racial group. The results of this inspection are
reported in Table 21.

Nine of the 35 biodata items favored neither Blacks nor Whites. Rather, the two
groups were differentially attracted to options which were not rationally keyed as the
correct response. Sixteen of the 35 items exhibiting differential item functioning
impeded the performance of Whites, and 10 items impeded Black performance on overall
test score.

Whites performed differentially poorly on several of the items regarding
community or civic involvements (e.g., items 69, 89, 94, 100, and 112). In contrast to
Blacks, Whites were less involved in civic organizations. Whites also performed
differentially worse on several items which reflect the themes of planning and working
for the future (e.g., items 23, 58, 66, and 87). Contrary to expectations, Whites were less
likely to stress the importance of prior planning. Interestingly, both of these groups of
items seem to be assessing job relevant skills and abilities for the job of police officer.
Therefore, despite the differential item functioning on these items, labeling the overall set
of biodata items biased seems inappropriate.

There was a small trend for Blacks to perform differentially worse on items
regarding attitudes toward work (items 23, 104, and 114). Specifically, Blacks expressed

less concern with hard work than Whites. For the most part however, items on which

149

Table 21:
Racial Group Encumbered by Flagged Biodata Items

 

Itemfl—___Difierentiallen£umber-ed

2 Whites
3 Blacks
6 Neither
8 Neither
10 Whites
l 1 Whites
12 Whites
18 Blacks
20 Blacks
23 Whites
33 Blacks
40 Neither
42 Neither
44 Neither
47 Neither
53 Whites
57 Whites
58 Whites
69 Whites
72 Blacks
74 Whites
77 Neither
83 Neither
87 Whites
89 Whites
94 Whites
95 Whites
99 Blacks
100 Whites
103 Blacks
104 Blacks
106 Blacks
l 10 Neither
l 12 Whites

1 14 Blacks

150

Blacks performed differentially worse failed to form rational groupings. However, there
was some evidence suggesting that biodata test constructors wishing to avoid racial
subgroup DIF might be advised to exclude items pertaining to personal child rearing
practices. Two such items were included in this biodata instrument (items 10 and 103).
Both items exhibited differential item functioning across Blacks and Whites. Whereas
Whites preferred options reflecting a reasoning or rewarding approach, Blacks preferred
punishment options. Similarly, there is some indication that items inquiring about how
others view the individual may also lead to difi’erential item fimctioning. Item 99, for
example, inquires about the degree to which the individual has been accused by others of
doing something solely because it was in their best self-interest. Despite the consistent
trend for Blacks to report greater civic and community involvement than Whites (e.g.,
items 69, 89, 94, 95, 100 and 112), Blacks reported that they had more often been
accused of doing things for their own self-interest than Whites. Given the greater
community involvement of the Black group, the finding that they perceive more
accusations of acting in their own self-interest may possibly be attributed to greater
perceived racism. The observed data are consistent with the interpretation that Blacks
may feel they are more often accused by others of negative behaviors.

The influence of the items exhibiting DIF on overall test score was examined in
three ways. Items which differentially encumbered Blacks were initially dropped from
the test, and the overall test scores were re-computed. The same procedure was repeated,
dropping only those items which were more difficult for Whites. Finally, all items which
were differentially more difficult for one of the 2 subgroups were dropped, and the
overall test scores were re-computed. Following each of these steps, t-tests were used to
examine differences between racial subgroups on the overall biodata test.

Table 22 reports the means and standard deviations of biodata test performance
for each racial subgroup for each of the t-tests described. As reported above, Blacks

151

Table 22:

T-tests of Overall Biodata Test Score by

Racial Subgroup Following Deletion of DIF Items

 

 

Biodata Items Black White
ExdudedimmAnalxsisJeaL SD Mm SD
No Items Excluded“ 90.77 14.09 86.33 13.97
Items More Difficult“

for Blacks 82.36 13.28 76.21 12.99
Items More Difficult

for White 76.61 12.16 75.90 12.41
Items Differentially More

Difficult for Either

Racial Subgroup 68.20 11.33 65.78 11.42

 

" indicates significant difference between groups, p<.05

152

outperformed Whites on the overall biodata test prior to deletion of items experiencing
DIF (t(203) = -2.27, p<.05). Logically, dropping biodata items which Blacks found more
difficult than Whites exacerbated this difference, (t(203) = -3.36, p<.01). Deleting those
biodata items which Whites found more difficult had the effect of removing the observed
disparity in overall test scores for racial subgroups (t(204) = -.41, p>.05). Similarly,
discarding all biodata items which were found to differentially encumber one of the two
racial groups also had the effect of removing the observed disparity in overall test score
(t(204) = -1.53, p>.05).
Summary

Blacks outperformed Whites on the biodata test used in the present study, when a
rational scoring key was developed for the job of police ofiicer (mean difference = .31).
In the absence of actual criterion information, it is unknown whether this subgroup
difference is attributable to job-related ability differences between the two racial groups.
At the item level of analysis, racial subgroups were found to be differentially encumbered
by numerous biodata items. As would be expected, discarding biodata items that either
Whites or both racial groups found more difficult had the effect of removing subgroup

differences on overall test score.

DISCUSSION

The present study introduced the concept of differential item functioning to biodata
employment testing. Figure 1 provided a conceptual model which guided the present
research. Examining the model in reverse order, differential responding to biodata items
by members of different racial subgroups was proposed to be influenced by cultural value
differences between racial subgroups. The adoption of cultural values was pr0posed to be
determined by the individual's racial identity. Racial identity, a psychological variable
assessing the individual's perceived racial identification, was thought to be influenced by
one's race. In order to discuss the results of analyses based on this model, the present
section will be divided into four parts. Initially, a brief description of each of the findings
regarding the hypotheses will be discussed. Next, the implications of these results will be
explored. The final two sections will address the limitations in the present study, as well
as future directions for this research.

Hmothesis I : Race & Racial Identity

A moderate relationship was expected between race and racial identity. It was
thought that rather than classification based on skin color or other physical information,
classification based on racial identity would provide a far greater understanding of the
individual's psychological nature. Thus, racial identity classification might provide more
information regarding the identity of an individual of one demographic race (e. g., Black)
who was raised in a neighborhood dominated by members of another race (e.g., Whites).
Results indicated that the demographic classification of race was identical to the racial

153

154

identity classification of nearly all the Blacks and Whites examined in the present
investigation (r = .93). Of the 200 individuals examined in the study, only 4 exhibited a
racial identity that was not consistent with their demographic racial classification. This
finding indicates that in future psychological investigations comparing Blacks and
Whites, the assessment of racial identity as a measure of classification of individuals
appears unwarranted. A single item assessment of self-reported race provides the same

information as a 15 item psychological scale.
Hypothesis 2: Race vs. Racial Identity, and Cultural Values

The present study adopted Kluckhohn and Strodtbeck's (1961) conception of
cultural values. According to this theory, there exist five cultural values (Human Nature,
Person-Nature, Time, Activity, and Relations), each of which has three separate
components. Based on the assumption that demographic race and racial identity would
be related but distinct concepts (Hypothesis 1), Hypothesis 2 argued that the adoption of
cultural values would be more strongly related to an individual's racial identity
classification than the individual's demographic race. This hypothesis was not
supported by the data. The strong relationship between demographic racial classification
and classification based on racial identity precluded any significant findings supporting
Hypothesis 2. As discussed below, the lack of empirical support regarding the concept of
racial identity is limited to racial identity classification. Racial identity strength was
found to influence the adoption of cultural values.

155

Hmothesis 3: Racial Subgroups and Cultural Values of the Typical American

The third hypothesis expressed two expectations. First, it was argued that the
cultural value pattemsdisplayed by Black and White racial subgroups would be identical.
Secondly, the cultural value patterns, or preference orderings, would be consistent with
theory-based expectations for the typical American. Empirical evidence suggests that
Black and White Americans share highly similar cultural value patterns.

The cultural value patterns produced for these two racial subgroups were not
completely consistent with theory for the typical American. Based on theoretical
expectations of the "typical" American, it was expected the dominant preference for the
basic view of human nature would be either a "mix " or "evil" view. Both Blacks and
Whites endorsed the "evil" view less frequently than either the " good" or "mixed"
alternatives. Similarly, the typical American was expected to most strongly endorse the
"mastery over nature" view of the relationship between humans and the natural world.
Both Blacks and Whites endorsed this view with the least frequency of the three
alternative views. However, inspection of the overall cultural value patterns produced by
Blacks and Whites provides empirical confirmation of many of the cultural value
preferences for the typical American espoused by theoreticians such as Kluckhohn and
Strodtbeck (1961).

Hypothesis 4: Cultural Value Dzflerences between Racial Subgroups

The greater interest of the present research was the examination of differences
between racial subgroups. Multiple analytical techniques were used to investigate the
possibility that Blacks and Whites differ in their endorsement of cultural values. Both t-

tests and Cohen's (1977) definition of a small effect size were used as criteria for

156

determining the presence of subgroup differences. Consistent with the writings of Mbiti
(1970) and McFadden and Gbekobou (1984), Blacks expressed greater preference for the
cultural values of Subjugation-to-Nature, Past, and Lineal. However, Blacks' preferences
for the cultural value alternatives of Evil, Future, and Doing were surprising based on
theory that indicated Whites would prefer these cultural values.

Hierarchical regression was used to examine several possible alternative
explanations for the findings regarding racial subgroup differences in cultural value
preferences. These analyses revealed that Blacks' preferences for the Doing alternative
could be attributed to socioeconomic differences between the two groups. Racial
subgroup differences in the tendency to respond in a socially desirable manner were not
found to account for the expressed difi‘erences in cultural values. However, the strength
of an individual's racial identity was found to moderate the relationship between race and
preferences for certain cultural values. Specifically, the Mastery over Nature alternative
was preferred by Whites with high racial identity. The previous finding that racial
subgroup differences existed in preferences for the Lineal alternative was qualified by the
finding that racial identity strength moderated this relationship. The Lineal alternative
was most likely to be endorsed by Blacks low in racial identity strength, but this
preference was not observed among Blacks high in racial identity strength. The
Individual cultural value alternative was endorsed most frequently by Whites low in
racial identity strength, and Blacks with very high racial identity strength.

Cultural value differences clearly exist between Blacks and Whites, although
these differences are not always consistent with theoretical expectations. In order to fully
understand differences in preferences for several cultural values, the strength of the

individual's racial identity must also be considered.

157

Hmothesis 5: Cultural Values Difl'erences and Biodata Responses

The fifth hypothesis examined whether cultural value differences could influence
responses to biodata employment items. Overall, 21% of the 107 biodata employment
items were found to be influenced by the cultural values of the respondent. The
individual's Human Nature orientation seemed especially likely to influence biodata item
responding - 4 of the 5 biodata items classified as reflecting the Human Nature
orientation were found to be influenced by the individual's endorsement of this cultural
value. Similarly, 6 of the 15 biodata items a priori classified as reflecting a Time
orientation were found to be influenced by the individual's endorsement of the cultural
values relating to Time. The individual’s Activity orientation was also found to influence
biodata responses. Of the 41 items classified a priori as reflecting one of the alternatives
of the Activity orientation, 8 items were found to be influenced by the cultural values of
the respondent.

In order to conduct the present study, biodata items fiom several databases were
scanned for biodata items which might reflect cultural values. Of these databases,
relatively few items were classified as reflecting the cultural values of Human Nature or
Time. In contrast, these databases contained many items which were classified as
reflecting the cultural values orientations of Activity and Relations. Responses to these
biodata items were found to be far less influenced by the cultural values of the

respondent.
Hypothesis 6: Racial Subgroups and Biodata Responses (& Cultural Values, too!)

After statistically controlling for overall test score differences, 1/3 of the biodata

items included in the present study were found to exhibit differential item functioning

158

between Blacks and Whites. Blacks and Whites were differentially attracted to response
options for a considerable number of biodata items. Although the finding that such a
large number of biodata items exhibited DIF is important in itself, the purpose of the
present study was to investigate a priori hypotheses regarding the influences of DIF on
biodata items. The racial subgroups were found to be differentially attracted to response
options of biodata items consistent with a priori expectations for 18 of the 107 biodata
items. Thus, although DIF was found in this dataset for approximately 33% of the items,
the DIF was consistent with theoretical expectations for half of these items (18 of 35), or
17% of the total set of items. Although these findings do not provide overwhelming
support for the hypothesis, they certainly indicate support for the idea that a theoretical
tmderstanding of cultural values can be used to predict subgroup DIF on biodata
instruments. The lack of overwhelming support for the hypothesis is attributable, in part,
to the inconsistency between theoretical expectations regarding subgroup differences for
cultural values, and actual empirical findings regarding these difl‘erences.

Sample size limitations precluded a direct analysis of the relationship between
cultural values, racial subgroup, and biodata option choice. However, 9 biodata items
were fotmd to be influenced by both the cultural values and race of the respondent.
Again, 4 of the 5 biodata items classified as reflecting a Human Nature orientation were
identified. It appears that the inclusion of biodata items which assess the individual's view
of basic human nature are likely to lead to significant response difl‘erences between
Blacks and Whites. The remaining 5 items which were found to be influenced by both
cultural values and race were composed of 2 items thought to reflect the Activity
orientation, 2 items reflecting the Relations orientation, and a single item assessing the

Time orientation.

159

Exploratory Analysis 1: Variance in Race

Citing the current controvery in psychology over use of the construct of race, the
present study investigated the possibility that measured demographic variables and
cognitive ability might account for the majority of the variance in the construct of race.
However, 5 of the cultural values subscales and 21 of the biodata items were shown to
explain additional variance in race after controlling for cognitive performance, sex, father
education, mother education, income, and urbanicity. This finding indicates that although
the demographic variables used in this study do account for part of the variance in race,

the construct of race assesses more than these simple demographic variables.

Exploratory Analysis 2: DIF and Overall Test Score

In order to examine the degree to which the observed difference in biodata test
score between Blacks and Whites was attributable to items that exhibited DIF, all DIF
items were removed from the computation of overall test score. Quite predictably, this
resulted in the removal of the significant difference in test score between racial
subgroups. Removal of items showing DIF that encumbered only Whites also removed
the subgroup difference in overall test score. Removal of items exhibiting DIF that
encumbered only Blacks increased the size of the subgroup mean difference to .43
standard deviation units. Note that in the absence of criterion information, it is
unknown whether the removal of such items would improve, impair, or even affect, the
validity of the biodata test

160

11"1Cl'

The results of the present study have several implications for three distinct
research areas. First, implications of findings regarding the construct of racial identity
will be discussed. Next, the implications of findings regarding cultural value differences
between racial subgroups will be considered. Finally, the impact of current study's results

on the examination of DIF on biodata tests will be examined.

Racial Identity

The utility of the construct of racial identity was examined in the present study.
Unlike previous research on the construct, racial identity was conceived here to have two
components: a classification component, based on the group of people the individual
identified as feeling closest to, spending the majority of their time with, and feeling most
comfortable among; and a second component which assessed the degree to which the
individual perceived their racial identity to be important to their sense of self. Based on
the results of this study, the assessment of racial identity for classification purposes
between Blacks and Whites appears unnecessary. Psychological assessment of racial
identity provided classification results that were virtually identical to self-reported race.
However, the present study's novel use of the strength of an individual's racial identity
was found to explain additional variance in the adoption of several cultural values.
Perhaps racial identity strength would be a theoretically useful construct for other
research investigations examining racial subgroup differences. The degree to which race
is perceived to be an important part of the individual's personality has been neglected in

research examining racial subgroup differences.

161

Racial Subgroups and Cultural Values

The present study clearly indicates that present theories of cultural values
proposed by theoreticians such as Kluckhohn and Strodtbeck (1961), Mbiti (1970), and
McFadden and Gbekobou (1984) require revision. The present study found that Blacks
are more likely than Whites to endorse an evil view of basic human nature, were more
likely to feel that they were at the whim of Nature, that the past is a good source of
instruction, and that thinking about the future is important. Blacks for whom their racial
identity was not very important were more likely than Whites to endorse the ideals that
societal authorities should be respected and obeyed. Blacks for whom their racial identity
was extremely important were more likely than Whites low in racial identity strength to
endorse the idea that individual goals should supersede group goals.

Several of these findings concerning preference differences between cultural
values are consistent with an earlier empirical study conducted by Carter (1990). This
researcher found that relative to Whites, Blacks viewed human nature as more evil, were
more subjugated to Nature, preferred the past, were more likely to adopt a long-term self-
development goal (i.e., preferred the Being-in-Becoming alternative), and were more
likely to feel power and authority should be gained through one's age and role in the
social hierarchy (i.e., Lineal alternative).

Taken together, these studies provide evidence that, contrary to theoretical
expectations, Blacks view human nature as more evil than Whites. Carter (1990)
suggests this difi‘erence may be attributable to Blacks' experience with racism.
Perceptions of racism in a White-dominated society may lead Blacks to believe that basic
human nature - no matter what the skin color - is evil.

Both studies also found Blacks believe relationships should be dominated by the
individual's role in society, though the present study qualified this finding. Blacks low in

162

racial identity strength are more likely than Whites or other Blacks to endorse this view.
Still, this finding is interesting, and suggests a greater willingness on the behalf of Blacks
to follow the judgment of authority, rather than their own goals. Blacks who feel very
strongly that their racial identity is important to their sense of self, on the other hand, are
less likely to feel endorse the Lineal perspective.

One admittedly speculative explanation for this finding is that Blacks who fail to
examine the impact of race on their own lives would likewise be unlikely to thoughtfully
consider the judgment of those in authority. Blacks who strongly believe that race is
important to their sense of self, on the other hand, have perhaps consciously considered,
and rejected, the idea that one should respect authority for authority’s sake. This
argument is strengthened by the possibility that those Blacks who consider race to be
extremely important are also likely to be those who perceive greater racism in society.
Individuals in authority may be viewed, either directly or indirectly, as the perpetrators of
racism. Blacks who are high in racial identity strength would thus be less willing to
heedlessly abide by the judgments of those in authority. Although alternative
explanations for this finding undoubtedly exist, it is clear that traditional theories which
assert that Blacks are likely to maintain a Lineal perspective of the Relations orientation
require revision. The empirical evidence indicates that only those Blacks that do not
strongly consider race to be an important part of their personality are likely to believe
authority should be based on age and positional power.

It is important to note that both the present study and Carter‘s earlier investigation
utilized the same measure of cultural values, though these scales were slightly modified
in the present study. If we are to better understand the cultural values of Blacks and
Whites in America, additiorml investigations of the assumed values of these groups is
required. Although the present study replicates several of Carter‘s findings, further

163

replications utilizing distinct measures may provide even greater understanding of the

cultural values of these groups.

Subgroup Differences in Biodata Responding

The observed cultural value differences between racial subgroups did not
consistently contribute to the observed subgroup differences in responses to biodata
items. This was evidenced by the finding that although the responses of 22 biodata items
were influenced by the cultural values of the respondent, the responses to only 9 of these
items were also influenced by the race of the respondent.

However, the explication of a priori hypotheses regarding the influence of cultural
values on biodata responding was far from futile. First, it was empirically demonstrated
that the respondent's cultural values can influence response choice selection -- in the
present study, 22/107 or 21% of the responses to biodata items were influenced by the
cultural values of the respondent. Second, the items in the present study were initially
selected for inclusion in the study only if it was thought to reflect 1 of the 5 cultural value
orientations. Racial subgroup difi'erences in response choice selecdon were found for 1/3
of the biodata items included in the study.

Although test bias cannot be equated with item bias or difi‘erential item
functioning, it is surprising that 1/3 of the items exhibited DIF between Blacks and
Whites. This finding is particularly surprising considering that biodata, at least on the
test level, have generally not exhibited race differences (Mumford & Stokes, 1992; Reilly
& Chao, 1982). One possible explanation for this conflict with previous studies is that
actual biodata tests used for employment purposes may include far fewer items reflecting
cultural values than were intentionally included in this study. However, even if the

presence of cultural value-laden items is not a widespread problem, it is a potential

164

problem. Thus, biodata items can, and perhaps should, be rationally screened for the
inclusion of cultural values prior to use in a selection instrument. Any item which is
found to reflect a cultural value known to differentiate racial subgroups would be flagged
for possible deletion. This step would be especially important if the validation strategy
for the biodata instrument did not include criterion-related instrument validation.
Consideration of the degree to which the biodata item accurately assesses a job-relevant
KSA could determine whether the item should be retained or omitted. Of course, the
item may be initially retained for empirical examination of criterion-related validity and
test fairness. Even here, however, sample constraints on the number of racial minorities
included in the validation study, or concerns with the number of biodata items to be
included in apre-test, may warrant the exclusion of biodata items which are judged a
priori to exhibit cultural values which are likely to lead to racial subgroup differences.

Which items are most likely to lead to racial subgroup differences? The results of
the present study provide a clear indication that items which inquire about the
respondent's view of basic human nature are extremely likely to lead to DIF between
Blacks and Whites.

Further, 39% of the biodata items which were a priori classified as reflecting a
Relations orientation were also found to exhibit DIF. Using the scoring key developed
for the job of police officer, 4 of the 14 biodata items which were classified as reflecting
the Relations orientation, and were actually found to exhibit DIF, did not impair
performance for either racial subgroup.

Three of the 14 Relations items were found to be difl‘erentially more difiicult for
Blacks. No consistent trends were detectable in the interpretation of these 3 biodata
items. The item “Would you rather play in a dance-band than be a soloist for an opera”
was intended to assess a propensity for teamwork vs. individual performance. Blacks

were formd to be more likely than Whites to report wanting to be a soloist. The second

165

item was commented upon above. Blacks reported a greater ntunber of instances of being
accused of acting in self-interest than Whites. The final Relations item which was
differentially more difficult for Blacks inquired about strategy when working on a very
important project. Whites received greater credit for their response “You would
minimally attend to other work responsibilities realizing that you could not give all your
time to the more important project”, than was awarded to Blacks for the response “You
would attempt to divide up your time equally”. This item had been classified as
reflecting the Relations orientation based on the expectation that Blacks would be
differentially attracted to the option, “You would ask your supervisor/boss to find
someone else to help you”. This result was not found.

Seven of the 14 Relations items were found to be differentially more difficult for
Whites. Several of these items assessed community/civic involvement, or recognition for
such involvement. Whites consistently reported less participation in civic affairs than
Blacks. This finding is consistent with Olsen’s (1970) review which reported that
regarding voluntary participation, “the general tendency for Blacks to be more active than
Whites is found to occur in every type of activity investigated” (p. 682). Whites also
scored lower than Blacks on several of the Relations items due to an apparent willingness
to report negative behaviors. Relative to Blacks, Whites reported more instances of
disrespect to teachers and hurting a friend in order to benefit themselves. Whites were
also likely to report fewer observations of people putting friendships and persorml goals
aside in order to achieve group goals. It is important to note that this study cannot
determine whether these difierences are attributable to actual subgroup differences, or
merely reflect differences in the way the two groups respond to employment items.
However, evidence supporting the latter possibility is provided by the finding that Blacks
achieved higher scores than Whites on both social desirability scales examined in the
present study.

166

The finding of Black-White subgroup differences in response to biodata items
assessing interpersonal relations is consistent with a previous study conducted by Schmitt
and Pulakos (under review). In their study of DIF in biodata items administered to
incumbents at a large federal agency, Schmitt and Pulakos found that 9 of 15 items
intended to assess interpersonal relations skills exhibited DIF. Unfortunately, the
increased use of team-based work processes is likely to require the inclusion of an
increased proportion of biodata items which assess interpersonal relations. The findings
in the present study of cultural value differences between Blacks and Whites regarding
the Lineal alternative may provide some insight into the types of biodata items to avoid.
Blacks low in racial identity strength were more likely to endorse the notion that power
and authority are gained by age and role in the social hierarchy. Perhaps these
individuals, then, would be expected to be more likely than Whites and Blacks for whom
racial identity is very important to select biodata options promoting vertical authority.

Biodata items classified as reflecting the Activity orientation which were most
frequently flagged for DIF often included elements referencing the degree to which the
individual enjoyed hard work, or experienced guilt when not working. Five biodata items
which included elements referring to hard work and/or guilt associated with a lack of hard
work exhibited DIF. Of these items, the scoring key used in this study found that 3 items
were equally difficult for both racial subgroups. The remaining 2 items were found to be
more difficult for Blacks. Blacks reported liking hard work less, and felt less bothered
when they were unable to accomplish a chore.

Finally, although only two items were included in the present study which
assessed attitudes toward disciplining a child, both items exhibited differential responding
between racial subgroups. Whereas Blacks preferred options regarding punishment,
Whites preferred reasoning with children. These differences may reflect actual racial

167

subgroup differences in child rearing practices. The inclusion of similar items on a

biodata instrument are likely to lead to DIF between Blacks and Whites.

Recommendations

The results of the present study provide several suggestions for constructors of
biodata instruments who wish to avoid differential item functioning between Blacks and
Whites. It is assumed biodata items are initially written or selected to assess specific
KSA dimensions as specified by job analysis. The results of the present study strongly
suggest the need for careful, rational inspection of each item under consideration for
possible inclusion in the biodata instrument. Biodata items which assess the individual’s
perception of basic human nature, involvement in civic events, attitudes toward hard
work or guilt associated with not working, and which reflect child rearing practices are
likely to lead to DIF between the two racial groups. These biodata items should be
dropped from the biodata instrument if racial subgroup DIF is a significant concern.
Special attention should also be given to biodata items which reflect the Doing
orientation, and the Future alternative of the Time orientation. Of course, validity must
remain the pivotal criterion for determination of whether or not to include a specific
biodata item in a test instrument. However, test makers can thoughtfully construct
biodata items which continue to assess job-related KSAs without including any elements
in the above list. Selection procedures which do not directly examine differential
validity, or require a very limited number of items for pre-testing, should be particularly

careful to avoid biodata items which contain any of the above elements.

168

SlI"'

Several limitations in the present study warrant a cautious interpretation of study
findings. As in the previous section, the limitations of the study will be discussed in
terms of the findings regarding racial identity, racial subgroups and cultural values, and
finally subgroup differences and biodata responding.

Racial Identity

The present study investigated the differences between Blacks and Whites in their
cultural values and in responses to biodata employment items. The conceptual model in
Figure 1 asserts that racial identity will be a distinct construct from demographic race,
and would be expected to provide greater information regarding the adoption of cultural
values than mere demographic classification. These assertions did not receive empirical
support. Blacks were found to nearly unanimously adopt an African American racial
identity, and Whites were found to nearly unanimously adopt a White American racial
identity. Would the same consistency between racial identity and demographic racial
classification also be found for Asians and Hispanics? The present study cannot address
this question, since only Blacks and Whites were assessed. However, the dark skin color
of Blacks has been argued as a barrier to assimilation into White culture. Other minority
groups such as Asians and Hispanics may have an easier time "passing" as a member of
the White majority group, and thus may be more likely than Blacks to choose a White
racial identity. The finding that classification of individuals based on racial identity
provides little information beyond demographic classification is limited to Blacks and
Whites.

169

Racial Subgroups and Cultural Values

A major limitation of the interpretation of results regarding cultural value
differences between racial subgroups is the disappointing internal consistency reliability
estimates for the scales used to assess cultural values. Although the 15 lO-item scales
proposed by Kohls, Carter, & Helms (1984) were subjected to principal components
analysis and rational inspection, the reliabilities of the resulting modified scales increased
only slightly from a mean of .66 to a mean of .69. These poor reliability estimates may
obscure additional differences in cultural values between Blacks and Whites. On the
other hand, the findings regarding racial subgroup differences for several cultural values
is all the more impressive considering the poor reliability of these scales.

The hypotheses examined in the present study were in large part based on
theoretical expectations of subgroup differences in the adoption of cultural values.
Combined with Carter’s (1990) findings, the results of the present study clearly suggest
that this theory is in need of revision. Revised theoretical explanations for the observed
differences in cultural values between racial subgroups is evidently required.

The composition of the sample also leads to a possible limitation of the findings
regarding subgroup differences between Blacks and Whites. The sample was composed
solely of college students. Are college students of each racial subgroup representative of
the actual population? Although this type of question is asked reflexively in the
Limitations section of any research project, it may have special significance for the
present study. Why? One possibility is that the members of the racial subgroups that go
to college may be very different from those who do not attend college. For instance,
consider the theoretically inconsistent finding that Blacks are more likely to be planners
of future events than are Whites. Perhaps this finding merely reflects difl’erences in
opportunities between the two racial subgroups. If Blacks are disadvantaged socio-

170

economically and educationally in comparison to Whites, only those individuals who can
successfully plan for the future would actually be expected to attend college. If Whites,
on the other hand, experienced greater opportunity to attend college, this racial subgroup
would not need to be particularly future-oriented in order to attend college. Thus, the
observed Black-White difference in future orientation may possibly be attributable to the
sample used - college students. The interpretation of findings regarding other
differences between racial subgroups must also be considered in the context of the

limitations of this sample.

Racial Subgroups and Biodata Responding

Perhaps the most obvious potential limitation to the results regarding subgroup
difierences in responses to biodata employment items is that subjects were administered
the biodata items in an experimental, instead of actual pre-employment, setting. Subjects
were asked to "imagine" they were applying for a job in a local law enforcement agency,
and were asked to respond as they would if they were applying for the job. To aid
motivation, participants were instructed that the 6 highest scorers would each receive
$25.00 cash. Despite these interventions, it is likely that the motivation of actual job
applicants would be considerably higher. The impact of these potential motivation
differences between research participants and actual job applicants on observed subgroup
difi’erences is unknown.

Results indicate that cultural values influenced option choice selection for 21% of
the biodata items examined. A potential criticism of this finding is that it may merely be
an empirical demonstration that two different measures of the same construct are related.
If valid, does this criticism diminish the results or implications of the present study? In

the opinion of the present author, it does not. From a practical standpoint, constructors of

171

biodata employment tests would hope to avoid any items which are likely to lead to racial
subgroup differences in employment testing when valid items without the subgroup
difference can be identified. Thus, knowledge of which cultural values should be avoided
in the construction of a biodata instrument is highly valuable information. From a
theoretical standpoint, this contribution is very exciting. Research examining differential
item functioning is routinely conducted post hoc on cognitive ability tests. This research
has failed to produce findings which can be used a priori to determine which items, if
any, might lead to DIF. The present study generated a priori hypotheses regarding which
biodata items would experience DIF - and findings indicate that fully 1/3 of the items
suspected to exhibit DIF, did!

A serious limitation of the present study is that biodata items were included only
if they were a priori classified as reflecting l of the 5 cultural value orientations. As just
stated, 1/3 of these biodata items examined demonstrated DIF between racial subgroups.
Unfortunately, the present study did not include "control" biodata items which were a
priori classified as reflecting none of the cultural values. Inclusion of such items would
allow a direct examination of DIF rates between biodata items thought to be neutral with
respect to Black-White cultural difi’erences, and those a priori classified as reflecting
cultural values. In the absence of the inclusion of such "control" biodata items, the
attribution of DIF to cultural values is severely weakened. However, the finding that 1/3
of the biodata items included in the present study exhibited DIF across racial subgroups
does seem particularly surprising, considering that researchers have argued that biodata
tests have rarely shown racial subgroup differences (Mumford & Stokes, 1992; Reilly &
Chao, 1982).

172

E 1]..

In the grand tradition of science, the present study asked a few questions, partially
answered some of them, and generated many, many more. The need for several
clarifications and future studies has been strongly suggested throughout this paper. This
final section will consider some areas of potentially fruitful future research.

The main goal of the present study was to provide an initial attempt at providing a
priori hypotheses regarding which biodata items might be expected to exhibit DIF
between racial subgroups. In order to accomplish this goal, cultural value differences
between racial subgroups were explored, and hypotheses were generated. The concept of
culture is enormous, and researchers have had a very difficult time limiting the definition
of the concept. As such, culture can mean many different things. Culture was defined in
the present study in terms of a people's system of underlying values. Other conceptions
of culture may generate additional interesting hypotheses worthy of investigation.
Differential item functioning research examining cognitive ability tests has demonstrated
that post hoc investigations provide little clarity or understanding. In order to understand
why racial subgroups are differentially attracted to different biodata response options, it is
imperative that researchers employ the use of a defensible theoretical framework.

The present study limited its focus to the examination of Black-White differences
in responses to biodata employment items. Other racial subgroups deserve research
attention as well. Indeed, some racial subgroups such as Asians and Hispanics which are

relative newcomers to American culture may be far more interesting to investigate.

173

Cultural differences between these groups and the "typical" American may be expected to
be far more disparate than between Blacks and Whites in America Further, our interest
in subgroup differences certainly need not be limited to race. Examination of the
burgeoning literature examining sex differences might also lead to the generation of a
priori hypotheses regarding male-female differences in responses to employment tests.
Additional employment tests should also be investigated for potential for DIF . The
present study revealed that Blacks and Whites strongly differ in their perception of
basic human nature. How would such a difference affect responses to integrity tests?
Would the type of integrity item asked influence DIF? For example, we might find
difi'erent rates of DIF for items which assess actual past behaviors, vs. perceptions of
frequency of such behaviors in the general public. Investigation of DIF on low

fidelity simulations (Motowidlo, Dunnette, & Carter, 1990; Motowidlo & Tippins,
1993) would also be potentially fruitful. Job applicants administered a low fidelity
simulation are presented with a series of ambiguous situations and for each are asked
which of several options they feel would be the best course of action. Subgroup
differences in response tendencies might be expected to be exhibited on subjective
employment tests such as these as well.

Future research might also include criterion data in DIF examinations. Such
information would allow direct examination of how subgroup differences at the item
level influence subgroup differences at the test level. The present study limited its
investigation to subgroup differences in responding to biodata items. Ultimately,

concerns with subgroup differences are most important when considered at the test

174
level in relation to validity. Although tests which exhibit differential validity across
racial subgroups are legally defensible in the presence of acceptable criterion-related
validity, such tests could be avoided if composed of valid items which do not exhibit
differential validity. Further research is needed to determine the relationship between
differential item functioning and test bias. The a priori classification of subjective
employment items which are likely to exhibit DIF may help test constructors in
avoiding problems associated with differential validity.

Despite recent criticisms of aflirmative action policies, it is clear that societal
concerns with issues of test fairness and test bias will remain at the forefiont of
personnel selection decisions. The examination of item-level subgroup difi'erences on
employment tests is just now in its formative stages. Future research investigations of
the influences of differential item functioning in employment items have the potential
to make an immediate impact on the selection procedures used to form the American

workforce.

APPENDIX A

Appendix A

Racial Identity Classification:

For each of the following, choose the group with whom you feel the stem most
closely applies:

1. I identify most strongly with:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

2. The heroes with whom I identify most tend to be:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

3. My most important role models are most often:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

4. The people with whom I share the closest emotional bonds tend to be:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

175

176

5. The entertainers whom I enjoy most in film, stage, and stand-up comedy tend to
be:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

6. I currently spend most of my social time with:
A. Afiican Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

7. If I had to choose a single racial group, I would say that my behaviors are
consistent with those associated with:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

8. If prior to being born I could select a racial identity, I would choose:
A. African American
B. Asian American
C. Hispanic American
D. Native American
E. White American

9. If given the opportunity, I'd spend most of my time with:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

10. I would prefer a romantic partner whose racial identity was:
A. African American
B. Asian American
C. Hispanic American
D. Native American
E. White American

 

177

11. Most of the sports figures I idolize tend to be:
A. African American
B. Asian American
C. Hispanic American
D. Native American
E. White American

12. The clubs and social organizations I belong to are composed primarily of:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

13. The musicians whom I enjoy most tend to be:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

14. Regarding the group of people I would refer to as my "community", I would say
most of these individuals would be:
A. African American
B. Asian American
C. Hispanic American
D. Native American
E. White American

15. If I were to choose a group of coworkers, I would probably choose a majority of
the individuals to be:
A. African Americans
B. Asian Americans
C. Hispanic Americans
D. Native Americans
E. White Americans

178

Racial Identity Strength:
Instructions: For the following items, indicate your level of agreement on the
following scale:
1 2 3 4 5 6 7
Strongly Strongly
disagree agree

8.‘

9.‘

10."I

ll.’

12."'

l3.

14.

15.

My race is an important part ofwho I am.
I tend to socialize with members of my own racial group.

I've never really given much thought to what it would be like to be a member of
some other race (R).

I've never given much thought to the racial experiences of others (R).
Iamproudtobeamemberofmy race.

I believe there are more positive characteristics associated with other races than my
own (R).

Generally I don't give much thought to racial issues (R).
Matters ofrace are veryimportantto me. -

I am strongly committed to my racial group membership.
I strongly identify with members of my own race.

I wish I were a member of some other race (R).

I rarely consider what race I am. (R)

If given the opportunity, I'd spend the vast majority of my time with members of
my own racial group.

I consciously seek out members of my own racial group for social activities.

I believe we should deemphasize racial difl’erences (R).

179

16. As a youth, I was primarily exposed to individuals of my own racial group.
17.‘ I think one's racial identity is an integral part of one's personality.
18. I generally feel the most "attach " to other members of my own race.

19. Some people think I spend too much time acting like people of some race other
than my own (R).

20. I generally value the opinions of individuals of my racial identity more than I value
the opinions of individuals of other racial groups.

 

* indicates item was included in final scale

d..-

 

180

Intercultural Values Inventory
Kohls, Carter, Helms, 1984

The following statements are statements of values. For each statement, indicate the
degree to which you believe the statement reflects a value that you hold, that is, a
value that you live your life or would raise your children by.

1 2 3 4 5 6 7
Strongly Strongly
disagree agree

3.* Most people can't be trusted. It pays to be suspicious and wary of
others. In the end, you can only trust W.

5.* You've got to constantly look out for your own welfare. It's a dog-
eat-dog world; it's everyone for themselves. If you don't look out
for yourself, no one else will.

8.* Once a thief, always a thief.

12."' Two words of advice on friends: Beware of false finds, and never
tell your best friend a secret.

15. A known criminal should be considered guilty until proven
innocent.

17.* Capital punishment is the only language criminals understand.
19.* Never trust a stranger.

23.‘ We need jails and prisons because people have an inclination toward
evil. No matter how you look at it, some people are born criminals.

26.‘ When the chips are down, most people will walk all over other
people if it will help them achieve their own goals. People can be
counted on to act in their own self-interest.

29.

181

Pe0ple are born as sinful creatures, and humans have a natural
inclination to follow after evil.

MD

1.*

ll.’

There are both evil people and good people in the world, and you
have to check people out to find out which they are. Only in that
way can you know which ones to associate with.

Each person's character is like a blank page. It waits to be written
upon, by the experiences each individual encounters in life. These
will ultimately differentiate prisoner from jailor.

By their fruits, you shall know them.
When your children are young, it pays to take time to give them

your attention and to take the effort to train them in the right ways.
Left alone, they can go wrong just as easily as right.

13. Birds of a feather flock together.

16.* To trust people, you have to get to know them first.

21 .* People, like coins, have two sides - good and evil.

24.‘ There will always be people who will extend a hand to help, and
there will also be those who will try to chop yours ofl’.

25.‘ People are not born "good" or "e ' "; instead, they are born like a
blank slate, and given certain experiences and the effects of external
forces, eventually develop into good or evil beings.

28. Some people are best suited to war and others to peace.

GOOD

2.* Most people are genuinely interested in the well being of their

fellow humans.

4. Human beings have been created in God's image.

182

9.‘ Every single person has within themselves the potential to do
great things.

10.* Deep down, I really believe that people are good at heart.

14.‘ As a general rule, I'm willing to trust people and give them a chance
until, through wrong actions, they prove themselves unworthy of my
trust.

183' A person should always be considered innocent until proven guilty.

20. I've never met a person I didn't like.

22.* There's always someone who will lend a helping hand when you're
in need.

27 3" Children need guidance and support from parents and adults, but
even at a very young age, they are generally capable of
distinguishing right from wrong.

30. Love conquers all.

183

B. W
W (and to the

supematural)?

S UBJUGA T ION

31 ." Life is generally determined by external forces, such as God, fate, or

36.‘

37.‘

40.

44."I

48.

50.*

54.‘

563*

60.*

genetics, over which people have little control.

A person cannot surpass the conditions life has set. What will be
will be.

We shall have a good harvest only if God sends the rains.

A person made humble by acknowledging one's own inferiority to
the elements is much more powerful than one who challenges them.

People should never be so presumptuous as to think that they are in
control of their own life.

Our parents have taught us the way their parents taught them. Their
methods have always worked in the past, and it honors their
memory to continue to use them.

People are but puppets on the great stage of life. We must accept, as
graciously as we can, the conditions life has set for us.

Catastrophes are a part or' life; they can't be avoided. If it's our
destiny, they will come, and nothing can stop them.

Life is like a leaf floating on a stream. It follows the cru'rent, and
sometimes it goes under.

My life is in God's hands.

184

HARMONY

33 ."‘ People should, in every way, attempt to live in complete harmony
with nature.

35.‘ Let nature be your teacher.
39.* People should never do anything to pollute their precious earth.

42."‘ Each person is but one component of nature and should, at all times,
respect the integrity of all other forms of life.

43. Humans and nature are interdependent.
46.‘ Human beings and nature in combination make a powerful team.
49.‘ We must dance to the song of the earth.

53.* People should always remember that we came from the earth, and
ultimately, we will retm'n to it.

55.‘ We are all roommates on Planet Earth, so let's take care of our
home.

58.* Life flows naturally in your direction when you are in balance with
yom'self and with nature. When you're one with nature, you're one
with yourself.

Ill/1ST ERY

32.‘ All natural resources were placed on this earth to be at people's
disposal.

34.‘ Humans are nature's greatest accomplishment, and people have
rightly been assigned the task of controlling and perfecting name.

3 8.* As people have learned more and more to conquer nature and reap
nature's treasures, the more comfortable life has become.

41. Each person is the Captain of their Ship, the Master of their Fate.

45. I predict that people will have overcome cancer within ten years'
time.

_._.._

185

47."' "Cloud seeding" to produce rain is one of the most important
innovations of our time. Now it is possible to do something about
the weather instead of just complaining about it.

51.‘ The human challenge is to conquer and control nature. Everything
from air conditioning to the " green revolution" has resulted from our
having met this challenge.

523'“ Our highly technological society has reached its high state of
civilization by ever striving to control the forces of nature which, in
their raw, unbridled state, stand in the way of human advancement.

57.‘ The human drive to control the forces of nature has led to the
extensive development of science and technology, and this has been
a good thing for humankind.

59.‘ Even "natural" calamities, such as volcanoes, could be controlled if
only we would apply enough efl‘ort, intelligence and money to this

end. It's only a matter of time until we can control all of nature for
our own good.

CW

1M . I If: [I ll?

PAST
63.* People should be guided by history and attempt to copy the past.

66. Mothers and grandmothers have been taking care of children for
thousands of years and doing a fine job of it. It would be better to
maintain this tradition than to place our children in day care centers
that are taught by strangers with degrees.

67.* Children should model their lives after those of their parents and
grandparents and after the heroes of our society.

72. The wisdom and sacrifices of our ancestors have made us a great
nation.

74.

77.‘

8O.‘

83.‘I

87.‘

89.

186

I look back with fondness on the days of my childhood. Those were
the happiest days of my life.

Let's build on our past, not destroy it. Change is destructive, for it
kills the traditions of our proud past.

Never question your elders. Follow their advice and you won't go
wrong. Tradition is the foundation of youth.

There are no real heroes today. For the really great men and women,
we must look to our golden past.

Since we've always done it that way, why change now? What was
good enough for my parents is good enough for me.

A person's roots are very important. Knowing who you were helps
you know who you are.

PRESENT

61."I

64.‘

68.‘

71.‘l

73."I

78.‘

81.‘

82.‘

86.’

88.‘

Don't worry about tomorrow; enjoy today.

Iwasbominthepast,andlwill die sometimeinthe future, but for
right now, I'm living for today.

Look to this day, for yesterday is but a dream and tomorrow is only
a vision.

Eat, drink, and be merry, for tomorrow you may die.
The present moment is everything. Let's make the most of it.
Live every day as if it were the only day that counts.

Be here now; for tomorrow is uncertain and yesterday is but a
memory.

Take time to stop and smell the flowers as you go through life. Live
every day to the fullest

Today is a new beginning for the rest of your life.

There is, in reality, no "yesterday" and no "tomorrow." Therefore, it
is important to experience today.

187

FUTURE

62.‘ Planning and goal-setting make it possible for people to accomplish
miracles.

65.‘ Plan your life. Set realistic yet challenging goals, and then spend the
rest of your life achieving them.

69."I Build a foundation today for a better tomorrow. It pays to scrimp
and save now so you will have something to live on in yom' old age.

70. Life flows in only one direction; therefore, you must keep up with
the flow.

75.‘ If you don't move forward, you're moving backwards. He who does
not look ahead is condemned to remain behind.

76.’ Wise people map out their plans for the future; they know what they
want to be doing five, ten and twenty years fiom now.

79.‘ Tomorrow will be a better day; the best is yet to come, so look to
the future—it promises more than the past.

84." The ability to look ahead and plan for the future makes humans
difi’erent from other animals.

85.‘ You should plan ahead for the unexpected, by putting aside a little
money for a "rainy day~.

90.‘ We should plan in such a way so as to conserve our resources for
future generations.

HACIDZIIIQRIENIAIIQN

WWW?
BEING

91. It's not necessary to accomplish great things in life to feel your life
has been worthwhile. It's enough to just "be".

96.‘ Relax, be yourself, and enjoy life as it comes.

188

97. Feelings, emotions, and the interpersonal relationships between
people are more important than anything else in life.

101.* To love is better than to do; to be is better than to have.

105* Every person, by definition, is a significant, worthwhile being,
whether they actually contribute anything to society or not.

107.‘ People's importance should not be judged by any acts they
perform. Rather their importance stems from their mere existence.

109.‘ Being a good person is more important than achieving success.
1 l4.* Being happy and contented is what life is really all about.

1 15. I just want to lead a quiet, simple life. I don't want to be rich or
famous.

119.‘ What a person is is more important than what they do.
BEING-IN-BECOAHNG
92.‘ The main goal in life is self-enlightenment.

94. Those who learn to live with loneliness come to know themselves
and to understand life in the process.

99." The happiness that people search for can only be found within
themselves.

100."' Know thyself.

104." It's more important to pay attention to your inner development than to try to " get
ahead" in life.

106. Everything I need to know to survive, I already have.
111. Do not look outward, but concern yourself only with the world within.
113." Self-actualization should be one of every person's primary goals in life.

116.‘ People's main purpose for being placed on this earth is for their own inner
development.

118. Meditation will enable a person to gain all of life's secrets.

189

DOING

93. If people work hard and apply themselves fully, their efforts will be
rewarded in material ways.

95.‘ To achieve anything, you have to at least make an effort. Anything
worth having is worth working for.

98. The happy person is the one who never sits still but who is
constantly involved in productive activity.

102.* You should continually work and strive to make this world a
better place in which to live.

103. Get out there and fight, and fight to win!

108.‘ Hard work and sacrifice are always rewarded.

1 10."' Establish your priorities; implement them; and results will occur.
112.’ Set your goals high and work hard to achieve them.

117.* Work is a virtue in and of itself. Hard work never hurt anyone.

120.‘ The best life is the active, productive life.

190

E-SQCIALDRIENIATIQN

“.1 l'l'E'l"lll I?

LINEAL

121. Some people are born to lead others. There are "leaders" and there
are "followers" in this world.

125." In times of difficulty, it's best to go to someone who has the power
to change the situation and implore them to help you.

129. Each person comes into this world with a different set of talents and
abilities. Each person must accept their "gifts" and limitations and
fit into their rightful place in society.

131.‘ The father (and/or mother) should be the "head" of the household;
every member of the family should show them the proper respect
and never question their authority.

134.‘ It is more efficient for decisions to be made by top management
and passed on down through the "chain of command".

138.‘ Never question your teachers. They know what is right.

140.‘ A good wife should be respectful and submissive to her husband.
She should follow his judgments, and decisions so that he can lead
the family.

l44."' Respect and obey your elders, for they are wise and their wisdom,
if followed, will serve us well.

146. It is important to make decisions which will meet the approval of
our elders.

149.‘ One should follow the requests of one's superiors without
questioning their authority.

191

C OLLA T ERAL

123.‘ Whenever I have a serious problem, I like to get my family or
close friends to help me solve it.

124.‘ The most satisfying and effective form of decision-making is
"group consensus". .

128.’ Management decisions which are made by consensus are always
better received.

130.‘ Thank God we don't have to stand on our own. We have the group
to support us and to sustain us.

135. Workers will achieve higher levels of productivity if they are part of
the decision-making process.

137. A person's life will be happier if it is spent in close proximity to
one's parents and relatives.

139. A married couple will be happiest if they discuss things and jointly

decide on a mode of action (which may include compromises but)
which will be to their mutual benefit.

142. What is not good for the swarm is not good for the bee.

145.‘ It is better for older children to live with their parents until they
many so they will always be part of the family.

147." Two heads are better than one.

HVDIVID UAL

122.‘ Ultimately, it is up to each individual to decide what is best for
themselves.

126. Any society which does not allow individuals to voice their dissent
is not a free society.

127. The creative tension which competition provides is healthy for an
organization or for a society. Competition brings out the best in
each individual.

192

132. People are flee to act as they see fit, and if their action doesn't have
a negative impact on others, they don't even need to let anyone else
know what they decided to do.

133. Civil disobedience is justifiable when the government's actions
conflict with one's own principles or one's own conscience.

136. Each person, one vote. Your vote counts as much as anyone else's.
141. All people are created equal.

143! People should be allowed to choose their own lifestyle and
develop their own unique identity.

148.‘ Every person should be encouraged to develop themselves to the
fullest extent, in whatever way they see fit.

150.* Dare to be different, and always be yoru'self.

 

" indicates item was included in final scale

193

Random Response Items:
Hough (1990)

1. If you wanted to fly an airplane, you would probably join:
A) The railroad

B) The airlines

C) A trucking company

2. Tomato plants are generally smaller than trees:
A) True

B) Not sure

C) False

3. If you were to test an experimental airplane, you would want to take along:
A) A map

B) A parachute

C) A flashlight

4. Groups of cub scouts are called:
A) Tribes

B) Troops

C) Classes

5. The branch of the Federal government that deals most with problems of employment
is the:

A) FBI

B) Department of Labor

C) Department of State

6. Have you ever gone to a doctor?
A) Yes

B) Not sure

C) No

7. Have you ever bought food in a restaurant or fast food place?
A) Yes

B) Not sure

C) No

194

Social Desirability Items:
Balanced Inventory of Desirable Responding (BIDR)
Paulhus (1984,1988)

Instructions: Indicate your agreement to each statement using the following scale:

1

2 3 4 5 6 7

Strongly Strongly
disagree agree

99°99‘99pr

My first impressions of people usually turn out to be right.

It would be hard for me to break any of my bad habits.

I don't care to know what other people really think of me.

I have always been honest with myself.

I always know why I like things.

When my emotions are aroused, it biases my thinking.

Once I've made up my mind, other people can seldom change my opinion.
Iamnotasafe driverwhenlexceedthespeedlimit.

I am fully in control of my own fate.

. It's hard for me to shut off a disturbing thought.

. I never regret my decisions.

. I sometimes lose out on things because I can't make up my mind soon enough.
. The reason I vote is because my vote can make a difference.

. My parents were not always fair when they punished me.

. I am a completely rational person.

I rarely appreciate criticism. .

. I am very confident of my judgments.

I have sometimes doubted my ability as a lover.

. It's all right with me if some people happen to dislike me.

. I don't always know the reasons why I do the things I do.

. I sometimes tell lies if I have to.

. I never cover up my mistakes.

. There have been occasions when I have taken advantage of someone.

I never swear.

. I sometimes try to get even rather than forgive and forget.

. I always obey laws, even if I'm tmlikely to get caught.

. I have said something bad about a friend behind his or her back.

. When I hear people talking privately, I avoid listening.

. I have received too much change flour a salesperson without telling him or her.
. I always declare everything at customs.

. When I was young I sometimes stole things.

. I never have dropped litter on the street.

. I sometimes drive faster than the speed limit.

. I never read sexy books or magazines.

35.
36.
37.
38.
39.
40.

195

I have done things that I don't tell other people about.

I never take things that don't belong to me.

I have taken sick-leave from work or school even though I wasn't really sick.

I have never damaged a library book or store merchandise without reporting it.
I have some pretty awful habits.

I don't gossip about other people's business.

196

Demographic Information Sheet
What is your age?
For each of the following, please circle your response.

What is your sex?
A) Female
B) Male

What is your race/ethnic group?

A) American Indian or Alaskan Native
B) Asian or Pacific Islander

C) Hispanic

D) Black (not of Hispanic origin)

E) White (not of Hispanic origin)

The place in which you spent the most time during your early life was a:
A) town of less than 2,000 people

B) town of between 2,000 and 10,000 people
C) city of 10,000 to 100,000 people

D) city larger than 100,000

How much formal education did your father receive?
A) Grade school or less

B) Some high school

C) High school graduate

D) Some college

E) Four-year college graduate

F) Advanced degree: M.S., M.B.A., Ph.D., etc.

—’i—

 

197

How much formal education did your mother receive?
A) Grade school or less

B) Some high school

C) High school graduate

D) Some college

E) Four-year college graduate

F) Advanced degree: M.S., M.B.A., Ph.D., etc.

What was your family's approximate anmnl household income while you were
gmwins up?

A) less than $20,000

B) $20,000 - $34,999

C) $35,000 - $44,999

D) $45,000 - $59,999

E) $60,000 - $80,000

F) Greater than $80,000

What was the cmmtry in which each of the following was m?
If you do not know the birthplace of one of these individual's, please respond "don't
know".

a) You?

 

b) yom' mother?

 

c) your father?

 

(1) your maternal grandfather?

 

e) your maternal grandmother?

 

f) your paternal grandfather?

 

g) your paternal grandmother?

 

198

If born in the United States, what was the state in which each of the following was
born?

If you do not know the birthplace of one of these individual's, please respond "don't
know".

a) You?

 

b) your mother?

 

c) yoru' father?

 

d) your maternal grandfather?

 

e) your maternal grandmother?

 

1') your paternal grandfather?

 

g) your paternal grandmother?

 

How long have you lived in the Midwest? years

I99

Subjective Employment Test Items & Their Hypothesized Relationship to Value
Orientation Alternatives

Snalcfindfi;

CLHI: Catalog of Life History Items (Owens, Glennon & Albright, 1966)
ABLE: Assessment of Background and Life Experiences (U .8. Army Research
Institute)

NEO-PI: NEO-PI-R

BM-C: Conscientiousness Scale (Barrick & Mount)

TDI: Transport Driver Inventory

Sales: Sales Aptitude Test

MBI: Rationally-keyed Biographical Data Items for Constructs Related to Sales
Performance (Mumford, 1993)

DRA: Biodata items from Domm-Richardson Associates

RGG: Russell, Green, & Griggs

MMPI: Minnesota Multiphasic Personality Inventory

F BI: Biodata items of the Federal Bureau of Investigation

200

Human Nature Orientation Items:
(-) indicates Good

(#) indicates Mixed

(+) indicates Evil

(CLHI)

What has been your experience with people?

1.
2.
3.
4.

(CLHI)

There is a lot of good in all people. (-)

There is some good in most people. (#)

People are about as good as they have to be. (+)

A smprising number of people are mean and dishonest. (+)

Which of the following statements best expresses your feelings concerning the
proctoring of college examinations?

l.

2.

3.

4.

Examinations should be closely proctored because few students are
completely honest in all situations. (+)

Examinations should be closely proctored although most students are
honest, a few need to be watched carefully. (#)

Close proctoring is not necessary, since cheating is not really much of a
problem. (-)

The best way to handle this problem is by use of the honor system, in
which students themselves are responsibly for each other. (-)

(TDD
Which one of the following techniques of disciplining a child would you use most
frequently?
1. Denying the child some material pleasure
2. Encouraging the child by pointing out good behavior (-)
3. Leaving decisions up to the child after discussion
4. Trying to reason with the child
5. Punishing or spanking the child, letting him know why he is being
prurished (+)
(FBI)

I tend to be cynical and skeptical of others' intentions.

9:59.“?

Strongly agree (+)

Agree

Neither agree nor disagree
Disagree

Strongly disagree (-)

201

(FBI)
I believe that most people will take advantage of you if you let them.

1. Strongly agree (+)

2. Agree

3. Neither agree nor disagree
4. Disagree

5.

Strongly disagree (-)

Person-Nature Orientation Items:
(-) indicates Subjugated to Nature
(#) indicates Harmony with Nature
(+) indicates Mastery over Natm'e

(ABLE)

Success is usually a result of:
1. More hard work then luck (+)
2. Some hard work and some luck
3. More luck than hard work (-)

(ABLE)
How often do you feel that you have very little control over the things that happen to
you?

1. Often (-)

2. Sometimes

3. Rarely or never (+)

(ABLE)

Getting a raise or a promotion is usually a matter of luck. Do you agree?
1. Yes (-)
2. Not sure
3. No (+)

(ABLE)

What happens to you in the long run is mostly under your control. Do you agree?
1. Yes (+)
2. Not sure
3. No (->

(ABLE)

Do you think that hard work makes the difference between who succeeds and who
doesn't?

1. Most of the difference (+)

2. Some of the difference

3. Little of the difl‘erences (-)

202

(ml)

How much do you tend to react to an unpleasant situation?
1. Generally try to react immediately and figure out the best solution.

2. Most of the time I put off a decision for a little while so I can think it
over.
3. Often want to "sleep on it" or put off a decision for quite a while.
4. I don't worry about it, things will take care of themselves. (#)
(TDI)

How do you feel about your share of luck in life?

1. Have had nothing but bad breaks. (-)

2. Have had about an even share of luck.

3. Have had more than your share of bad breaks.

4. Have had more good breaks than bad ones.

5. Luck has been your way practically all the time.
(FBI)

When there are unexpected changes at work or school,
1. You usually feel upset, and your mood may be affected.
2. You just try to "deal with it".
3. You feel challenged by the new situation. (+)

(F 131)
If you have a m important meeting in a city one hour away, but the roads are
treacherous due to an ice storm, which of the following are you most likely to do?

1. Call to try to reschedule the appointment.

2. Leave early and drive with great caution to the appointment. (+)

3 Attempt to drive to the appointment, but turn around if the roads are too

4. Attempt to get a fiiend to drive you to the appointment

203

Time Orientation Items:
(-) indicates Past

(#) indicates Present

(+) indicates Future

(CLHI)
Concerning your present activities, do you
1. Make rather precise and detailed plans (+)
2. Make broad and general plans, but not detailed ones

3. Make few plans, let "nature take its course"
(CLHI)
If you visit a museum, how do you feel?

1. Bored

2. Thoroughly interested (-)

3. Neither bored nor interested

4. Don't go to museums

5. It's a good place to take children
(ABLE)
How often do you lose sleep over your worries?

l. Often (+)

2. Sometimes
3. Rarely or never

(ABLE)

You believe in:
1. Living for the moment (#)
2. Not sure .
3. Working for the future (+)

(MEI)
Relative to your coworkers how much time have you spent preparing for meetings?
1. Much more time than most people (+)
2. More time than most people
3. Less time than most people
4. Much less time than most people

—a ‘-

 

 

204

(MED
How often have you been late for meetings?
1. Very often
2. Often
3. Sometimes
4. Seldom
5. Never (+)
(MEI)
How likely have you been to recognize that you do not have a piece of equipment or part
before you actually need to use it?
1. Much more likely than most people (+)
2. Somewhat more likely than most people
3. About as likely as others
4. Somewhat less likely than most people
5. A good bit less likely than most people
(MBI)
To what extent did you limit certain career choices because you thought you wouldn't do
very well?
1. Great extent (+)
2. Large extent
3. Moderate extent
4. Slight extent
5. Not at all
(FBI)

When working on an assignment with a deadline, what do you normally do?

1.

2.

3.

4.

(FBI)

Work only on the assignment with a deadline to get it done before all
other assignments.

Wait until the deadline approaches and then work feverishly on the
assignment to get it done. (#)

Ask for an extension of the deadline in order to accomplish the
assignment.

Work on all assignments, keeping in mind the deadline.

When reviewing your schedule,

1.
(+)
2.
3.
4.

You like to have things written down on your calendar weeks in advance.

You like to have your schedule written out days in advance.
You think ahead, but keep most of the information in your head.
You rely on others to remind you of events or responsibilities.

_._i_

 

205

(FBI)
When buying gifts for friends or relatives, how early do you typically make the purchase
prior to giving the gift?
1. Purchase the gift following the occasion for which it was intended.
2. One day or less
3. Two to 6 days
4. One week or more (+)
(FBI)
Do you generally employ the use of a date book, daily planner, or other scheduling aid?
1. Yes (+)
2. No
(F 31)
At what time do you generally arrive at scheduled appointments?
1. A few minutes earlier than the schedule time. (+)
2. At the scheduled time.
3. A few minutes later than the scheduled time.
(FBI)
When you take a vacation, which do you prefer?
I. Like to plan it down to the last detail. (+)
2. Like to make general plans, but let the details take care of themselves.
3. Like to take spontaneous trips and recreation.
(FBI)

To what extent do you plan your day?
1. I do a lot of planning before starting each day. (+)
2. I plan my day in a general way, but don't spend much time at it. -
3. I don't do a lot of planning.

206

Activity Orientation Items:

(-) indicates Being

(#) indicates Being-in-Becoming
(+) indicates Doing

(CLHI)
What do you feel has been your major accomplishment, outside of work?
1. Family activities
2. Community activities
3. Development of yourself (#)
4. Development of your social activities

(CLHI)
How many hours a day do you spend in constructive work?
1. About 4
2. 5 or 6
3. 7 to 9
4. over 9 (+)
(CLHI)
What do you think is the most important thing a person should get out of college?
1. Training for a profession
2. General cultural knowledge
3. Personal maturity (#)
4. Social polish
(CLHI)

In the course of a week, which of the following gives you the greatest satisfaction?

1. Being told you have done a good job. (+)
2. Helping people solve their problems.
3. Being with your family and close fliends.
4. Having flee time to use as you please.
5. None of these

(CLHI)

How greedy disturbed are you if something is left unfinished?
1. Slightly
2. Moderately
3. Considerably
4- Highly (+)

207

(CLHI)
Fantasies are
1. Something to avoid
2. Something to enjoy (#)
3. Something to build constructively upon (+)

(CLHI)
Which of the following is most important to you?

1. Professional status or authority (+)
2. Money
3. Family and fliends
4. Religion
5. Recreation (-)
(CLHI)

How much do you enjoy work, as opposed to spare time activities?

1. Enjoy work much more (+)

2. Enjoy work more

3. About the same

4. Enjoy spare time activities more

5. Enjoy spare time activities much more (-)
(ABLE)

How often do you skip class for something more fun?
1. Many times (-)
2. A few times
3. Rarely or never

(ABLE)
How do you feel about hard work?
1. It makes you feel good (+)

2. You don't mind it now and then
3. You don't really like it

(ABLE)

Rewards will come to people who work for them. Do you agree?
1. Yes (+)
2. Not sure

3. No

A

 

208

(ABLE)
Planning things a lot ahead of time makes life boring. Do you agree?
1. Yes (-)
2. Not sure
3. No
(TDD
When you go on a vacation trip in your car and are going to as particular place, do you:
1. Dislike to make unnecessary stops until you get where you are going (+)
2. Plan to stop only at planned points along the way
3. Plan the trip so you can detour or stop whenever something interest you
4. Sometimes get sidetracked and don't get to your original destination
(TDD
Which one of these characteristics bothers you most in people you meet?
1. Bragging
2. Shyness
3. Lack of initiative (+)
4. Trying to get something for nothing
5. Being very competitive
(TDD

When you were a child and broke something flying to see "what makes it tick", what did
your parents do?
1. Became angry and punished you
2. Tried to explain to you that it was wrong, becoming angry only in certain
instances
3. Usually said little or nothing about it
4. Tried to help you find the answers you were looking for (#)

(TDI)
How upset do you get if you have to leave a chore unfinished?
1. It bothers me a lot (+)
2. I bothers me
3. I don't get too concerned about it
4. It really doesn‘t bother me

(TDD

What would be the most important factor in your recommending an employee for an
increase or promotion?

Ability to get the work out (+)

Quality of work or technical competence

Ability to get along with people

Ability to come up with new ideas

:“P’NS‘

209

(MBI)
How often have you held off talking about a problem because the timing wasn't right?
1. Very often
2. Often
3. Sometimes
4. Seldom
5. Never (-)
(MBI)

How often have you neglected other areas of your life in order to accomplish more at
work?

1. Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never
(MBI)
How often have you been upset at the end of a workday because you felt you hadn't
accomplished enough?
1. Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never
(MBI)

To what extent has your satisfaction with life depended upon what's been going on at
work or school?

1. Great extent (+)
2. Large extent
3. Moderate extent
4. Slight extent
5. Not at all

(MEI)

How often do you end up creating work for yourself to keep flom getting bored?
1. Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never

210

(MEI)

How often have you felt guilty when you took time off just to relax?
Very often (+)

Often

Sometimes

Seldom

Never

MPP’N?‘

(MBI)

Relative to others, how often have you found yomself comparing your achievements
with those of your fliends?

Very ofien (+)

Often

Sometimes

Seldom

Never

V‘PP’NS‘

(MBI)

How often have you read books just to learn something?
Very often (#)

Often

Sometimes

Seldom

Never

P‘PP’NE‘

(MBI)
How often have you chosen classes, project, or assignments simply to learn something
new?

1. Very often (#)
2. Often
3. Sometimes
4. Seldom
5. Never
(MB!)

How often have you attended cultural events when you were uncertain of whether or not
you would enjoy them?
1. Very often (#)
2. Often
3. Sometimes
4. Seldom
5 Never

211

(M31)
How comfortable have you been discussing your strengths and weaknesses with others?

1. Very comfortable (#)
2. Comfortable
3. Neither comfortable nor uncomfortable
4. Uncomfortable
5. Very uncomfortable
(DRA) -

How often did you make a "to do" list in school to make sure you accomplished
everything?

1 Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never
(RGG)
How often have you done something that made you really stop and think about what was
important in yom' life?
1 Very often (#)
2. Often
3. Sometimes
4 Seldom
5 Never

(FBI)
I don't always know the reasons why I do the things I do.
1. Very true (-)

2 Mostly true

3. Somewhat true

4 Mostly untrue

5 Very untrue (+)

(FBI)
How would your fliends describe you when faced with decisions?
1. As one who acts impulsively. (-)
2. As one who considers more than one alternative before acting.
3. As one who is very cautious and takes a long time to arrive at a decision.

212

(FBI)
On weekends, you typically
1. Plan your activities in general terms
2. Plan your activities specifically and draw up a schedule
3. Plan some activities, but leave most of the weekend for "spontaneous"
activity.(-)
4. Plan very few or no activities, simply do what you feel like at the time.
(-)
(FBI)
Do you generally do your best
1. At whatever job you are doing (+)

2. Only in what you are interested
3. Only when it is demanded of you

(F Bl)
When you are reading and come across a word you do not know, what do you usually
do?

1 Keep right on reading.

2. Immediately look it up in the dictionary. (#)

3. Sometimes look it up depending on the context in which it is used.

4 Make a mental note to look the word up at a later date.

(FBI)

When you have been away on vacation and it's about time to return, do you:
1. Stay as long as you can and sometimes don't start back as soon as you
should.
2. Feel ready to return although you don't leave until it is necessary.
3. Become anxious to get home and leave when you had originally planned. (+)
4. Often start home earlier than you actually have to.

(FBI)
If you were given more assignments than you could complete in a given period of time,
what would you do?

1. Complete as many assignments as possible in the given time. (+)

2. Prioritize the assignments according to importance, and then complete as

much as possible in the given time.
3. Ask a supervisor to decrease the workload to a manageable level.
4. Ask a co-worker to help you complete the assignments.

(FBI)

In college, which type of class did you enjoy most?
1. Lecture
2. Laboratory

3. Discussion (#)

L‘

213

(FBI)
How do you arrange your study time in college?
1. According to a regular, set schedule.

2. A fairly regular arrangement, but not according to any schedule.
3. Sporadic, whenever convenient and/or necessary. (-)

(FBI)

When working on simple tasks,
1. You like to think about solutions to other work-related problems.
2. You concentrate on the task at hand.

3. You like to daydream or think of things outside of work. (-)

(FBI)

When you take a vacation, which do you prefer?
1. Like to plan it down to the last detail.
2. Like to make general plans, but let the details take care of themselves.
3. Like to take spontaneous trips and recreation. (-)

(FBI)

To what extent do you plan your day?
1. I do a lot of planning before starting each day.
2. I plan my day in a general way, but don't spend much time at it.
3. I don't do a lot of planning. (-)

(FBI)
I don't like to waste my time daydreaming

1. Strongly agree (+)

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

(FBI)
I waste a lot of time before settling down to work.

1. Strongly agree (-)

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

214

(FBI)

1 work hard to accomplish my goals.
Strongly agree (+)

Agree

Neither agree nor disagree
Disagree

Strongly disagree

9993?)!“

(FBI)

I have little interest in speculating on the nature of the universe or the human condition.
Strongly agree

Agree

Neither agree nor disagree

Disagree

Strongly disagree (#)

P'PP’N?‘

(FBI)
1 am a productive person who always gets the job done.

1. Strongly agree (+)

2. Agree

3. Neither agree nor disagree
4. Disagree

5.

Strongly disagree

215

Relations Orientation Items:
(-) indicates Lineal

(#) indicates Communal

(+) indicates Individual

(CLHI)
What do you feel has been your major accomplishment, outside of work?
1. Family activities
2 Community activities (#)
3. Development of yourself
4 Development of your social activities

(CLHI)
Which of the following is most important to you?
1. Professional status or authority
2. Money
3. Family and fliends (#)
4. Religion
5. Recreation
(ABLE)
How often were you disrespectful to your high school teachers?
1. Often
2. Sometimes

3. Rarely or never (-)

(ABLE)

Do you respect people in positions of authority?
1. Yes, almost all of them (-)
2. Yes, some of them
3. Yes, but only a few of them

(ABLE)

How often have you gone against your parents'lguardians wishes?
I. Quite often (+)
2. Sometimes
3. Rarely or never (-)

(ABLE)

Do you ignore a rule if it gets in the way of what you want to do?
1. Yes, usually (+)
2. Yes, sometimes
3. No, never

216

(ABLE)
People should have greater respect for authority. Do you agree?
1. Yes (-)
2. Not sure
3. No
(NEO-PI)
I usually prefer to do things alone.
1. Strongly Agree (+)
2. Agree
3. Neither Agree nor Disagree
4. Disagree
5. Strongly Disagree
(NEO-PI)

I prefer jobs that let me work alone without being bothered by other people.

1. Strongly Agree (+)
2. Agree
3. Neither Agree nor Disagree
4. Disagree
5. Strongly Disagree
(TDI)
Who do you feel was most important in teaching you about life and how to live?
1. My parent(s)
2. My religious teachers
3. My close friends
4. My own observations and thinking (+)
(IDI)
As a young person when you did something wrong, whose praise did you value most?
1. A fliend
2. A teacher
3. Your parents
4. Someone else
5. Did things well for your own satisfaction (+)
(TDI)

What would be the most important factor in your recommending an employee for an
increase or promotion?

1. Ability to get the work out

2. Quality of work or technical competence

3. Ability to get along with people (#)

4. Ability to come up with new ideas

217

(Sales)

Would you rather play in a dance-band than be a soloist for the opera?
1. Yes (#)
2. No (+)

(MBI)

Relative to others, how much time do you devote to social service or charity
organizations?

1. Much more time than others (#)
2. More time than others
3. About as much time as others
4. Less time than others
5. Much less time than others
(MBI)
How often have you put in a lot of extra efi‘ort to buy some independence?
1. Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never
(MBI)
In previous jobs, to what extent have you tried to blend in and become part of the team?
1. Great extent (#)
2. Large extent
3. Moderate extent
4. Slight extent
5. Not at all (+)
(MBI)

How often have you gone along with a decision you did not agree with in order to
maintain group harmony?

1. Very often (#)
2. Often

3. Sometimes

4. Seldom

5. Never (+)

218

(MBI)

How important has it been for you to be in community/volunteer work?

1. Extremely important (#)

2. Very important

3. Somewhat important

4. Slightly important

5. Not very important
(MBI)

How often have you found yourself thinking about how the consequences of your actions
will affect others?

1. Very often (#)
2. Often
3. Sometimes
4. Seldom
5. Never
(RGG)

How often have you seen people put fliendships and personal goals aside to try and
achieve a greater good for the larger group?

1. Very often (+)
2. Often
3. Sometimes
4. Seldom
5. Never
(RGG)

How often have you hurt a fliend to benefit yourself?
1. Very often (+) '

2. Often

3 . Sometimes

4. Seldom

5. Never
(RGG)

How often have you been accused by others of doing something solely because it was in
your best interest?

1. Very often (+)
2. Often

3. Sometimes

4. Seldom

5. Never

219

(FBI)

How much time do you volunteer to help service groups in your community?
1. Very few hours monthly
2. Several hours monthly
3. Several hours weekly (#)

(FBI)
What do you consider to be the major motivating force in your life?
1. Prestige
2. Material gains
3. To invent something new
4. To gain a position of security
5. To help others (#)
(FBI)

How often have you reflained fl'om doing something because you thought that the action
might reflect negatively on other members of an organization of which you are a member
(e.g., work, school, religious group, etc)?

1. Often (#)
2. Sometimes
3. Rarely

4 Never

(FBI)
If you were given more assignments than you could complete in a given period of time,
what would you do?

1. Complete as many assignments as possible in the given time.

2. Prioritize the assignments according to importance, and then complete as

much as possible in the given time.
3. Ask a supervisor to decrease the workload to a manageable level.
4. Ask a co-worker to help you complete the assignments. (#)

(FBI)
To how many civic organizations (i.e., Circle K, CADA, Listening Ear, etc) do you
belong?

l. 0

2. l

3. 2 or 3

4. 4 to 6 (#)

5. 7 or more (#)

220

(FBI)
How would you react if your parents or a close relative asked you to come home for an
important family event when had planned a weekend trip with fliends?
1. Tell your parents you can't go because of the previous engagement. (+)
2. Tell your fliends you can't join them and go home. (-)
3. Try to reschedule the event with your friends.

(FBI)

Would your choice of an ideal occupation be one which
1. Allowed a great deal of interaction with other people. (#)
2. Would require working with a small group of people.
3. Would allow you to work closely with one other person.
4. Would allow you to work by yourself. (+)

(FBI)
When someone around you is disturbed about a personal problem, which one of the
following do you usually do?
1. Leave the person alone, avoid the subject.
2. Offer advice and suggest a possible solution.
3. Sympathize with the person.
4. Encourage the person to talk it out with you. (#)
(FBI)

Which of the following situations do you most prefer?
1. Being alone (+)
2. Being alone with pets
3. Being with one other person
4 Being in large groups of people (#)

(FBI)
Since high school have you ever received a commendation for civic involvement (i.e.,
good citizenship awards, community service awards, etc.)?

1. Yes (#)

2. No

(FBI)

The work accomplishments with which you have been most satisfied have been
1. Largely due to your own thought and effort. (+)
2. The result of team effort with your colleagues. (#)

221

(FBI)
If you attended a meeting, dinner or lecture (outside of school or work) where the
speaker said something with which your organization would disagree, but with which
you personally agree,
1. You would express your agreement. What you do on your own time is your
business. (+)
2. You would express your agreement, but hope no one from your
organization was there to witness it.
3. You would not react. You feel that you represent your organization at all
times. (#)

(FBI)
If you had to work on a very important project at work,
1. You would concentrate all of your energy on it, even if that meant missing
a deadline on a less important project.
2. You would minimally attend to your other work responsibilities realizing
that you could not give all your time to the more important project.
3. You would attempt to divide up your time equally.
4. You would ask your supervisor/boss to find someone else to help you.

(#)

(FBI)
I would rather cooperate with others than compete with them.

1. Strongly agree (#)

2. Agree

3. Neither agree nor disagree
4. Disagree

5.

Strongly disagree (+)

(FBI)

I believe we should look to our religious authorities for decisions on moral issues.

1. Strongly agree (-)

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

APPENDIX B

Appendix B

2. What has been your experience with people? (Human Nature: Good)

1. There is a lot of good in all people.
2. There is some good in most people.
3. People are about as good as they have to be.
4. A surprising number of people are mean and dishonest.
3. Which of the following statements best expresses your feelings concerning the

proctoring of college examinations? (Human Nature: Evil)

1. Examinations should be closely proctored because few students are
completely honest in all situations.

2. Examinations should be closely proctored although most students are
honest, a few need to be watched carefully.

3. Close proctoring is not necessary, since cheating is not really much of a
problem.

4. The best way to handle this problem is by use of the honor system, in
which students themselves are responsible for each other.

6. On weekends, you typically (Activity: Doing)

1. Plan your activities in general terms

2. Plan your activities specifically and draw up a schedule

3. Plan some activities, but leave most of the weekend for "spontaneous"
activity.

4. Plan very few or no activities, simply do what you feel like at the time

11. I tend to be cynical and skeptical of others' intentions. (Human Nature: Evil)

1. Strongly agree

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

15. When reviewing your schedule, (Time: Future)
You like to have things written down on yotu' calendar weeks in advance.
You like to have your schedule written out days in advance.
You think ahead, but keep most of the information in your head.
You rely on others to remind you of events or responsibilities.

99°F!"

222

223

21. To what extent do you plan your day? (Time: Future)
1. I do a lot of planning before starting each day.
2. I plan my day in a general way, but don't spend much time at it.
3. I don't do a lot of planning.

25. If you visit a museum, how do you feel? (Time: Past)

1. Bored
2. Thoroughly interested
3 . Neither bored nor interested
4. Don't go to museums
5. It's a good place to take children
26. How many hours a day do you spend in constructive work? (Activity: Doing)
1. About 4
2. 5 or 6
3. 7 to 9
4 over 9
28. People should have greater respect for authority. Do you agree? (Relations:
Lineal)
1. Yes
2. Not sure
3. No

33. How do you feel about hard work? (Activity: Doing)
1. It makes you feel good
2. You don't mind it now and then
3. You don't really like it

47. I believe we should look to our religious authorities for decisions on moral issues.
(Relations: Lineal)

1. Strongly agree
2. Agree
3. Neither agree nor disagree
4. Disagree
5. Strongly disagree
59. Which one of these characteristics bothers you most in people you meet? (Activity:
Doing)

1. Bragging

2. Shyness

3. Lack of initiative

4. Trying to get something for nothing
5 Being very competitive

63.

68.

71.

73.

75.

224

How greatly disturbed are you if something is left unfinished? (Activity: Doing)

1.
2.
3.
4

Slightly
Moderately
Considerably

Highly

How often have you chosen classes, project, or assignments simply to learn
something new? (Activity: Being-in-Becoming)

1
2.
3.
4
5

Very often
Often
Sometimes
Seldom
Never

I work hard to accomplish my goals. (Activity: Doing)

.V'PP’NS"

Strongly agree

Agree

Neither agree nor disagree
Disagree

Strongly disagree

How likely have you been to recognize that you do not have a piece of equipment or
part before you actually need to use it? (Time: Future)

9:591"?

I

am
1.
2.
3
4
5

Much more likely than most people
Somewhat more likely than most people
About as likely as others

Somewhat less likely than most people
A good bit less likely than most people

a productive person who always gets the job done. (Activity: Doing)

Strongly agree

Agree

Neither agree nor disagree
Disagree

Strongly disagree

Yes, almost all of them
Yes, some of them

Do you respect people in positions of authority? (Relations: Lineal)
1.
2.
3.

Yes, but only a few of them

225

77. I believe that most people will take advantage of you if you let them. (Human

Nature: Evil)
1. Strongly agree
2. Agree
3. Neither agree nor disagree
4. Disagree
5. Strongly disagree
87. You believe in: (Time: Present)
1. Living for the moment
2. Not sure

3. Working for the future

96. Do you generally employ the use of a date book, daily planner, or other scheduling
aid? (Time: Future)
1. Yes
2. No

l 10. Which of the following situations do you most prefer? (Relations: Collateral)
1. Being alone
2. Being alone with pets
3. Being with one other person
4. Being in large groups of people

APPENDIX C

Appendix C

2. What has been your experience with people? (Human Nature)

1. There is a lot of good in all people.
2. There is some good in most people.
3. People are about as good as they have to be.
4. A surprising number of people are mean and dishonest.
3. Which of the following statements best expresses your feelings concerning the

proctoring of college examinations? (Human Nature)

1.

2.

3.

4.

Examinations should be closely proctored because few students are
completely honest in all situations.

Examinations should be closely proctored although most students are
honest, a few need to be watched careflrlly.

Close proctoring is not necessary, since cheating is not really much of a
problem.

The best way to handle this problem is by use of the honor system, in
which students themselves are responsible for each other.

6. On weekends, you typically (Activity)

1.
2.
3.

4.

Plan your activities in general terms

Plan your activities specifically and draw up a schedule

Plan some activities, but leave most of the weekend for "spontaneous"
activity.

Plan very few or no activities, simply do what you feel like at the time

8. Which of the following is most important to you? (Relations)

9:593"?

Professional status or authority
Money

Family and friends

Religion

Recreation

226

10.

ll.

12.

18.

20.

23.

33.

227

Which one of the following techniques of disciplining a child would you use most

frequently? (Human Nature)

1 Denying the child some material pleasure

2 Encouraging the child by pointing out good behavior

3. Leaving decisions up to the child after discussion

4 Trying to reason with the child

5 Punishing or spanking the child, letting him know why he is being
punished

I tend to be cynical and skeptical of others' intentions. (Human Nature)

1. Strongly agree

2 Agree

3 Neither agree nor disagree
4. Disagree

5 Strongly disagree

If you have a m important meeting in a city one hour away, but the roads are

treacherous due to an ice storm, which of the following are you most likely to do?

(Person-Nature)

1. Call to try to reschedule the appointment.

2. Leave early and drive with great caution to the appointment.

3. Attempt to drive to the appointment, but turn around if the roads are too
rcy.

4. Attempt to get a fliend to drive you to the appointment

When you take a vacation, which do you prefer? (Time)

1. Like to plan it down to the last detail.
2. Like to make general plans, but let the details take care of themselves.
3. Like to take spontaneous trips and recreation.

Would you rather play in a dance-band than be a soloist for the opera?

(Relations)
1. Yes
2. No

Concerning you present activities, do you (Time)

1. Make rather precise and detailed plans
2. Make broad and general plans, but not detailed ones
3. Make few plans, let "nature take its course"

How do you feel about hard work? (Activity)

1. It makes you feel good
2. You don't mind it now and then
3. You don't really like it

40.

42.

47.

53.

57.

228

How often have you been upset at the end of a workday because you felt you
hadn't accomplished enough? (Activity)

1. Very often

2. Often

3. Sometimes

4. Seldom

5. Never

How often do you end up creating work for yourself to keep from getting bored?
(Activity)

1. Very often

2. Often

3. Sometimes

4. Seldom

5. Never

How often have you felt guilty when you took time off just to relax? (Activity)
1. Very often

2. Often

3. Sometimes

4. Seldom

5. Never

I believe we should look to om- religious authorities for decisions on moral issues.

(Relations)

1. Strongly agree

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

When you are reading and come across a word you do not know, what do you
usually do? (Activity)

Keep right on reading.

Immediately look it up in the dictionary.

Sometimes look it up depending on the context in which it is used.
Make a mental note to look the word up at a later date.

PP’NE‘

In college, which type of class did you enjoy most? (Activity)
1. Lecture

2. Laboratory

3 Discussion

58.

69.

72.

74.

77.

229

When you go on a vacation trip in your car and are going to as particular place, do

you: (Activity)

1. Dislike to make unnecessary stops until you get where you are going

2 Plan to stop only at planned points along the way

3. Plan the trip so you can detour or stop whenever something interest you
4 Sometimes get sidetracked and don't get to your original destination

To how many civic organizations (i.e., Circle K, CADA, Listening Ear, etc) do
you belong? (Relations)

1. 0

2. l

3. 2 or 3

4. 4to 6

5. 7 or more

I have little interest in speculating on the nature of the universe or the human
condition. (Activity)

1. Strongly agree

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

How often were you disrespectful to your high school teachers? (Relations)
1 Many times
2. A few times
3 On rare occasions
4 Never

I believe that most people will take advantage of you if you let them. (Human

Nature)

1. Strongly agree

2. Agree

3. Neither agree nor disagree
4. Disagree

5. Strongly disagree

83.

87.

89.

94.

95.

99.

230

How often have you put in a lot of extra effort to buy some independence?

(Relations)
1. Very often
2. Often
3. Sometimes
4. Seldom
5. Never

You believe in: (Time)

1. Living for the moment
2. Not sure

3. Working for the future

How important has it been for you to be in community/volunteer work?
(Relations)

1 . Extremely important
2. Very important

3. Somewhat important
4 Slightly important

5 Not very important

How often have you seen people put fliendships and personal goals aside to try
and achieve a greater good for the larger group? (Relations)

1. Very often
2. Often

3 . Sometimes
4. Seldom

5. Never
How often have you htut a fliend to benefit yourself? (Relations)
1. Very often
2. Often

3. Sometimes
4. Seldom

5. Never

How often have you been accused by others of doing something solely because it
was in your best interest? (Relations)

1. Very often
2. Often

3. Sometimes
4. Seldom

5. Never

100.

103.

104.

106.

110.

112.

231

How much time do you volunteer to help service groups in your commrmity?

(Relations)

1. Very few hours monthly
2. Several hours monthly
3. Several hours weekly

4. Very little or not at all

When you were a child and broke something trying to see "what makes it tick",

what did your parents do? (Activity)

1. Became angry and punished you

2. Tried to explain to you that it was wrong, becoming angry only in certain
instances

3 Usually said little or nothing about it

4. Tried to help you find the answers you were looking for

How upset do you get if you have to leave a chore unfinished? (Activity)
1. It bothers me a lot
2. I bothers me
3. I don't get too concerned about it
4. It really doesn't bother me

What happens to you in the long run is mostly under your control. Do you agree?

(Person-Nature)

1. Yes

2. Not sure

3. No

Which of the following situations do you most prefer? (Relations)
1. Being alone

2. Being alone with pets

3. Being with one other person

4. Being in large groups of people

Since high school have you ever received a commendation for civic involvement
(i.e., good citizenship awards, community service awards, etc.)? (Relations)

1. Yes

2. No

 

114.

232

If you had to work on a very important project at work, (Relations)

1. You would concentrate all of your energy on it, even if that meant missing
a deadline on a less important project.

2. You would minimally attend to your other work responsibilities realizing
that you could not give all your time to the more important project.

3. You would attempt to divide up your time equally.

4. You would ask your supervisor/boss to find someone else to help you.

APPENDIX D

 

Red-fie...

3.3+
. 333+

 

 

 

8.28 3.8.3.; .223

.. Q..-

I
.po

«6

 

was

r. 0.0

4 . 0.9

 

233

234

 

 

 

 

. 6.”???
Ti .
. .651?» :3
0.0
f...
30
aegis-3....

8.2.0 8.193.... 1.3.8

235

 

 

3622?.»
6E.—sail!
8.3+,

 

8.9.0 glac> .2330

 

236

 

_

>0 .m. if..-
Bill...
Baal-

 

 

 

 

 

 

.
:2.

:3

 

 

237

 

L

.. v.0

Tum

 

 

22..
"2.86 38.5....
.228 £928.. 35......

 

 

238

W ..

 

 

352..-?
.65.-I...
823+,

 

 

 

 

239

 

II
It
I
I

 

 

 

. 35...?»
Bil-r

‘ B s \

 

 

 

 

 

 

 

241

 

 

6 g 5.1...
>0

5+
33+,

 

 

’I’O’O
IO’IO’O
’0‘

 

 

.228 5.192....» .223

 

..-6
..«6
..Q6
..Q6
.1 0.0

..Eo

242

 

 

 

 

 

 

 

243

ii‘
+++

 

 

 

8.353883...
32.053.83.83".

 

‘ \11
“.“ .
-‘
“o‘
‘R
‘8...
“
‘0 ‘0‘
" “

Ar ‘.°

 

 

 

 

 

 

 

 

 

246

Film“.

 

 

5.3.3.8..-
6.58:1...
. as ‘

 

 

 

go’oo‘o£

 

 

 

 

 

248

 

 

 

_

5.3.1.1.,
5595+
633+,

 

ii

 

 

249

 

>0 52...?»
653+
- 6.5+,

 

 

 

 

.r 5.0

250

 

 

 

arm-fit?
63.31.!
5.5+!

 

 

 

251

 

.65-1.1...
333+

 

 

div

 

j

 

 

 

Ea
3.2.5 8.823...
.398 8:83.; .233

grqo

 

252

 

 

‘
‘Is
5“
‘\
sy“\s|
“
Q‘
“
‘\

 

5+

BELT,
>0
‘ 3.3+

 

"21.58.85...
8.28 8.193.; .223

OO’OOIIC
’
ocioopfn.‘ .....
uo‘utgcc’:
00’.-

 

:«d

: .6

L. 06

gr boa

 

253

 

:v6

..—6

. N6

L. 3

 

 

 

_

352...?»
BEG-2+
6877‘

 

O
..uo

mm

0
..oo

. Ed

..90

 

gr .oo

254

 

 

_

>0 3.3+
633‘}

 

 

3 3.0
8.120 5.828...
.328 3.8.8.; .2230

 

 

APPENDIX E

r:

 

 

 

 

, ,
1.3+

 

 

 

 

 

 

 

‘ ,
3.3+

 

 

 

 

257

£+

1285193..

 

258

 

 

 

 

 

 

 

3285195..

 

 

260

 

 

8.398293:

262

 

 

 

 

0.03

{83:93

 

 

 

3
6
2
3+
4.2: o ‘

 

 

 

0.82.0 5.10.80:

 

gfl

 

 

 

 

. 5.8.8....
.226

 

 

 

265

 

.r 3

.vqg

t. 0.0

At '.°

 

 

 

8.2.0 sci—0.8.2

 

93......
I; I

 

 

 

 

 

‘ ‘
3.3+

 

 

 

 

 

 

 

mum

 

 

 

“25519.8:

 

269

 

 

 

guitar
x£ ‘

 

“228 8.198.:

.. 3.6

 

.. 3.0

4‘ v.0

 

. . uvd

 

5 A
+MI

271

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

274

 

 

3+4
1.! ‘

 

 

 

 

 

 

275

 

 

 

 

 

 

 

 

 

276

 

 

 

 

nil-III
853+-

 

53
8.28 5.198....

 

 

278

 

 

 

 

hoe-n

8.2.0 .50—10.8.:

 

 

 

279

 

H

fi

 

 

8.2.0 :3:—6.8.3

 

 

 

 

 

 

 

 

 

 

 

 

quill..-
1.3+

 

3208.198...

 

 

 

 

 

I
..-o

O
:20

At “.0

282

«w

I
:00

I
:80

 

t v.6

 

I
:30

 

tr “.0

“2295298..

 

283

 

 

 

 

 

sail-I—
lilo}

 

 

+fl~l
+m A

Rae. Opllon
Inna-won G :
Blot“

 

285

 

 

 

 

286

 

 

833
8.3.6 5:832.
8.26 eggs:

 

.3
:3
:3
:3
. no

r Ho

 

 

 

find

287

w

. v.0

0
inc

 

8.35 33.5....
8.28 8.198.:

0
too

288

 

 

 

 

 

289

 

3'30

3205198.:

 

LIST OF REFERENCES

LIST OF REFERENCES

Allport, G. W. (1961). W. New York: Holt,
Rinehart, & Winston.

Angoff, w. H. (1972, Sept.). W
W. Paper presented at the annual meeting of the American Psychological

Association, Honolulu.

Angofl‘, W. H. (1993). Perspectives on difi‘erential item functioning

methodology. In P. Holland and H. Wainer, eds. W.
Lawrence Erlbaum Associates: Hillsdale, NJ.

Astin, A. W. (1964). Criterion-centered research. Educationaljnd
BsxchologicalMeasnremem, 25 807-822.

Atolagbe, E. (1980). Further study on values of Black and White children: Are
they different? MW .42. 448-452.

Barrick, M. R. & Momt, M. K. (1991) The big five personality dimensions
and job performance: A meta-analysis. W, 55, 1-26.

Beutler, L. E., &Bergan,J. (1991). Value changeincounseling and

psychotherapy: A search for credibility Wm 18 16-
24.

Butler-Omololu, C., Doster, J ., & Lahey, B. (1984). Some implications for
intelligence test construction and administration with children of different racial

groups. WWW. 19(2), 63-75.

Cardall, C., & Coffman, W. E. (1964). A method for comparing the
performance of different groups on the same items of a test. W

W, 2, 64-65. Princeton, NJ: Educational Testing Service.

Carter, R. T. (1990). Cultural value difi'erences between African Americans
and White Americans. Went, 3.1. 7149.

Carter, R. T., & Helms, J. E. (1990). The intercultural values inventory (ICV).
WWW. Princeton, NJ: Educational Testing Service.

290

291

Carter, R. T., & Helms, J. E. (1987). The relationship between Black value-

orientations and racial identity attitudes Meastmmemandfixaluationjnflmmselina
andDexelmeent, 12,185 195.

Cascio, W. F. (1975). Accuracy of venfiable biographical information blank
responses loumalflAnplmdPsxchologx 69 767-769

Chandler, C. R (1971). Value orientations among Mexican Americans m a
southwestern city W, 5.8, 262-271

Chapman, A. D. (1981). Value-orientation analysis: The adaptation of an

anthropological model for counseling research. WM,
5900), 637-642.

Clark, K. B. & Clark, M. P. (1939). The development of consciousness of self
and the emergence of racial identity 1n Negro preschool children. Mal

125191191291 19, 591-599.

Cleary, T. A., & Hilton, T. J. (1968). Aninvestigation of item bias.
EducationaLandPsxchologicaLMeasuremenLi 115-124.

Cohen. J- (1977). StatisticalhweLAnalxsisfoLtheBehafiomLSeim New
York: Academic Press.

Cohen, J., & Cohen, P. (1983)AppliedMultip1r.RegressionLCorrr1ation
AnalxmsfontheBehmomlfimnces New Jersey: Lawrence Erlbaum Associates

Cole, N. S. (1973). Bias in selection. loumaloflEducationalMeasrmem, 19.
237-255.

Cole, N. s. (1981). Bias in testing. Amerieaufisysholozist. 16, 1076-1077.

Cole, N. S., &Moss, P. A. (1989). Biasintestuse. InR.L. Linn(Ed..)
EdngationaLMeasuremem (3rd Ed. pp. 201-219). New York: American Council on
Education/Macmillan.

Cross, W. E., Jr. (1978). Models of psychological nigrescence: A literature
review. lmmalofBlasszxchologx. 5.13-31.

Darlington, R. (1971). Another look at "cultural fairness". lmnmlgf
EducationalMeasnrement, 8, 71-82.

Dole, A. A. (1995). Why not drop race as a term? WW
59(1). p 40-

292

Dorans, N. J. & Kulick, E. (1986). Demonstrating the utility of the
standardization approach to assessing unexpected differential item performance on the

Scholastic Aptitude Test loumathEducationaLMcasunemem, 23, 355-368

Dorans, N. J., Schmitt, A. P., & Bleistein, C. A. (1992). The standardization
approach to assessing comprehensive differential item functioning. Jgnmalgf

EducationalMeasnremem, 22(4) 309-319

Downie, N. M., & Heath, R. W. ( 1970). W. New York:
Harper and Row.

Bells, K. W. ,Davis, A, Havighurst, R. J, Herrick, V. E., & Tyler, R. W.
(1951) Intelligemandmnrramifferences. Chicago. University of Chicago Press.

Fish, J. M. (1995). Why psychologists should learn some anthropology.
Amefimfiflcholosisls, 59(1) 44-45

Fleishman, E. A. (1988). Some new frontiers in personnel selection research.

Wm. 41, 679-701-

Freedle, R. & Kostin, I. (1990). Difficulty of four verbal item types and an
index of differential item functioning for Black and White examinees. lonmalgf

EdncationalMeasurcmcm, 21(4) 329-343
Freedle. R. Kostin.1.. &SCth (1987) W

1:-..11rl1°.“111‘-11 1111 .‘. ‘-11==1- 1°=1°1.1

WW (Research Rep No RR-37-48)

Princeton, NJ: Educational Testing Service.

Gaertner, S. L. (1976). Nonreactive measures in racial attitude research: A

focus on "liberals". In P. A. Katz, (Ed). Imrdsmefiliminationcfkacism (pp.
183-211). New York: Pergamon Press.

Green,B. F., Crone, C. R., &Folk,V. G. (1989). Amethod forstudying

differential dimmer functioning lmrmalcfEducationalMeasmcm, 26(2) 147-
1 60.

Gross, N. C., Mason, W. S., & McEachern, A. W. (1958). Explorations in role
analysis; Studies of the school superintendency role. New York: Wiley.

Hambleton, R. K., Swaminathan, H., & Rogers, H. J. (1991). Eundammflalsgf
W. Sage Publications: Newbury Park, CA.

 

293

Hartigan, J. & Wigdor, A. (1989). Eaimcssjnfimrzlcxmemlesfina National
Academy Press: Washington, DC.

Helms, J. E. (1984). Toward a theoretical explanation of the effects of race on
counseling: A Black and White model. WW5}, 12,153-165.

Helms, J. E. (1990). Blackanthituacialidcntimmteseamhand
practice. Greenwood Press: New York.

Helms, J. E. (1992). Why is there no study of cultural equivalence in
standardized cognitive ability testing? Amcncanflstchcloaist. .41, 1083-1101.

Hough, L. M, Eaton, N. K., Dunnette, M. D. Kamp, J. D., & McCloy, R. A.
(1990). Criterion-related validities of personality constructs and the effect of

response distortion on those validities. WM, 15, 581-595.

Hunter, J. & Hunter, R. (1984). Validity and utility of alternative predictors of
job performance Psychologicalflulletin, 26 72-98.

Inclan, J. (1985). Variations in value orientations in mental health work with
Puerto Ricans. Psychotherapy, 2, 324-334.

James, L. R & Brett, J. M. (1984). Mediators, moderators, andtests for
mediation. MW, 69, 307-321.

Jensen, A. R (1995). Psychological research on race differences. American
25319119193151, 59(1). 41.

Kaplan, B. (1963). Issues in educating the culturally disadvantaged. PhLDelta
Kaman, 25.

Katz, J. H., & Ivey, A. E. (1977). White awareness. The frontier of racism
awareness training BersmneLandfiuidancclonmal 55 485-488

Kohls, R, Carter, RT., &Helms, J. E. (1984). WWW
Cited m R. T. Carter, The relationship between Black American college students'

value orientations and their racial identities. ExalnatimandMeasmmemjn
Counselinundflexclopmcm, 12(4) 185-195

Kluckhohn, C. (1952). Values and value orientations in the theory of action. In

T- Parsons & E- A- Shils (Eds) MW. Harvard
University Press: Cambridge.

294

Kluckhohn, F. R, & Strodtbeck, F. L. (1961). W.
Row, Peterson, and Company: Evanston, IL.

Kovel. J- (1970). W New York: Pantheon-

Kroebcr. A. L.. & Kluckhohn. C. (1952). WWW
mdjgfimfions. Peabody Museum of American Archaeology and Ethnology:

Cambridge, MA.

Ledvinka, J. (1979). The statistical definition of fairness in the federal selection

guidelines and its implications for minority selection. W, 32,
5 5 1 -5 62.

Lindell. M. K. & PerrY. R W- (1992). Behaidoralinmdationufflommunim
Emmfllanning. Hemisphere Publishing Company: Washington.

Lippmann, W. (1922). A future for tests. W, 3}.

Lord, F. M. (1952). A theory of test scores. WW (Whole
No. 7), Richmond, VA: William Byrd Press.

Lord. F. M., & Novick. M. R. (1968). W.
Reading, MA: Addison-Wesley.

Malpass, R. S., & Symonds, J. D. (1974). Value preferences associated with
social claSS. sort. and race. WW. 5. 282-299-

Maslow, A. H. (1970). W (2nd ed.). New York: Harper
& Row.

MbitiJ. S- (1970). Afimmlizimandnhilosophies. Garden City,NY
Anchor Books, Doubleday.

McFadden, J, & Gbekobou, K. N. (1984). Counseling African children 1n the
United States. W. 18(3). 225-230

MeROY.R- G-.&Zurchcr.L-A. (1983). W
W. Charles Thomas: Illinois.

WW (1974). (ed. by H. B. Woolf). Simon &

Schuster: New York.

 

295

Milliones, J. (1980). Construction of a Black consciousness measure:

Psychotherapeutic implications. WM. 11.
175-182.

Motowidlo, S. J, Dunnette, M. D., & Carter, G. W. (1990). An alternative

selection procedure: The low-fidelity simulation. immalpfiAppliedEsxchplogy,
15(6), 640-647.

Motowidlo, S. J., & Tippins, N. (1993). Further studies of the low-fidelity

simulation 1n the form of a situational inventory. humalpflfleeupatmnalfl
Qrganizationallsxchologx 66(4). 337-344

Mumford. M. D- (1993). WWW
Lelatedjesalesperformanee. Unpublished manuscript. Fairfax, VA: George Mason
University.

Mumford, M. D, & Owens, W. A. (1987). Methodology review: Principles,
procedures, and findings 1n the application of background data measures. Applied

RascholcgicaLMeasntementn, 1-31.

Mumford, M. D. & Stokes, G. S. (1992). Developmental determinants of
individual action: Theory and practice in applying background measures. In M.

Dunnette and L Hough (Eds ) Handbookcflndusnialandflrganimional

W. Consulting Psychologists Press, Inc.: Palo Alto,
CA.

Nagle, B. F. (1953). Criterion development. WW, 6, 271-289.

Nobles, W. W. (1980). African philosophy: Foundations for black psychology.
In R. L. Jones (ed.). 31W (2nd ed.). p. 23-36. New York: Harper &

Row.

Novick, M. R, & Ellis, D. D. (1977). Equal opportunity in educational and
employment selection. Amefieanflsyehplpgist, 22, 306-320.

Olsen, M. E. (1970). Social and political participation of blacks. Amm'iean
MW, 35, 682-697.

Osterlind, S. J. (1983). W. Newbury Park, CA: Sage.

Owens, W. A., Glennon, J. R, & Albright, L. E. (1966). AME—1113191!
items. Greensboro, NC: Center for Creative Leadership.

296

Parham, T. A. & Hehns, J. E. (1981). The influence of Black students' racial
identity attitudes on preference for counselor's race. WW

Emholm 28. 250-257

Paulhus D L (1988) Assessingselfideccptimandjmpressicnmamaemmjn

WWW Unpublished manual,
University of British Columbia, Vancouver, Canada.

Paulhus, D. L. (1991). Measurement and control of response bias. In M. D.

Dunnette& L M. Hough(Eds..) Measueufpcrsonalitundsocialmxcholcaical
attitudes (pp. 17-59). New York: Academic Press.

Reilly, R. R, & Chao, G. T. (1982). Validity and fairness of some alternative
employee selection procedures. EemenneLRsyehplegy, 35, 1-62.

Rogers, H. J, & Kulick, E. (1987). An investigation of unexpected differences in
item performance between Blacks and Whites taking the SAT. In A. P. Schmitt & N.

J Dorans (EdS) WW8 (RM-
87-1). Princeton, NJ: Educational Testing Service.

Rokeach, M. (1973). W. The Free Press: New York.
Rokeach, M. (1976). W. Jossey-Bass: Washington.

Rudner, L., Getson, P. & Knight, D. (1980). Biased item detection techniques.
loumalprducationaLStatistics, 5, 213-223.

Rushton, J. P. (1995). Construct Validity, Censorship, and the Genetics of Race.
Americanllsxchclcaist. 59(1). p. 40.

Scheuneman, J. D. (1978). ethnic group bias 1n intelligence test items. In S. W.

Lundsteen (Ed) Qumtalfactcrsdnlramumandmstmcucn (Diversity Series N0 56
p. 65- 77). New York: ERIC Clearinghouse on Urban Education.

Scheuneman, J. D., & Gerritz, K. (1990). Using differential item functioning
procedures to explore sources of rtern difficulty and group performance
characteristics loumalpffdncaticnalerrcmem. 22.109-131

Schmidt, F., Hunter, J, & Caplan, J. (1981). Validity generalization results for
jobs 1n the petroleum industry. Wen, M, 609-626.

Schmidt, F. Pearlman, K., & Hunter, J. (1980). The validity and fairness of
employment and educational tests for Hispanic Americans: A review and analysis.

Personnellsxchologx, 33, 705-724.

 

297

Schmitt, A., P. (1988). Language and cultural characteristics that explain
differential item functioning for Hispanic examinees on the Scholastic Aptitude Test.

loumalciEducationalMcasutemmt. 25(1). 1-13

Schmitt.A P. &BleistcimC A (1987) Wm
11 111 1 :z :111-1 11 11:1 .1 11‘11.11° ' 11 (RR-87-23).

Princeton, NJ: Educational Testing Service.

Schmitt, A., P, & Dorans, N. (1990). Differential item functioning for minority
examinees on the SAT lmnnalthducaticnaLMeasntcmem. 22(1) 67-31

Schmitt, N. W., & Pulakos, E. D. (under review). Biodata and differential
prediction: Some reservations.

Schmitt, N., Gooding, R, Noe, R, & Kirsch, M. (1984). Meta-analyses of
validity studies published between 1964 and 1982 and the investigation of study

characteristics. W1 31, 407-422.
Shepard, L. A. (1981). Identifying biasintestitems. InB. F. Green(Ed..) New

11‘ 1111 111°:1111'.-11‘-111'-11 ‘1-111 111° - 1: 1.1° 1 r 111211

was, No. 11 (pp. 79-104). San Francisco: Jossey-Bass.

Shepard, L. A., Camilli, G., & Averill, M. (1981). Comparison of procedures for
detecting test-item bias with both internal and external ability criteria. loumaigf

EducationaLStatistics. 6, 317-375.

Solorzano, D. G. (1992). - An exploratory analysis of the efi‘ects of race, class, and

gender on student and parent mobility aspirations. WM, 6.1,
30-44.

Spencer, M. B. (1984). Black children' s race awareness, racial attitudes, and self-

concept. A reinterpretation lmunachChildEsxcholouandPsxchim. 25. 433-
441.

Stricker, L. (1982). Identifying test items that perform differentially 1n

population subgroups: A partial correlation index. WM
Measurement, 6(3) 261-273

Stuit, D. B, & Wilson, J. T. (1946). The efi‘ect of an increasingly well defined
criterion on the prediction of success at naval training school (tactical radar). .Ipnmal

cprpliedPsxchologx, 39, 614-623.

298

Tatsuoka, K. K., Linn, R L. ,Tatsuoka, M. M. & Yamamoto, K. ( 1988)
Differential item functioning resulting fi'om the use of different solution strategies.

lmrmalcflEducationaLMeasruMfiM). 301-319

Terry, R. W. (1977). W. Grand Rapids, MI: William B.
Eerdmans.

Terry, R W. (1981). The negative impact of on White values. In B. P. Bowser

&R G Hunt(Eds..).lmpacts_o£racism_0nMitc.Americans(pp 119-151) Beverly
Hills, CA. Sage Publications.

Thissen, D., Steinberg, L., & Gerrard, M. (1986). Beyond group-mean
differences: The concept of item bias. EsyehemgieaLBnlletin, 99(1), 118- 128.

Thorndike, E. (1927). Ihemeasmementpfinteliigence. New York: Teachers

College Bureau of Publications.

Thorndike, R (1971). Concepts of culture-faimess. Wynn!
Measurement, 8. 63-70.

Thurstone, L. L. (1925). A method of scaling educational and psychological
tests lmrmalflEducationalPsxcholm. 16. 263-278

Veale, J. R, & Foreman, D. I. (1983). Assessing cultural bias using foil response
data: Cultural variation loumalpffiducaticnalMeasutement, 29(3). 249-258

Vontress, C. F. (1971). Racial differences: Impediments to rapport. Jpnmalpf
MW, 18, 7-13.

Williams, J. B., & Roberson, J. K. (1967). A method of assessing racial attitudes
in preschool children. EAucationalPsxcholmMeasurement. 22. 671-689-

Wonderlic Personnel Test. Inc. (1992). Eondeflicketscnnrllcstfificholastic
LexeLExamllscL’sManual. Libertyville, IL.

Yee, A. H. (1983). Ethnicity and race: Psychological perspectives. Educational
Psychologist. 19. 14-24.

Yee, A. H. ,Fairchild,H. H, Weizmann,F., &Wyatt,G. E. (1993). Addressing
psychology’ sproblems withrace. Amerimfisyehelegist, 18(11), p. 1132-1140.

 

 

 

 

"11111111111111111111