I .‘~ .
i 3.952%. skkmuﬂéwVUIoH-v envy":
. ‘

oo'v.‘-o lod- -
.'

x5343“. . o

12.3
1

tr
’0!
02.: Ill
(19:. 01:,
1.31:0"l. \
a V

o till-.12.. p no A:
.uulo‘.isvuptl¢
vi 'pllilr’iv u...
o' .1. I!:lu.obl .
4’1“!le

o

s I 4
1.1....) .4!
.10.. 901‘.
.(r.vt£..t¢ ’5'!»
:31}!
,3. x;r 7 f.
.y :31 a . . . ..
1.7.32. «:1!

. o .
#5:...
:43: Q . A
L («lie
I 1....4...
«23..., v. .1 0"}.
I. : . .;.4.o¢r...':’
. '0'! I.-"'I’l
arVn.
ll”.
vb“; ‘0
a 7...!
1."!

D

(at 3:0!

u'.‘l I... .
-- 1141.04.

 

 

III

   

t. . .. I».
. .1... 2445:955f... .
. ., s: 55.2...
531 . .

Illlll‘llllllllllIlll‘lllllllllllllllll

3 1293 00794

This is to certify that the

dissertation entitled
INMATE GRADUATION FROM
AN ACADEMIC COLLEGE PROGRAM:
How It AFFects Their Rates 0F Recidivism
Upon Parole From Prison

presented by

Arthur Kirk

has been accepted towards fulﬁllment
of the requirements for

Ph.O. degreein Education ~ -

 

 

/

771:1 1'>1‘Afrffrl p112
. ‘ ;
Major professor

Dr. Eldon Nonnamaker

Date May 20, 1992

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

i LIBRARY
Michigan State
i University

 

 

*__

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

____1 :1

MSU Is An Affirmative Action/Equal Opponunlty Institution

 

 

 

 

 

 

 

 

 

 

 

 

INMATE GRADUATION FROM
AN ACADEMIC COLLEGE PROGRAM:
How It Affects Their Rates Of Recidivism
Upon Parole From Prison

by
Arthur Kirk

A DISSERTATION

Submitted to
Michigan State University
In partial Iultlllment oi the requirements
tor the degree at

DOCTOR OF PHILOSOPHY

Department of Educational Administration
1992

ABSTRACT

INMATE GRADUATION FROM AN ACADEMIC COLLEGE PROGRAM:
How It Affects Their Rates of Recidivism Upon Parole From Prison

by
Arthur Kirk

The main purpose of this ex-postfacto observational study was to examine the
relationship between inmate graduation with an associate degree from the College
Opportunity - Prison Extension (C.O.P.E.) Program (offered by Montcalm Commu-
nity College of Sidney, Michigan) and recidivism rates. Simply examining these two (2)
factors (education and recidivism) alone would not have allowed for consideration of
the many other conditions (factors) which research and/ or common belief suggest are
simiﬂcantly related to recidivistic behavior. Ignoring these other inﬂuences would not
have allowed a determination as to whether the study ﬁndings were attributable to the
education variable or to one or more of the other factors. Thus. one of the important
tasks of this study was to identify and statistically evaluate the major biological.
psychological. personal. and environmental factors believed to be signiﬁcantly related
to recidivistic behavior.

Of the forty-two (42) factors (variables) examined. four (4) were identiﬁed as having
a major predictive/ causative relationship with recidivism. They were:

1) Age Upon Parole

2) History Of Substance Abuse

3) Type 01' Instant Oﬂensem

4) Academic Educational Level At Time Of Instant Offensets)

 

As for the inﬂuence of post-secondary education, results of the study
indicated no overall statistically signiﬁcant relationship between completion
qfthe C.O.P.EPrograrn qfstudy and reduced rates qfrecidivismforpartici-
pants. Harmer, certain sub-groups within the C.O.P.E. Group appeared to have
lower recidivism rates as a direct result of completing the C.O.P.E. Program of
study: speciﬁcally, those inmates who are less than twenty-six (26) years dage
upon paroleﬁ'om prison. who do not have a history qfsrdrstance abuse, who are
sent to prisonfor a violent qﬂ'ense. and/or who enter prison without a high
school diploma or GED Certificate.

In regard to the full inﬂuence of academic edumtion on recidivism. it was
determined that completion of a high school education is critically important. For of
the study subjects who were paroled without a high school diploma or GED Certiﬁcate.
seventy-eightpoint three percent ( 78.3%) recidivated within a two (2) year period of time.

 

© Copyright by Arthur Kirk I992
All Rights Reserved

 

DEDICATION
To those who languish in our prisons

 

 

ACKNOWLEDGEMENTS

I would like to sincerely thank my major adviser. Dr. Eldon Nonnamaker. for
his warm support and skillful guidance throughout this study. I am also most
grateful to Dr. Louis Hekhuis, Mr. Zolton Ferency, Dr. Vincent Hoﬂinan. and Dr.
Betsy Becker for their encouragement and assistance.

The inmate-students. teachers and administrators in the College Opportu-
nity - Prison Extension (C. O.P.E.) Program gave freely of their time and talent.
Without their cooperation. this study would not have been possible. A special
expression of thanks goes to Mr. Danny Herman. C. O.P.E. Director. who did
much to help and encourage me in the conduct of this study.

A great amount of gratitude is extended to Mr. R. Douglas Kosinski.
Supervisor. Special Studies Unit of the Michigan Department of Corrections for
his extensive assistance in the design of this study and in the statistical
analyses of the data. Without his guidance and direction this study would not
have been completed.

I am most appreciative of the cooperation provided me by Mr. William Kime,
RetiredDeputyDirectorofPrograms andPlanningfor the MichiganDepartment
of Corrections. A special note of thanks is extended to Mr. Terry Murphy, Chief
of Research for the Michigan Department of Con'ections, and his associates Mr.
JeﬂAnderson, Ms. June Daman, Mr. Steve DeBor, Ms. Peggy Kersey. Ms. Carole
Rankin and Ms. Mary Volakakis for their unending patience and assistance.

My family and friends gave me the support and encouragement to complete
this study. To them I will always be indebted.

 

vi

PREFACE

A self-written justiﬁcation for the development and furtherance of correc-
tional education could not begin to approach the humaneness and plain
common sense expressed in this plea by Winston Churchill:

The mood and temper of the public in regard to the treatment
of crime and criminals is one of the most unfailing tests of any
country. A calm. dispassionate recognition of the rights of the
accused and even of the convicted criminal against the state:
a constant heart-searching by all charged with the duty of
punishment: a desire and eagerness to rehabilitate in the world
of industry those who have paid their due in the hard coinage
of punishment: tireless efforts towards the discovery of cura-
tive and regenerative processes: unfailing faith that there is a
treasure. if only you can ﬁnd it, in the heart of every man: these
are the symbols which, in the treatment of crime and the
criminal. mark and measure the stored-up strength of a nation
and are sign and proof of the living virtue in it.

 

TABLE OF CONTENTS

Page
HST OF TABLES 00.000000000000000...000000000...COOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO d

”ST OF FIGURES OOODOOOOOOOOOOOOOOOO0.0.. 0000000000 OOOOOOOODOOOOOOOOOOOOO ........ OOOOOOOOOOOOOOOOOO ‘1

Chapter

I.

Introduction
Statement of the Problem
Statement of Purpose
Signiﬁcance of the Study .............. ........ ...........
Theoretical Framework ............... . .......
Sampling of Subjects
Limitations
Dclimitations

”.mption. OOOOOOOOOOOOOOOOOOOOOOOOCOOOOOOOO...OCO0.00.0.0.0...0......C.0.0000000000000000

Deﬁnition of Terms

REVIEW or LITERATURE .....
Recidivism in the United States
General Statistics
Related Factors
The C.O.P.E. Program at Montcalm Community College
Brief History
Current Operational Structure

RESEARCH DESIGN OOOOOOOOOOOOOOOOOOOOOOI0..0.0.00000000000000000000...OOOOOOOOOOOOOOOOOO.

mmducuon OO0.0.0.000...OOOOOOOOOOOO...00.0.00....OOOOOOOOOOOOOOOOOOO0.0000000000000000000000D

~I~I~IHH

0‘1

1 1
12
13
14
15
15

88 23388388

 

viii

Sample ..... ................................ ..... . ....... 37
Controls 39
Data Gathering Tools 45
Data Collection Procedures 46
Statistical Analyses 47

PRESENTATION AND ANALYSIS OF DATA 65
Raw Data 65
Group Comparisons...... ............................ ....... ...... 78
Matching Variables 78
Prisoner/Demographics Background 80
Criminal History 82
Current Offense/Sentence ............ ............ ..... 86
Institutional History 86
Education History SS
Recidivism ........ 91
Predictive/Causative Factors 98
Testing of Hypothesis 123

SUMMARY. CONCLUSIONS AND RECOMMENDATIONS 127

Testing of Hypothesis 134
Conclusions 136
Recommendations 136
Suggestions for Future Research 137
Concluding Observations 137

 

ix

APPENDICES

A.

p 9

2‘5"

grasses

LETTER REQUESTING PRELIMHVARY STUDY APPROVAL
LETTER FROM MCC GRANTING

PRELIMINARY STUDY APPROVAL . ....... 140
RESPONSE TO LETTER GRANTING PRELIMINARY

MCC STUDY APPROVAL ... 141
LETTERS FROM MCC TRANSMITTING PROPOSAL
LETTER REQUESTING STUDY APPROVAL FROM MDOC ..."...m...” 145
LETTER OF STUDY APPROVAL FROM MDOC ...... 147
LETTERS REQUESTING STUDY APPROVAL FROM UCIRI-IS 148
LETTERS OF STUDY APPROVAL FROM UCRIHS 153
LETTER OF FINAL STUDY APPROVAL FROM MCC......................... 155
MDOC DATA PROCESSING SERVICE REQUEST ............................. 156
ATTITUDINAL SURVEY INSTRUMENT mm...................................... 181

 

Table

4.1
4.2

4.3

4.4

4.5

4.7
4.8
4.9
4.10

LIST OF TABLES

Page
GROUP COMPARISONS ON THE MATCHING VARIABLES 79
GROUP COMPARISONS ON THE PRISONER

DEMOGRAPHICS [BACKGROUND VARIABLES .. .............. s 1
GROUP COMPARISONS ON THE

CRIMINAL HISTORY VARIABLES ........ ......... . ........... 83
GROUP COMPARISONS ON THE

CURRENT OFFENSE/SENTENCE VARIABLE ............ 85
GROUP COMPARISONS ON THE

mSTITUTIONAL HISTORY VARIABLES .......................... 88

GROUP COMPARISONS ON THE

EDUCATIONAL HISTORY VARIABLES .... 90
GROUP COMPARISONS ON RECIDIVISM ............................... 98
PREDICTIVE/CAUSATIVE FACTORS OF RECIDIVISM........... 110
MAJOR MODEL VARIABLES .................................................. 117

 

LIST OF FIGURES

Figure Page
1.1 VICTIMS OF VIOLENT CRIMES

Nelve-Year-Old Children .......... . ..... . ..... 2
3.1 GROUP I

Selection Process .................................. . .............. 38
4. 1 RECIDIVISTS BY GROUP

High School Education — Prison Entry 92
4.2 RECIDIVISTS BY GROUP
4.3 RECIDIVISTS BY GROUP

Within Two-Year Period of Time 97
4.4 RECIDIVISTS BY HIGH SCHOOL GRADUATION AT PRISON ENTRY

Samples Combined 99
4.6 RECIDIVISTS BY GROUP

High School Education — Upon Parole 107
4.8 RECIDIVISTS BY OFFENSE TYPE

Samples Combined 112
4.7 RECIDIVISTS BY SUBSTANCE ABUSE

Samples Combined 114
4.8 RECIDIVISTS BY AGE UPON PAROLE

Samples Combined 115
4.9 TOTAL RECIDIVISM
5.1 RECIDIVISTS — COMPARISON GROUP

High School Education Upon Parole ....... ............... 131

 

 

INTRODUCTION TO THE STUDY

 

Introduction

Crime in the United States has reached serious levels. Our response to the problem
has been to inwcerate more people per capita than any other free nation on earth.
A Justice Department Survey released in early April 1984 reported 438.830 men and
women behind prison bars. and about 2 10.000 in jails around the country— an
increase of one-hundred ﬁfteen percent (1 15%) over 10 years (Gest. 1984). In a more
current survey covering the ﬁrst six months of 1986. the Justice Department reported
the nation's prison population increased to a record total Of 528.945. Of that number.
44.330 werein federal prisons and 484.6 15 in state institutions. The ﬁve (5) states with
the largest prison populations were:

 

 

 

 

1) California 55.238
2) Texas 37,780
3) New York 38.100
4) Florida 29,712
5) Ohio 21.942

 

The ﬁve (5) states with the smallest prison populations were:

 

 

 

1) North Dakota 411
2) Vermont 701
3) New Hampshire 732
4) Wyoming 888

 

5) South Dakota 1.089

 

2

VICTIMS OF VIOLENT CRIME
Twelve-Year-Old Children

 

89%

 

 

 

0 20 40 60 80 100
Percentage

l — During Lifetime I

Figure 1.1

 

 

As of June 30. 1986. the 25. 192 female inmates made up 4.72% of the total prison
population. The latest racial breakdown ( 1984) showed 5 1 .7% white. 45.3% black. 1%
American Indian. and the rest Asians and other groups (“Business Is Booming,"
1987). Projection studies dealing with incarcerated individuals convincingly point out
the number of people conﬁned in our prisons and jails will continue to dramatically
increase over the next ten (10) years and beyond.

According to Justice Department ﬁgures released on March 8. 1987. eighty-three
percent (83%) of 12-year Old children in the United States will be victims Of violent
crimes at least once in their lifetime. Fifty-two percent (52%) will be victims of such
crimes more than once (“Study Finds Crime Awaits Many Ost.” 1987). The study.
based on ﬁgures compiled by the Govemment’s National Crime Survey from 1975
through 1 984. went on to report that eighty-nine percent (8996) of 12-year-Old boys will
face one or more violent crimes or attempted crimes. and seventy-three percent (73%)
of the girls (see Figure 1 . 1. above). Further. forty-ﬁve percent (45%) Of black males will
become victims of violent crime three (3) or more times— almost double the possibility
for black females (24%) and triple the likelihood for white females (13%). Thirty-seven

Chapter I: Introduction to the Study

percent (3 796) of white males are likely to be victimized three (3) or more times during
their lifetime (Cassata. 1987). Based on current crime rates. the study reached these
projections:

0 One (1) out of every one-hundred thirty-three (133) Americans will become a
murder victim. Among black men. the estimate is dramatically higher: one (1)
out of thirty (30).

0 One (1) out of every twelve (12) women will be the victim of a rape or attempted
rape. The rate for black women is one (1) out of nine (9).

0 Nearly everyone will be the victim of a personal theft at least once. and eighty-
seven percent (8 79b)will be personal theft victims three (3) or more times (“Study
Finds ....” 1987).

The Bureau publishes crime victimization rates based on twice-a-year interviews
which involve 101.000 persons in 49.000 households. This particular report was
drawn from approximately two (2) million interviews conducted during the ten (10)
years ending in 1984. The rape statistics. however. were projected from 1973- 1982
(Cassata. 1987).

At the end of 1988. a record 627.402 people were incarcerated in federal and state
prisons. an increase of seven point four percent (7.4%) over 1987. ‘The 1 988 increase
translates into a nationwide need for more than 800 new prison bed Spaces a week. "
said Lawrence Greenfeld. Corrections Unit Chief for the Justice Department’s Bureau
of Justice Statistics (“1988 Saw Record Prison Population.” 1989).

Based on these reports and other like evidence. it seems fair to conclude we have
not done a very admirable job of identifying. much less dealing with. the root causes
of crime in this country. Our main “solution” to the problem of crime. namely stiﬁ‘er
penalties. has become a problem in itself. The newly constructed prisons we ﬁnd in
many states ﬁll up as soon as they open. leaving the numerous problems associated
with overcrowding looming over our heads. In these overcrowded institutions. we may
well be breeding criminals rather than restraining them.

 

Chapter I: Introduction to the Study

4

Of the many volumes of data we have collected in the ﬁeld of adult corrections. none
more glaringly reﬂects our failure to bring this problem of crime under reasonable
control than those dealing with recidivism (repeat crime). The literature is replete with
support for the contention that a very large percentage of the total number of crimes
committed in this country are perpetrated by a comparatively small number of repeat
oﬁ‘enders (Goldfarb and Singer. 1 977). It follows. then. that by putting forth an eﬁort
to identify and statistically evaluate the major conditions (factors) which seem to affect
recidivistic behavior. we develop a starting point from which to more eﬁ'ectively deal
with the general problem of crime in this country.

With this purpose in mind. one of the important tasks of this study was to identify
and statistically evaluate the mayor biological. psychological. personal. and environ-
mental factors which appeared to have a signiﬁcant relationship with recidivistic
behavior. in either a positive (decreasing) or negative (increasing) fashion. This was
accomplished through an extensive review of criminal justice literature (adult
corrections) in an eﬂ'ort to determine the ﬁndings of other researchers in this area. The
factors identiﬁed through this search. and others which were commonly believed to be
related to recidivistic behavior. were subjected to a series of statistical analyses using
these study subjects as the basis for the data. A determination was then made as to
which if any of these factors were signiﬁcantlyrelated to recidivistic behavior. in regard
to this study population.

With this task accomplished. the researcher addressed the main focus of this ax-
postfacto observational study: an examination of the relationship between post-
secondary academic attainment levels and rates of recidivism. This phase of the
study allowed for examination and testing of the widely held belief. especially among
those in the ﬁeld of correctional education. that increased participation in academic
programs on the part of inmate-students leads them to lower rates of recidivism once
they are released from prison (Gaither. 1983). Those who support this notion contend
that increased understanding .of human behavior leads these inmate-students to
personal insights which earlier were not at hand. and in an occupational sense.

 

Chapter I: Introduction to the Study

5
provides them with options which previously did not exist. All this. the supporters say.

leads to lower rates of recidivism.

In the interest of exploring this belief. the study examined two (2) Groups of former
inmates. The primary subjects made up Group I. and consisted of those inmate-
students who were awarded an associate degree from the College Opportunity -
Prison Extension Program during their present prison commitment (the incarcera-
tion period under study). and were paroled to the "free community" between 1 980 and
1984 inclusive. In order to have earned a degree from the C.O.P.E. Program these
subjects must have spent some or all of their incarceration period in one or more of
the following prisons located in Ionia. Michigan: Michigan Refomiatory. Michigan
Training Unit, or the Riverside Correctional Facility. Two (2) other prisons are now
operational in the Ionia area: the Ionia Temporary Facility. and the Ionia Maximum
Facility. However. these him (2) prisons are quite new and were not in operation during
the time period under study.

The members of Group 11 were viewed as the secondary (comparison) subjects.
They differed from members of Group I in regard to their levels of academic attainment
upon parole ﬁ'om their prison commitment. Namely. they individually had less than
a completed post-secondary education at the time of parole. where completed
meant the earning of an academic degree. Like the members of Group 1. these subjects
were also paroled between 1980 and 1984. In addition. they served part or all of their
present commitment in one of the three (3) prisons then located in Ionia. Michigan This
requirement provided reasonable assurance their incarceration clcperiences were
similar to those subjects in Group I. and they had pre- and post-secondary educational
opportunities equivalent to members of Group 1.

Additionally. a one-to-one matching procedure on six (6) selection factors
(study variables) was used informulating Group II. to ensure that like subjects in
regard to these reportedly important variables were observed (selected). These
matching variables were chosen for that purpose because the research literature

 

Chapter I: Introduction to the Study

8
indicated they were signiﬁcantly related to recidivistic behavior. they were commonly
believed to be related to recidivism. or because they assisted in the control of the study
design.

By way of brief background. the C. O.P.E. Program is a community college prison
extension program offered by Montcalm Community College of Sidney. Michigan (near
Ionia). Through the College. inmates can take classes leading to an associate degree
in General Studies. an associate degree in Arts and Sciences. or an associate degree
in Applied Arts and Sciences with a concentration in: Business Administration.
Accounting. Business DataProcessing .or Food Service Technology. The Program. in its
basic form. ﬁrst became operative at the Michigan Reformatory in 1 968. and has grown
and expanded steadily since that time. It now provides post-secondary educational
oﬁ'erings at the three (3) prisons mentioned earlier. and one (1) of the two (2) new Ionia
prisons. (These prisons comprise what is referred to as the Ionia Complex.) Further.
the C. O.P.E. Program is now operative in the Thumb Regional Correctional Facility in
Lapeer County. Michigan. (This facility was not in operation during the time period
under study.)

As previously mentioned. recidivism is a real problem in this country. There is
widespread acknowledgment among professionals. in the ﬁeld of criminal justice and
among members of the general public as well that “recidivism provides a staggering
societalproblern. bothintennsqfhumansutferingandintemisofwastedhumanand
economic resources" (Gaither. 1983. p. 86). Roberts (1973) was responding to this
problem when he wrote: ‘Correctional education should strive to impart to the inmate
the skills. knowledge. and attitudes necessary for attaining its primary objective-
successful adjustment to society" (p. 51).

Therefore. in an attempt to make a small contribution to the goal of reduced
recidivism. this study examined the eﬁ'ect the C. O.P.E. Program had on that undesired
activity. Additionally. the scope of the study was broadened to allow for an
examination of the overall relationship between academic education and
recidivistic behavior. It was felt this added depth. insight. and fuller meaning to

speciﬁc ﬁndings of the study.
Chapter I: Introduction to the Study

 

Statement of the Problem

Do outside factors which precede and follow a prison sentence such as: home
environment. demographics. education. biological conditions. personality traits. etc.
have an identiﬁable relationship with recidivistic behavior? Ifso. can these inﬂuences
be statistically measured as to their individual and collective strength?

Statement of Purpose

The primary purpose of this study was to determine if graduation with an associate
degree from the C.O.P.E. Program provided a positive (reducing) inﬂuence on the
recidivism rates of its participants once they were paroled from prison.

Signiﬁcance of the Study

Recidivism research is an important factor in understanding criminal behavior
(Nacci. 1978). Nacci pointed out that by reducing the criminal activities of chronic
oﬂ‘enders. one may signiﬁcantly change the total crime picture. Ifcollege programs for
prison inmates can signiﬁcantly reduce the rates of recidivism of the participants. then
it follows that programs such as C. O.P.E. maybe useful in reducing not only recidivism.
but the overall crime rate as well.

To be certain. recidivism is a complex phenomenon which cannot be explained in
basic terms. A simple analysis of acknowledged factors related to criminal recidivism
is not enough. because there are so many other factors which impact on the
relationship under study (McCollum. 1977: Hoffman & Beck. 1984: 1985). There-
fore. this study was designed to amplify a full scope of possible predictive] causative

factors. and to test their impact on recidivism empirically.

Theoretical Framework

This study was based on the theories of a number of researchers: I-Ioﬂman and
Beck (1985) provided a “salient factor score” and ﬁve-year follow-up as methodology

 

Chapter I: Introduction to the Study

S

for realistically measuring recidivism. Hoﬂ'man and Stone-Meierhoefer (1980)
provided a ﬂexible criterion measure for recidivism. one which they suggested gives
realism to the measure and accounts for conﬂicting reports of recidivism rates. As to
the conﬂicting reports of recidivism rates. Griswold (1978) reported that recidivism
measures are not equally valid or reliable because different measures can and do
produce discrepant ﬁndings. Hoffman and Stone-Meierhoefer (1980) pointed out
there is considerable conﬂict and uncertainty as to even crude estimates of recidivism
rates of released inmates.

With respect to recidivism’s association with academic education. Craig ( 1983)
found that data analyses indicated no evidence of a relationship between participation
in educational programs and rates of criminal recidivism. Havilaud (1 982) found there
was not a signiﬁcant difference in the rate of recidivism between those inmates who
gaduated from a two-year college program while incarcerated and those inmates who
had not gaduated from a two-year progam while incarcerated.

Blackburn (1981). on the other hand. found that a reduction in the absolute
recidivism rate appeared to have developed as a result of participation in an academic
college progam. He reports those inmates who participated in the College Program (at
Hagerstown Junior College in Hagerstown. Maryland) exhibited an overall recidivism
rate one-third lower than those inmates who did not participate.

In preliminary research relating recidivism and academic education. Note and
Holloway (1986) reported signiﬁcantly lower rates of recidivism for inmate-students
who g'aduated from an associate degee prison progam. They also studied other
groups of inmates with lesser levels of academic education. concluding the recidivism
rate is dependent on the education variable.

Recidivism‘s association with numerous variables was stressed by McCoIlum
(1977). She pointed to the key role of the total prison experience and other outside
factors (life history. demogaphics. personality traits. and others) in measuring the

causes of recidivism.

 

Chapter I: Introduction to the Study

Research Design

Sampling of Subjects

The source of data for this study was former Michigan state prison inmates. all of
whom were paroled by the Michigan Department of Corrections between 1 980 and 1 984
(inclusive). They were formulated into two (2) study Groups:

GROUP I — Inmates 'who during their present prison commitment
graduated with an associate degree from the C.O.P.E. Program.

GROUP II — Inmates who had less than a completed post-secondary
education at the end of their present prison commitment.

Members of Group I. the primary observational goup. numbered one-hundred
sixteen (n= 1 1 6). They were selected from records maintained by Montcalm Community
College and the Michigan Department of Corrections. They had as their entry] inclusion
criteria:

1) Graduated from the C.O.P.E. Program during their present prison commit-
ment. where graduation meant they were awarded an associate degree.

2) Were paroled during the period 1980 to 1984.
S) Remained alive for two (2) years following parole.

Members of Group 11. the secondary (comparison) observational goup. also
numbered one-hundred sixteen (n=1 1 6) so as to have a balanced study desigr. They
had as their entry/ inclusion criteria:

I) Served all or part of their incarceration period in one of the three (3) Ionia
prisons.

2) Were paroled during the period 1980 to 1984.

 

Chapter I: Introduction to the Study

10

8) Had less than a completed post-secondary education upon parole from
present prison commitment. where completed referred to the awarding of
a degree from an accredited college or university.

4) Met all of the matching criteria. within established ranges. on a one-to-
one basis with members of Group I.

6) Remained alive for two (2) years following parole.

The subset of six (6) study variables which served as the basis for the matching
process (in a prioritized fashion) were:

1) Age upon parole from present prison commitment
2) Academic educational level at time of instant oﬁense(s)
3) Ernployment status at time of instant oﬁenseis)

4) Michigan Department of Corrections (MDOC) assaultive risk classiﬁcation
at time of parole from present prison commitment

6) Race
8) Marital status at time of instant oﬂ'enseis)

The data related to the subjects (ex-inmates) of this ex-postfacto observational
study came from records maintained by Montcalm Community College. the Michigan
Department of Corrections. the Michigan State Police (MSP). and the Federal Bw'eau of
Investigation (FBI). To insure the corUldentiality of all personal data obtained. reporting
was done only in collective (group) form. Individual identities of the study subjects
through name. prison (institutional) number. social security number. address. or any
other identiﬁer is known only to this researcher and the parties mentioned above. and
will remain unreported and unavailable to members of the general public!

 

Chapter I: Introduction to the Study

11

Data Analyses

All the study data was organized. analyzed. and printed through the use of a
computer. The following software applications were employed: WordPerfect. dBase IH
Plus. Foxbase. Lotus 1 -2-3. Quattro. Stats-2. Statistical Package for the Social Sciences
(SPSS). SPSS/PC Plus. PageMaker. Microsoft Windows. Harvard Graphics. Excel.
In‘a‘Vision. Designer and ProComm. Where it was deemed helpful. the data collected
and analyzed was presented in table or ﬁgure form.

One of the distinguishing characteristics of the scientiﬁc method is the formulation
and testing of hypotheses concerning population parameters. Tests of statistical
hypotheses require a priori formulation of decision rules as well as knowledge of
sampling distributions of test statistics.

Thus. the ﬁrst major step in data analyses was a series of crosstabulations
performed on all but the information-type study variables. This is one of the most
important sampling distributions in the behavioral sciences. and is most useful irn
attempting to draw inferences about variability as well as measures of central
tendency. The type(s) of crosstabulation analyses applied (standard. dichotomous.
collapsed. collapsed with a control. three-way with a control. and/ or four-way with a
control) varied depending upon the form(s) which proved most useful in the conduct
of the study. Further. an analysis qf variance (ANOVA) was performed on certain
study variables where the examination of means (averages) proved useful.

The purposes behind all these analyses was to determine the conditions on which
the two (2) study Groups signiﬁcantly (.05 level) differed. and to help determine which
of the independent variables shared a signiﬁcant (. 05 level) relationship with
recidivism. Because the dependent (y) variable (recidivism) was qualitative and had
numerous categories. and because many of the study variables lent themselves to
multicollinear relationships. a discriminant function analysis model with a

forward selection (stepwise) algorithmwas used as the major statistical procedure.
Such a statistic helped determine. on an individual basis and irn cluster form. the major

 

Chapter I: Introduction to the Study

12

predictive] causative factors associated with recidivism. Further. this model allowed
for the examination and measurement of the lone relationship between post- secondary
academic education and recidivism.

A determination as to whether or not study subjects recidivated was made by
running Law Enforcement IrJormation Network (L.E.I.N.). checks on them. This
information. which was provided by the Michigan State Police (MSP). revealed whether
any of these subjects were arrested for a felony class crime in Michigan. and the data
supplied to the network by other states and the Federal Bureau of Investigation (FBI)
provided such information on a national scale. Any felony class crime committed
outside US. Tenitory went undetected and unreported.

To add to the internal validity of the study. an attitudinal survey instrumentwas
designed by the researcher and was mailed to all members of Group I. It was used. irn
part. to explore the reason(s) members of Group I decided to attend and. irn turn.
graduate from the C. O.P.E. Program. The major purpose of this instrument. however.
was to determine if an attitudinal change relative to criminal activity took place and.
if so. whether that change took place before they entered the C. O.P.E. Program. while
they were irn it. or after they gaduated.

It was recognized that such an approach was far from scientiﬁc. but the conduct
of human behavior seems not to lend itself very well to scientiﬁc scrutiny. Simply. it
was felt that ignoring the possibility that such an attitudinal element existed. and was
a part of the behavioral interplay of their actions following parole from prison. would
have been a very serious oversight.

Null Hypothesis

The below listed null hypothesis was formed for the purpose of testing the major
assumptions underlying the study design:

There are no overall statistically signiﬁcant diﬁerences in the rates of
recidivism between Group I and Group II.

 

Chapter I: Introduction 'to the Study

- '—-—-..

13

In addition to the hypothesis. a signiﬁcant research question was formulated for
the purpose of gaining insight into the subjective element of attitudinal change on the
part of Group 1 participants. That research question was:

Are there detectable attitudinal changes regarding criminal activity on
the part of Group I (C.O.P.E.) members as a result of their college
experiences?

Limitations

Sample sizewas a major limitation of this study. Each of the two (2) Groups had
one-hundred sixteen (n=1 1 6) subjects. thereby supplying a total sample (N=232)
sufﬁcient for conducting signiﬁcant statistical analyses. Even so. a large set of samples
from various prisons around the country who also have college progams would have
added geatly to the internal and external validity of the study.

The selection qfsubiect matching variables was limited to those factors which
were available through the computerized information system maintained by the
Michigan Department of Corrections. The vast majority of this data was related to entry
information. gathered when the subjects were ﬁrst committed to prison. An eﬂ‘ort by
the MichiganDepartment of Corrections is underway to broaden the base of inmate data
available through the automated system. but because this study extends back many
years. the limitation existed.

Further. there existed a possibility that matching the two (2) Groups on all six
(6) study variables would not be possible. due to the limited size (1 .933) of the Pool
from which the comparison subjects were chosen. This proved to be the case. Thus.
in seeking statistically non- signiﬁcant diﬂ'erences between the two (2) Groups on these
variables. the number of matching variables had to be reduced to ﬁve (5).

Getting a good return on the questionnaires sent to members of Group I was
another signiﬁcant limitation of the study. The address listings available in regard to
these irunate-students were in some cases eight (8) years old. making the current
whereabouts of the subjects difﬁcult to determine. Additionally. many of the subjects

 

Chapter I: Introduction to the Study

14

contacted choose not to respond. possibly because of their desire to leave remem-
brances and tracings of their incarceration in the past.

Delimitations

The selection of a deﬁnitionfor “recidivism" was the major delimitation of the
study. As previously pointed out. recidivism rates may be artiﬁcially high or low.
depending upon how one uses criteria such as arrest(s). conviction(s). prior
incarceration(s). parole violation(s). etc. Also. criminal recidivism rates will vary
geatly depending upon how many years beyond release. parole. or transfer to a

correction center the study extends.

This researcher sought to avoid many of these problems by ﬁrst operationally
deﬁning recidivism as: arrested for a felony class crime following parole to the
free community. Secondly. parole violations which resulted in a return to prison
were not treated as recidivistic behavior unless the violation involved arrest for a
felony class crime. Further. the study incorporated a two-yearfollow-up period for
each subject. That is. the behavior of each study subject was tracked for a period of
two (2) years following parole from prison. Since all the subjects were paroled to the
”free community" between 1980 and 1984. .the cut-off date for data gathering was
December 3 1 . 1986. Thus. even the subjects who were paroled the last day of 1984
were tracked for a two-year period of time.

By not including subjects who were paroled after 1 984. assurances were provided
that all study participants were “on the streets" long enough for a reasonable test of
their behavior. Consideration was given to including subjects who were paroled
through December 31. 1985.:but it was deemed not advisable since data related to
recidivistic behavior which occurred irn late 1987 might well not get into the L.E.I.N.
system until mid- 1988. This would have geatly delayed the conduct of this study.

The study was delimited to err-inmates who appeared in the records
maintained by Montcalm Community College. the Michigan Department of
Corrections. the Michigan StatePolice. and theFederalBureau ofInvestigation.

 

Chapter I: Introduction to the Study

15

The data analyzed was conﬁned to the dependent variable (recidivism) and the major
independent variable (academic education). in conjunction with the secondary and
transfomled variables listed in Appendix K (p. 158) of this document.

Assumptions

The study assumed the data used was accurate and reliable. Further. when
involved in the one-to-one matching process between members of Group I and
members of Group II. it was assumed there was an insigniﬁcant and balancing
difference between: 1) marital status at time of instant oﬂenseIs) and marital
status at arrest. 2) employment status at time of instant oﬂ’enseis) and
occupation at arrest. and 3) academic educational level at time of instant oﬁense
and highest grade at entry (into prison).

It was also assumed the C.O.P.E. Program of instruction is suﬂlciently sound
tojustify using it as the data sourcefor the major independent varidrle. Ifthe
Program is not sound. then ﬁndings of the study would be vitiated because the
association between education and recidivism would be weakened: in effect. one would
be measuring the association not between education and recidivism. but between poor
education and recidivism.

Deﬁnition of Terms

A number of terms and acronyms are of special importance to the study. They are
deﬁned as follows:

Analysis Of Variance (ANOVAJ A method of dividing the total variation of
observations irnto components which can be attributed to or associated with particular

sources of variation. e. g. . the difference between goups or classes used in classifying
the observations.

 

Chapter I: Introduction to the Study

16

ASCII An acronym which stands for American Standard Code Information Inter-
change. It is a standard code for representing characters as binary numbers. used on
most microcomputers. computer terminals. and printers. In short. an ASCII ﬁle
appears irn English form. and can be read by non-technical people.

Attribute Attributes are qualitative or functional characteristics of individuals.
objects or goups. as distinguished from quantiﬁable characteristics. For example.
age. height. weight and wealth of individuals can all be regarded as variables because
they can be quantiﬁed. but sex. country of origin. and political persuasion can be
regarded as attributes.

Balanced Sample A sample which has some predeﬁned characteristics in
common with the population from which it is drawn.

Bias A systematic and non-random (but not necessarily intentional) distortion irn

a result or sample.

Biased Sample A sample selected using a pre-selected or favored (biased)
sampling method. The term is somewhat unsatisfactory since it is the sampling
method which is biased rather than the sample itself.

Categorical OutcomeA qualitative factor which falls into a category such as yes/
no. single/ married/ divorced. recidivated / did not recidivate. etc.

Cell A subclass or subcategory in a two-way or multiway frequency classiﬁcation.

Cell Frequency The number of observations which fall in a particular cell of a
frequency classiﬁcation. .

Chi-Square Analysis A statistic used to determine if the observed frequencies of
occurrence of the categorical values of a qualitative variable allow us to reject a
hypothesis about the expected or theoretical frequencies of occurrence.

 

Chapter I: Introduction to the Study

17

CMIS An acronym for Correctional Management Information System. A manage-
ment information system maintained by the Michigan Department of Corrections. in
which detailed data is available on current inmates and those who have been released
since 1980.

CodingT‘he procedure of coding involves three (3) stages. The ﬁrst stage is to decide
how to categorize the responses: the second stage is to allocate numerical or symbolic
values to the categories; the third stage is to allocate each individual response to the
appropriate category.

Comparison Subjects (Group) In experimental testing. a goup of persons or
objects used as a standard of comparison or accepted norm with which to evaluate
others among which a new process or method. or set of processes and methods. has
been implemented.

Confounding Variable An intervening variable which is intertwined or confused
with a concomitant variable and is also related to the criterion variable. Researchers
seek to eliminate the inﬂuence of such a variable on the criterion variable. so that any
observed differences in the means on the criterion variable can be attributed to the
predictor variable itself. rather than to the concomitant variable which is confounded
with it.

C.O.P.E. An acronym which stands for College Opportunity - Prison Extension: an
academic post- secondary education extension progam offered irn four (4) Michigan
prisons by Montcalm Community College of Sidney. Michigan.

Crosstabulation This statistical procedure produces tables showing the joint
distribution of two or more variables that each have a limited number of distinct values.
Cell counts. cell percentages. expected values. residuals. and various mea-sures of
association can be obtained. One can also specify the treatment of missing values.
obtain measures of association without printing tables. and request an index of tables.

Chapter I: Introduction to the Study

18

Data The plural of the Latin word ‘datum’ (=given). The word can mean any
information which is ‘given’ or provided for the solution of a problem.

Degrees Of Freedom The number of independent goups or subcategories into
which a sample or population may be divided.

Dependent Variable A variable which can be predicted by reference to other
variables. The term is used in regession analysis to indicate the variable which is likely
to have resulted from. or may be predicted by. one or a number of other variables.

Dichotomy The division of constituents of a sample. set or population into two (2)
goups.

Discriminant Function Analysis Discriminant analysis. ﬁrst introduced by Sir
Ronald Fisher. is the statistical technique most commonly used to identify the
variables that are important for distinguishing among goups. and to develop a
procedure for predicting goup membership for new cases whose goup membership
is undetermined. It is used as the major statistic in this study (with a forward selection
algorithm) in an effort to identify predictive/ causative factors of recidivism. and from
that to predict which future parolees are most likely to recidivate.

In many situations. discriminant analysis. like multiple regession analysis. is
used as an exploratory tool. In order to arrive at a “good” model. avariety of potentially
useful variables are included in the data set. It is not known in advance which of these
variables are important for goup separation and which are. more or less. extraneous.
One of the desired end-products of the analysis is identiﬁcation of the “good” predictor
variables. In usirng a forward selection (stepwise) algorithm with this statistic. the ﬁrst
variable included in the analysis has the largest acceptable value for the selection
criterion. In this study. the value of strength was measured by a series of crosstabular
analyses.

 

Chapter I: Introduction to the Study

19
After the ﬁrst variable is entered. the value of the criterion is reevaluated for all
variables not in the model. and the variable with the largest acceptable criterion value
is entered next. At this point. the variable entered ﬁrst is reevaluated to determine
whether it meets the removal criterion. If it does. it is removed from the model.

The next step is to examine the variables not in the equation for entry. followed by
examination of the variables in the equation for removal. Variables are removed until
none remain that meet the removal criterion. Variable selection terminates when no
more variables meet entry or removal criteria.

Ex-Post Facto An action conducted retrospectively.

Hierarchical Log-LinearAnalysisA special class of statistical techniques. called
log-linear models. has been formulated for the analysis of categorical data. These
models are useful for uncovering the potentially complex relationships among the
variables in a multiway crosstabulation. Log-linear models are similar to multiple
regession models. However. in log-linear models. all variables that are used for
classiﬁcation are indeperndent variables. and the dependent variable is the number of
cases in a cell of the crosstabulation.

This study used a fully saturated hierarchical log-linear analysis model. which
included a backward elimirnation algorithm. for conﬁrmatory evidence. This conﬁr-
matory evidence was used to lend support to the ﬁndings of the major model (a

discriminant function analysis) relative to the identiﬁcation of predictive] causative
factors of recidivism.

A fully saturated model of this sort uses the natural logs of the cell frequencies.
rather than the actual counts. The backward elimination algorithm starts with all
effects iii a model. and then removes those that do not satisfy the criterion for
remaining in the model. In this irnstance. the hierarchical scheme goes from least
signiﬁcant model variable to most signiﬁcant model variable. This elimination process
continues until a point is reached where the next variable elimination dramatically

Chapter I: Introduction to the Study

alters the clustered affect of the independent variables on the outcome variable. In
other words. the elimination process continues until the statistical worth of the model
breaks down to the point where it no longer serves as an eﬁ‘ective analytical tool. At
this point. the last variable removed is reinserted in the model. and the analysis is then
considered complete.

Independent VariableA term used in regression analysis to mean one of a number
of predictor. causal. or explanatory variables.

Instant Qﬂ‘enseThe criminal oﬂ'ense for which the subjects were incarcerated. and
from which they were granted their ﬁrst parole during the years 1980 through 1984.
For members of Group I. this was the ﬁrst parole following graduation from the
C.O.P.E. Program For members of Group II. it was the ﬁrst parole during the time
period 1980 through 1984.

Inverse Relationship An association where an increase in the value of one factor
results in a decrease in the value of another factor.

Ionia Complex A group of ﬁve (5) prisons loeated in Ionia. Michigan. They are:
Michigan Reformatory. Michigan Training Unit. Riverside Correctional Facility. Ionia
Temporary Facility. and the Ionia Maximum Facility.

LEJN. An acronym for Law Enforcement Information Network. A police informa-
tion system by which a determination can be made whether or not an individual was
ever charged and/ or convicted of a felony class crime.

Hatching Variables Factors such as age. race. marital status at time of instant
oﬁ'ense(s). etc. . upon which study subjects are paired or matched. The purpose behind
this process is to insure that diﬂ'erences in performance or condition can properly be
attributed to the independent variable(s) (factors) under study.

 

Chapter I: Introduction to the Study

21

Model A model is generally an attempt to summarize the complexity of the real
world in the form of simpliﬁed statements or relationships.

Multicollinearity (collinearityIA particularly vexing problem in the application
of multiple regression analysis or logistic regression analysis. in which two or more
predictor variables are very highly correlated with each other. thus making it diﬂlcult
to determine their individual inﬂuence on the outcome variable.

Nuisance Variable In statistics. a variable which causes an undesired source of
variation in a study and. in turn. adversely affects the measurement of the dependent
variable.

Null Hypothesis A particular hypothesis being tested. as distinguished ﬁ'om any
alternative hypotheses that may be considered in the context. In statistical usage. the
term often means a hypothesis that there is no difference between the sample mean

and the mean of a parent group. or between the means of two samples.

Parole Released from prison to the "free community" under the supervision of ﬁeld
service personnel from the Michigan Department of Corrections. prior to the expiration
of the court imposed maximum sentence for the committed oﬁ'ense.

Population The word population. when used by a statistician. may refer to any
speciﬁed collection of objects. people. organizations. etc.

Preeem Prison Commitment The incarceration period under study. In this case.
that from which the subjects received their ﬁrst parole during the years 1980 through
1984.

Primary Suluects (Group) The main subjects in a study. Those who are of most

concern and] or interest to the researcher.

RandomA method of selecting a sample may be said to be random if it gives to each
element in the population an equal chance of being selected.

 

Chapter I: Introduction to the Study

Recidivism As it relates to this study: arrested for a felony class crime following
parole from prison.

Recode Substituting new codes for the original coding of the data.

Released Discharged from prison without parole conditions or community center
placement. as a result of having served the maximum sentence. A person with this
status is not required to be under supervision by staff from the corrections department.

Sample Any subgroup of the population can be called a sample.

Secondary Subjects Study subjects which are used to compare and contrast
against the primary subjects. They are often referred to as the comparison group. or
in experimental studies. the control group.

Signiﬁcance Level In hypothesis testing it is usual to obtain from a given set of
sample data a test statistic calculated for the purpose of the test. This test statistic can
only be used if its distribution under the null hypothesis is known.

If the test statistic falls in a range of values (known as the critical region) which.
in total. have a small probability of occurrence under the null hypothesis. that
hypothesis will be rejected. This small probability is called the signiﬁcance level.

The most commonly used value is 0.05 (5%). although any other level may be
chosen.

Table A table is a systematic summary presentation of data.

Variables In this study. the dependent variable was recidivism. The key
independent variable was academic educational level upon parolefrom present
prison commitment. Also of concern for the purpose of determining mediating
inﬂuences on recidivism were the following secondary independent variables: date of
birth. race. date of prison entry for instant oﬁ‘ense(s). academic educational level at

———

Chapter I: Introduction to the Study

23

time of instant offense(s). type of instant offense(s). date of ﬁrst arrest. prior adult
felony conviction(s). criminal profile— juvenile property offense(s). criminal proﬁle—
juvenile drug offenses(s). criminal profile— juvenile violent offense(s). criminal
proﬁle— adult property offense(s). criminal proﬁle— adult drug offense(s). criminal
proﬁle— adult violent offense(s). in the community at least three years prior to prison
commitment for instant offense(s). marital status at time of instant oﬂ‘ense(s).
employment status at time of instant offense(s). history of substance abuse. evidence
of a serious physical illness or disability at time of instant offense(s). evidence of a
serious emotional or psychological problem at time of instant offense(s). upbringing.
financial status of upbringing environment. academic educational level of mother at
time of subject’s instant offense(s). academic educational level of father at time of
subject’s instant offense(s). family emotional support system at time of subject’s
instant offense(s). number of non-bondable major misconduct reports in prison for
which the subject was found guilty during his present prison commitment. date of ﬁrst
parole for instant offense(s). place from which paroled. parole placement. academic
educational level upon parole from present prison commitment. WOO assaultive risk
classiﬁcation at the time of parole. evidence of a serious physical illness or disability
at time of parole from present prison commitment. evidence of a serious emotional or
psychological problem at time of parole from present prison commitment. age at ﬁrst
arrest. age at time of present prison commitment. year of graduation from C. O.P.E..
age upon parole from present prison commitment. length of time served for instant
offense(s). academic educational attainment level increase during present prison
commitment. criminal recidivism— property oﬁ’ense(s). criminal recidivism— drug
offense(s). criminal recidivism— violent offense(s).

Wilks' LambdaThis statistic is sometimes called the Ustatistic. and served as the
selection criterion of the major model (a discriminant ﬁrnction analysis) of this study.

When variables are considered individually. lambda is the ratio of the within-
groups sum of squares to the total sum of squares. A lambda of one (1) occurs when

all observed group means are equal. Values close to zero (0) occur when within-groups

 

Chapter I: Introduction to the Study

24

variability is small compared to the total variability; that is. when most of the total
variability is attributable to differences between the means of the groups. Thus. large
values of lambda indicate that group means do not appear to be different. while small
values indicate that group means do appear to be different.

 

Chapter I: Introduction to the Study

 

REVIEW OF LITERATURE

Recidivism in the United States
General Statistics

Much of the literature reveals that recidivism rates for persons released from prison
are very high. Moseley (1976) reported that of those inmates released on parole.
seventy-one percent (71%) were involved in repeat criminal activity. Another report
noted that eighty percent (80%) of all felonies are committed by repeaters (Chamber
of Commerce. 1972). Goldfarb and Singer (1977) indicated that while statistics are
incomplete and conclusions may be drawn from them only tentatively. recidivism rates
are between ﬁfty (50) and eighty (80) percent. “The average prisoner is back in society
within three (3) years, repeating crimes within a year" (p. 9).

Other reports. however. provide lower estimates of recidivism rates. In an article
by Wallerstedt (1984). Steven R. Schlesinger. Director of the Bureau of Justice
Statistics said: “... close to a third of State prisoners released retumed to prison within
3 years and more thanaquarter were back in 2 years or less” (p. 1). Saxbe (1974) stated
that two (2) out of every three (3) oﬁ‘enders released from the Federal Prison System
did not return to prison for a serious offense within a two-year period. a rate which
”... certainly refutes the charges we keep hearing about a 70 or 80 percent recidivism
rate for all prison systems" (p. I).

Hoffman and Stone-Meierhoefcr (1 980) explained such variance in reports as
due to the application of different criterion measures. In their study of recidivism rates
for 1 .806 released federal prisoners. the authors found that recidivism rates depended

 

26
upon data based on four (4) criteria: arrest(s). conviction(s). prior incarceration(s) of
60 days or more. and prison commitment(s). Further. they found that recidivism rates
varied signiﬁcantly. depending upon the number of years that passed since release.
The recidivism rates one (1) year after release for each criterion were: 29.0. 1 5.4. 12.6.
and 8.7 percent respectively. These four (4) rates changed (increased) when computed
ﬁve (5) years later to 60.4. 41.7. 34.3. and 27.5 percent respectively. Thus. the
calculation of recidivism rates according to the authors depend in part whether one
is thinking in terms of arrest(s) . conviction(s) . prior incarceration(s) of 60 days or more.
prison commitment(s) . as well as the time which has passed since release from prison.

Griswold (1978) reported: ‘... that all recidivism measures are not necessarily
equally valid or reliable and that the use of diﬁerent measures can produce discrepant
ﬁndings " (p. 247) . Hoﬁnan and Stone-Meierhoefer ( I 980) stated: “A lthough the topic
qf recidivism elicits much interest. there appears to be considerable conﬂict and
uncertainty as to even crude estimates of the recidivism rate for persons released from
prison" (p. 53). They went on to suggest that:

Even with the use of omcial records (such as FBI data). there are a
large number of ways in which recidivism may be deﬁned. For
example. one might deﬁne recidivism as any of the following: any
new arrest. new felony arrest only. any new conviction. new felony
conviction only. any new commitment of sixty days or more. or new
prison commitment only. Return to prison for administrative
parole violation (e.g.. abscondlng) might be excluded. while admin-
istrative return to prison as a parole violator in lieu of prosecution
for a new offense might be counted. Or. one might wish to include
or exclude all types of parole violation. In addition. if other than an
arrest criterion is used. one must decide how pending charges or
unknown dispositions are to be counted. Clearly. for comparative
purposes it is essential that any recidivism rate reported be
accompanied by an explicit operational deﬁnition of the criterion
used. (Hoﬁman and Stone-Meierhoefer. 1 980. pp. 55-56)

 

Chapter II: Review of Literature

27

Wallerstedt (1984). who serves as the Social Science Analyst for the Bureau of
Justice Statistics (BJS). U.S. Department of Justice. deﬁned recidivism as:

the multiple occurrence of any of the following key events in the
overall criminal justice process:

0 commission of a crime

0 arrest

0 charge

0 conviction

0 sentencing

. incarceration

In the order given. these six (6) phases represent an increasingly
deeper penetration by oﬁenders into the criminal justice system.
and each is an important target for criminal justice statistics
programs. recidivism refers to reincarceration or the return of
released sentenced oﬁenders to the custody of State correctional
authorities. (p. 1)

Rates of criminal recidivism vary greatly when one considers oﬁ'ense patterns. On
this issue Wallerstedt (1984) said:

Released prisoners who go back to prison diner signiﬁcantly when
grouped according to their original oﬁ’ense. property oﬁ’enders
are more likely to return to prison (a median of 36.8%) than are
violent oﬁ'enders (31 .596). The median recidivism rate for burgla-
ries is the highest of till speciﬁc oﬁ’enses. followed by robbery and
theft. The lowest rate is for illicit drugs. followed by homicide.
forgery] fraud] embezzlement. and sexual assault. (p. 3)

Thus. it became quickly apparent as one read the literature relative to the subject
of criminal recidivism that term deﬁnition. factor examination. standards of factor
measurement and. in turn. conclusions regarding the scope and nature of criminal

 

Chapter II: Review of Literature

recidivism. vary greatly. In an article on criminal recidivism research. Nacci (1978)
concluded with a hopeful and highly global view:

As a very young science. criminological research must necessarily
grope for facts. never certain where important advances will be
made. Clinical psychologists measuring changes in self-concept.
unit managers assessing the eﬁectiveness of a unit's drug program.
researchers studying demographic characteristics and eﬁects on
criminal activity. and academic sociologists building complex
system simulations will all make contributions to the burgeoning
pool of data. Let us not jump epochs too quickly. Science eventually
will winnow the chaff from the wheat. but at this point in time it is
diﬂcult to tell one from the other. (p.25)

Related Factors

Homnan and Beck (1984) cited empirical studies which reported an association
between age at time of release from prison and recidivistic behavior. For example. they
referred to a report by the National Parole Institutes which concluded:

One of the most ﬁrmly established pieces of statistical knowledge
is that the older a man is when released from prison. the less likely
he is to return to crime. Such ﬁndings have been reported for many
decades. and in numerous jurisdictions. both in the 0.8. and abroad.
the easiest interpretation of this ﬁnding is that people become less
criminal as they become more mature. (National Parole Institutes.
I 964)

The authors (Homnan and Beck. 1984) noted the magnitude of the association
between age and recidivism must be established by controlling for the aﬁ'ects of other
variables known to be associated with recidivism. such as prior criminal record. They
studied this issue using data from a large (N=6.248) group of federal prisoners. The
results indicated that recidivism rates declined with increased age. and the

 

Chapter II: Review of Literature

association was not diminished when statistical control was exercised for the aﬁ'ect of
prior criminal record.

Wallerstedt (I 984) reported that age is signiﬁcantly related to criminal recidivism.
and the younger the age at release. the higher the likelihood ofbeing retumed to prison.
In examining the race variable. he found that “... consistently lower rates are observed
for white releasees” (p. 5). He also found in examining the factor of sex that: “... the
portion of recidivists among males was substantially higher than for female releasees”
(p. 4). However, he cautioned that: ‘... for both raceand sex. it is not known the degree
to which compositional diﬂ'erences across these groups (such as age. offense, or
criminal history) may be contributing to the obsemed difference in recidivism rates” (p.
5).

In another study by Homnan and Beck (1985) . recidivism among released federal
prisoners was measured by means of a 'salient factor score” and ﬁve-year follow-up.
The ﬁnal score was obtained by adding up six (6) intermediate scores (ranging in value
from zero (0) to three (3) points) for each of the following six (6) factors: prior
commitmentls) of more than thirty (30) days as an adult or juvenile: age at the time
of current oﬁ‘ense; age at time of prior commitment(s); recent commitment free period
(three (3) years): probation/parole/conﬁnement/ escape status violator in regard to
the current offense: and heroin] opiate dependence. According to the researchers. the
salient factor score retained predictive power when the follow-up period was extended
to ﬁve (5) years and the deﬁnition of recidivism was restricted to those cases that
sustained a new sentence of imprisonment exceeding one (1) year: the outcome
measure thus focused on the most serious known instances of recidivism (p. 506).

Seashore. Habefelde. Irwin. and Baker (1976). in a follow-up study involving
released inmates who had participated in a variety of college programs while in prison.
examined a large number of background characteristics in connection with the factor
of recidivism. including the following: age at release: time served this sentence: type
of present oﬁense (violent. property. drugs. etc): prior arrests: prior felony convictions:
record of excessive use of drugs or alcohol; education completed prior to present

 

Chapter II: Review of Literature

commitment: education upon release from present commitment: tested grade level:
social class: race: and work history. According to the authors. “... the purpose of these
analyses was to determine the relative success of the ex-prisoners in this study.
comparing participants in ditferent programs with each other and with the comparison
and control groups of nonparticipants" (p. 87).

Gaither (1983) reported that several studies examined post-release criminal
behavior of inmates who had participated in prison academic programs. and that some
of these studies reported a nonsigniﬁcant relationship between participation in the
program and recidivism. He cited these examples:

0 There is no clear evidence that education programs reduce
recidivism. Martinson. 1974

o A slightly negative relationship exists between participation in
educational programs and success on parole. Coombs. 1 965

c There is no relationship between parole success and years of
school completed. Kassebaum at at. 1971: Arizona Department of
Corrections. 1 976; and New York State Division of Parole. l 964

o Inmates enrolled in prison education programs had a higher
recidivism rate than prisoners who were never enrolled in such
programs. Glaser. 1 964

o Inmates who participated in prison college programs generally
returned to their previous lifestyle after leaving prison. Lewis.
1 973

However. Gaither (1983) pointed out: “... not all the research is negative. Other
research indicates that a positive relationship exists between participation in education
programs and recidivism” (p. 84). He cited these examples:

. A signiﬁcant relationship exists between education and success
aﬁer release from prison. Schnurr. 1 948: and Lanne. 1 935

 

Chapter II: Review of Literature

31

c There is a signiﬁcant diii’erence in parole success rates between
students and non-students that favored students. Saden. I 962

0 Individuals who had made satisfactory educational progress had
lower recidivism rates than controls: however. the differences
were not signiﬁcant. Kusuda and Babst. 1964

o Attainment of a certain educational level appears to lower the
likelihood, of recidivism. Waldron. 1974; and Thomas. 1957

In his own study. Gaither (1983) took a look at a number of community college
programs operating in some of the Texas prisons. He studied a group of 710 former
inmates. “... 360 of whom had not participated in the junior college program" (p. 84).
He reported that fourteen percent (1 4%) of the college group were recidivists. and
thirty-two percent (32%) of the non-college group were recidivists.

In another recently completed study. Craig (1983) examined the question: Does
participation by inmates in college-level academic programs reduce the rate of
recidivism? He reported that data analysis gave no evidence of a relationship between
participation in educational programs and recidivism. He did. however. ﬁnd a
sigiiﬁcant relationship between graduation from an associate degree program and
criminal recidivism.

Haviland (1982) reported: ‘There was not a signiﬁcant difference in the rate of
recidivism between those inmates who had been graduated from a two-year college
program while incarcerated and those inmates who had not graduatedfroma two-year
program while incarcerated. " (abstract).

Blackburn (198 1) studied the relationship between recidivism and participation
in a program oﬁ'ered by Hagerstown Junior College for incarcerated oﬁ‘enders at the
Maryland Correctional 'D'aining Center (MCTC) in Hagerstown. Maryland The number
of participants available was ﬁve-hundred sixty-one (561). Two (2) groups of two-
hundred forty-three (243) subjects each (n=243) were matched on the basis of six (6)
variables. including race. age at time of release. and date of release. Over the period

 

Chapter II: Review of Literature

32

of the study (1970- 1 978). one-hundred eighty-nine (189) cases were ﬁnally subjected
to data analyses. A major ﬁnding of the study was that participation in the college
program while conﬁned at MC'IC had an affect on recidivism in two (2) major regards:

1) A reduction in the absolute recidivism rate appears to have developed as
a result of program participation.

2) An increase in the ratio of release time also appears to have developed
from program participation.

Hoke and Holloway (1986) reported on preliminary research relating the two (2)
variables. recidivism and education. Three (3) groups were studied: 1) one-hundred
(100) inmate- students (at the Lebanon Correctional Institution in Ohio) who graduated
with an associate degree from Wilmington College of Ohio. and were paroled during
1982 and 1983 (n=95): 2) one-hundred (100) inmates who had a high school diploma
or GED Certiﬁcate and attended no more than two (2) quarters of the associate degree
program. and were paroled during 1 982 and l 983 (n= 1 6): and 3) one-hundred inmates
(100) from the general prison population who reported no high school diploma or GED
Certiﬁcate and had no contact with the associate degree program. and were paroled
during 1982 and 1983 (n=1 13) (Wilmington College of Ohio - Oﬂce of Continuing
Education. 1986).

'I‘hepurposeofthestudywastodetermineifthecollegegraduates wouldupon
parole reintegrate more successfully than their non-college counterparts. " Recidivism
rates for the three (3) groups were 1 l .6. 15.5. and 28. 3 percent. respectively. indicating
the recidivism rate was dependent on the education variable. Other variables were also
measured for each group. including employment status at arrest and at the end of the
ﬁrst year on parole. parole performance. prior 1 uvenile incarceration(s) . and prior adult
incarceration(s). Such information was drawn from their prison records held by the

Ohio Adult Parole Authority (Wilmington College of Ohio - Oﬂce of Continuing
Education. 1986).

 

Chapter II: Review of Literature

33

“The data showed that linear relationships exist between educational attainment
in prison. employment on parole. and freedom from re-incarceration. " ”This ﬁnding
reinforces the thesis that the more education one receives in prison. the better his
chances for successful reintegration . .. " into society upon parole from prison. However.
'the relatively small number of people included in the study. in coryunction with the fact
that an experimental methodology was not used. means the research results must be
viewed with caution". “Nevertheless. as adescriptive study of three hundred oﬂenders.
the recidivism research is illuminating. The noteworthy differences in the post-release
performance of the comparison groups provide empirical support for the proposition that
trwesting in correctional education is a prudent use of public resources." (Wilmington
College of Ohio - once of Continuing Education. 1986).

By way of concluding remarks on this topic. McCollum (1977) indicated that it is
unrealistic to measure the effectiveness of a particular prison program in terms of
recidivism alone. She emphasized the importance of total prison experience and
various other factors such as a person's life history and the quality of that life
at the time of incarceration. “Additionally. postrelease family and other socioeco-
nomic connections. ifany. access to opporttmity systems. mental and physical health,
and a host of other variables contribute substantially to an individual's behavior on
release from incarceration." (p. 32).

The C.O.P.E. Program at Montcalm Community College
Brief History

Montcalm Corrununity College opened in 1967 to serve residents of Montcalm
County. It is located in the west-central region of the lower peninsula of Michigan. The
County is primarily rural. and has a current population of about ﬁfty-seven thousand
(5 7,000) persons. The College is the only post-secondary educational institution within
a ﬁfty (50) mile radius of Sidney. the town in which the College is located. Being a
community college. there is a heavy emphasis on and commitment to vocational
education (Community Corrections Resource Programs. Inc.. 1976).

 

Chapter II: Review of Literature

34

In the ﬁscal year 1968. several classes were offered to prison inmates at the
Michigan Reformatory (L02). located in Ionia. Michigan. In 1969. Montcalm Community
College extended college credit to those inmates who took classes through what
became termed the College Opportunity - Prison Extension (C.O.P.E.) Program The
Michigan Department of Corrections provided funding for these courses. and continued
to provide ﬁnancial support to the program up until 1972.

In that same year (1972). the Department of Health. Education. and Welfare (HEW)
awarded funds to Montcalm Community College for the expansion of the C. O.P.E.
Program In addition to offering more courses at the Michigan Reformatory. the C. O.P.E
Program opportunities were extended to inmates at the Michigan Training Unit. also
located in Ionia. Michigan. Further. inmates at these two (2) prisons were oﬁ'ered the
opportunity to earn an associate degree from MCC (Community Corrections
Resource Programs. Inc.. 1976). In 1977. the Programwas extended to the Riverside
Correctional Facility. the third prison to be located in Ionia. Michigan. As of the fall of
1986. three-hundred twenty-three (323) inmates had graduated from the Montcalm
Community College C.O.P.E. Program with an associate degree.

Current Operational Structure

As of 1987. two (2) new prisons became operational in the Ionia area: the Ionia
Temporary Facility. and the Ionia Maximum Facility. The C. O.P.E. Program expanded
to the Ionia Temporary Facility. and now serves four (4) of the ﬁve (5) prisons in the area.
The Ionia Maximum Facility houses inmates who are highly assaultive. and are seldom

released from their cells. Thus. educational opportunities were not extended to
inmates in that facility.

Through the C.O.P.E. Program. inmates can take classes leading to an associate
degree in General Studies. an associate degree in Arts and Sciences. or an associate
degree in Applied Arts and Sciences with a concentration in: Business Administration.
Accounting. Business Data Processing. or Food Service Technology. The College also
offers one-year certiﬁcates in certain vocational areas such as Food Service. In the
spring of 1989. some sixty-seven (67) courses were offered the inmates of the four (4)

 

Chapter II: Review of Literature

35

prisons in areas including: English, Speech, Advertising. Social Science. Psychology.
Blueprint Reading. Small Business Management. Stress Management. Human Rela-
tions in Business. Reading. Writing. Typing. Introduction to Business. Problem Solving.

Hw'nanities. Algebra. Legal Research and Writing. Marketing and Estates. Wills. and
Trusts.

 

Chapter II: Review of literature

 

RESEARCH DESIGN

 

Introduction

The main focus of this ex—post facto observational study was to examine the
relationship between inmate graduation with an associate degree from the College
Opportunity - Prison Extension (C.O.P.E.) Program and rates of recidivism. Simply
examining these two (2) factors (post-secondary education and recidivism) alone
would have been senseless because it would not have allowed for consideration of the
many other conditions (factors) which research and/ or common belief indicated are
signiﬁcantly related to recidivistic behavior. Under such a simplistic approach. one
could not determine if the study ﬁndings were attributable to the education variable

or to one or more of the other factors.

Thus. a statistical design was developed which helped identify. individually and in
cluster form. the predictive/ causative factors associated with recidivism. Further. the
design also allowed for the examination and measurement of the lone relationship
between post-secondary academic education and recidivism. This was done through
the application of a series of statistical techniques such as crosstabulation analyses
(standard. dichotomous, collapsed. collapsed with a control. three-way with a control.
and/orfour-way with a control). analysis of variance (ANOVA). and the main study

statistic - a discriminant function analysis with a forward selection (stepwise) algo-
rithrn.

What follows is an explanation in categorical form of the research design relative
to sample selection. study controls. data collection tools. procedures for data collection.
and statistical analyses performed on the data:

 

36

37

Sample

The primary study subjects formed Group I. and totaled one-hundred sixteen
(n=1 1 6) former inmates. They were selected from a list of all inmates who had
graduated from the C. O.P.E. Program with an associate degree as of September 9.
1986. With proper assurances that conﬁdentiality would be maintained. this list was
provided by the Director of the C. O.P.E. Program. Mr. Danny Herman. and consisted
of three-hundred twenty-three (323) potential study subjects. Of the three-hundred
twenty-three (323). two-hundred seven (207) were eliminated for the reasons listed
below (see Figure 3. 1 . p. 38): ’

1)

2)

3)

4)

Death in the institution or while on parole - 3 eliminated

One (1) of the potential subjects died while in prison. and the other two (2) died
while on parole. 'lhose who died while on parole were eliminated because they
had not remained alive for a period of two (2) years following parole.

Discharged by the court - 2 eliminated

These two (2) potential subjects had their convictions reversed by the court. and

were discharged from prison. Thus. they were no longer considered convicted
felons.

Received an outstate parole - 15 eliminated

These ﬁfteen (15) potential subjects received paroles to states other than
Michigan. Because of the varying conditions of parole supervision around the
country. itwas determined they would not make proper study subjects. Namely.
they would bring to the study a set of factors different from those who served
their parole in the state of Michigan.

Still in prison as of December 31. 1984 - 96 eliminated

These potential subjects had not received paroles as of December 3 1 . l 984. and
thus could not be indiuded in the study.

 

Chapter III: Research Design

C.O.P.E. GRADUATES
Study Selection Process

Death/Discharge
5

Selected Subjects
1 16

In Prison
96

 

No Records
47 Outside S‘tzrdy Limits

Figure 3.1

 

5) Outside study time limits - 44 eliminated

These potential subjects received paroles either prior toJanuary 1 . 1980 or after
December 31. 1984. Thus. they were outside the time limits of the study.

6) Physical records not available - 47 eliminated

The Michigan Department of Corrections maintains institutional and central
omce records (physical) on former prisoners for ﬁve (5) years following
completion of parole. with the exception of individuals who have extended
paroles (more than two years). After that time. only computer ﬁles can be
retrieved. These computer ﬁles contain only sketchy information relative to

 

Chapter III: Research Design

39
these former prisoners. most of which is related to conditions at the time of their
entry into prison. Thus. forty-seven (47) potential subjects were eliminated
because physical records from which the study data was to be collected were
not available.

Determination as to the ﬁrst ﬁve (5) sets of conditions were made using the
Conectional Management Information System (CMIS) This computerized information
system. which is maintained by the Michigan Department of Corrections. provides basic
entry data and institutional activity data on each active prisoner. and on those inmates
or former inmates who are within a ﬁve (5) year period following parole. These computer
ﬁles are periodically purged so as to rid the CMIS of outdated information: the purged
information is then archived on computer tape.

The ﬁnal group of forty-seven (47) potential subjects were eliminated from the
original pool upon discovering. through a check with the Michigan Department of
Corrections (records department). that physical records were not available on these
individuals. Only the archived information was available on these subjects. and it
could not provide the background information necessary for the conduct of this study.

Controls

The secondary (comparison) subjects consisted of one-hundred sixteen (n=1 1 6)
former inmates. and made up Group 11. They matched on a one-to-one basis on certain
study variables (factors) with the subjects in Group I. The ﬁrst step in the matching
process began with a service request submitted to the Michigan Department of
Corrections (data processing) on March 2. 1987 (see AppendixJ. p. 156). The request
asked for a listing on computer tape and printout of all persons who met these three
(3) criteria:

1) Received a parole from January 1. 1980 through December 31. 1984
(inclusive).

 

Chapter III: Research Design

40

2) Served all or part of their incarceration at an institution in Ionia.
Speciﬁcally:

a. MTU. the Michigan Training Unit

b. RCF. the Riverside Correctional Facility

c. RMI. Ionia Reformatory

3) Did not have an academic educational level equal to two (2) years of

college or above upon parole.

In a phone conversation on July 15. 1987 between Mr. Terry Murphy. Chief of
Research for the Michigan Department of Corrections. and Mr. Larry Walker. Analyst
with the Data Processing Section of the Michigan Depariment of Corrections. it was
determined the following information could be included as part of the output without
delaying the original request (see Appendix J. p. 157):

1)
3)
3)
4)
5)
6)
7)
8)
9)
10)

1 1)
12)

Prisoner 1.D.

LD. Preﬁx

Ionia institution placement history (locations and dates)
Highest grade at prison entry

Date of birth

Date received at prison

Race

Marital status at arrest

Occupation at arrest

Parole date (if multiple paroles in applicable term. include all of the
dates)

Assaultive risk classiﬁcation

Academic educational level at commitment (highest grade)

With the data relative to the forty-two (42) study variables collected for members
of Group I and the requested tape and printout in hand. the initial steps in the matching

Chapter III: Research Design

41

process were begun. The subset of six (6) study variables involved in the matching
process were (in order of priority):

1) Age at parole from present prison commitment

2) Academic educational level at time of instant offense(s)

3) Employment status at time of instant oﬁense(s)

4) Michigan Department of Corrections (MDOC) assaultive risk classiﬁcation
at time of parole from present prison commitment

5) Race

6) Marital status at time of instant offense(s)

It was determined from the printout that data in regard to one-thousand nine-
hundred and eighty-three (1 .983) potential Group II subjects was available on the
tape. It was anticipated that each potential subject had one (1) Master Record. one (1)
Identiﬁcation Record. and one (1) or more 'Il'ansit Records. The Master Record provides
such data as prison preﬁx. prison number. date of entry into prison. birth date. sex.
and race. The Identiﬁcation Record provides data such as marital status at time of
arrest. highest academic grade at time of prison entry. occupation at time of arrest.
and the individual’s Michigan Department of Corrections assaultive risk classiﬁcation
at time of parole. The Transit Record(s) contains data as to which institutions the
person resided in during their incarceration period. and date of parole. The tape ﬁle
contained a total of twelve-thousand seven-hundred and thirty-four (12.734) ﬁles (not
to be mistaken as representing that many individuals). The ﬁrst one-hundred (100)
records were printed out in order to provide reasonable assurance the information

requested was contained on the tape.

It was discovered the tape contained numerous “orphan” 'n’ansit Records. where
“orphan” meant there was no Master Record or Identiﬁcation Record which by prisoner
identiﬁcation number (prison number) matched. Through the use of the Statistical
Package for the Social Sciences (SPSS). a program was written to rid the ﬁle of “orphan”
records. to consolidate the three records per individual into a single record. and to
output the ﬁeld names and column positions of the data contained on the tape (see

 

Chapter III: Research Design

42

AppendixM. p. 1 79). In order to become a potential subject. each individual had to have
a consolidated record consisting of a Master Record. an Identiﬁcation Record. and one
or more Transit Records. This purging process yielded one-thousand nine-hundred
thirty-three (1.933) potential study subjects.

Another SP$ program was written which pulled data relative to the six (6)
matching variables from the master tape provided by the Michigan Department of
Corrections. and placed that data in an output ﬁle (see Appendix M. p. 180). A third
SPSS program was written which created a new variable (OCCSTA'I), which was related
to the occupational status of the potential subjects at the time they were arrested for
the crime which resulted in their incarceration.

The reason for the creation of the new variable was that the occupation coding
format used by the Michigan Department of Corrections in the Identiﬁcation Records
was not directly translatable to the occupation variable coding used in the study.
Speciﬁcally. the Michigan Department of Corrections uses a three digit code which
reveals the type and / or nature of the work performed by the subject prior to
incarceration. The study. on the other hand. was concerned with the work history of
each subject. Namely. the study was concerned with whether the subjects had work
historiea. and if so whether they worked full-time. part-time. were unemployed. a
student. etc. Thus. the new variable was created by a procedure which converted
occupation type to work history. Further. those who were disabled. a student. or whose
work history was unknown or unavailable were identiﬁed and coded accordingly.

Following the occupation variable conversion. the program wrote the data related
to the six (6) matching variables to an output ﬁle on a diskette. and output the data
in hard copy form as well.

Next. this SPSS output ﬁle (in ASCII form) was imported into dBase HI Plus. and
the coding used by the Michigan Department of Correctionswas converted to the coding
format used by the study. For example. in regard to the variable of race: the Michigan
Department of Corrections uses“ B for black. W for white. M for Mexican; whereas the
study codes were set at one (1) for black. two (2) for white. and three (3) for Hispanic.

 

Chapter m: Research Design

43

Great care was taken in the conversion of the codes to be certain the conversion did
not affect the integrity of the data.

Using this data ﬁle and the data ﬁle collected on Group 1 subjects. a program was
written in dBase III Plus (a microcomputer database program) for the purpose of
matching each subject in Group I with a counterpart (on the six study variables) from
the Group II Subject Pool. The ﬁrst run of the program called for a literal matching of
subjects. and yielded thirty-eight (38) matches. On the second run. the matching
routine on the age variable was changed to allow a matching of one year either side
of Group I subject age. This run provided another twenty-one (2 1). for a total of ﬁfty-
nine (59) matches. The third run allowed for a change of one year either side of the
educational level variable. except for those with a completed high school diploma or
a GED certiﬁcate. In those two instances. no change (range) was permitted. This run
yielded another fourteen (14) matches. for a total of seventy-three (73).

The fourth run established some ranges for matching on age and educational level.
Those subjects in Group I whose age was between nineteen (19) and twenty-two (22)
were matched with subjects twenty-two (22) years of age and under: those subjects in
Group 1 whose age was between twenty-three (23) and thirty-four (34) were matched
with subjects who fell within that range: those subjects in Group I whose age was
thirty-ﬁve (35) through forty-six (46) were matched with subjects who were thirty-ﬁve
(35) and over.

As for the educational level matching: subjects in Group I with an educational level
at time of arrest between eight (8) and eleven (1 l) were matched with subjects whose
educational level fell within that range: those subjects in Group I whose educational
level was twelve (12) were matched with subjects with an educational level of twelve
(12): those subjects in Group I who had a GED certiﬁcate were matched with subjects
who had a GED certiﬁcate: and those Group I subjects whose educational level was
between thirteen (l3) and sixteen (16) were matched with subjects who had
educational levels of thirteen (13) and above. but did not possess a post-secondary

 

Chapter III: Research Design

44

degree. This matching run yielded another eleven (11) matches. for a total of eighty-
four (84).

The remaining thirty-two (32) subjects (1 16-84) were matched by hand. giving
strict attention to four (4) variables: age at parole. academic educational level at time
of arrest. Michigan Department of Corrections assaultive risk classiﬁcation. and race.
The hand matching process was conducted with great care and concern. and the
researcher was conﬁdent the matching reﬂected the best result possible.

Where possible. a match was made between a member of Group I and a member
of the Group II Subject Pool when they matched exactly on the six variables. Where
a number of potential subjects matched the Group I subject exactly. a random
selection process was employed in choosing the Group 11 matching subject. In the case
of subjects who matched as a result of widened parameters. the best match was
selected manually. Again. in the case of identical potential subjects. a random

selection process was employed.

Despite a giant effort to match each subject in Group I (C. O.P.E.) with a counterpart
from the Group II (comparison) Subject Pool. this was not fully possible. The
application of various crosstabulation analyses. and an analysis of variance (ANOVA)
(see “Statistical Analyses.” p. 47 and “Group Comparisons.” p. 78) clearly
indicated the two (2) Groups were properly matched on only ﬁve (5) of the six (6)
variables. The matching procedure used did not result in Group matchings on
“academic educational level at time of instant offense(s)." to the point where the
diﬁ'erences between the two (2) Groups proved non-signiﬁcant at the .05 level. The
inability to match the Groups on this variable related to the hand-matching step which
garnered the remaining thirty-two (32) subjects from the Group 11 Pool.

It proved impossible to match the two (2) Groups on all six (6) of the matching
variables. due to the limited size (1.933) of the Group 11 Pool. It was decided this non-
matched variable. “academic educational level at time of instant oﬁense(s)."
could be properly dealt with by applying other statistical controls.

 

Chapter 111: Research Design

45

Data Gathering Tools

The instruments used in data collection consisted of a variables dictionary. coding
sheet. and attitudinal survey - all of which were developed by this researcher.

The variables dictionary (Appendix K. p. 158) provided general instructions to the
coders: the name of each variable; a description in commentary form of each variable:
a data source listing which explained where in the inmate records the speciﬁc data for
each variable could be found: a numeric coding scheme which assigned a numeral
(numeric code) to every possible condition for each gven variable: and. three (3)
appendices which provided the coders with: l) a listing of property/ drug/ violent
oﬁenses by name and crime category. 2) a listing of common misdemeanor offenses
so the coders would not mistake them for felony offenses. and. 3) a listing of non-
bondable offenses which might be committed by an inmate while in a prison setting.
The latter are called misconduct reports by prison ofﬁcials. but are commonly called
“tickets” by the inmates.

A coding sheet (AppendixL. p. 1 77) was developed for each subject involved in the
study. and simply allowed for the entry of numeric codes for each of the forty-two (42)
variables. There were also spaces provided to enter each subj ect's name. prison
number. Social Security number. and address at time of parole. This latter information
was not entered in the computer ﬁles. and was collected for the purpose of running
L.E.I.N. checks and. in the case of Group I members. to send out copies of the
attitudinal survey. Lastly. a comments section was included to provide a place where
coding problems could be addressed.

Coders were cross-checked in two (2) ways: 1) this researcher randomly selected
coded records and recoded them to ensure accuracy of the data. and 2) records
checked by one (1) coder were given to another coder to ensure intercoder reliability.
Where discrepancies were noted. diﬁ'erences were discussed in group form with this
researcher providing the group discussion leadership.

 

Chapter III: Research Design

46

The coders were instructed to set aside records which presented them with some
form of problem such as missing data. conﬂicting data. etc. Further. they were
instructed to describe the nature of the problem in the comments section of the coding.
Then. this researcher checked and resolved these problems on a daily basis during the
data collection period.

Once the coding sheets were completed. a data entry clerk entered the data into
dBase m Plus. The accuracy of data entry into dBase HI Plus was checked on a random
record basis. In all. thirty (30) records in each Group were manually checked by this
researcher. item for item. Also. a hard copy of the complete data set for both groups
was secured and an “eyeball” check was made for any codes which appeared to be

inaccurate.

The attitudinal survey instrument included a cover letter (Appendix N. p. 181)
directed at each member in Group I. explaining the purpose and intent of the study.
It also provided these subjects with directions on how to ﬁll out and return the survey
to this researcher. The main intent of the instrument was to gather information on why
they decided to further their education while in prison. and to also determine if
obtaining a college degree proved in their opinion helpful upon their release to the "free
community.”

Data Collection Procedures

The Michigan Department qf Corrections. Records Bureau. supplied the physical
records on all the subjects. These records were made available at their omce site in
the Steven T. Mason Building. Lansing. Michigan. They were also kind enough to
provide omce space where the manual extraction of data took place.

Many of the subjects were inactive. meaning they had completed the requirements
of parole. Thus. their records were not in the active ﬁle collection at the Steven T. Mason
Building. Those records were secured by a Michigan Department of Corrections
employee from the main records depository in Lansing. Michigan and brought to the

 

Chapter 111: Research Design

47

coding site (Steven T. Mason Building). Upon completion of the coding. all records were
returned to the records department so they could be placed back on ﬁle.

On a weekly basis. the data on the coding sheets was entered into the database by
a data entry clerk. The entire data collection effort took about three (3) weeks per
Group.

The attitudinal survey was sent only to members of Group I. This was accomplished
through the mail merge function of an electronic word processor (WordPerfect). The
secondary ﬁle consisted of the names and addresses of all members in Group I (and
was available only to this researcher). Prior to setting up this secondary ﬁle. all
members of Group I were checked on the OMS (by an MDOO employee) to determine
if they were active. where active meant they had been returned to prison. In the case
of active subjects. the attitudinal survey was sent to them at their prison address. This.
it was felt. resulted in a better questionnaire response than would have otherwise been
the case. The primary ﬁle consisted of the cover letter and the attitudinal instrument
itself. Aretum envelope with postage afﬁxed and addressed to this researcher at a post
omce box located on the campus of Michigan State University was sent to each Group
I subject.

The questionnaire consisted of twenty-ﬁve (25) questions. with ﬁve (5) possible

responses to each question: strongly agree. agree. disagree. strongly disagree.
and undecided. The returns were tabulated and a data entry clerk entered the data

into a dBase m Plus database. structured speciﬁcally for that purpose.

Statistical Analyses

The full data setwas imported into Lotus 1-2-3 from dBase 111 Plus. and ahard copy
related to each of the two (2) Groups was printed.

The full data set containing data on both Groups of subjects was also brought into
SPSS/PO+. using the translate utility that accompanies that application.

 

Chapter 111: Research Design

48
To verify the accuracy of both translations. the data from seven (7) ﬁelds in each
of twenty-four (24) (twelve from each Group) records contained in the SPSS/PO+ ﬁle
was output in hard copy form. These two (2) hard copy data sets were manually
compared. and they matched perfectly with each other and with the original data set.
providing strong evidence that a loss of data integrity had not been experienced.

The opening task in the analyses process was to examine the full set of variables.
with the intent of reducing their numbers so as make those which remained more
manageable. It was determined the eight (8) variables listed below served only as
information variables (many were used in the calculation of other variables) and could
thus be eliminated from further consideration and analyses:

#01 Subject Number

#02 Date of Birth

#04 Date Of Prison Entry For Instant Oﬁensem

#07 Date Of First Arrest (Age at ﬁrst arrest remained)
#27 Date Of First Parole For Instant Oﬁenseis)

#28 Place From Which Paroled

#29 Parole Placement

#38 Year Of Graduation From C.O.P.E. (Group I only)

Next. numerous statistical analyses were run on the remaining thirty-four (34)
variables to help determine the conditions on which the two (2) study Groups
signiﬁcantly diﬁ'ered. and those on which they were alike. A description of those

analyses follows.

The ﬁrst statistic run was a series of crosstabulations on the remaining thirty-four
(34) study variables:

#03 Race

#05 Academic Educational Level At Time Of Instant Offenseis)
soc Type or Instant Oﬁ‘ense(s)

#08 Prior Adult Felony Conviction(s)

 

Chapter 111: Research Design

#09
#10
#11
#12
#13
#14
#15

#10

#17

#18

#19

#20

#21

#22

#23

#24

#26

#30

#31
#32

49

Criminal Proﬁle — Juvenile Property Oﬁense(s)

Criminal Proﬁle — Juvenile Drug Offense“)

Criminal Proﬁle — Juvenile Violent Oﬁenseis)

Criminal Proﬁle -- Adult Property Oﬁ’ense(s)

Criminal Proﬁle — Adult Drug Oﬁense(s)

Criminal Proﬁle — Adult Violent Oﬁense(s)

In The Community At Least Three Years Prior To Prison Commitment
For Instant Oﬁense(s)

Marital Status At Time Of Instant Oﬁense(s)

Employment Status At Time Of Instant Oﬁense(s)

History Of Substance Abuse

Evidence Of A Serious Physical Illness Or Disability At Time Of Instant
Oﬁense(s)

Evidence Of A Serious Emotional Or Psychological Problem At Time Of
Instant Oﬁense(s)

Upbrlnsins

Financial Status Of Upbringing Environment

Academic Educational Level Of Mother At Time Of Subject's Instant
Offense“)

Academic Educational Level Of Father At Time Of Subject’s Instant
05st)

Family Emotional Support System At Time Of Subject’s Instant
Offense(s)

Number of Non-Bendable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment

Academic Educational Level Upon Parole From Present Prison Com-
mitment

unoc Assaultive Risk Classiﬁcation At The Time or Parole
Evidence Of A Serious Physical Illness Or Disability At Time Of Parole
From Present Prison Commitment

 

Chapter III: Research Design

#33

#34
#35
#37
#33
#39

#40
#41
#42

hidence Of A Serious Emotional Or Psychological Problem At Time Of
Parole From Present Prison Commitment

Age At First Arrest

Age At Time Of Present Prison Commitment

Age Upon Parole From Present Prison Commitment

Length Of Time Served For Instant Offense

Academic Educational Attainment Level Increase During Present
Prison Commitment

Criminal Recidivism — Property Offense(s)

Criminal Recidivism — Drug Offense(s)

Criminal Recidivism — Violent Oﬁ'ense(s)

Next. a dichotomous (yes/ no) crosstabulation was run on the ten (10) study

variables listed below:

#09 Criminal Proﬁle — Juvenile Property Oﬁense(s)

#10 Criminal Proﬁle — Juvenile Drug Offense(s)

# 1 1 Criminal Proﬁle — Juvenile Violent Oﬁense(s)

# 12 Criminal Proﬁle — Adult Property Offense(s)

# 18 Criminal Proﬁle — Adult Drug Oﬁ’ense(s)

# 14 Criminal Proﬁle — Adult Violent Oﬂense(s)

#28 Number of Non-Bendable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment

#40 Criminal Recidivism — Property OﬁenseIs)

#41 Criminal Recidivism — Drug Oﬁense(s)

#42 Criminal Recidivism — Violent Oﬁense(s)

In an effort to simplify the dependent variable (recidivism). a new variable called
RECIDALL was formed. By employing a dichotomous yes / no) crosstabulation
statistic. this researcher was able to determine if subjects in either Group recidivated.
Thus. instead of examining the three (3) possible types of recidivistic behavior deﬁned

 

Chapter III: Research Design

51

in this study (see AppendixK.. p. 1 68). one could simply determine on ayes or no basis
if the subj ect(s) recidivated. Thus. a yes/ no answer in regard to recidivistic behavior
became available when needed.

Some of the variables had outcome categories too numerous to serve the study in
a practical and meaningful way. Thus. collapsed bivariate crosstabulations were run
on twelve (12) of the study variables:

#03
#05
#06
#16
#17
#18
#22
#25

#26

#30

#31
#34

Race

Academic Educational Level At Time Of Instant Oﬁense(s)

Type Of Instant Oﬁense(s)

Marital Status At Time Of Instant Oﬁ'ense(s)

Employment Status At Time Of Instant Oﬁense(s)

History Of Substance Abuse

Financial Status Of Upbringing Environment

Family Emotional Support System At Time Of Subject's Instant
Oﬁense(s)

Number of Non-Bendable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment

Academic Educational Level Upon Parole From Present Prison Com-
mitment

MDOC Assaultive Risk Classiﬁcation At The Time Of Parole

Age At First Arrest

A three-way crosstabulation statistic controlling for “oﬁense type“ (non-violent or

violent) by Group (I or II) was run on the three (3) outcome variables related to
recidivistic behavior. which were:

#40 Criminal Recidivism — Property Offense“)

#41

Criminal Recidivism — Drug Oﬁense(s)

#42 Criminal Recidivism — Violent Offense(s)

 

Chapter III: Research Design

A three-way crosstabulation controlling for “oﬁense type” (non-violent or violent)
by Group (I or II) was also run on RECIDALL. the generated variable which allowed
for a yes/ no determination of recidivism.

The variables dealing with “academic educational level at time of instant
oﬁ’enseis)” and “academic educational level upon parole” were made dichotomous
by dividing subjects into high school graduates and non—graduates. Thus. a dichoto-

mous crosstabulation analysis by group (I-IS - Y/ N) was performed on the following
variables:

#05 Academic Educational Level At Time Of Instant Oﬁense(s)
#30 Academic Educational Level Upon Parole From Present Prison Com-
mitment

A four-way crosstabulation statistic controlling for “education at time of instant
oﬂense(s)” (HS - Y/N). and “education at time of parole” (HS - Y/N) by Group (I or
II) was run on RECIDALL. This was done to provide insight into the impact the C.O.P.E.
Program had on those who entered prison with a high school diploma or equivalent
(GED Certiﬁcate).

Another four-way crosstabular analysis was run on RECIDALL. controlling for
“age upon parole“ and “history of substance abuse" (< 26. no history of substance
abusel> 26. with a history of substance abuse) by Group (I or II). and “age upon
parole” and “academic education at time of instant oﬁense(s)" (< 26 with no high
school diploma or GED Certiﬁcate / > 26 with a high school diploma or GED Certiﬁcate)
by Group (I or II). This was done to test the factors which seemed to beneﬁt those who
successfully completed the C.O.P.E. Program of study. An analysis of variance (ANOVA)
was run on nineteen (19) of the study variables:

#05 Academic Educational Level At Time Of Instant Oﬁ’ense(s)
#09 Criminal Proﬁle — Juvenile Property Oﬁ’ense(s)

# 10 Criminal Proﬁle — Juvenile Drug Offense“)

# 1 1 Criminal Proﬁle — Juvenile Violent Oﬁense(s)

 

Chapter III: Research Design

#12
#13
#14
#23

#24

#26

#30

#34
#35
#37
#36
#39

#40
#41
#42

53

Criminal Proﬁle — Adult Property Oﬂense(s)

Criminal Proﬁle — Adult Drug Oﬁense(s)

Criminal Proﬁle — Adult Violent Oﬁense(s)

Academic Educational Level Of Mother At Time Of Subject’s Instant
Oﬁ'ense(s)

Academic Educational Level Of Father At Time Of Subject's Instant
Oﬁ'ense(s)

Number of Non-Bondable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment

Academic Educational Level Upon Parole From Present Prison Com-
mitment

Age At First Arrest

Age At Time Of Present Prison Commitment

Age Upon Parole From Present Prison Commitment

Length Of Time Served For Instant Oﬁ‘enseis)

Academic Educational Attainment Level Increase During Present
Prison Commitment

Criminal Recidivism — Property Oﬁense(s)

Criminal Recidivism — Drug Offense(s)

Criminal Recidivism — Violent Oﬁ‘ense(s)

The statistics run on the data up to this point provided clear evidence that six (6)
of the variables were of no further value to the study. since they offered incomplete
and / or insufﬁcient data to allow for a meaningful contribution to the project. The lack
of complete and sumcient data can be evidenced and supported by examining Chapter
N (“Raw Data. " p. 65). The six (6) study variables eliminated at this point were:

#19

#20

Eyidence Of A Serious Physical Illness Or Disability At Time Of Instant
Oﬁ'enseis)

hidence Of A Serious Emotional Or Psychological Problem At Time Of
Instant Oﬁense(s)

 

Chapter III: Research Design

#23

#24

#32

#33

54

Academic Educational Level Of Mother At Time Of Subject's Instant
Oﬁense(s)

Academic Educational Level Of Father At Time Of Subject's Instant
Oﬁense(s)

Evidence Of A Serious Physical Illness Or Disability At Time Of Parole
From Present Prison Commitment

midence Of A Serious Emotional Or Psychological Problem At Time Of
Parole From Prcsent Prison Commitment

It should be noted at this point that three (3) of the remaining twenty-eight (28)
variables were of the dependent variety. That is. they related to the study outcome -
recidivism. They were:

#40
#41
#42

Criminal Recidivism — Property Oﬁenseis)
Criminal Recidivism — Drug Oﬁ'ense(s)
Criminal Recidivism — Violent Oﬁense(s)

With that circumstance in mind. the next phase of the statistical analyses of the
data set was to determine which of the remaining twenty-ﬁve (25) independent
variables were signiﬁcantly (.05 level) related to the outcome (dependent) variable
(recidivism).

First. three-way crosstabulation analyses of recidivism (through the variable
RECIDALL). controlling for the independent variable (in dichotomous or collapsed
form) by Group (I and II) were performed on the ﬁve (5) variables listed below:

#05
#08
#12
#26

#38

Academic Educational Level At Time Of Instant Offense“)
Prior Adult Felony Convictionis).

Criminal Proﬁle — Adult Property Oﬁense(s)
Number of Non-Bendable Major Misconduct Reports In Prison For

Which The Subject Was Found Guilty During His Present Prison
Commitment

Length Of Time Served For Instant Oﬁense(s)

 

Chapter 111: Research Design

Avariable called prior arrest(s) (PRIORARR) was generated at this point in the data
analyses. It incorporated the variables: juvenile property oﬁenseis). juvenile drug
offense(s). juvenile violent offense(s). adult property offense(s). adult drug
oﬁ'ense(s). and adult violent oﬁense(s) to assist in determining if individual subjects

had any prior arrests.

A three-way crosstabulation analysis of recidivism (through the variable RE-
CIDAIL) controlling for PRIORARR (in dichotomous form - Y/ N) by Group (I or II) was
performed.

Next. crosstabulation analyses of recidivism (through the variable RECIDALL). by
each of the independent variables (e.g. . race) were performed on the remaining twenty-
ﬁve (25) independent variables listed below:

#03
#05
#06
#08
#09
#10
#11
#12
#13
#14
#15

#16
#17
#18
#21
#22

Race

Academic Educational Level At Time Of Instant Oﬁ'ense(s)
Type Of Instant Oﬁenseis)

Prior Adult Felony Conviction(s)

Criminal Proﬁle — Juvenile Property Oﬁ’ense(s)

Criminal Proﬁle — Juvenile Drug Offense(s)

Criminal Proﬁle — Juvenile Violent Oﬁenseis)

Criminal Proﬁle — Adult Property Oﬁense(s)

Criminal Proﬁle — Adult Drug Oﬁ’enseis)

Criminal Proﬁle — Adult Violent Oﬁenseis)

In The Community At Least Three Years Prior To Prison Commitment
For Instant Oﬁense(s)

Marital Status At Time Of Instant Oﬁ'ense(s)

Employment Status At Time Of Instant Oﬁense(s)

History Of Substance Abuse

Upbringing
Financial Status Of Upbringing Environment

 

Chapter 111: Research Design

#25

#26

#30

#31
#34
#35
#37
#38
#39

58
Family Emotional Support System At Time Of Subject's Instant
Oﬁ‘enseis)
Number of Non-Bondable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment
Academic Educational Level Upon Parole From Present Prison Com-
mitment
MDOC Assaultive Risk Classiﬁcation At The Time Of Parole
Age At First Arrest
Age At Time Of Present Prison Commitment
Age Upon Parole From Present Prison Commitment
Length Of Time Served For Instant Oﬁ'ense(s)
Academic Educational Attainment Level Increase During Present

. Prison Commitment

At this point it was determined that sixteen (16) of the twenty-ﬁve (25) independent
variables tested were not signiﬁcantly (at the .05 level) related to the outcome variable
(recidivism). They were thus eliminated from further consideration and analyses.
Those eliminated variables were:

#03
#09
#10
#11
#13
#14
#15

#16
#17
#21
#22

Race

Criminal Proﬁle — Juvenile Property Oﬁense(s)
Criminal Proﬁle — Juvenile Drug Offense(s)
Criminal Proﬁle — Juvenile Violent Oﬁense(s)
Criminal Proﬁle — Adult Drug OﬁenseIs)

Criminal Proﬁle — Adult Violent Oﬁensels)

In The Community At Least Three Years Prior To Prison Commitment
For Instant Oﬁenseis)

Marital Status At Time Of Instant Oﬁense(s)
Employment Status At Time Of Instant Oﬁ’ense(s)
Upbrlnslns

Financial Status Of Upbringing Environment

 

Chapter III: Research Design

57

#26 Number of Non-Bendable Major Misconduct Reports In Prison For
Which The Subject Was Found Guilty During His Present Prison
Commitment

#31 MDOC Assaultive Risk Classiﬁcation At The Time Of Parole

#34 Age At First Arrest

#35 Age At Time Of Present Prison Commitment

#37 Age Upon Parole From Present Prison Commitment

One (1) of the nine (9) independent variables. which proved in a statistical sense
to be signiﬁcantly related to recidivism. was eliminated due to the confounding effect
it was having on other predictor variables. That variable was:

# 12 Criminal Proﬁle — Adult Property Oﬁ'enseis)

Speciﬁcally. it was determined this variable was part of a multicollinearity eﬁ'ect
involving other independent variables. including “prior adult felony conviction(s)”
and “prior arrest(s).” Because this potential predictor variable proved to be highly
intertwined with these other two (2) similar variables. to the point where it would be
extremely Mean to determine its individual inﬂuence on the outcome (dependent)
variable (recidivism). it was eliminated from further consideration and analyses (see
deﬁnition of “Confounding Variable" on p. 1 7. and of “Multicollinearity” on p. 21).

Another of the nine (9) independent variables which proved to be signiﬁcantly (.05
level) related to the outcome variable (recidivism) was eliminated from further
consideration and analyses. That variable was:

#25 Family Emotional Support System At Time Of Subject's Instant
Oﬁense(s)

This independent variable was used as a crosstabular control in looking at
recidivistic behavior (through RECIDALL) by reducing its original ﬁve (5) response
categories (see Appendix K. p. 1 65) to two (2). thus making it a dichotomous (strong
support] some support) variable. In doing so it was discovered that only twenty-eight
(28) of the two-hundred thirty-two (N=232) study subjects came under the strong

 

Chapter III: Research Design

support category. Because of this low 11. the variable was eliminated as a possible ﬁnal
predictor variable.

Athird independent variable which proved to be signiﬁcantly related in a statistical
sense to recidivistic behavior (through RECIDALL). was eliminated from the study.
That independent variable was:

#30 Academic Educational Level Upon Parole From Present Prison Com-
mitment

It was eliminated because it was clearly Group speciﬁc. That is. members of Group
I (C. O.P.E.) all had a minimum academic educational level of at least fourteen (l 4) years
upon parole. simply by virtue of having earned an associate degree while in prison.
Conversely. members of Group II all had an academic educational level below fourteen
(14) years upon parole. as a result of the selection process used in this study. Because
the two (2) Groups diﬂ'ered widely on this variable due to the study design, it was
eliminated at this point as a possible predictor variable. However. the condition which
the variable reﬂected (academic educational level upon parole) was not lost to the
study. Rather. it was represented by the factor of Group. which by study design was
a “given” variable included in the major statistical model.

A fourth independent variable which proved to be signiﬁcantly related to recidivis-
tic behavior (through WALL) in a statistical sense. was also eliminated from the
study. That independent variable was:

#39 Academic Educational Attainment Level Increase During Present
Prison Commitment

It was evident this variablewas also part of a multicollincarity effect. involving some
of the independent variables associated with the study subjects. including: “academic
educational level upon parole” (#30). and “time served” (#38). As an information
supplyingvariable it served its purpose well. but fell far short of consideration as a ﬁnal
predictor variable to be included in the major model because of its powerful tie with
members of the C.O.P.E. Group (1) in particular.

 

Chapter III: Research Design

69
It was decided that one (1) of the independent variables which proved not to be
signiﬁcantly (.0639) related to recidivism in a statistical sense (at the .05 level). was
nevertheless to be included in the major model (a discriminant ﬁrnciion analysis with
afonrrard selection (stepwise) algorithm) as a possible predictor variable. That variable

was:
#37 Age Upon Parole From Present Prison Commihent

The somewhat arbitrary age of twenty-six (26) was chosen as a dichotomous cut-
oﬁ' point in the crosstabular analyses because the Michigan Department of Corrections
uses that age as a categorical division point in their risk classiﬁcation system.
Secondly. a cut-off point such as twenty-one (2 1) years of age yielded only four (4)
subjects in the under twenty-one (2 1) category. certainly too small a goup (an n) from

which to develOp meaningful ﬁndings.

The lack of statistical signiﬁcance relative to this independent variable was due to
the small size of the population (N=232). Because of a one-to-one relationship between
sample size and the magritude of chi square. the crosstabular analyses proved not to
be signiﬁcant in this instance. However. by simply increasing the study population to
two-hundred sixty (260) instead of two-hundred thirty-two (232). the chi square
analysis would display a statistically signiﬁcant relationship with recidivistic behav—
ior. assuming a constant distribution. More importantly. the variable was included
in the major model because prior research clearly and ﬁrmly establishes age upon
parole as a signiﬁcant factor relative to recidivistic behavior. It was felt that to ignore
the inﬂuence of this factor would be to introduce a major fault into the study design.

In a ﬁnal statistical effort to identify the independent variables suitable for
inclusion in the major statistical model. the remaining six (6) independent variables
(25- 16-44- 1:6). along with the generated variable “prior arrest(s)" (PRIORARR) . were
subjected to a series of three-way crosstabular analyses. These crosstabular analyses
took the form of examining the outcome variable (recidivism) (through the variable

 

Chapter 111: Research Design

RECIDAIL) by Group (I or II) with the independent variable (in dichotomous or
collapsed form) as a control. The seven (7) variables examined were:

#05 Academic Educational Level At Time Of Instant Offensets)
#06 Type Of Instant Oﬁensets)

#08 Prior Adult Felony Conviction(c)

#18 History Of Substance Abuse

#37 Age Upon Parole From Present Prison Commitment

#38 Length Of Time Served For Instant Oﬁense(s)

- Prior Arrest(s) (PRIORARR)

Two (2) of the seven (7) remaining independent variables were eliminated from the
study at this point. due to the results of the three-way crosstabular statistic applied
to them. They were:

#38 Length Of Time Served For Instant Oﬁense(s)
- Prior Arrest(s) (PRIORARR)

The variable “length of time served for instant oﬁensels)" (#38) was eliminated
because there were too few cases of subjects who served less than twenty-four (24)
months in prison. With twenty-four (24) months set as the dichotomous division point.
only twenty-two (22) subjects out of the entire population of two-hundred thirty-two
(N=232) served less than twenty-four (24) months in prison. Changing the twenty-four
(24) month division point would have been simply an arbitrary act. with no support
from the literature for such a decision. Further. the three-way analysis did not reveal
or specify a signiﬁcant or meaningful relationship between time served and recidivism.
Thus. the variable was eliminated from further analyses and consideration.

Also eliminated was the variable “prior arrest(s).” The three-way analysis
performed on the variable at this point revealed a multicollincar relationship between
this variable and that of “prior adult felony conviction(s) " (#08) . with the latter being
the stronger outcome predictor of the two (2). Thus. the variable “prior arrest(s)” was
eliminated from the study.

 

Chapter III: Research Desigr

61

The remaining ﬁve (5) variables (7 -2=5) listed below were built into the major model.
a discriminant function analysis model with a forward selection (stepwise) (Wilks'
Lambda) algorithm:

#06 Academic Educational Level At Time Of Instant Oﬁense(s)
#06 Type Of Instant Oﬁense(s)

#08 Prior Adult Felony Conviction(c)

#18 History Of Substance Abuse

#37 Age Upon Parole From Present Prison Commitment

It needs to be noted at this point that primary or secondary Group (I or II)
membership was “built” into the major model as an independent (predictor) variable
by virtue of the general study design. That is. the major thrust of the study was directed
at determining if completion of the C. O.P.E. Program signiﬁcantly reduced the overall
recidivism rates of that study Group (1). Therefore. Group (I or II) membership became
a “given” in the list of variables included in the major model. So. in examining the ﬁve
(5) independent variables selected for inclusion in the major model. one must also
understand that Group (I or II) membership was the controlling sixth (6) variable.

Both a discriminant function analysis and a hierarchical log-linear analysis were
considered as major models for this study. It was decided a discriminant function
analysis with aforward selection algorithm was the best choice when it became clear
that six (6) independent variables were to be used in the major model. Using six (6)
variables in a hierarchical log-linear model would result in a minimum of 2° (64) cells.
with two-hundred thirty-two (232) cases available for analysis. It was felt that such
a statistical model would not yield worthwhile results. A preliminary run of this
statistical technique yielded empty cells in three-fourths (3 / 4) of the cases. and only
four (4) cells with frequencies of ﬁve (5) or more (the usual standard for determining
sigriﬁcance).

Therefore. a discriminant ﬁrnction analysis with a forward selection (stepwise)
algorithmwas employed because it was desigred to use all cases for all estimates. and
was accordingly not subject to the problems of low cell n’s. However. the hierarchical

 

Chapter III: Research Design

log-linear analysis model with a backward elimination algorithm was used for
conﬁrmatory purposes. It was determined that this statistical technique would add
validity to the major model ﬁndings. once the major model (a discriminant function
analysis) pared the six (6) variables down to something less than that number. Thus.
it was concluded the end result of the major study ﬁndings would reﬂect a prepon-
derance of the evidence technique. and would represent the most reliable ﬁndings

possible under these study conditions.

The justiﬁcation for including a hierarchical log-linear analysis model as a
conﬁrmatory statistical technique in this study related to the fact that the
advantages of statistical models which summarize data and test hypotheses are well
recogrized and accepted in the ﬁeld of research. Regession analysis. for example.
examines the relationship between a dependent variable and a set of independent
variables. Analysis of variance techniques provide tests for the effects of various factors
on a dependent variable. But neither technique is appropriate as the major model for
categorical data. where the observations are not from populations that are normally
distributed with constant variance. as is the case with the two (2) Groups (I and II) in
this study. Rather. a special class of statistical techniques called log-linear models has
been formulated for the analysis of categorical data. These models are useful for
uncovering the potentially complex relationships among the variables in a multiway
crosstabulation. in a way which is more effective and reliable than other analyses like
those mentioned above.

However. even when attention is restricted to hierarchical models. many diﬁ'erent
types are possible for a given set of variables. The rules for selection require that the
model should “ﬁt" the data. be substantively interpretable. and as simple (parsimo-
nious) as possible. The strategy used in this study was to systematically test the
contribution to the model made by all the terms which proved to be signiﬁcantly related
to the dependent variable through other statistical analyses (like crosstabulation
analyses). and then to back them out in a hierarchical fashion. The hierarchical system
used in connection with this data thus employed a backward elimination algorithm.
Under this procedure. the eﬂ'ect whose removal resulted in the least-sigriﬁcant change

 

Chapter III: Research Design

in the likelihood-ratio chi-square was eligible for elimination. provided the observed
signiﬁcance level (.05) was larger than the criterion for remaining in the model. To
ensure a hierarchical model. only effects corresponding to the generating class were
examined at each step.

In this study. the generating class as determined by the major model (a discriminant
ﬁrnction analysis) was: ACADEMIC EDUCATIONAL LEVEL AT TILE OF lNST‘ANT
OFFENSEIS) " TYPE OF HVS’I‘ANT OFFENSE(S) ‘ HISTORY OF SUBSTANCE ABUSE ‘
AGE UPON PAROLE ‘ GROUP ‘ PRIOR ADULT FELONY CONVICTIONIS)

These variables were inserted irnto the major model (a discriminant function
analysis with a forward selection algorithm) in the order of their measured strength
relative to their association with the outcome variable (RECIDALL) (see above listing).
Four (4) of the six (6) variables proved to be signiﬁcantly related to the outcome
variable. They were:

#37 Age Upon Parole From Present Prison Commitment

# 18 History of Substance Abuse

#06 Type Of Instant Oﬁ‘ense(s)

#06 Academic Educational Level At Time Of Instant Oﬁense(s)

These four (4) variables along with the “given” variable Group were then built into
the conﬁrmatory model (a hierarchical log-linear analysis with a backward elimina-
tion algorithm) in a further eﬁ‘ort to measure the signiﬁcance of their relationship
relative to the outcome variable (RECIDALL).

The generating class for this statistical procedure was as follows: ACADEMIC
EDUCATIONAL LEVEL AT TIME OF INSTANT OFFENSElS) ‘ TYPE OF mSTANT
OFFENSES) " mSTORY OF SUBSTANCE ABUSE ‘ AGE UPON PAROLE ‘ GROUP

The ﬁrst step was to examine only the ﬁfth-order interaction. In other words. the
ﬁrst eﬁ'ect (variable) to be backed out of the model was “Group.” because it was the one
which displayed the least amount of statistical strength in the independent variable
cluster. as related to the outcome (recidivism). The backward elimination process

Chapter III: Research Design

ﬂ .

continued until the model contained those interactions where the overall analytical
results displayed and reﬂected the “best” model.

The ﬁrnal goup of effects (variables) which remained after this elimination process
was completed were determined to be the major predictive/ causative factors associ-
ated with the outcome (recidivism). In the order of their statistical strength (strongest
one ﬁrst) relative to their association with the outcome variable (recidivism). they were:

1) Age Upon Parole From Present Prison Commitment

2) History of Substance Abuse

3) Type Of Instant Oﬁense(s)

4) Academic Educational Level At Time Of Instant Oﬁense(s)

Their individual and cluster form effect on the outcome is discussed under the
section titled “Statistical Summary. " on page 124 of Chapter N.

In addition. the major and conﬁrmatory models also allowed for the examination
and measurement of the lone relationship between post-secondary academic educa-

tion and recidivism. which is discussed under the same section: “Statistical
Summary.” on page 124 of Chapter IV.

 

Chapter III: Research Design

 

PRESENTATION AND ANALYSIS OF DATA

Raw Data

The crosstabulation analyses run on thirty-four (34) of the forty-two (42) study
variables provided some interesting insights into the similarities and diﬁ'erences
regarding the two (2) Groups. With no effort at this juncture to determine signiﬁcant
differences. the data is presented irn raw form for those who need and/ or desire such
detail.

Race (#03) — Seventy-two (72) (62. 1%) of the C.O.P.E. Group (Group I) members
were black. and seventy-two (72) (62. 1 96) of the Comparison Group (Group II) members
were black. Forty-three (43) (37.1%) of the C.O.P.E. Group (Group I) members were
white. and one (1) (0.9%) was other (Indian). Forty-four (44) (3 7. 99b) of the Comparison

Group (Group II) members were white.

Note the small rounding error associated with the percentage ﬁgures of. in this
case. Group 1. Such errors go unreported beyond this point in the data presentation.

Academic Educational Level At Time Of Instant Offense“) (#05) —At the time
they committed the criminal offense for which they were irncarcerated. the C.O.P.E.
Group (Group I) members ranged in educational level (academic) from eight (8) to
sixteen (16) years of education. with fourteen ( 1 4) (1 2. 1 96) of them having earned a GED
Certiﬁcate. Two (2) (1.7%) C.O.P.E. Group (Group I) members had eight (8) years of
academic education. seven (7) (6%) had nine (9) years. eighteen (18) (16.4%) had ten
(10) years. forty-three (43) (37.1%) had twelve (12) years. four (4) (3.4%) had thirteen
(I 3) years. four (4) (3.4%) had fourteen (14) years. four (4) (3.4%) had ﬁfteen (l 5) years.
and one (1) (0.9%) had sixteen (16) years of academic education. The Comparison

 

66

Group (Group II) members ranged in educational level (academic) from ﬁve (5) to
thirteen (13) years of education. with twenty-one (2 1) (18.196) of them having earned
a GED Certiﬁcate. Two (2) (I . 796) Comparison Group (Group II) members had ﬁve (5)
years of academic education. one (1) (0.996) had seven (7) years. thirteen (13) (1 1 .296)
had nine (9) years. twenty-four (24) (20. 796) had ten (10) years. twenty- seven (27)
(23.396) had eleven (1 1) years. twenty-ﬁve (25) (21 .696) had twelve (12) years. and three
(3) (2.6%) had thirteen (13) years of academic education.

Type Of Instant Oﬁ‘ense(s) (#06) — Eleven (11) (9.596) of the C.O.P.E. Group
(Group I) were incarcerated for a property offense. three (3) (2.696) were drug offenders.
one-hundred one (101) (87.196) were violent offenders. and one (1) (0.996) was a
property and violent oﬁ‘ender. Among the members of the Comparison Group (Group
II): thirty-ﬁve (35) (30.2%)were property offenders. three (3) (2.696) were drug oﬁ'enders.
seventy-ﬁve (75) ( 64. 796) were violent offenders. and three (3) (2.696) were property and

violernt oﬁ'enders.

Prior Adult Felony Conviction(s) (#08) — Seventy-eight (78) (67.296) of the
C.O.P.E. Group (Group 1) members had no prior adult felony convictions. while ﬁfty-
eight (58) (5096) of the Comparison Group (Group II) were in this category. Seventeen
(17) (14.796) members of the C.O.P.E. Group (Group I) had one (1) prior felony
conviction. while the same was true for twenty-eight (28) (24. 1 96) of the Comparison
Group (Group II) members. Eight (8) (6. 996) C.O.P.E. Group (Group I) members had two
(2) prior felony convictions. while sixteen (16) (13.896) Comparison Group (Group 11)
members were in this category. Five (5) (4.396) members of the C.O.P.E. Group (Group
I) had three (3) prior adult felony convictions. and eight (8) (6.996) members of the
Comparison Group (Group II) were irn this category. Five (5) (4.396) of the C.O.P.E.
Group (Group I) members had four (4) or more prior adult felony convictions. while six
(6) (5. 196) of the Comparison Group (Group II) were so classiﬁed. Accurate and
complete data relative to this variable was not available for three (3) (2.696) of the
C.O.P.E. Group (Group I) members.

Criminal Proﬁle — Juvenile Property Offense(s) (#09) — Eighty-nine (89)
(76. 796) of the C.O.P.E. Group (Group I) members had no history of a juvernile property

 

Chapter IV: Presentation and Analysis of Data

67

oﬁ'ense. while seventy-one (71) (61 .296) of the Comparison Group (Group 11) members
were in this category. Of those subjects who had this type of history. they clustered
around one (1) to three (3) offenses. Nine (9) (7.8996) C.O.P.E. Group (Group I) members
and twenty-one (2 l) (1 8. 1 96) Comparison Group (Group II) members had one (1) offense
on their record. Seven (‘7) (696) C.O.P.E. Group (Group 1) members and eight (8) (6. 996)
Comparison Group (Group 11) members had two (2) offenses. four (4) (3.496) C.O.P.E.
Group (Group I) members and six (6) (5.2%) Comparison Group (Group 11) members
had three (3) offenses on their records. One (1) (0. 996) C.O.P.E. Group (Group I) member
and one (1) (0.996) Comparison Group (Group II) member had four (4) such offenses
on their records. ’

One (1) (0.996) C.O.P.E. Group (Group 1) member had ﬁve (5) such oﬂenses. one ( 1)
(0.996) had six (6). one (1) (0.9%) had seven (7). one (1) (0.996) had eight (8). and two (2)
(1 . 796) had incomplete information regarding this variable in their records. As for the
Comparison Group (Group II) . two (2) (1 . 796) had ﬁve (5) such offenses on their records.
two (2) (1 . 796) had six (6) . and ﬁve (5) (4.3%) had incomplete record references in regard
to this variable. In total. twenty-ﬁve (25) (21 . 796) C.O.P.E. Group (Group I) members
and forty (40) (34.596) members of the Comparison Group (Group II) had a history of
juvenile property oﬂenses.

Criminal Proﬁle — Juvenile Drug Oﬁ'ense(s) (it 10) — One-hundred twelve (l 12)
(96.696) members of the C.O.P.E. Group (Group I) had no record of a juvenile drug
oﬁ‘ense. while one-hundred nine (109) (9496) members of the Comparison Group
(Group II) were irn this category. Two (2) (1 . 796) members of the C.O.P.E. Group (Group
I) and three (3) (2.696) members of the Comparison Group (Group II) had a sirngle (l)
oﬁ'ense of this type. The remaining six (6) subjects. two (2) (1 . 796) in the C.O.P.E. Group
(Group I) and four (4) (3.496) in the Comparison Group (Group II). had irncomplete
record references in this area. ,

Criminal Proﬁle — Juvenile Violent Oﬁense(s) (# I I) — Ninety-four (94) (8196)
members of the C.O.P.E. Group (Group I) had no history of a juvernile violent oﬂ'ense.
while ninety-three (93) (80.296) members of the Comparison Group (Group II) were in

 

Chapter IV: Presentation and Analysis of Data

68

this category. Sixteen (16) (13.896) members of the C.O.P.E. Group (Group I) had one
(1) oﬂ'ense of this type. while ten (10) (8. 696) of the Comparison Group (Group II) had
one (1) offense. Four (4) (3.496) members of the C.O.P.E. Group (Group I) had two (2)
oﬁ’enses of this type. while ﬁve (5) (4.396) members of the Comparison Group (Group
II) had two (2) oﬁ'enses. two (2) (1 . 796) had three (3) offenses. and one (1) (0.996) had four
(4) offenses of this kind. The remaining seven (7) subjects. two (2) (1.796) in Group I
and ﬁve (5) (4.396) in Group II. had irncomplete record references in regard to this
variable.

Criminal Proﬁle — Adult Property Oﬁense(s) (#12) — Eighty-ﬁve (85) (73.396)
Group 1 members had no history of prior adult property oﬁ'enses. rnineteen (l 9) (1 6.496)
had one (1) such prior oﬁense. ﬁve (5) (4.396) had two (2) such offenses. three (3) (2.696)
had three (3). one (1) (0.996) had ﬁve (5). one (1) (0.996) had nine (9). and two (2) (1 .796)
had incomplete record references in this area. As for Group II. sixty-four (64) (55.2%)
had no such history. twenty-ﬁve (25) (21 .696) had one (1) such prior offense. eleven (1 1)
(9.5%) had two (2) . six (6) (5.296) had three (3). ﬁve (5) (4.396) had four (4). four (4) (3.4%)
had ﬁve (5). and one (1) (0.996) had irncomplete record references.

Criminal Proﬁle —Adult Drug Oﬁence(s) (#18) — One-hundred ten (1 10) (94.896)
members of the C.O.P.E. Group (Group 1) had no prior criminal history of this kind.
while one-hundred thirteen (1 13) (97.496) members of the Comparison Group (Group
II) fell into this category. Four (4) (3.496) members of the C.O.P.E. Group (Group I) had
one (1) prior oﬁ'ense of this kind. while two (2) (1 .7%) members of the Comparison
Group (Group II) had a single (1) prior oﬁ‘ense. Two (2) (1 . 796) members of Group I. and
one ( 1) (0.996) member of Group II. had incomplete record references in regard to this
study variable.

Criminal Proﬁle — Adult Violent Offensels) (#14) — Eighty-nine (89) (76. 796)
members of the C.O.P.E. Group (Group I) had no prior oﬂ'ense of this kind. while ninety-
two (92) (79.396) members of the Comparison Group (Group II) were irn this category.
Twelve (12) (1 0.396) members of the C.O.P.E. Group (Group I) had one (1) prior offense
of this kind. while severnteen (l7) (1 4. 796) members of the Comparison Group (Group

 

Chapter IV: Presentation and Analysis of Data

II) were in this category. Nine (9) (7.896) members of the C.O.P.E. Group (Group I) had
two (2) prior oﬁ‘enses. one (1) (0.996) had three (3). one (1) (0.996) had four (4). and one
(1) (0.9%) had ﬁve (5). In the Comparison Group (Group II). ﬁve (5) (4.396) had two (2)
prior oﬁ‘enses of this kind. and one (1) (0.996) had three (3). Three (3) (2.696) members
of the C.O.P.E. Group (Group I). and one (1) (0.996) member of the Comparison Group
(Group II) had incomplete record references in regard to this variable.

In The Community At Least Three Years Prior To Prison Commitment For
Instant Oﬁense(s) (#15) — Eighty-one (81) (69.896) of the C.O.P.E. Group (Group I).
and sixty-eight (68) (58.696) of the Comparison Group (Group II) were in the community
at least three (3) years prior to being incarcerated for their instant (current) oﬁ‘ense.
Thirty-ﬁve (35) (30.296) members of the C.O.P.E. Group (Group I). and forty-eight (48)
(41 .496) members of the Comparison Group (Group II) were in the commurnity less than
three (3) years prior to being incarcerated for their current offense.

Marital Status At Time Of Instant Oﬂense(s) (#16) — Ninety-two (92) ( 79.396)
members of the C.O.P.E. Group (Group I) were single. ﬁfteen (15) (12.996)were married.
two (2) (1 . 796) were separated. and seven (7) (696) were divorced. Of the Comparison
Group (Group II) members: ninety-six (96) (82.896) were single. ten (10) (8. 696) were
married. two (2) (1 .796) were separated. and eight (8) (6.996) were divorced.

Employment Status At Time Of Instant Oﬁ'ense(s) (#17) — Six (6) (5.296)
members of the C.O.P.E. Group (Group I) and two (2) ( 1.796) members of the
Comparison Group (Group II) had no work histories. Twenty-three (23) (1 9.896)
members of the C.O.P.E. Group (Group I) and twenty-ﬁve (25) (21 .696) members of the
Comparison Group (Group II) worked full-time when they committed the oﬁ’ense for
which they were sent to prison. Eight (8) (6. 996) members of the C.O.P.E. Group (Group
I) and eleven (1 1) (9.596) members of the Comparison Group (Group 11) worked part-
time. Nineteen (l9) (1 64%) members of the C.O.P.E. Group (Group I) and twenty-ﬁve
(25) (21 .696) members of the Comparison Group (Group II) worked intermittently.

 

Chapter IV: Presentation and Analysis of Data

70

Two (2) (1.796) members of the C.O.P.E. Group (Group I) and four (4) (3.496)
members of the Comparison Group (Group II) were laid-oﬁ. Fifty (50) (43. 1 96) members
of the C.O.P.E. Group (Group I) and forty-four (44) (3 7.996) members of the Comparison
Group (Group II) were unemployed at the time they committed the instant oﬁ'ense.
Eight (8) (6.996) members of the C.O.P.E. Group (Group I) and four (4) (3.496) members
of the Comparison Group (Group II) were students. Records for one (1) (0. 996) member
of the Comparison Group (Group II) were incomplete irn regard to this variable.

History Of Substance Abuse (# 18) - Forty-two (42) (36.296) members of the
C.O.P.E. Group (Group I) had a history of substance abuse. and seventy-three (73)
(62.996) had no such history. In the Comparison Group (Group II). forty-ﬁve (45)
(38.896) had a history of substance abuse. sixty-eight (68) (58.696) did not. The
remaining four (4) subjects. one (1) (0.996) in the C.O.P.E. Group (Group I) and three
(3) (2.696) irn the Comparison Group (Group II) . had incomplete records in regard to this
variable.

Evidence of A Serious Physical Illness Or Disability At Time Of Instant
Oﬁense (#19) — One (1) (0.996) member of the C.O.P.E. Group (Group I) and seven (7)
(696) members of the Comparison Group (Group II) evidenced this kind of history.
Whereas. one-hundred ﬁfteen (l 15) (99.196) members of the C.O.P.E. Group (Group
I). and one-hundred nine (109). (94%) members of the Comparison Group (Group II).
did not have a history of this kind.

Evidence Of A Serious Emotional Or Psychological Problem At Time Of
Instant Oﬁ'ense(s) (#20) — Eleven (1 1) (9.596) members the C.O.P.E. Group (Group
I) and seventeen (1 7) (14.796) members of the Comparison Group (Group II) evidenced
this kind of history. Whereas. one-hundred ﬁve (105) (905%) members of the C.O.P.E.
Group (Group I) and ninety-nine (99) (85.396) members of the Comparison Group
(Group II) did not have this kind of history.

Upbringing (#2 I) —Eighty-six (86) (74. 1 96) members of the C.O.P.E. Group (Group
I) were raised by their natural parent(s). while ninety (90) (7 7.696) of the Comparison
Group (Group II) members were in this category. Two (2) (1 . 796) members of the

 

Chapter IV: Presentation and Analysis of Data

71

C.O.P.E. Group (Group I) were raised by stepparent(s) . twenty (20) (1 7.296)by a natural
parent and a stepparent. ﬁve (5) (4.396) by relative(s). one (1) (0. 996) by guardian(s) . and
two (2) (1 . 796) were raised by foster parent(s). As for the Comparison Group (Group 11):
ten (10) (8.696) were raised by stepparent(s). twelve (12) (10.3%) by a natural parent.
and a stepparent. one (1) (0.996) by relative(s). one (1) (0.996) by foster parent(s). and
two (2) (1 . 796) had incomplete records in regard to this variable.

Financial Status Of Upbringing Environment (#22) — One (1) (0.996) member of
the C.O.P.E. Group (Group I) and two (2) (1 . 796) members of the Comparison Group
(Group 11) came from wealthy homes where the family income was at least $50.000 per
year. Eighty-ﬁve (85) (73.396) members of the C.O.P.E. Group (Group I) and eighty-eight
(88) (75.996) members of the Comparison Group (Group II) came from homes where the
family income was between $ 15.000 and $50.000 per year. TWenty-six (26) (22.4%)
members of the C.O.P.E. Group (Group I) and twenty-four (24) (20. 796) members of the
Comparison Group (Group 11) came from homes where the yearly income was below
$15,000. Four (4) (3.496) members of the C.O.P.E. Group (Group I). and two (2) (1 .796)

members of the Comparison Group (Group II) had incomplete records in regard to this
variable.

Academic Educational Level OfMotherAt Time Of Subject's Instant Oﬁ'ense(s)
(#23) — In the case of one-hundred (100) (86.296) C.O.P.E. Group (Group I) members
and one-hundred ﬁve (105) (90.596) Comparison Group (Group 11) members. data in
regard to this variable was not available. The remaining sixteen (16) (13.896) C.O.P.E.
Group (Group I) subjects had mothers whose academic educational level ranged from
eight (8) years to sixteen (16) years. As for the other eleven (1 1) (9.596) members of the
Comparison Group (Group II) . their mothers had an academic educational level range
from six (6) years to sixteen (16) years.

Academic Educational Level Of Father At Time Of Subject's Instant Oﬁense(s)
(#24) — Data in regard to one-hundred four (104) (89. 796) C.O.P.E. Group (Group I)
members and one-hundred eight (108) (93. 196) members of the Comparison Group
(Group II) was not available for this variable. The remairning twelve (12) (1 0.396)

 

Chapter IV: Presentation and Analysis of Data

72

C.O.P.E. Group (Group I) members had fathers whose academic educational level
ranged from eight (8) years to ﬁfteen (15) years. The remaining eight (8) (6.9%)
Comparison Group (Group 11) members had fathers whose acadennic educational level

ranged from seven (7) years to thirteen (13) years.

Family Emotional Support System At Time Of Subject's Instant Oﬁ’ense(s)
(#25) — Forty-eight (48) (41 .496) of the C.O.P.E. Group (Group I) and ﬁfty-seven (57)
(49. 1 96) members of the Comparison Group (Group II) received strong emotional
support from members of their immediate family. Forty-seven (47) (40.596) of the
C.O.P.E. Group (Group II) members and forty-eight (48) (41.496) members of the
Comparison Group (Group II) received some support from one (1) or more members
of his immediate family (mother. father. brother. sister. wife. etc). Fifteen (15) (12.996)
members of the C.O.P.E. Group (Group I) and eleven (1 1) (9.596) members of the
Comparison Group (Group II) received no support from any member of their immediate
family. The remaining six (6) (5.196) members of the C.O.P.E. Group (Group I) had
incomplete record references in regard to this variable.

Number Of Non-Bondable Major Misconduct Reports In Prison For Which The
Subject Was Found Guilty During His Present Prison Commitment (#26) — Sixty-
seven (67) (5 7.8) members of the C.O.P.E. Group (Group I) and forty-eight (48) (41 .496)
members of the Comparison Group (Group 11) did not receive a non-bondable major
misconduct report during their incarceration period. TWenty-two (22) (1 996) members
of the C.O.P.E. Group (Group I) and thirty-one (3 l) (26. 796) members of the Comparison
Group (Group II) received one (1) non-bondable major misconduct report during their

incarceration.

In the C.O.P.E. Group (Group I). twelve (12) (1 0.396) received two (2) non-bondable
major misconduct reports while incarcerated. ﬁve (5) (4.396) received three (3) such
reports. ﬁve (5) (4.396) received four (4). two (2) (1 .796) received ﬁve (5). two (2) (1 .796)
received seven (7). and one (1) (0. 996) received nine (9) non-bondable reports.

In the Comparison Group (Group II). eight (8) (6.996) received two (2) such reports.
eleven (1 1) (9.596) received three (3). nine (9) (7.896) received four (4). one (1) (0.996)

 

Chapter IV: Presentation and Analysis of Data

73

received ﬁve (5). two (2) (1 .796) received six (6). one (1) (0.996) received seven (7). one
(1) (0.996) received nine (9). one (1) (0.996) received ten (10). one (1) (0.996) received
thirteen (13). one (1) (0.9%) received sixteen (16). and one (1) (0.996) received seventeen
(17) non-bondable reports.

Academic Educational Level Upon Parole From Present Prison Commitment
(#30) — In the C.O.P.E. Group (Group I). ninety-six (96) (82.896) had fourteen (l 4) years
of acadennic education upon parole. three (3) (2.6%) had ﬁfteen (1 5) years. fourteen (14)
(12.196) had sixteen (16) years. one (1) (0.9%) had seventeen (17) years. and two (2)
(1 . 796) had eighteen (18) years of academic education. As for the Comparison Group
(Group II). forty-ﬁve (45) (38.8%) had GED Certiﬁcates. one (1) (0.996) had ﬁve (5) years.
one (1) (0.9%) had seven (7) years. eight (8) (6.9%) had nine (9) years. ten (10) (8.6%)
had ten (10) years. sixteen (16) (13.8%) had eleven (1 1) years. twenty-three (23) (19.896)
had twelve (12) years. and twelve (12) (10.3%) had thirteen (13) years of academic
education upon parole. A

MDOC Assaultive Risk Classiﬁcation At The Time Of Parole (#31) — In the
C.O.P.E. Group (Group I) . one (1) (0. 996) subject was classiﬁed very low. three (3) (2.696)
were low. forty-seven (47) (40.596) were middle. forty-nine (49) (42.296) were high. ﬁfteen
(15) (1 2.996) very high. and one (1) (0. 996) unknown. In the Comparison Group (Group
11). three (3) (2.696) were very low. thirteen (13) (1 1 .296) low. forty (40) (34.596) middle.
forty-seven (47) (40.5%) high. and thirteen (13) (I 1 .296) very high.

Evidence Of A Serious Physical Illness Or Disability At Time Of Parole From
Present Prison Commitment (#32) — One (1) (0.996) member of the C.O.P.E. Group
(Group I) evidenced this kind of problem. one-hundred ﬁfteen (1 15) (99. 1 96) did not.

No (0) (096) members of the Comparison Group (Group II) evidenced this ﬁnd of
problem.

Evidence Of A Serious Emotional Or Psychological Problem At Time Of Parole
From Present Prison Commitment (#33) — None (0) (096) of the C.O.P.E. Group

(Group 1) members evidenced this ﬁnd of problem. Only one (1) (0. 996) member of the

Comparison Group (Group II) evidenced this ﬁnd of problem: one-hundred ﬁfteen
(l 15) (99.196) did not.

Chapter IV: Presentation and Analysis of Data

74

Age At First Arrest (#34) — The age of ﬁrst arrest in regard to the C.O.P.E. Group
(Group I) ranged from ﬁve (5) to forty (40). with the geatest concentration being
between twelve (12) and twenty-two (22) years of age. The range irn regard to the
Comparison Group (Group II) was from seven (7) to thirty-six (36). with the geatest
concentration being between twelve (12) and twenty-two (22) years of age. In the case
of the C.O.P.E. Group (Group I). eleven (1 1) (9.596) were ﬁrst arrested at age fourteen
(14). ten (10) (8.696) at age ﬁfteen (15). ten (10) (8.696) at age sixteen (16). twenty-one
(21) (18.1%) at age seventeen (17). seventeen (17) (14.7%) at age eighteen (18). and
thirteen (13) (1 1 .296) at age nineteen (19). As for the Comparison Group (Group II).
twelve (12) (10.396) were ﬁrst arrested at age fourteen (l 4). twelve (12) (10.396) at age
ﬁfteen (15). eight (8) (6.9%) at age sixteen (16). ﬁfteen (15) (12.996)at age seventeen(17).
eleven (11) (9.596) at age eighteen (18). and ten (10) (8.696) at age nineteen. Further
speciﬁcs in regard to this variable are available from this researcher.

Age At Time Of Present Prison Commitment (#35) — Three (3) (2.696) members
of the C.O.P.E. Group (Group I) were sixteen ( l 6) years of age when they entered prison.
thirteen (l3) (1 1 .296) were seventeen years of age. twenty (20) (1 7.296) were eighteen
years of age. eighteen (18) (15.596) were nineteen (19) years of age. sixteern (16) (13.8%)
were twenty (20) years of age. fourteen (14) (1 2.1 96) were twenty-one (2 1) years of age.
thirteen (13) (1 1 .296) were twenty-two (22) years of age. three (3) (2.696) were twenty-
three (23) years of age. four (4) (3.496) were twenty-four (24) years of age. one (1) (0.996)
was twenty-ﬁve (25) years of age. two (2) (I . 796) were twenty-six (26) years of age. two
(2) (1 . 796) were twenty-seven (27) years of age. one (1) (0.996)was twenty-eight (28) years
of age. one (1) (0.996) was twenty-nine (29) years of age. one (1) (0.996) was thirty (30)
years of age. one (1) (0.996) was thirty-two (32) years of age. one (I) (0.996) was thirty-
three (33) years of age. and two (2) (1 . 796) were thirty-nine (39) years of age at the time
they entered prison.

As for the Comparison Group (Group II). eight (8) (6. 996) were seventeen (17) years
of age. thirteen (13) (1 1 .296) were eighteen (18) years of age. thirteen (l3) (1 1 .296) were
rnineteen (19) years of age. ﬁfteen (15) (1 2.996) were twenty (20) years “of age. eleven (1 1)
(9.596) were twenty-one (2 1) years of age. ten (10) (8. 696) were twenty-two (22) years of

 

Chapter IV: Presentation and Analysis of Data

75

age. eleven (1 1) (9.596) were twenty-three (23) years of age. seven (7) (696) were twenty-
four (24) years of age. seven (7) (696) were twenty-ﬁve (25) years of age. six (6) (5.296)
were twenty-six (26) years of age. four (4) (3.4%) were twenty- seven (27) years of age.
ﬁve (5) (4.396) were twenty-eight (28) years of age. one (1) (0. 996) was twenty-nine (29)
years of age. three [3) (2.696) were thirty (30) years of age. and two (2) (1 . 796) were thirty-
eight (38) years of age at the time they entered prison.

Age Upon Parole From Present Prison Commitment (#37) — One (1) (0. 996)
member of the C.O.P.E. Group (Group I) was nineteen (19) years of age when he was
paroled from prison. one (1) (0.996) was twenty (20) years of age. four (4) (3.496) were
twenty-one (2 1) years of age. three (3) (2.696) were twenty-two (22) years of age. twelve
(12) (1 0.3%) were twenty-three (23) years of age. twelve (12) (1 0.396) were twenty-four
(24) years of age. ﬁfteen (15) (12.9%) were twenty-ﬁve (25) years of age. sixteen (16)
(13.896) were twenty-six (26) years of age. ﬁfteen (15) (12.9%) were twenty-seven (27)
years of age. ten (10) (8.696) were twenty-eight (28) years of age. ten (10) (8.696) were
twenty-nine (29) years of age. ﬁve (5) (4.3%) were thirty (30) years of age. two (2) (I .796)
were thirty-one (3 1) years of age. three (3) (2.696) were thirty-two (32) years of age. one
(1) (0.996) was thirty-three (33) years of age. one (1) (0.996) was thirty-four (34) years
of age. two (2) (I . 796) were thirty-ﬁve (35) years of age. one (1) (0.996)was thirty-six (36)
years of age. one (1) (0.996) was forty-four (44) years of age. and one (1) (0. 996) was forty-
six (46) years of age upon parole from prison.

As for the Comparison Group (Group II). one (1) (0.996) was rnineteen (19) years of
age. one (1) (0. 996) was twenty (20) years of age. ﬁve (5) (4.396) were twenty-one (2 1)
years of age. one (1) (0.996) was twenty-two (22) years of age. eighteen (18) (15.596) were
twenty-three (23) years of age. fourteen (14) (1 2. 1 96) were twenty-four (24) years of age.
sixteen (16) (13.8%) were twenty-ﬁve (25) years of age. nine (9) ( 7.896) were twenty-six
(26) years of age. eighteen (18) (1 5.596) were twenty-seven (27) years of age. nine (9)
( 7.896) were twenty-eight (28) years of age. eight (8) (6. 996) were twenty-nine (29) years
of age. two (2) (l .7%) were thirty (30) years of age. ﬁve (5) (4.396) were thirty-one (3 1)
years of age. three (3) (2.696) were thirty-two (32) years of age. one (1) (0.996)was thirty-
three (33) years of age. one (1) (0.996) was thirty-four (34) years of age. two (2) (1.7%)

 

Chapter IV: Presentation and Analysis of Data

76

was thirty-ﬁve (35) years of age. and two (2) (1 . 796) were forty (40) years of age upon
parole from prison.

Length Of Time Served For Instant Oﬁense (# 38) —The C.O.P.E. Group (Group
I) ranged in time served from twenty (20) months to one-hundred seventy-two (I 72)
months. The Comparison Group (Group II) ranged from two (2) months to one-hundred
thirty-seven (137) months. There were no unusual concentrations in either goup.
Speciﬁc details in regard to this variable are available from this researcher upon

request.

Academic Educational Attainment Level Increase During Present Prison
Commitment (#39) — Five (5) (4.396) members of the C.O.P.E. Group (Group I) did
not show a full (1) year of increase. four (4) (3.4%) showed one (1) year of increase. forty-
one (41) (35.396) showed a two (2) year increase. twenty (20) (1 7.296) showed a three
(3) year increase. thirty (30) (25.996) showed a four (4) year increase. nine (9) (7.8%)
showed a ﬁve (5) year irncrease. four (4) (3.496) showed a six (6) year increase. one (1)
(0.9%) showed a seven (7) year increase. and two (2) (1 .796) showed an eight (8) year
increase. As for the Comparison Group (Group II) . ninety-ﬁve (95) (81 .996) showed less
than a one (1) year increase. sixteen (16) (13.896) showed a one (1) year increase. one
(1) (0. 996) showed a two (2) year increase. two (2) (1 . 796) showed a three (3) year increase.
and two (2) (1 .796) showed a four (4) year increase.

Criminal Recidivism — Property Oﬁense(s) (#40) — Ninety-six (96) (82.896)
members of the C.O.P.E. Group (Group I) were not arrested for a felony class property
crime (did not recidivate). during the two (2) year period following their parole.
Fourteen (14) (12.196) members of the C.O.P.E. Group were arrested for this type of
offense one (1) time during the two (2) year follow-up period. ﬁve (5) (4.396) were arrested
twice (2). and one (1) (0. 9%) was arrested three (3) times. As for the Comparison Group
(Group II). eighty-one (8 1) (69,896) were not arrested for this type of oﬁ‘ense durirng the
two (2) year follow-up period. twenty-ﬁve (25) (21 .696) were arrested one (1) time. six
(6) (5.296) were arrested twice (2). and four (4) (3.496) were arrested three (3) times.

 

Chapter IV: Presentation and Analysis of Data

77

Criminal Recidivism -— Drug Oﬁ‘ense(s) (#41) -— One-hundred ten (1 10) (94.896)
members of the C.O.P.E. Group (Group I) were not arrested (did not recidivate) for this
type of offense during the two (2) year period following their parole. Five (5) (4.396)
members of the C.O.P.E. Group (Group I) were arrested one (1) time. and one (1) (0.996)
was arrested twice (2). One-hundred ten (110) (94.896) members of the Comparison
Group (Group II) were not arrested for this type of offense during the two (2) year period
following their parole. Five (5) (4.396) members of the Comparison Group (Group II)

were arrested once (1). and one (1) (0.996) was arrested three (3) times.

Criminal Recidivism - Violent Oﬁense(s) (#42) — Ninety-six (96) (82.896)
members of the C.O.P.E. Group (Group I) were not arrested (did not recidivate) for this
type of felony crime during the two (2) year period of time following their parole. Eleven
(1 1) (9.596) members of the C.O.P.E. Group (Group I) were arrested one (1) time. six
(6) (5.296) were arrested twice (2) . and three (3) (2.696) were arrested three (3) times. One-
hundred ﬁve (105) (90.596) members of the Comparison Group (Group II) were not
arrested for this type of crime during the two (2) year period of time following their
parole. Six (6) (5.296) members of the Comparison Group (Group II) were arrested once
(1). three (3) (2.696) were arrested twice (2). one (1) (0.996) was arrested three (3) times.
and one (1) (0.996) was arrested four (4) times.

These initial standard crosstabulation analyses (in raw form) served the study in
three (3) ways: First. they revealed the fact that data relative to six (6) of the study
variables was incomplete and / or insufficient to allow for a meaningful contribution
to the intent and purpose the project. These six (6) variables were thus eliminated from
any further consideration and/ or analyses. Those six (6) study variables were:

# 19 Evidence Of A Serious Physical Illness Or Disability At Time Of Instant
Oifense(s)

#20 Evidence Of A Serious Emotional Or Psychological Problem At Time Of
Instant Offense(s)

#23 Academic Educational Level Of Mother At Time Of Subject's Instant
Oﬁense(s)

 

Chapter IV: Presentation and Analysis of Data

78

#24 Academic Educational Level Of Father At Time Of Subject‘s Instant
Oﬁense(s)

#32 Evidence Of A Serious Physical Illness Or Disability At Time Of Parole
From Present Prison Commitment

#33 Evidence Of A Serious Emotional Or Psychological Problem At Time Of
Parole From Present Prison Commitment

Secondly. the initial crosstabulation analyses revealed the need to collapse some
of the categories (cells) in order to add meaningfulness to the data at hand. Thirdly.
the crosstabulation analyses brought notice to the fact that certain other statistics
such as dichotomous crosstabulation analyses. collapsed bivariate crosstabulation
analyses. collapsed crosstabulation analyses with a control. three-way crosstabula-
tion (controllirng for oﬁ'ense type and/ or Group) analyses. and analyses of variance
(AN OVA) needed to be applied to some of the study variables in order for the data
associated with them to make a meaningful and worthwhile contribution to the study.
Thus. after running such statistics on some variables. the complete results were used
to make comparisons between the two (2) Groups. and speciﬁcally to determine the
signiﬁcance (.05 level) of their diﬂ‘erences. Those comparisons follow:

Group Comparisons
Matching Variables

Race (#03) — The race factor was one of the six (6) matching variables. Thus. as
one might expect. the two (2) Groups were not signiﬁcantly (at the .05 level) diﬁ'erent
under a collapsed bivariate crosstabulation analysis when it came to this biological
trait. Seventy-two (72) (62. 196) members from each Group were black. and forty-four
(44) (37.9%) were white (one (1) Indian in Group 1).

Marital Status At Time Of Instant Oﬂensels) (#16) — Ninety-two (92) (79.396)
members of Group I (C.O.P.E.) and ninety-six (96) (82.896) members of Group 11
(Comparison) had never been married at the we they committed the criminal offense

Chapter IV: Presentation and Analysis of Data

 

79

Table 4.1
Group Comparisons On The Matchlng Variables

 

 

 

 

 

 

Study Significant
lemmas Elﬂatences
Race (#03) _ No
Marital Status (#16) - - No
Employment Status (#17) __ _ No
MDOC Assaultive Risk Classlflcatlon (#31) No
Age Upon Parole From Prison (#37) - No

 

for which they were imprisoned. Twenty-four (24) (20. 796) of the Group 1 members and
twenty (20) (1 7.296) of the Group II members were either married at the time they
committed the instant oﬁ'ense. or had at some poirnt earlier been married. Again. this
was one ( l) of the six (6) matching variables. so the non-signiﬁcant (at the .05 level)
diﬁ'erences under a collapsed bivariate crosstabulation arnalysis between the two (2)
Groups was a plarnned and expected outcome.

unployment Status At Time Of Instant Oﬁense(s) (#17) — In regard to the
employment status of the subjects at the time they committed the criminal oﬁense for
which they were imprisoned. ﬁfty-eight (58) (5096) members of Group I and ﬁfty (50)
(43.5) members of Group II were not employed. Twenty-three (23) (19.8%) members
of Group I were employed full-time. twenty-seven (27) (23.396) were employed less than
full-time. and eight (8) (6.996) were students. As for members of Group II. twenty-ﬁve
(25) (21 . 796) were employed full-time. thirty-six (36) (31 .396) were employed less than
full-time. and four (4) (3.596) were students. Being one (1) of the six (6) matching
variables. the non-signiﬁcant diﬁ'erences under a collapsed bivariate crosstabulation
analysis between the two (2) Groups was a planned and expected outcome.

MDOC Assaultive Risk Classiﬁcation At The Time Of Parole (#31) — One (1)
(0.996) member of Group I had an MDOC assigned risk classiﬁcation at the time of

 

Chapter IV: Presentation and Analysis of Data

80

parole of very low. three (3) (2.696) were low. forty-seven (47) (40.996) were middle. forty-
rnine (49) (42.6%) were high. and ﬁfteen (15) (1396) were very high. Three (3) (2.696)
members of Group II were classiﬁed very low. thirteen (l3) (1 1 .296) were low. forty (40)
(34.5%) were middle. forty-seven (47) (40.596) were high. and thirteen (1 3) (1 1 .296) were
very high. While the differences between the two (2) Groups under a collapsed bivariate
crosstabulation analysis proved non-signiﬁcant at the .05 level by virtue of this being
one (1) of the matching variables. they were of substantive irnterest.

Age Upon Parole From Present Prison Commitment (#37) - The C.O.P.E.
Group (Group I) ranged in age upon parole from nineteen (19) through forty-six (46).
Those in the Comparison Group (Group II) ranged in age from rnineteen (19) through
forty (40) years of age upon parole from prison. An analysis of variance (AN OVA)
performed on the data revealed the Group I mean age upon parole to be twenty-six (26)
(26.6379) years of age. and the Group 11 mean age upon parole to be twenty-six (26)
(26.2759) as well. The mean for the entire population (both Groups) was calculated to
be twenty-six (26) (26.4569) years of age upon parole from prison. The differences
between the Groups proved to be non- signiﬁcant at the . 05 level. This being a matching
variable. the results were an expected outcome.

Prisoner Demographics/Background

History Of Substance Abuse (#18) — Forty-two (42) (36.596) members of Group
I had a history of substance abuse. while seventy-three (73) (63.596) members of this
Group had no such history. Forty-ﬁve (45) (39.896) members of Group II had a history
of substance abuse. sixty-eigrt (68) (60.296) had no such history. Thus. the diﬁ'erences
between the two (2) Groups under a collapsed bivariate crosstabulation analysis were
non-signiﬁcant at the .05 level.

Upbringing (#21) — In regard to the matter of upbringing. eighty-six (86) (74.196)
members of Group I and ninety (90) (78.996) members of Group II were brought up by
one (I) or both (2) natural parents. Twenty (20) (I 7.296) members of Group I and twelve
(12) (1 0.596) members of Group II were not brought up by their natural parent(s). Ten
(10) (8.696) members of Group I and twelve (12) (1 05%) members of Group II were

 

Chapter IV: Presentation and Analysis of Data

81

Table 4.2
Group Comparisons On
The Prlsoner Demographics/Background Varlables

 

Study Significant
Marlablas W
History Of Substance Abuse (#la) ................................................. No
Upbringlng (#21) ............................................................................. No
Flnanclal Status Oi Upbringlng Environment (#22) ....................... No
Famlly Emotional Support System (#25) ........................................ No

 

brought up by one (1) natural parent and one (1) step-parent. The differences between
the two (2) Groups under a crosstabulation analysis proved non-signiﬁcant at the .05
level.

Financial Status Of Upbringing Environment (#22) — Through the application
of a collapsed bivariate crosstabulation analysis. the data regarding the ﬁnancial
status of the upbringing environment was divided into two (2) categories: non-poor and
poor. Eighty-six (86) (76.896) of the Group I members and rninety (90) ( 78.996) of the
Group 11 members came from non-poor homes. where the family income was $ 15.000
a year or above. Twenty-six (26) (23.296) of the Group I members and twenty-four (24)
(21 . 1 96) members of Group 11 came from poor homes. where the family income was
below $15,000 per year. The differences between the two (2) Groups under this
arnalysis developed as non-signiﬁcant at the .05 level.

Family Emotional Support System At Time Of Subject's Instant Oﬁense(s)
(#25) — Again. through the use of a collapsed bivariate crosstabulation analysis. the
data regarding the family emotional support system at the time the subject committed
the criminal offense for which he was incarcerated was divided irnto three (3) categories:
strong support. some support. and no support. Forty-eight (48) (43.696) members of
Group I received strong support. forty-seven (47) (42. 796) received some support. and

 

Chapter IV: Presentation and Analysis of Data

ﬁfteen (15) (1 3.696) received no emotional support from their immediate families. Fifty-
seven (57) (49.196) of the Group 11 members received strong support. forty-eight (48)
(41 .496) received some support. and eleven (1 1) (9.596) received no emotional support
from their families. The differences between the two (2) Groups under this analysis
proved non-signiﬁcant at the .05 level.

Crlmlnal Hlstory

Prior Adult Felony Conviction(s) (#08) — In regard to prior adult felony
convictions. seventy-eight (78) (67.296) members of Group I and ﬁfty-eight (58) (5096)
members of Group II had no history of prior adult felony convictions. Seventeen (17)
(14. 796) members of Group I had one (1) prior adult felony. eight (8) (6.996) had two (2)
prior adult felony convictions. ﬁve (5) (4.396) had three (3) prior felony convictions. two
(2) (1 . 796) had four (4) prior adult felony convictions. two (2) (1 . 796) had ﬁve (5) prior
adult felony convictions. one (1) (0. 996) had thirteen (l 3) prior adult felony convictions.
and data on this variable was unavailable for three (3) members of Group I.

As for Group II. twenty-eight (28) (24. 1 96) members had one (1) prior adult felony
convictions. sixteen (1 6) (13.8%) had two (2) prior adult felony convictions. eight (8)
(6.996) had three (3) prior adult felony convictions. four (4) (3.496) had four (4) prior adult
felony convictions. and two (2) (1 . 796) had. ﬁve (5) prior adult felony convictions.
Although statistically the differences between the two (2) Groups proved non-
signiﬁcant under a crosstabulation analysis. they were of substantive interest.

Criminal Proﬁle — Juvenile Property Oﬁense(s) (#09) — The differences
between the two (2) Groups on the matter of property offenses as a juvenile proved
signiﬁcant (.0288) at the .05 level. under a dichotomous crosstabulation analysis.
Eighty-nine (89) (78. 1 96) of the Group 1 members had no prior history of juvenile
property offenses. while only seventy-one (7 1) (6496) of the Group 11 members were in
this category. Nenty-ﬁve (25) (21 .996) of the Group 1 members had one or more
property oﬁ'enses as a juvenile. while forty (40) (3696) members of Group II were irn this
category. An arnalysis of variance (AN OVA) performed on the data for this variable
indicated the mean number of juvenile property offenses for members of Group I to be

 

Chapter IV: Presentation and Analysis of Data

Table 4.3

 

Group Comparisons On
The Criminal History Variables
Study Significant
Mounties cram
Prlor Adult Felony Convlctton(s) (#08) .......................................... No
Juvenlle Properly Ollensets) (#09) ............................................... Yes
Juvenlle Drug Oilense<s> (#10) ...................................................... No
Juvenlle Vlolent Ottenseis) (#l l) .................................................. No
Adult Properly Oltensets) (# t 2) .................................................... Yes
Adult Drug Ottenseis) (#13) ............................................................ No
Adult Vlolent Oltense(s) (#14) ........................................................ No
In The Communlty Three Years (#15) ............................................ No
Age At Flrst Arrest (#34) .................................................................. No

 

.57. and for members of Group II to be .73 (rounded). The mean number of oﬁ'enses
for both Groups was calculated at .65 (rounded). Thus. we see that under an arnalysis
of variance (AN OVA) the between- Groups differences proved non-signiﬁcant at the .05
level. However. it was clear that members of Group II in a general collective sense had
a more extensive history of juvenile property oﬂ’enses.

Criminal Proﬁle — J uvenile Drug Oﬁense(s) (#10) — One-hundred twelve (1 12)
(98.296) members of Group I had no history of a juvenile drug oﬁ'ense. while one-
hundred nine (109) (97.396) members of Group II were irn this category. Two (2) (1 .896)
members of Group I and three (3) (2.796) members of Group 11 did have such a history.
The diﬁ’erences between the two (2) Groups under a dichotomous crosstabulation
analysis were non-signiﬁcant at the .05 level. Under an arnalysis of variance (AN OVA) .
the means for the two (2) Groups proved to be minuscule. and the between-Groups

differences were highly non-signiﬁcant. Neither of these two (2) Groups had much of
a history of juvenile drug offenses.

 

Chapter IV: Presentation and Analysis of Data

84

Criminal Proﬁle — Juvenile Violent Oﬁense(s) (# 1 1) — Ninety-four (94) (82.596)
members of Group I had no history of a violent juvenile offense. while twenty (20)
(1 7.596) members did have such a history. As for Group II. ninety-three (93) (83.8%)
had no such history. while eighteen (18) (16.296) did have a history of one or more
violent juvenile offenses. The diﬂ‘erences between the two (2) Groups under a
dichotomous crosstabulation analysis proved non— significant at the .05 level. Under
an analysis of variance (AN OVA). the means of the two (2) Groups were minuscule. and

the between-Groups differences were clearly non-significant.

Criminal Proﬁle — Adult Property Oﬁ'ense(s) (#12) — Under a dichotomous
crosstabulation analysis. the two (2) Groups were signiﬁcantly (.0042) different at the
.05 level on the matter of prior adult property offenses. Eighty-ﬁve (85) (74.696)
members of Group I did not have such a history. while twenty-nine (29) (25.496) had
a history of one or more such prior offenses. Only sixty-four (64) (55.796) members of
Group 2 had no such history. while ﬁfty-one (5 1) (44.396) had one or more prior adult
property offenses on their records. An analysis of variance (AN OVA) performed on this
data revealed the Group I mean to be .46 (rounded). and the Group 11 mean to be .91 .
The mean for the entire population was .69 (rounded). The differences between two (2)
Groups under this analysis proved significant (.0062) at the .05 level. with a clear
indication that Group 11 members had a more extensive history of prior adult property

offenses.

Criminal Proﬁle — Adult Drug Oﬁense(s) (#13) — Neither Group had much of
an adult drug offense history. One-hundred ten (1 10) (96.596) of the Group I members.
and one-hundred thirteen (1 13) (98.396) members of Group II. had no history of prior
adult drug oﬁ'enses. Four (4) (3.596) members- of Group I. and two (2) (1 .796) members
of' Group II had one or more such prior oﬁ'enses. The diﬁ’erences between the two (2)
Groups under a dichotomous crosstabulation analysis proved non-significant at the
.05 level. Under an analysis of variance (AN OVA). the diﬁ'erences in the two (2)

independent Group means proved to be minuscule. and the between-Groups diﬁer-
ences were thus non-significant.

 

Chapter IV: Presentation and Analysis of Data

 

. Table 4.4
Group Comparlsons On
The Current Offense/Sentence Varlable
Study Signltlcant
Mariam: Diagrams:
Type or Instant Oftense(s) (#06) ................................................... Yes

 

Criminal Proﬁle — Adult Violent Oﬁense(s) (#14) — Eighty-nine (89) (78.896)
members of Group I. and ninety-two (92) (8096) members of Group 11 did not have a
history of prior adult violent offenses. Twenty-four (24) (21 .296) members of Group I.
and twenty-three (23) (2096) members of Group II had one or more such prior offenses.
The differences between the two (2) Groups under a dichotomous crosstabulation
analysis proved non-signiﬁcant at the .05 level. Under an analysis of variance
(ANOVA). the Group I mean was .37 and the Group II mean was .26. The entire
population (both groups) mean was .32. The between-Groups differences under this
analysis proved non-significant at the .05 level.

In The Community At Least Three Years Prior To Prison Commitment For
Instant Oﬁense(s) (15) — Eighty-one (8 1) (69.896) members of Group I. and sixty-eight
(68) (58. 696) members of Group II. were in the community (free of a jail or prison term)
for at least three (3) years prior to the commission of’ their current (instant) oﬁ'ense.
Thirty-ﬁve (35) (30.296) members of Group I, and forty-eight (48) (41 .496) members of
Group II. were not in the community at least three years prior to committing their
current offense. The differences between the two (2) Groups under a crosstabulation

analysis were non-significant at the .05 level. but a substantive interest existed.

Age At First Arrest (#34) — Twenty-six (26) (22.496) members of Group I. and
thirty-one (3 1) (28.796) members of Group II. were ﬁrst arrested while under the age
of ﬁfteen (15). Ninety (90) (77.696) of the Group I members. and. seventy-seven (77)
(71 .396) members of Group II. were ﬁfteen (l 5) years of age or older when ﬁrst arrested.

 

Chapter IV: Presentation and Analysis of Data

86
The diﬁ'erences between the two (2) Groups under a collapsed bivariate crosstabula-
tion analysis proved non-signiﬁcant at the .05 level. Under an analysis of variance
(ANOVA). the mean age at ﬁrst arrest was 17.2 for Group I. and 16.8 for members of
Group II. The man for the entire population was seventeen (17) years of age. The
differences between the two (2) Groups under this analysis proved non-signiﬁcant.

Current Offense/Sentence

Type Of Instant Oﬁensets) (#06) — Of' the one-hundred sixteen (1 16) members
of Group I. eleven (1 I) (9.596) were sent to prison for a property oﬁense. three (3) (2.696)
for a drug offense. one-hundred one (101) (87.196) for a violent oﬁ‘ense. and one (1)
(0. 996) for a property and violent offense. As for Group II members. thirty-ﬁve (35)
(30.296) were sent to prison for a property oﬁ'ense. three (3) (2.696) for a drug oﬁ‘ense.
seventy-ﬁve (75) (64. 796) for a violent offense. and three (3) for a property and violent
oﬁ‘ense. The diﬁ'erences between the two (2) Groups under a collapsed bivariate
crosstabulation analysis proved highly signiﬁcant (.0006) at the .05 level. It was
apparent through this analysis that Group I had a much higher percentage of violent

oﬁ'enders.

lnstltutlonal Hlstory

Number Of Non-Bondable Major Misconduct Reports In Prison For Which The
Subject Was Found Guilty During His Present Prison Commitment (#26) — Sixty-
seven (67) (5 7.896) members of Group I had no non-bondable misconduct reports
‘ during their period of incarceration. twenty-two (22) (1 996) had one (1) such miscon-
duct report. and twenty-seven (27) (23.396) had two (2) or more such reports. Forty-
eight (48) (41 .496) members of Group II had no non-bondable misconduct reports
during their period of incarceration. thirty-one (3 I) (26. 796) had one such report. and
thirty-seven (37) (31 .996) had two (2) or more such reports.

The differences between the two (2) Groups under a collapsed bivariate analysis
proved signiﬁcant (.0444) at the .05 level. Looking at non-bondable misconduct
reports in dichotomous form. we found that sixty- seven (67) (5 7.896) members of Group

 

Chapter IV: Presentation and Analysis of Data

87
I did not receive such a report during their period of incarceration. while forty-nine (49)
(42.296) did receive such reports. Under this analysis. forty-eight (48) (4 1 .496) members
of Group II did not receive a non-bondable misconduct report while incarcerated. while
sixty-eight (68) (58. 696) did receive such reports.

The Group diﬁ'erences under a dichotomous crosstabulation analysis proved
highly signiﬁcant (.0181) at the .05 level. Under an analysis of variance (ANOVA). the
calculated mean for the entire population was 1.38 (rounded). For members of Group
I the mean was .98. and for Group II it was 1 .77. The differences between the two (2)
Groups proved signiﬁcant (.01 25) at the .05 level. So we see that even though members
of Group I spent considerably more time in prison (see “Length Of Time Served". p.
88). they received signiﬁcantly fewer ,non-bondable (serious) misconduct reports.

Age At Time Of Present Prison Commitment (#35) —Three (3) (2.6%) members
of the C.O.P.E. Group were sixteen (16) years of age when they entered prison. thirteen
(l3) (1 1 .296)were seventeen years of age. twenty (20) (1 7.296)were eighteen years of age.
eighteen (18) (15.596) were nineteen (19) years of age. sixteen (16) (13.896) were twenty
(20) years of age. fourteen (14) (1 2. 196) were twenty-one (2 1) years of age. thirteen (13)
(1 1 .296)were twenty-two (22) years of age. three (3) (2.696) were twenty-three (23) years
of age. four (4) (3.496) were twenty-four (24) years of age. one (1) (0.996) was twenty-ﬁve
(25) years of age. two (2) (1.796) were twenty-six (26) years of age. two (2) (1 .796) were
twenty-seven (27) years of age. one (1) (0.996) was twenty-eight (28) years of age. one
(I) (0.996)was twenty-nine (29) years of age. one (1) (0.996) was thirty (30) years of age.
one (I) (0.996)was thirty-two (32) years of age. one (1) (0.996) was thirty-three (33) years
of age. and two (2) (1 .796) were thirty-nine (39) years of age at the time they entered
prison.

As for the Comparison Group. eight (8) (6. 996) were seventeen (17) years of age.
thirteen (l3) (1 1 .296)were eighteen (18) years of age. thirteen(l3)(1 1 .296)werenineteen
(19) years of age. ﬁfteen (15) (12.996) were twenty (20) years of age. eleven (1 1) (9.596)
were twenty-one (21) years of age. ten (10) (8.696) were twenty-two (22) years of age.
eleven (1 1) (9.596) were twenty-three (23) years of age. seven (7) (696) were twenty-four

 

Chapter IV: Presentation and Analysis of Data

 

 

Table 4.5
Group Comparisons On
The Institutional History Varlables

Study Significant
Marianas utterances

Number or Non-Bondable Misconduct Reports (#26) ............... Yes

Age At Tlme Ot Present Prison Commltment (#35) - Yes

Length Ol‘ Tlme Served (#38) -- - ....... Yes

 

 

(24) years of age. seven (7) (6%) were twenty-ﬁve (25) years of age. six (6) (5.296) were
twenty-six (26) years of age. four (4) (3.496) were twenty-seven (27) years of age. ﬁve (5)
(4.396) were twenty-eight (28) years of age. one (1) (0.996) was twenty-nine (29) years
of age. three (3) (2.696) were thirty (30) years of age. and two (2) (1 . 796) were thirty-eight
(38) years of age at the time they entered prison.

The diﬁ'erences between the Groups under a crosstabulation analysis proved
signiﬁcant (.0427) at the .05 level. Under an analysis of variance (ANOVA). the mean
age upon entry into prison for members of Group I was 20.7 (rounded). and for those
inn Group II it was 22. l (rounded). The mean for the entire population was 2 1.4
(rounded). The diﬁerences between the two (2) Groups under this analysis proved
signiﬁcant (.0069). That is. the members of the C.O.P.E. Group (Group I) were
signiﬁcantly younger when they went to prison.

Length Of Time Served For Instant Oﬁ’ense(s) (#38) — The C.O.P.E. Group (I)
ranged in time served from twenty (20) months to one-hundred seventy-two (172)
months. The Comparison Group (II) ranged from two (2) months to one-hundred thirty-

severn (137) months. There were no unusual concentrations in either Group.

Under a crosstabulation analysis. the diﬁ‘erences between the two (2) Groups
proved signiﬁcant (.0036) at the .05 level. An analysis of variance (AN OVA) performed

 

Chapter IV: Presentation and Analysis of Data

on the data revealed the mean time served for members of Group I to be 76.08 months.
and for members of Group II. the mean was 49.27 months. The entire population mean
for time served was 62.62 months. The diﬁ'erences between the two (2) Groups under
this arnalysis proved highly signiﬁcant (.0000) at the .05 level. So we saw that members
of the C.O.P.E. Group (I) in a collective sense served considerably more time in prison
(slightly more than two (2) years on average) than their counterparts in Group 11.

Educational Hlstory

Academic Educational Level At Time Of Instant Oﬁ’ense(s) (#05) — Under a
collapsed bivariate crosstabulation analysis. it was determined that nine (9) (7.896)
members of the C.O.P.E. Group (I) had an academic educational level at gade nine
(9) or below when they entered prison (time of instant offense). thirty seven (37) (31 .996)
had completed ten (10) or eleven (1 1) years of academic education. ﬁfty-seven (57)
(49. 196) had ﬁnished high school (gade 12) or had earned a GED Certiﬁcate. four (4)
(3.4%) had thirteen (13) years of education. four (4) (3.496) had fourteern (14) years of
education. four (4) (3.496) had ﬁfteen (15) years of education. one (1) (0. 996) had sixteen
(16) years of educatiorn.

As for those in Group II. sixteen (16) (13.896) had an academic educational level at
grade nine (9) or below when they entered prison. ﬁfty-one (51) (4496) had completed
ten (10) or eleven (1 1) years of academic education. forty-six (46) (39. 796) had ﬁrnished
high school (gade 12) or had earned a GED Certiﬁcate. and three (3) (2.696) had
thirteen (13) years of academic education.

Under tints analysis. the diﬁ'erences between the two (2) Groups proved signiﬁcant
(.0245)at the .05 level. Ananalysis of variance (ANOVA) performed on the data revealed
a Group 1 mean of 1 1 .52 (rounded) years of academic education upon entry into prison:
for Group II. the mean is 10.88 years. The Group 1 range was eight (8) years to sixteen
(16) years of academic education. The members of Group II ranged from ﬁve (5) to
thirteern (13) years of education. The diﬁ'erences under this analysis proved signiﬁcant
(.0006) at the .05 level.

 

Chapter IV: Presentation and Analysis of Data

 

 

 

Table 4.6
Group Comparisons On
The Educational History Varlables

Study Significant
Munchies Differences

Educational Level At Instant Ottensem (#05) Yes

Educational Level Upon Parole (#30) Yes

Educatlonal Attainment Level Increase (#39) Yes

 

 

An analysis of this variable under a dichotomous crosstabulation (graduated HS-
YIN) by Group revealed that seventy (70) (60.396) members of Group I (C.O.P.E.)
gaduated from high school prior to entry into prison. and forty-six (46) (39. 796) had
not gaduated from high school prior to entry into prison. As for the members of Group
11 (Comparison). forty-nine (49) (42.296) gaduated from high school prior to entry irnto
prison. and sixty-seven (67) (57.896) did not graduate from high school. The diﬁ'erences
between the two (2) Groups under this analysis proved signiﬁcant (.0058) at the .05
level. This was originally a matching variable. but the limited size of the Comparison
Group (11) Pool prevented this eﬁ'ort from materializing.

Academic Educational Level Upon Parole Prom Present Prison Commitment
(#30) — A collapsed bivariate crosstabulation analysis run on the data pertaining to
academic educational level upon parole revealed that ninety-six (96) (82.896) members
of Group I had fourteen (1 4) years of education upon parole from prison. three (3) (2. 696)
had ﬁfteen (15) years of education. fourteen (14) (12.196) had sixteen (16) years of
education. one (1) (0.9%) had seventeen (17) years of edusation. two (2) (1.7%) had
eighteen (18) years of education.

As for Group 11 members. ten (10) (8.696) had nine (9) or fewer years academic
education upon parole from prison. twenty-six (26) (22.496) had ten (10) or eleven (1 1)
years of education. sixty-eight (68) (58. 696) had twelve (12) years of education or a GED

 

Chapter IV: Presentation and Analysis of Data

91

Certiﬁcate. and twelve (12) (1 0.3%) had thirteen (13) years of academic education.
Under this analysis. the differences between the two (2) Groups proved highly
signiﬁcant (.0000) at the .05 level. Under an analysis of variance (AN OVA) performed
on the data. we found the educational range upon parole for Group I to be fourteen
(14) to eighteen (18) years of education. The range for Group II was ﬁve (5) to thirteen
(13) years. The differences between the two (2) Groups under this analysis proved to
be highly signiﬁcant (.0000).

Academic Educational Attainment Level Increase During Present Prison
Commitment (#39) —As one might well have expected. the differences between the
two (2) Groups in academic educational increase during the incarceration period
proved signiﬁcant (.0000) under a crosstabulation analysis. The range of increase for
Group I (C.O.P.E.) ran from zero (0) (their number of years of academic education did
not increase. but they did earn an associate degee) to eight (8) years. The Group 11
(Comparison) range ran from zero (0) to two (2) years. Under an analysis of variance
(AN OVA). the Group I mean was more than three (3) (3.0862) years. and the Group 11
mean was considerably less than one (1) (.2759) year. The differences between Groups
was signiﬁcant (.0000) under this analysis as well.

Recidivism

Group Statistics

Academic Educational Level At Time Of Instant Oﬁense(s) (#05) — Under a
three-way crosstabulation analysis of recidivism (through the variable RECIDALL) by
Group (I and II) controlling for the variable “academic educational level at time of
instant oﬁense(s)" in dichotomous form (HS graduate- Y/ N). we found that of those
irn Group I (C.O.P.E.) who did recidivate (35) during the two (2) year period of time.
nineteen (19) (54.3%) were high school gaduates and sixteen (16) (45. 796) were not
high school gaduates when they entered prison (see ﬁgure 4. 1 . next page). Of those
in Group II who recidivated (46). twelve (12) (26. 1 96) were high school gaduates. and
thirty-four (34) (73.996) were not high school gaduates when they entered prison (see
ﬁgure 4. 1 . next page).

Chapter IV: Presentation and Analysis of Data

RECIDIVISTS BY GROUP
High School Education — Prison Entry

    

 

 

lml
P
O 80‘
r
C
e 60‘
n
t 40-
a
9 .
0 2n
8

0‘

C.O.P.E. (i) Comparison (ll)

 

 

Non-Graduates I Graduates I

Figure 4.1

 

Of those in Group I (C.O.P.E.) who did not recidivate (81) during the two (2) year
period of time. ﬁfty-one (51) (6396) were high school graduates. and thirty (30) (3796)
were not high school graduates upon prison entry. As for those in Group 11
(Comparison) who did not recidivate (70). thirty-seven (37) (52.996) were high school
graduates. and thirty-three (33) (47.196) were not high school gaduates when they
entered prison. Under this analysis. the differences between the two (2) Groups (I &
II) proved not to be signiﬁcant at the .05 level.

In a three-way crosstabulation analysis of recidivism (under RECIDALL) by the
variable “academic educational level at time of instant offense(s)" in dichotomous
form (HS gaduate- Y/ N) controlling for Group (I and II). we found irn Groupl (C.O.P.E.)
nineteen (19) (54.396) of those who recidivated (35) were high school graduates. and
sixteen (16) (45.796) were not high school gaduates (see Figure 4. 1 , above). The
differences between the high-school graduates and the non-graduates proved signiﬁ-
cant (.3805) at the .05 level. Of those in Group I who did not recidivate (8 1). ﬁfty-one

 

Chapter IV: Presentation and Analysis of Data

RECIDIVISTS BY GROUP
Ottense Type

“30)

37V

    

coca-300-O'u

 

 

   

C.O.P.E. (I) Comparison (ll)
liver»: I Non-Vlolent'

 

Figure 4.2

 

(51) (6396) were high school gaduates and thirty (30) (3796) were not high school
graduates.

In the Comparison Group (II) of those who recidivated (46). twelve (12) (26.1 96) were
high school gaduates and thirty-four (34) (73.9%) were not high school graduates. Of
those inn Group 11 (Comparison) who did not recidivate (70). thirty-seven (37) (52.996)
were high school graduates and thirty-three (33) (47.196) were not high school
gaduates (see ﬁgure 4. 1, p. 92). The differences between Group I and Group II proved
to be signiﬁcant (.0043) at the .05 level.

Type Of Instant Oﬁ'ense(s) (#06) — In looking at recidivistic behavior (through
RECIDALL) while controlling for “oﬁense type” in dichotomous form (non-violent and
violent) by Group (I & II). we found that of the thirty-ﬁve (35) members of Group Iwho
did recidivate during the two (2) year follow-up period. nine (9) (25.796) were sent to
prison for a non-violent offense. and twenty-six (26) (74.396) were violent offenders (see

 

Chapter IV: Presentation and Analysis of Data

94

Figure 4.2. p. 93). Of the forty-six (46) recidivists inn Group II. seventeen (17) (37%) were
sent to prison for a non-violent offense. while twenty-nine (29) (6396) were violent
oﬁ‘enders (see Figure 4.2. p. 93).

Of the eighty-one (8 1) members of Group I who did not recidivate. ﬁve (5) (6.296)were
non-violent offenders. and seventy-six (76) (93.896) were violent oﬁ‘enders. As for the
seventy (70) members of Group II who did not recidivate. twenty-one (2 1) (3096) were
non-violent oﬁ'enders. and forty-nine (49) ( 7096) were violent offenders. Under this type
of analysis. differences between the two (2) Groups (I and II) proved non-signiﬁcant at
the .05 level.

Criminal Recidivism — Property Oﬁ’ense(s) (#40) — Ninety-six (96) (82.896)
members of Group I. and eighty-one (8 1) (69.896) members of Group II were not arrested
for a felony class property offense during the two (2) years following their parole.
Twenty (20) (1 7. 296) members of Group I. and thirty-ﬁve (35) (30.296) members of Group
II were arrested one (1) or more times for a felony class property offense during the two
(2) year follow-up period. These differences in the two (2) Groups under a dichotomous
crosstabulation analysis proved signiﬁcant (.0307) at the .05 level.

In- looking at this type of recidivistic behavior under a crosstabulation analysis
while controlling for oﬁ'ense type (non-violent and violent). we ﬁnd that of the ninety-
six (96) members of Group I who were not arrested for a felony class property oﬁ'ense
during the two (2) year follow-up period. eight (8) had been sent to prison for a non-
violent offense. and eighty-eight (88) had been sent to prison for a violent oﬁ‘ense. Of
the eighty-one (81) members of Group II who were not arrested for a felony class
property offense during the two (2) year follow-up period. twenty-four (24) had been
sent to prison for a non-violent oﬁ‘ense. while ﬁfty-seven (57) had been sent to prison
for a violent offense.

Of the twenty (20) members of Group I who were arrested one (1) or more times for
a felony class property oﬁ‘ense. six (6) were non-violent offenders and fourteen (14)
were violent oﬁ'enders. Of the thirty-ﬁve (35) members of Group II who were arrested
one (1) or more times for a felony class property oﬁ'ense. twenty-one (2 1) had been sent

 

Chapter IV: Presentation and Analysis of Data

95
to prison for a non-violent oﬁ'ense while fourteen (14) had been sent to prison for a
violent oﬁ'ense. The differences between the two (2) Groups in this analysis proved non-

signiﬁcant (.05 level) for the non-violent offenders. but signiﬁcant (.042 7) for the
subjects who were sent to prison for violent offenses.

Under an analysis of variance (AN OVA). the Group I mean for recidivistic behavior
involving one or more property offenses was .46. whereas it was .9 1 for members of
Group II. The entire population mcan was .69. Under this analysis. the diﬂ'erences
between the two (2) Groups proved signiﬁcant ( . 0062). Thus. it was concluded that
members of Group II committed a signiﬁcantly higher number of felony class property
oﬁ'enses during the two (2) year follow-up period than was true for those irn Group I
(C.O.P.E.). '

Criminal Recidivism — Drug Oﬁense(s) (#41) — One-hundred ten (1 10) (94.896)
members of Group I. and one-hundred ten (1 10) (94.896) members of Group II were not
arrested for a felony class drug oﬁ‘ense during the two (2) years following their parole.
Six (6) (5.296) members of Group I. and six (6) (5.296) members of Group II were arrested
one (1) or more times for a felony class drug offense during the two (2) year follow up
period. The diﬁ'erences between the two (2) Groups under a dichotomous crosstabu-
lation arnalysis proved non-signiﬁcant at the .05 level.

In looking at this type of recidivistic behavior while controlling for oﬁ‘ense type (non-
violent and violent). we found that of the one-hundred ten (1 10) members of Group I
who were not arrested for a felony class drug oﬁ'ense during the two (2) year follow-
up period. thirteen (13) were non-violent oﬁennders and ninety-seven (97) were violent
offenders. Of the one-hundred ten (1 10) members of Group II who were not arrested
for a felony class drug oﬁ'ense during the two (2) year follow-up period. thirty-six (36)

were non-violent offenders and seventy-four (74) were violent offenders.

A Of the six (6) members of Group I who were arrested one (1) or more times for a felony
class drug oﬁ'ense. one (1) was a non-violent oﬁ’ender and ﬁve (5) were violent oﬁ'enders.
As for the six (6) members of Group II who were arrested one (1) or more times during
the follow-up period for a felony class drug offense. two (2) were non-violent oﬁ‘enders

 

Chapter IV: Presentation and Analysis of Data

and four (4) were violent offenders. The diﬁ‘erernces between the two (2) Groups under
this analysis proved non-signiﬁcant (.05 level). in the case of both the non-violent and

the violent offenders.

Under an analysis of variance (AN OVA). the mean number of felony class drug
oﬁ'enses for members of Group I during the two (2) year follow-up period was .04. and
for those in Group II it was .02. The diﬁ'erences between the two (2) Groups under this
analysis proved non-signiﬁcant. Again. we see this entire population (botln Groups) did
not have a history of drug oﬁ'enses. either as juverniles or adults.

, Criminal Recidivism — Violent Oﬁ’ense(s) (#42) — Ninety-six (96) (82.896)
members of Group I. and one-hundred ﬁve (105) (90.596) members of Group II were
not arrested for a felony class violent offense during the two (2) year period following
their parole. Twenty (20) (1 7.296) members of Group I. and eleven (1 1) (9.596) members
of Group II were arrested one (1) or more times for a felony class violent oﬁ'ense during
the two (2) year follow-up period. These diﬂ'erences between the two (2) Groups under
a dichotomous crosstabulation analysis proved non—signiﬁcant at the .05 level.

In looking at this type of recidivistic behavior while controlling for oﬁ'ense type (non-
violernt and violent). we found that of the ninety-six (96) members of Group I who were
not arrested for a felony class violent offense during the two (2) year follow-up period.
nine (9) had been sent to prison for a non-violent oﬁ'ense and eighty—seven (87) had
been sent to prison for a violent offense. Of the one-hundred ﬁve (105) members of
Group II who were not arrested for a felony class violent oﬁ‘ense during the two (2) year
follow-up period. thirty-four (34) had been sent to prison for non-violent offenses. while
seventy-one (7 I) had been sernt to prison for a violent offense.

0f the twenty (20) members of Group I who were arrested one (I) or more times for
a felony class violent offense during the two (2) year follow-up period. ﬁve (5) had been
sernt to prison for a non-violent oﬁ'ense. while ﬁfteern ( 15) had been sent to prison for
a violernt oﬁ'ense. Of the elevern (1 1) members of Group II who were arrested one (I) or
more times for a felony class violent oﬁernse during the two (2) year period. four (4) had
been sent to prison for a non-violent oﬁ‘ense. while seven (7) had been sernt to prison

 

Chapter IV: Presentation and Analysis of Data

97

RECIDIVISTS BY GROUP
Within Two-Year Period of Time

‘

38888

    

ﬁOOO-‘SOO-IO'U

 

 

   

C.O.P.E. (I) Comparison (II)

 

IE Non-Reeldivlsts I Recldlvlsts I

Figure 4.3

 

for a violent offense. The differences irn this analysis between the two (2) Groups in
regard to both the non-violent and violent offenders. proved non-signiﬁcant at the .05
level. Under an analysis of variance (ANOVA). we ﬁnd the mean for Group I in regard
to an arrest for a felony class violent crime during the two (2) year period to be .37. and
for those in Group II it came out to be .26. The differences between these two (2) Groups
under this analysis proved to be non-signiﬁcant.

In looking at recidivistic behavior in a. collective sense under a dichotomous
crosstabulation analysis (through the variable RECIDALL) by Group (I & II). thirty-ﬁve
(35) (30.296) members of Group I (C.O.P.E.) did recidivate during the two (2) year period.
and eighty-one (81) (69.896) did not recidivate (see Figure 4.3, above). In the case of
Group 11. forty- six (46) (39. 796) recidivated during the two (2) year period. while seventy
(70) (60.396) did not recidivate (see Figure 4.3. above). The differences between the two
(2) Groups proved non-signiﬁcant at the .05 level. Under an analysis of variance

 

Chapter IV: Presentation and Analysis of Data

Table 4.7
Group Comparisons On Recidivism

 

 

 

 

Study Significant
liariabiee Differences
Recidivism - Properly Otlense(s) (#40) - Yes
Recidivism - Drug Otlense(s) (#41) No
Recidivism - Vlolent Otlensets) (#42) No
RECIDALL (Generated Varlable) No

 

 

(AN OVA). the Group 1 mean for recidivistic acts was .57 (rounded). and for Group II
it was .66 (rounded). The between-Groups differences proved non-signiﬁcant under

this analysis.

While the differences betweern the two (2) Groups under a dichotomous crosstabu-
lation analysis and an analysis of variance (ANOVA) proved non-signiﬁcant (.05 level)
in a statistical sense. the 9.5% difference in recidivism rates between the two (2)
Groups was of substantive interest. We must be cautioned. however. that the overall
differences between the two (2) Groups in regard to recidivistic behavior had to be
considered in conjunction with other variables of inﬂuence.

Predictive/Causative Factors

In iderntifying predictive] causative factors of recidivism. the ﬁrst step was to
measure the statistical signiﬁcance of the remaining twenty-ﬁve (25) independent
variables in relationship to recidivism. This was done by looking at recidivism in a
collective form (not divided into property. drug. or violent oﬁenses) through the
generated variable RECIDALL. and by examining the indeperndernt variables under
crosstabulation analyses irn various forms (standard. dichotomous. collapsed.
and] or three-way with a control). The results of those analyses were as follows:

Race (#03) — Under a crosstabulation arnalysis of recidivism (through the variable
RECIDALL) controlling for “race“ in dichotomous form (non-white] white). ﬁfty-ﬁve

Chapter IV: Presentation and Analysis of Data

 

RECIDIVISTS BY HIGH SCHOOL EDUCATION

AT PRISON ENTRY
Samples Combined

40‘

    

GOOD-taﬂﬂ-IO'D

 

 

Graduates (I I9) Non-Graduates (I I3)

 

I: Did Not Recldivate I Did Recldlvate I

Figure 4.4

 

(55) (37.996) of the non-white subjects recidivated and rninety (90) (62.196) did not
recidivate. As for the white subjects. twenty-six (26) (29.996) recidivated and sixty-one
(61) (70.196) did not recidivate. The factor of “race" proved to be not signiﬁcantly (at
the .05 level) related to recidivism.

Academic Educational Level At Time Of Instant Offense“) (#05) — In an
examination of recidivism (through the variable RECIDALL) under a crosstabulation
analysis controlling for the variable “academic educational level at time of instant
oﬁense(s) " irn dichotomous form (HS gaduate- Y/ N) . itwas determined that of the total
number of high school graduates (l 19) in the entire population (232). thirty-one (3 1)
(26.196) recidivated and eighty-eight (88) ( 73.996) did not recidivate (see Figure 4.4.
above). Of those who were not high school gaduates (1 1 3). ﬁfty (50) (44.296) recidivated
and sixty-three (63) (55.896) did not recidivate within the two (2) year period of time the

 

Chapter IV: Presentation and Analysis of Data

100
subjects were tracked (see Figure 4.4. p. 99). The variable “academic educational
level at time of instant offense(s)" proved to be signiﬁcantly (.0037) related to
recidivism at the .05 level.

me Of Instant Oﬂense(s) (#06) — In a crosstabulation of RECIDAIJ. by “type
of instant oﬁense(s)" irn dichotomous form (non-violent/violent), twenty-six (26)
(5096) of the non-violent offenders recidivated and twenty-six (26) (5096) did not
recidivate. As for the subjects who were sent to prison for a violent oﬁ'ense. ﬁfty-ﬁve
(55) (30.696) recidivated and one-hundred twenty-ﬁve (1 25) (69.496) did not recidivate.
The variable “type of instant oﬁense(s)" proved to be signiﬁcantly (.0096) related to
recidivism at the .05 level.

Prlor Adult Felony Conviction(s) (#08) — In a crosstabulation of RECIDALL by
“prior adult felony conviction(s)” irn dichotomous form (no prior/ one or more prior).
forty-one (4 1) (30. 1 96) subjects who had no prior adult felony convictions recidivated.
while ninety-ﬁve (95) (69.996) subjects in this category did not recidivate. Of those with
one or more prior adult felony convictions. forty (40) (4396) subjects recidivated. and
ﬁfty-three (53) (5796) did not recidivate. The variable “prior adult felony conviction(s)“
proved to be signiﬁcantly (.0456) related to recidivism at the .05 level.

Criminal Proﬁle — Juvenile Property Oﬁ’ense(s) (#09) — In a crosstabulation
of RECIDAIL by “juvenile property offense(s)" in dichotomous form (yes/ no). ﬁfty-
seven (57) (35. 696) of the subjects with no j uvenile property offense(s) recidivated. while
one-hundred three (103) (64.496) did not recidivated. Of those subjects with one or
more juvenile property offense(s). twenty-four (24) (36.996) recidivated. and forty-one
(4 1) (63. 1 96) did not recidivated. The variable “juvenile property offense(s)“ proved
not to be signiﬁcantly related to recidivism at the .05 level.

Criminal Proﬁle — Juvenile Drug Oﬁense(s) (# 10) — In a crosstabulation of
RECIDALL by “juvenile drug oﬁense(s)" in dichotomous form (yes/ no). seventy-
eight (78) (35.396) of the subjects who did not have a juvenile drug offense recidivated.
while one-hundred forty-three (143) (64. 796) of the subjects in this category did not

 

Chapter IV: Presentation and Analysis of Data

101

recidivate. Of those subjects with one or more juvenile drug offense(s). three (3) (6096)
recidivated. and two (2) (4096) did not recidivate. The variable “juvenile drug
oﬁense(s)” proved not to be signiﬁcantly related to recidivism at the .05 level.

Criminal Proﬁle — J uvenile Violent Oﬁ’enseIs) (# 1 1) — In looking at RECIDALL
by “juvenile violent offense(s)” in dichotomous form (yes / no). sixty-three (63)
(33. 796) of the subjects who did not have a juvenile violent offense recidivated. while
one-hundred twenty-four (124) (66.396) did not recidivate. Of those subjects with one
(1) or more juvenile violent oﬁ'ense(s) . eighteen (1 8) (4 7.496) recidivated. and twenty (20)
(52.696) did not recidivate. The variable “juvenile violent oﬁensets)” proved not to be
signiﬁcantly related to recidivism at the .05 level.

Criminal Proﬁle — Adult Property Oﬁense(s) (#12) — In an examination of
RECIDALL by “adult property oﬁense(s)” in dichotomous form (yes/no). forty-four
(44) (29.596) subjects who did not have a prior adult property offense recidivated. while
one-hundred ﬁve (105) (70.596) did not recidivate. Of those subjects with one (1) or
more prior adult property offense(s). thirty-seven (37) (46.396) recidivated. and forty-
three (43) (53.896) did not recidivate. The variable “adult property oﬁ’ense(s)“ proved
to be signiﬁcantly (.01 16) related to recidivism at the .05 level.

Criminal Proﬁle — Adult Drug Oﬁ'ense(s) (#13) — In looking at RECDDALL by
“adult drug oﬁense(s)” irn dichotomous form (yes / no). seventy-eight (78) (3596)
subjects who did not have a prior adult drug oﬁ'ense recidivated. while one-hundred
forty-ﬁve (145) (6596) did not recidivate. Of those subjects with one (1) or more prior
adult drug offense(s) . three (3) (5096) recidivated. andthree (3) (5096) did not recidivate.
The variable “adult drug oﬁense(s)” proved not to be signiﬁcantly related to
recidivism at the .05 level.

Criminal Proﬁle — Adult Violent Oﬁense(s) (#14) — In an examination of
RECIDALLby “adult violent oﬁense(s)” in dichotomous form (yes / no) . sixty-one (6 1)
(33. 796) of the subjects with no prior adult violent oﬁense recidivated. while one-
hundred twenty (120) (66.396) did not recidivate. Of those subjects who had one (I) or
more prior adult violent oﬁ'ense(s). twenty (20) (42.696) recidiVated. and twenty-seven

 

Chapter IV: Presentation and Analysis of Data

102

(27) (5 7.496) did not recidivate. The variable “adult violent oﬁensets)” proved not to
be signiﬁcantly related to recidivism at the .05 level.

In The Community At Least Three Years Prior To Prison Commitment For
Instant Oﬁense(s) (#15) — In examining RECIDALL by the variable “in the
community at least three (3) years prior to prison commitment for instant
offense(s)” in dichotomous form bres/ no). forty-eight (48) (32.296) subjects who were
in the community for tlnree (3) years prior to their irnstant offense(s) recidivated. while
one-hundred one (101) (67.896) of the subjects in this category did not recidivate. Of
those who were not in the connrnunity for three (3) years prior to their instant offense(s) .
thirty-three (33) (39.896) recidivated. and ﬁfty (50) (60.296) did not recidivate. The
variable “in the community at least three (3) years prior to prison commitment
for instant oﬁ’ense(s)” proved not to be signiﬁcantly related to recidivism at the .05

level.

Marital Status At Time Of Instant Oﬁensets) (it 16) — In looking at RECIDALL
by “marital status at time of instant oﬁensets)” in dichotomous form (never
married lever married). sixty-seven (67) (35.696) subjects who were never married
recidivated. while one-hundred twenty-one (12 1) (64.496) subjects in this category did
not recidivate. Among those subjects who at one time or another were married.
fourteen (14) (31.896) recidivated. and thirty (30) (68.296) did not recidivate. The
variable “marital status at time of instant offense(s)” proved not to be signiﬁcantly
related to recidivism at the .05 level.

Employment Status At Time Of Instant Oﬁense(s) (#17) — In an examination
of RECIDALL by “employment status at time of instant offense(s)” in collapsed
form (not working/working] student). thirty-seven (37) (34.396) of those who were not
working recidivated. while seventy-one (7 1) (65. 796) did not recidivate. As for those
subjects who were working. thirty-nine (3 9) (35. 1 96) recidivated, while seventy- two (72)
(64.996) did not recidivated. As for the study subjects who had a student status at the
time of their instant offense. ﬁve (5) (4 1.7%)recidivated. while seven (7) (58.396) did not

 

Chapter IV: Presentation and Analysis of Data

103

recidivate. The variable “employment status at time of instant offense(s)” proved
not to be signiﬁcantly related to recidivism at the .05 level.

History Of Substance Abuse (# 18) — In looking at RECIDALL by the variable
“history of substance abuse” in dichotomous form (yes / no). thirty-eight (38) (43. 796)
of the subjects who had a history of substance abuse recidivated. while forty-nine (49)
(56.396) of the subjects in this category did not recidivate. Of those subjects who did
not have a history of substance abuse. forty-one (41) (29.196) recidivated. and one-
hundred (100) (70.996) did not recidivate. The variable “history of substance abuse"
proved to be signiﬁcantly (.0244) related to recidivism at the .05 level.

Upbringing (#21) — In an examination of RECIDALL by the variable “upbringing”
in dichotomous form (natural parent(s) / no natural parent). seventy-two (72) (34.696)
subjects who were raised by at least one (1) natural parent recidivated. while one-
hundred thirty-six (136) (65.496) of the subjects irn this category did not recidivate. Of
those who were not raised by at least one (1) natural parent. nine (9) (40.996)
recidivated. and thirteen (l 3) (59.1 96) did not recidivate. The variable “upbringing”
proved not to be signiﬁcantly related to recidivism at the .05 level.

Financial Status Of Upbringing Environment (#22) — In examining RECIDALL
by the variable “ﬁnancial status of upbringing environment” irn dichotomous form
(non-poor] poor). sixty (60) (34. 1 96) of the subjects ﬁ'om non-poor homes (irncome of
$ 15.000 or more per year) recidivated. while one-hundred sixteen (l 16) (65.996) did
not recidivate. Of tlnose subjects from poor homes (income below $ 1 5.000 per year).
twenty (20) (4096) of the subjects recidivated. and thirty (30) (6096) did not recidivate.
The variable “ﬁnancial status of upbringing environment" proved not to be
signiﬁcantly related to recidivism at the .05 level.

Family Emotional Support System At Time of Subject's Instant Oﬁense(s)
(#25) — In examinirng RECIDALL by the variable “family emotional support system
at time of subject's instant offense(s)" in dichotomous form (strong support] some
support) ﬁfteen (15) (53.696) of subjects who had strong family support recidivated.
while tlnirteen (13) (46.496) subjects in this category did not recidivate. As for those

 

Chapter IV: Presentation and Analysis of Data

104

subjects who had some family support at the time they committed their irnstant oﬁ'ense.
sixty-six (66) (32.496) recidivated. while one-hundred thirty-eight (1 38) (67.696) did not
recidivate. The variable “family emotional support system at time of subject's
instant oﬁense(s)” proved to be signiﬁcantly (.0272) related to recidivism at the .05
level.

Number of Non-Bendable Major Misconduct Reports In Prison For Which The
Subject Was Found Guilty During His Present Prison Commitment (#26) — In
looking at RECIDALL by the variable “number of non-bendable major misconduct
reports in prison for which the subject was found guilty during his present prison
commitment” in dichotomous form (none/ one or more). thirty-four (34) (29.696) of tlne
subjects who did not have a non-bendable misconduct report recidivated. and eighty-
one (8 1) (704%) subjects in this category did not recidivate. Of those subjects who
received one (1) or more non-bondable misconduct reports. forty-seven (47) (40.296)
recidivated. while seventy (70) (59.896) did not recidivate. The variable “number of
non-bendable major misconduct reports in prison for which the subject was
found guilty during his present prison commitment” proved not to be signiﬁcantly
related to recidivism at the .05 level.

Academic Educational Level Upon Parole From Present Prison Commitment
(#30) — In looking at recidivism (through the generated variable RECIDALL) by the
independent variable “academic educational level upon parole from present
prison commitment” in dichotomous form (HS gaduate- Y/ N). all one-hundred
sixteen (n=1 1 6) (1 0096) members of Group I were higln school gaduates at the time of
their parole from prison. Whereas. eighty (80) (6996) members of Group II were high
school gaduates at the time of their parole from prison. and thirty-six (36) (31 96)
members were not high school graduates at the time of their parole from prison.

MDOC Assaultive Risk Classiﬁcation At The Time Of Parole (#31) — In an
examination of RECIDALL by the variable “MDOC assaultive risk classiﬁcation at
the time of parole” in dichotomous form (low or middle] high). thirty-one (3 1) (2996)
of those who were classiﬁed as a low or middle assaultive risk recidivated. while
seventy-six (76) (71 96) of the subjects in this category did not recidivate. As for the

 

Chapter IV: Presentation and Analysis of Data

105

subjects who were classiﬁed as high assaultive risks. forty-nine (49) (39.596) recidi-
vated. and seventy-ﬁve (7 5) (60.596) did not recidivate. The variable “MDOC assaultive
risk classiﬁcation at the time of parole” ‘ proved not to be signiﬁcantly related to
recidivism at the .05 level.

Age At First Arrest (#34) -— Irn examirning RECIDALL by the variable “age at ﬁrst
arrest” in dichotomous form (under 15/ ﬁfteen and older). twenty (20) (35.196) of the
subjects under ﬁfteen (15) years of age recidivated. while thirty-seven (37) (64.996)
subjects in this category did not recidivate. Of tlnose subjects who were ﬁfteen (15)
years of age or older at the time of their ﬁrst arrest. sixty (60) (35.996) of the subjects
recidivated. while one-hundred seven (107) (64. 1 96) did not recidivate. The variable
“age at ﬁrst arrest” proved not to be signiﬁcantly related to recidivism at the .05 level.

Age At Time Of Present Prison Commitment (#35) — In examining RECIDALL
by the variable “age at time of present prison commitment” in dichotomous form
(under 21/ 21 and over). forty-one (41) (34.596) of those subjects who were under
twenty-one (2 1) when they entered prison recidivated. while seventy-eight (78) (65.596)
subjects in this category did not recidivate. As for the subjects who were twenty- one
(2 1) years of age or older when they entered prison. forty (40) (35.496) recidivated. and
seventy-three (73) (64.696) did not recidivate. The variable “age at time of present
prison commitment” proved not to be signiﬁcantly (at the .05 level) related to the
outcome variable (recidivism).

Age Upon Parole From Present Prison Commitment (#37) — In a crosstabular
arnalysis of recidivism (through the variable RECIDALL) by the independent variable
“age upon parole from present prison commitment” in dichotomous form (under
26/ 26 and over). forty-three (43) (41 .396) subjects who were under twenty-six (26)
when they were paroled from prison recidivated. while sixty-one (6 1) (58. 796) subjects
irn this category did not recidivate. As for the subjects who were twenty-six (26) ycars
of age or older when they were paroled from prison. thirty-eight (38) (29. 796)
recidivated. while nirnety (90) (70.396) subjects irn this category did not recidivate (see
Figure 4.8. p. 1 15). The variable “age upon parole from present prison commit-
ment” proved under this analysis not to be signiﬁcantly (.6748) related to recidivism.

 

Chapter IV: Presentation and Analysis of Data

106

Length Of Time Served For Instant Oﬁ'ense(s) (#88) — In a crosstabular
examination of recidivism (through the variable RECIDALL) by the independent
variable “length of time served for instant oﬁense(s)” in dichotomous form (under
24 months / 24 months or more). three (3) (13.696) subjects who served under twenty-
four (24) months in prison recidivated. while nirneteen (19) (86.496) subjects irn this
category did not recidivate. As for those subjects who spent twenty-four (24) or more
months in prison. seventy-eight (78) (3 7. 1 96) recidivated. and one-hundred thirty-two
(1 32) (62.996) did not recidivate. The independent variable “length of time served for
instant oﬁensem” proved to be signiﬁcantly (.0278) related to the outcome variable-
recidivism.

Academic Educational Attainment Level Increase During Present Prison
Commitment (#39) — In an examination of the dependent variable recidivism
(through the variable RECIDALL) by the independent variable “academic educa-
tional attainment level increase during present prison commitment“ in dichoto-
mous form (less than 2 years / 2 ycars or more). forty-ﬁve (45) (37.596) subjects who had
an academic educational increase of less than two (2) years recidivated. while seventy-
ﬁve (75) (62.596) subjects 111 this category did not recidivate. Of tlnose subjects who had
an academic educational level increase of more than two (2) years. thirty-six (36)
(32. 1 96) subjects recidivated. while seventy- six (76) (67. 996) subjects did not recidivate.
The indeperndent variable “academic educational attainment level increase during
present prison commitment” proved to be signiﬁcantly (.3924) related to the
outcome variable- recidivism.

The variable RECIDALL by the generated variable “prior arrest(s)" (PRIORARR)
was examined under a crosstabulation analysis. Of those subjects who had no prior
felony arrests either as a juvernile or an adult. thirty-three (33) (26.896) of the subjects
recidivated and ninety (90) (73.296) did not recidivate. Of those subjects who had one
(1) or more prior felony arrests. forty-eight (48) (4496) of the subjects recidivated. and
sixty-one (61) (5696) did not recidivate. The variable “prior arrest(s)” (PRIORARR)
proved to be signiﬁcantly (.0061) related to recidivism at the .05 level.

A four-way crosstabular analysis controlling for “education at time of instant
offense(s)” (HS- YIN). and “education at time of parole“ (HS- Y/N) by Group (I or

 

Chapter IV: Presentation and Analysis of Data

107

RECIDIVISTS BY GROUP
High School Education — Upon Parole

d

88888

    

«coo-onetiao-e

 

 

.

C.O.P.E. (0 Comparison (ll)

 

 

Did Not Recldlvate I Did Recldlvate I

Figure 4.5

 

II) was run on RECIDALL. From this analysis. we saw that of the seventy (70) (60.396)
members of Group I (C.O.P.E.) who entered prison with a high school diploma or GED
Certiﬁcate and had a high school diploma or GED Certiﬁcate upon parole. nineteen
(19) (2 7. 196) recidivated and ﬁfty-one (5 1) (72.996) did not recidivate. Of the forty-nine
(49) (42.296) members of Group H (Comparison) who entered prison with a high school
diploma or GED Certiﬁcate and had a high school diploma or GED Certiﬁcate upon
parole. twelve (12) (24.596) recidivated and thirty-seven (37) (75.596) did not recidivate.

On the other hand. of the forty-six (46) (39.796) members of Group I (C.O.P.E.) who
entered prison without a high school diplonna or GED Certiﬁcate and had a high school
diploma or GED Certiﬁcate upon parole. sixteen ( 1 6) (34.896) recidivated and thirty (30)
(65.296) did not recidivate (see Figure 4.5. above). Of the thirty-one (3 1) (26.796)
members of Group 11 (Comparison) who entered prison without a high school diploma
or GED Certiﬁcate and had a high school diploma or GED Certiﬁcate upon parole.
seventeen (17) (54.896) recidivated and fourteen (14) (45.296) did not recidivate (see

 

Chapter IV: Presentation and Analysis of Data

108

Figure 4.5. previous page). Of the thirty-six (36) (31 . 196) members of Group II
(Comparison) who entered prison without a high school diploma or GED Certiﬁcate
and did not have a high school diploma or GED Certiﬁcate upon parole. seventeen (17)
(47.296) recidivated and nineteen (19) (52.896) did not recidivate.

This analysis showed that recidivism rates for those subjects who entered prison
withahighschmtdiplonmorGEDCerdﬁcatedidmtdiﬂ'ersignmcamtybeaueen
Groups (I orH). However.for those subjects in Group I who entered prison without
ahighschwldipbmawGEDCaiiﬂcataparticipationinanacademic
pmgramseemedtohaveanqﬂ‘ectontheirratesqfrecidivism.

Another four-way crosstabular analysis was run on RECIDALL. controlling for
“age upon parole” and “history of substance abuse“ (< 26. no history of substance
abusel> 26. with a history of substance abuse) by Group (I or II). arnd “age upon
parole“ and “academic education at time of instant oﬁense(s)“ (< 26 with no high
school diploma or GED Certiﬁcate / > 26 with a higln school diploma or GED Certiﬁcate)

by Group (I or II).

From this analysis. we saw that of the subjects in Group I (C.O.P.E.) who were under
twenty-six (26) years of age upon parole from prison and had no history of substance
abuse. six (6) (2496) recidivated and nineteen (19) (7696) did not recidivate. Of the
twenty-six (26) members of Group 11 (Comparison) in tlnis category. thirteen (13) (5096)
recidivated and thirteen (13) (5096) did not recidivate. The differences between the two
(2) Groups was signiﬁcant at the .05 level (.0549). Of the subjects in Group I (C.O.P.E.)
who were over twenty-six (26) years of age and had a history of substance abuse.
twenty-nine (29) (31 .996) recidivated and sixty-two (62) (68. 196) did not recidivate. Of
the ninety (90) members of Group II (Comparisorn) in this category. thirty-three (33)
(36. 796) recidivated and ﬁfty-seven (57) (63.396) did not recidivate. The differences
between the two (2) Groups did not prove signiﬁcant at the .05 level (.4964).

Of the nirneteen (1 9) members of Group I (C.O.P.E.) who were under twenty-six (26)
years of age upon parole from prison and did not have a high school diploma or GED
Certiﬁcate at the time of their instant oﬁense(s). six (6) (31 .696) recidivated and thirteern

 

Chapter IV: Presentation and Analysis of Data

109

(13) (68.496) did not recidivate. Of the twenty-eight (28) members of Group 11
(Comparison) who were irn this category. severnteen (l 7) (60. 796) recidivated and eleven
(1 1) (39.396) did not recidivate. The diﬁ‘erences between the two (2) Groups was
signiﬁcant at the .05 level (.0499).

Of the subjects in Group I (C.O.P.E.) who were over the age of twenty-six (26) upon
parole from prison and had a high school diploma or GED Certiﬁcate at the time of their
instant oﬁ'ense(s) . twenty-nine (29) (29.996) recidivated and sixty-eight (68) (70. 1 96) did
not recidivate. Of the eighty-eight (88) members of Group 11 (Comparison) in this
category. twenty-nine (29) (3396) recidivated and ﬁfty-nine (59) (6796) did not recidivate.
The diﬁ'erences between the two (2) Groups were not signiﬁcant at the .05 level (.6544).

This analysis showed the C.O.P.E. Program seemed to be ofmost bandit to
younger inmates under the age qftwenty-six (26) uponparole. who did not have
ahistorgru’substance abuse. andwhoenteredprisonwithoutahighschool
diplomaorGEDCertiﬂcate. Whereas. those who did notmeet thiscriteria

,experiencediittlemeasurabiebeneﬂtﬁ'om completioantheProgr-am.

At this point in the data analyses. the sixteen (16) indeperndent variables which
proved not to be signiﬁcantly (.05 level) related irn a statistical sense to the outcome
variable (recidivism) (see Table 4.8. next page). were eliminated from further analyses
and consideration. Of the nine (9) remaining independent variables (25- 1 6:9). four (4)
of them which proved to be signiﬁcanntly (.05 level) related to recidivism (through the
variable RECIDALL). were eliminated from the study for reasons outlined in Chapter
111 (pgs. 57-58) of this document. Further. one (1) variable which did not reveal a
statistically signiﬁcant (at the .05 level) relationship with recidivism was included in
the goup of possible predictor variables for reasons outlined in Chapter III (p. 59) of
this document.

These six (6) indeperndernt variables (25- 16-3+ 1) . along with the gernerated variable
“prior arrest(s)” (PRIORARR). were subjected to a series of three-way crosstabular
analyses. the results of which follow:

 

Chapter IV: Presentation and Analysis of Data

110

Table 4.8
Predictive/Causative Factors Ot Recidivism

 

 

 

Study Significant
Mariebles Differences
Race (#03) ....................................................................................... No
Educational Level At Instant Offense(s) (#05) ............................. Yes
Type Of Instant Offense(s) (#06) ................................................... Yes
Prlor Adult Felony Conviction(s) (#08) .......................................... No
Juvenlle Property Offense(s) (#09) ................................................ No
Juvenile Drug Offense(s) (#10) ...................................................... No
Juvenlle Vlolent Offense(s) (#1 I) .................................................. No
Adult Property Offense(s) (#12) .................................................... Yes
Adult Drug Offenses (#13) .............................................................. No
Adult Vlolent Offense(s) (#14) ........................................................ No
In The Community Three Years (#15) ............................................ No
Marital Status At Instant Offense(s) (#16) ...................................... No
Employment Status At Instant Offense(s) (#17) ............................ No
History Of Substance Abuse (#18) ................................................ Yes
Upbringing (#21) ............................................................................. No
Financial Status Of Upbringing Environment (#22) ....................... No
Family Emotional Support System (#25) ....................................... Yes
Non-Bondable Misconduct Reports In Prison (#26) ..................... No
Educational Level Upon Parole (#30) ........................................... Yes
MDOC Assaultive Risk Classification At Parole (#31) .................. No
Age At First Arrest (#34) .................................................................. No
Age At Time Of Prison Commitment (#35) _ -- No
Age Upon Parole (#37) ................................................................... No
Length Of Time Served (#38) Yes
Educational Level Increase (#39). _ - _ Yes
Prlor Arrest(s) (PRIORARR) (Generated) ....................................... Yes

 

 

Chapter IV: Presentation and Analysis of Data

111

Academic Educational Level At Time Of Instant Offense(s) (#05) — Inn a three-
way crosstabular analysis of RECIDALL by Group (I or II) with the independent
variable “academic educational level at time of instant oﬁense(s)" irn dichotomous
form (HS gaduate- Y/ N) as a control. it became evident there existed mirnimal
differences in recidivism rates between Group I and Group 11 among those who entered
prison with a high school diploma or above. Nineteen (19) (2 7. 196) of the seventy (70)
(60.396) members of Group I who entered prison with a high school diploma or GED
Certiﬁcate recidivated. while twelve (12) (24.596) of the forty-nine (49) (42.296) members
of Group II who were in this academic educational category recidivated.

Among the forty-six (46) (39. 796) members of Group I who entered prison with
something less than a high school diploma or GED Certiﬁcate. sixteen (16) (34.896) of
them recidivated. Of the sixty-seven (67) (5 7. 896) members of Group II who were irn this
academic category. thirty-four (34) (50. 796) recidivated. The overall recidivism rate for
the members of this study population who entered prison with a high school diploma
or a GED Certiﬁcate was twenty-six poirnt one percent (26. 1 96). whereas for those who
entered prison with something less than a high school diploma. the recidivism rate was
forty-four point two percent (44. 296) (see Figure 4.4. p. 99).

Type Of Instant Oﬁ'ense(s) (#06) — Irn a three-way crosstabular analysis of
RECIDALL by Group (I or II) with the independent variable “type of instant
offense(s)" in dichotomous form (non-violent] violent) as a control. nine (9) (64.396) of
the fourteen (14) non-violent offenders inn Group I recidivated. Of the thirty-eight (38)
non-violent offenders in Group II. seventeen (17) (44. 796) recidivated.

As for the one-hundred two (102) (87.996) violent offenders 111 Group I. twenty-six
(26) (25.596) recidivated. Of the seventy-eight (78) (67. 196) violent offenders in Group
II. twenty-nine (29) (3 7.296) recidivated.

The ﬁndings revealed signiﬁcant differences between members of Group I and
members of Group II on this variable. The analysis of these diﬁ'erences. however. were
mixed in that Group I members displayed a higher percentage of recidivism than Group
11 members among the non-violent goup. but the reverse was true when one examined

 

Chapter IV: Presentation and Analysis of Data

112

RECIDIVISTS BY OFFENSE TYPE

Samples Combined

100“
P
e -
r so
C
e -
n

r 40.
O
9 .
e 20
S

 

 

 

Non-Vlolent VIOIOI'II

 

I! Did Not Recldlvate I Did Recldlvate I

Figure 4.6

 

the violent goup results. Thus. the statistical results in this irnstance were quite open
to interpretation. While the results appeared to irndicate little or no connection between
this variable (factor) and academic pursuits (Group I or Group II membership). a
relationship between this variable and recidivistic behavior was strongly indicated.

These ﬁndings. based on a limited size data set. revealed that ﬁfty percent (5096)
of the non-violent offenders recidivated. while only thirty point six percent (30.696) of
the violent oﬁ'enders recidivated (see Figure 4.6. above). These ﬁndings supported the
contention commonly found in the literature related to adult corrections. that non-
violent qﬂ'enders are more likely to recidivate than are violent Qﬂ'enders.

Prior Adult Felony Conviction(s) (#08) — In a three-way crosstabular analysis
of RECIDALL by Group (I or II) with the independent variable “prior adult felony
conviction(s)" in dichotomous form (no prior / one or more prior) as a control. nineteen
(19) (24.496) of the seventy-eight (78) members of Group I with no prior adult felony
conviction(s) recidivated. Of the ﬁfty-eight (58) members of Group II who had no prior

 

Chapter IV: Presentation and Analysis of Data

113

adult felony conviction(s). twenty-two (22) (3 7. 996) recidivated. Of the thirty—ﬁve (35)
members of Group I who had one (I) or more prior adult felony conviction(s). sixteen
(16) (54.396) recidivated. Among the ﬁfty-eight (58) members of Group II who had one
(1) or more prior adult felony conviction(s). twenty-four (24) (41 .496) recidivated.

The differences between Group I and Group II membership on this variable were
mixed. and gave no indication of a relationship between this variable and Group I or
Group II membership. The ﬁndings irn regard to this study population irndicated that
subjects who had one (1) or more prior adult felony conviction(s) were more likely to
recidivate. due it appeared to an established pattern of criminal acts.

History Of Substance Abuse (#18) — In a three-way crosstabular analysis of
RECIDALL by Group (I or II) with the independent variable “history of substance
abuse" in dichotomous form (yes/ no) as a control. nineteen (19) (45.296) of the forty-
two (42) members of Group I who had a history of substance abuse recidivated. Of the
forty-ﬁve (45) members of Group II who had a history of substance abuse. nineteen
(19) (42.296) recidivated.

As for the seventy-three (73) members of Group I who did not have a history of
substance abuse. sixteen (l 6) (21 .996) recidivated. An‘nong the sixty-eight (68) members
of Group II who did not have-a history of substance abuse. twenty-ﬁve (25) (36.896)
recidivated.

The differences between Group I and Group 11 members in regard to this variable
were mixed. Among tlnose with a history of substance abuse. Group (I or II) diﬁ'erences
were minimal. There was. however. a sizeable difference (21 .996versus 36.896) betweern
Group I and Group 11 members among those who did not have a history of substance
abuse. Thus. a relationship between this variable and Group I or Group II membership
was not evident from the analysis.

However. the dataset associated with this study population clearly reported
asignificant diﬂ‘erenceinrecidivismratesbetween thosesubiectswitha
history of substance abuse and those with no such history. The overall rate of
recidivism among subjects with a history of substance abuse was forty-three poirnt

 

Chapter IV: Presentation and Analysis of Data

114

RECIDIVISTS BY SUBSTANCE ABUSE

    

 

 

 

 

Samples Combined

100i
P
a a
, so
C
e 50‘
n
r .
0 4o
9
e 20‘

o‘ ..
History No History
E Did Not Recldlvate I Recldlvated I
Figure 4.7

 

seven percent (43.796). whereas the recidivism rate for those without a history of
substance abuse was twenty-nine point one percent (29.196) (see ﬁgure 4. 7, above).

Age Upon Parole (#87) — In a three-way crosstabular analysis of RECIDALL by
Group (I or II) with the independent variable “age upon parole" in dichotomous form
(under 26/ 26 or more) as a control. ten (10) (30.396) of the forty-eight (48) members
of Group I who were under twenty-six (26) at the time they were paroled recidivated.
Among the ﬁfty-six (56) members of Group II who were under twenty-six (26) years of
age upon parole. twenty-one (21) (52.596) recidivated.

0f the shay-eight (68) members of Group I who were twenty-six (26) years of age
or more upon parole. twenty-ﬁve (25) (30. 1 96) recidivated. As for the sixty (60) members
of Group II who were twenty-six (26) years of age or more upon parole. twenty-ﬁve (25)
(32.996) recidivated. Further. the ﬁndings irn regard to this study population irn
combined form indicated that forty-one point three percent (41.396) of those who were
under twenty-six (26) years of age upon parole recidivated. and that only twenty-nine

 

Chapter IV: Presentation and Analysis of Data

115

RECIDIVISTS BY AGE UPON PAROLE

 

 

 

 

 

Samples Combined
p
S
I
C
6
n
r
O
s
S
s , _
Under 26 Over 26
IE Did Not Recldivate I Recldlvated I
Figure 4.8

 

point seven (29.7%) of those who were over twenty-six (26) years of age upon parole
recidivated (see Figure 4.8, above). Tlnis ﬁndirng was consistent with clear indications
in the criminal justice adult corrections literature that inmates who are older when
releasedfrom prison are less likely to recidivate.

Length Of Time Served (#38) — In a three-way crosstabular analysis of
RECIDALL by Group (I or II) with the independent variable “length of time served"
in dichotomous form (less than 24 months/ 24 months or more) as a control. two (2)
(5096) members of the four (4) members of Group I who served less than twenty-four
(24) months in prison recidivated. Of the eighteen (18) members of Group II who served
less than twenty-four (24) months in prison. one (1) recidivated. Among the one-
hundred twelve (1 12) members of Group I who served more than twenty-four (24)
months in prison. thirty-three (33) (29.596) recidivated. As for the ninety-eight (98)
members of Group II who served more than twenty-four (24) months irn prison. forty-
ﬁve (45) (45.996) recidivated.

 

Chapter IV: Presentation and Analysis of Data

118

The extremely small number of study subjects who served less than twenty-four
(24) months in prison (22) invalidated any ﬁndings relative to either an association
between this independent variable and Group (I or II) membership. as well as any
relationship between this variable and the outcome variable- recidivism. Selection of
some other dichotomous division poirnt (expressed inn months) would have been purely
arbitrary and would hear no support in criminal justice (adult corrections)
literature.Thus. tints independent variable was eliminated from further analyses and
consideration for this. and other. reasons (see Chapter IH, p. 60).

Prior Arrest(s) — In a three-way crosstabular analysis of RECIDALL by Group (I
or II) with the generated independent variable “prior arrest(s)" in dichotomous form
(no prior arrest(s)/ one or more prior arrest(s)) as a control. sixteen (16) (22.296) of the
seventy-two (7 2) Group 1 members who had no prior arrest(s) recidivated. Among the

_ ﬁfty-one (5 1) members of .Group II who had no prior arrest(s). seventeen (17) (33.396)
recidivated.

As for the forty-four (44) members of Group 1 who had one (1) or more prior arrest(s) .
nineteen (19) (43.296) recidivated. Among the sixty-ﬁve (65) members of Group II who
had one (1) or more prior arrest(s). twenty-nine (29) (44.696) recidivated.

A relationship between “prior arrest(s)" and Group I or Group II membership
could not be established from the analysis performed. In other words. tlnere existed
no evidence that this independent variable was associated with post-secondary
education. Further. because of a multicollinear relationship between “prior arrest(s)”
and “prior adult felony conviction(s).” this variable was eliminated from further
analyses and consideration (see Chapter 13.. p. 60).

Table 4.9 (next page) lists the ﬁve (5) independent variables included irn the major
model. a discriminant ﬁrnction analysis with a forward selection (stepwise) algorithm.

As noted in Chapter III (p. 61). primary or secondary Group (I or II) membership was
“built” into the major model as a predictor variable by virtue of the general study

design. So. in examining the ﬁve (5) independent variables (see Table 4.9. p. 1 1 7)

 

Chapter IV: Presentation and Analysis of Data

117

, Table 4.9
Major Model Varlables

 

Study
MDIIDDIQS

Academic Educational Level At Instant Offense(s) (#05)
Type Of Instant Offense(s) (#06)

History Of Substance Abuse (#18)

Age Upon Parole From Present Prison Commitment (#37)
Prior Adult Felony Convictions (#08)

 

selected for inclusion in the major model. one must understand that Group (I or II)
membership served as the controlling sixth (6th) variable.

Further. it should be understood that the true strength of the variable Group
(I or II) on the major model outcome was masked due to the fact that Group I
included subjectsfor whom 'such membership had little or no measurable affect
on their rates of recidivism. That is. seventy (70) members of Group I entered prison
with a high school diploma or GED Certiﬁcate. and data analyses conducted up to this
point clearly indicated the C. O.P.E. Program provided little measurable beneﬂtfor these
subjects. Therefore. it was anticipated that the variable Group would not prove
to be signiﬁcantly related to the outcome (recidivism) under the major model
analysis. and would not become one of the predictor variables.

The generating class as determined by the major model (a discriminantfunction

analysis) in their order of statistical strength (signiﬁcance levels) relative to the
outcome (recidivism) was:

1) Academic Educational Level At Time Of Instant Oﬁenseis) (.0012)
2) Type Of Instant Offense(s) (.0080)
3) History Of Substance Abuse (.0244)

 

Chapter IV: Presentation and Analysis of Data

1 18
4) Age Upon Parole Prom Present Prison Commitment (.0933)
5) Group Membership (I or II) (.1788)
8) Prior Adult Felony Conviction(s) (.2377)

These six (6) variables were then inserted into the major model one at a time in
the order listed above. Serving as the selection criterion of the major model. the Wilks’
Lambda statistic (see p. 23 of ChapterI) was used to introduce the variables into the
model in a stepwise fashion. Beginning with step zero (0) . a point at which all variables
were removed from the model. the variable “academic educational level at time of
instant oﬁense(s)” was introduced into the model. This process continued until the
F level or tolerance was insumcient for further computation.

At this point. four (4) of the six (6) variables proved to be signiﬁcantly related to the
outcome variable (recidivism). and the remaining two (2) variables could not be
meaningfully introduced into the model (as suspected. Group membership was one
of them). Those four (4) variables were in the order of their statistical strength
(signiﬁcance level):

1) Age Upon Parole Prom Present Prison Commitment (.0000)

2) History Of Substance Abuse (.0000)

3) me Of Instant Offense(s) (.0001)

4) Academic Educational Level At Time Of Instant Oﬂ’ense(s) (.0012)

These four (4) variables. along with the “given” variable Group (I or 11) (now viewed
purely as a design variable). were then built into the conﬁrmatory model (a
hierarchical log-linear analysis with a baclavard elimination algoritlun) in a further
effort to measure the signiﬁcance of their relationship relative to the outcome variable
(RECIDAIL). The initial generating class as determined by the conﬁrmatory model
(a hierarchical log-linear analysian their order of statistical strength (signiﬁcance
levels) relative to the outcome (recidivism) was:

1) Age Upon Parole
2) History Of Substance Abuse
3) Type Of Instant Oﬁense(s)

 

Chapter IV: Presentation and Analysis of Data ’

119

4) Academic Educational Level At Time Of Instant Oﬁense(s)
5) Group Membership

The ﬁrst step was to examine only the ﬁfth-order interaction. In other words. the
ﬁrst eﬁ'ect (variable) to be backed out of the model was “Group.” because it was the
one which displayed the least amount of statistical strength in the independent
variable cluster. as related to the outcome (recidivism). The backward elimination
process continued until the model contained those interactions where the overall
analytical results displayed and reﬂected the “best” model.

The ﬁnal group of effects (variables) which remained after this elimination process
was completed were determined to be the major predictive/ causative factors associ-
ated with the outcome (recidivism). In the order of their statistical strength (strongest
one ﬁrst) relative to their association with the outcome variable (recidivism) . they were:

1) Age Upon Parole

2) History Of Substance Abuse

3) Type Of Instant Oﬁ‘ense(s)

4) Academic Educational Level At Time Of Instant Oﬁense(s)

Their individual and cluster form affect on the outcome is discussed under the
section titled “Statistical Summary.” on page 124 of this chapter.

In addition. the major and'conﬁrmatory models also allowed for the examination
and measurement of the lone relationship between post-secondary academic educa-
tion and recidivism. which is also discussed under “Statistical Summary.“ on page
124 of this chapter.

Survey Results

The results of the Group IAttitudinal Survey are listed in Table 4. l 0 (beginning on
page 1 21). The labels used in the table are: SA (Strongly Agree). A (Agee). D (Disagree).
SD (Strongly Disagree). and U (Undecided). Forty-four (44) (3 7. 99b) of the one-hundred

 

Chapter IV: Presentation and Analysis of Data

120
sixteen (n=1 1 6) subjects in Group I (C.O.P.E.) responded. Their responses are shown
as percentage ﬁgures to the right of each question.

It needs to be emphasized that the attitudinal survey was a purely subjective
inclusion in this study. It was not part of the primary data analyses (see p. 12 of
Chapter 1). Rather. it was included for the purpose of getting a “subjective feel” for the
reasons members of Group I (C. O.P.E.) decided to attend and eventually graduate from
the C. ORE. Program. whether any attitudinal changes regarding personal behavior
was detectable. and whether attainment of a college degree proved helpful to them

upon release from prison.

Because of the subj ective nature of the survey. no attemptwas made to develop and
report the ﬁndings of a non-response bias analysis. Ifthis instrument (the survey form)
and its ﬁndings were part of primary data analyses. such a report (the non-response
bias analysis) would have proved critical to the ﬁndings: for such a report would have
outlined the composition of those who responded to the survey. Such an analysis
would have allowed one to judge the objectivity of the collective responses.

In this instance such an instrument was not included for a series of reasons: First
off. this researcher extended a promise of complete anonymity to the involved
institution (Montcalm Community College). to the Michigan State University Committee
on Research Involving Human Subjects (UCRH-IS). and to the subjects (C.O.P.E. Group)
themselves via the cover letter (seeAppendixN on p. 1 81) sent to them. Further. a trace
code was not included because it may well have reduced the number of responses (a
promise of complete anonymity could not have been extended to the prospective
respondents). and because this researcher felt ethically bound not to trace the
respondents.

In recognition of the fact that we examined purely contextual data. the following
analysis is presented:

0 A high percentage of C. O.P.E. participants entered the Program to impress the
Parole Board (7896). or to “kill time“ (8596). but their interests later turned to self-
improvement (98%).

 

Chapter IV: Presentation and Analysis of Data

121

Table 4.10
Survey Results

 

 

 

 

 

SA A 12 SD ll

1. I entered the C.O.P.E. Program to impress the Parole Board .......... 42 36 I3 09 00
2. I entered the C.O.P.E. Program to kill time ....................................... 57 28 O9 02 M
3. I entered the C.O.P.E. Program to improve myself .......................... 87 I 1 00 00 O2
4. I first entered the C.O.P.E. Program to Impress the Parole Board

and/or to kill time. but my interests later turned

to self-improvement ..................................................................... 89 O9 01 00 Oi
5. My C.O.P.E. studies helped me to better understand myself ......... 86 12 01 00 01
6. My C.O.P.E. studies helped me to better deal with

my Incarceration ................ 92 06 00 00 02
7. My C.O.P.E. studies helped me to better understand

other people .................................................................................. 72 2O 01 01 06
8. Participation in the C.O.P.E. Program helped me

to more clearly define my personal goals ................................. 81 18 00 00 01
9. The COPE Program helped me become a better person ............... 93 O6 00 00 01
10. I decided not to become Involved In further criminal actIvlty

prior to entering the C.O.P.E. Program............. ........................... 57 12 I3 02 16
II. I decided not to become involved in further criminal activity

while In the C.O.P.E. Program ...................................................... 37 22 12 O8 21
12. I decided not to become involved In further criminal activity

after I graduated from the C.O.P.E. Program ............................. 42 18 09 03 28
I3. I never decided against being further Involved

In criminal activity _ ............ 21 16 42 O2 19
14. My graduation from the C. O. P. E. Program helped me

gain release from prison" ............ 91 08 00 00 01
15. My graduation from the C. O. P. E. Program helped me

better adjust to parole and/or release from prison .................... 83 12 02 00 03
16. My graduation from the C.O.P.E. Program helped me

get work once i was released from prison _ 62 11 08 03 I6
17. Graduation from the C.O.P.E. Program has caused me

to view myself In a more positive way 92 08 00 00 00

 

 

Chapter IV: Presentation and Analysis of Data

122

SA A 12 50 ll

18. Graduation from the C.O.P.E. Program has given me
more self-confidence ................................................................... 82 12 00 00 06
19.1 have or wlll continue my college education even further ........... 57 I3 07 00 23

20. I think all prisons should have a college program like C.O.P.E. 92 08 0000 00
21. If all prisons had a program like C.O.P.E., there would be

less trouble among the Inmates .................................................. 99 01 00 00 00
22. If all prisons had a program like C.O.P.E.. many inmates

would be able to Improve their chance

of successfully completing parole .............................................. 72 21 00 01 06
23. If all prisons had a program like C.O.P.E.. there would be

less repeal crime .......................................................................... 61 23 00 00 16
24. Without the benefits of the C.O.P.E. Program. I probably

would have continued with my same old life patterns ............. 64 29 00 00 07
25. WIthout the benefits of the C.O.P.E. Program. I probably would

be back in prison -_ - - - _ 42 23 04 01 30

 

 

- As for attitudinal changes regarding criminal activity. sixty-nine percent (6996)
said they decided not to become involved in further criminal activity prior to
entering the C. O.P.E. Program. ﬁfty-nine percent (5996) said they decided not to
become involved in further criminal activity while in the C. O.P.E. Program. and
seventy percent ( 7096) said they decided not to become involved in further
criminal activity after they graduated from the C. O.P.E. Program.

While there is a lot of overlap in response (the three ﬁgures add up to
considerably more than 100%). clearly the respondents reﬂect a strong
attitudinal change regarding further criminal activity as a result of the
C. O.P.E. Program inﬂuence.

0 The respondents felt that graduation from the C. O.P.E. Program helped them to

better adjust to parole (9596). and further to get a job upon release from prison
(7396).

¥

Chapter IV: Presentation and Analysis of Data

128

From this instrument. one can gain some subjective measure of the value of the
C. O.P.E. Program. . Its participants overwhelmingly felt it was a beneﬁcial experience.
one which better prepared them for their readjustment to the “free community.” In
light of the relatively low cost of this Program (C.O.P.E.) to the Michigan Department of
Corrections (about $200,000 per year). the subjective merits alone seem tojustify

the continuation qf the Program. and other like programs around the State. This
conclusion carries with it the caution that it is based on self-reported evidence.

Testing Of Hypothesis

The following primary research hypothesis was formulated: people who success-
fully complete the C.O.P.E. Program of study will have signiﬁcantly lower rates

qf recidivism than other like groups of inmates (generalized). The hypothesis was
tested in the following manner.

The relationship between completion of the C. O.P.E. Program (the earning of an
associate degree from Montcalm Community College) and reduced rates of recidivism
was evaluated through a multi-step process. This approach was taken in full
awareness that the study involved a very small sample and. because of that. no single
piece of statistical evidence would prove compelling. The primary tests of the
hypothesis were accomplished through a series of crosstabular analyses designed to
assess the bivariate relationship between recidivism and a wide range of independent
variables. the major one being completion of the C.O.P.E. Program Additionally. the
correlation of the various independent variables was assessed.

Following completion of these steps. the analyses of the C. O.P.E./ recidivism
relationship were reﬁned by the inclusion of a series of control variables in a
multivariate contingency table analysis.

Despite the fact these analyses failed to reveal a statistically signiﬁcant
relationship between completion at the C.O.P.E. Program and recidivism. there
was presented strong evidence linking certain of these independent variables
with the outcome (recidivism). Further. there was evidence presented linking

 

Chapter IV: Presentation and Analysis of Data

124

certain sub-groups within the C.O.P.E. Group (Group 1) to an qﬂirrnative answer
relative to the research hypothesis. In other words. certain sub-groups within
the C.O.P.E. Group appeared to have made a significant gain relative to their
rates qf recidivism (they appeared to be significantly lower) as a direct result
of completing the C.O.P.E. Program qf study.

Thus. after the data set had been reduced to a considerably smaller set of potential
predictors of recidivism. a discriminant junction analysis was employed in order to
evaluate the multivariate relationship between recidivism and the following independ-
ent variables: academic education at entry into prison (HS- YIN). type of oﬁ'ense
(violent/non-violent). history of substance abuse (YIN). age upon parole (<26/
>26 years of age). and Group (I or II). As a ﬁnal and conﬁrmatory step. a hierarchical
log-linear analysis was emplbyed in the hope of reﬁning and / or conﬁrming the results
of the discriminant function analysis (the major model).

It should be noted that all the above procedures were severely constrained by the
relatively small sample size available for analysis. Thus. many of the ﬁndings. while
apparently large in absolute terms. were not statistically signiﬁcant and must be
viewed with caution. This is particularly critical when evaluating any of the multiva-
riate analyses. It should also be noted that the matching process for sample selection
did not provide any measure of control for problems of self-selection. and it was thus
impossible to assess the contribution of prisoner motivation on the study outcomes
(despite the impressions one can garner from the attitudinal survey). Similarly.
variables not included in the matching process were not controlled for. and the effects
of differences on such variables could not be evaluated.

Statistical Summary

A crosstabular analysis indicated that those who completed the C. O.P.E. Program
of study (Group I) had an overall recidivism rate of thirty point two percent (30.296) (35
of l 1 6) . while the Comparison Group (Group II) recidivated at a thirty-nine point seven
percent (39. 7%) rate (46 of l 16) (see Figure 4.9, p. 1 25). The difference of nine point ﬁve
percent (9.5%) was not statistically signiﬁcant at the .05 level. The sizeable difference

 

Chapter IV: Presentation and Analysis of Data

125

TOTAL RECIDIVISM
By Group

“000-0-300-‘01
B 8 S 8

 

 

 

O

C.O.P.E.,(i) Comparison (II)

 

LNon-Recldlvists I Recldivlsts I

Figure 4.9

 

in percentage terms between the two (2) Groups was not suﬁicient to report with any
degree of conﬁdence that the C. O.P.E. Program positively reduced rates of recidivism.

in a purely bivariate sense. for those who completed the Program (graduated with an
associate degree). There emerged. however. some evidence to suggest that completion
of the C.O.P.E. Program may be predictive of parole success for selected sub-groups of
the prison population.

The use of control variables indicated that inmates who were under the age
of twenty-six (26) upon parole. who did not have a history of substance abuse.
who were incarcerated for a violent Qﬂ'ense. and those who entered prison
without a high school diploma or GED Certiﬁcate seemed to beneﬁt signiﬁ-
cantlyfrom completion qf the C.O.P.E. Program. At the very least. these factors (age
upon parole. history of substance abuse. type of offense. and academic educa-
tional level at time of prison entry) seemed to play a mitigating role in the
relationship between completion of the C.O.P.E. Program of study and reduced rates
of recidivism.

 

Chapter IV: Presentation and Analysis of Data

126

Further analyses employing historical and demographic variables as statistical
controls indicated the relationship between completion of the C. O.P.E. Program of study
and recidivism may be contingent on factors which precede entry into prison. The
principal variables in this category are history of substance abuse and age upon

parole.

The control for history of substance abuse indicated a specifying effect. While
there was no signiﬁcant relationship between completion of the C. O.P.E. Program of
study and recidivism for prisoners with a pattern of substance abuse. prisoners with
no such history showed aﬁfteen percent (1 5%) difference in recidivism ratesfor
C.O.P.E. versus non-C.O.P.E. prisoners (21 .9% vs. 38.8%). This observed difference
was signiﬁcant at the .05 level.

The control for age upon parole demonstrated a stronger specifying eﬁ‘ect.
Prisoners twenty-six (26) years qf age or older upon parolefrom prison showed
no effect on recidivismfrom C.O.P.E. Program completion. However. prisoners
under the age qftwenty-six (28) upon parolefrom prison displayed a twenty-two
point six percent (22.6%) dw’erence in recidivism rates (29.2% C.O.P.E. vs.
51.8% man-C.O.P.E.). The difference was signiﬁcant at the .05 level.

The ﬁnding that any relationship between C. O.P.E. Program completion and
recidivism was strongly dependent on prisoner characteristics gave j ustiﬁcation to the
conduct of a multivariate analysis. When a discriminantfunction analysis employing
Group (C. O.P.E. or non- C. O.P.E.). academic educational level at time of prison
entry (HS or GED Certiﬁcate- YIN). oﬁense type (violent/non-violent). history of
substance abuse (YIN). age upon parole from prison (<26I >26). and prior adult
felony conviction(s) (YIN) was performed. age upon paroleﬁ'orn prison. history
of substance abuse. qﬂ'ense type. and academic educational level at time qf
prison entry emerged as signiﬁcant predictors Qf recidivism. Group (C.O.P.E.l
non- C. OPE.) membership did not emerge as a signiﬁcant overall predictor of
recidivism. This may have been attributable to a true lack of association. or it might
have occured because its eﬁ'ects were limited to a comparatively small subset of the
Group I (C. O.P.E.) population.

 

Chapter IV: Presentation and Analysis of Data

 

SUMMARY, CONCLUSIONS,
AND RECOMMENDATIONS

 

Summary
Group Comparisons

The matching variables were selected on the basis of an extensive review of the
crlrninal justice (adult corrections) literature. They were:

1) Race

2) Marital Status At Time Of Instant Offense(s)

3) Employment Status At Time Of Instant Offense(s)

4) MDOC Assaultive Risk Classiﬁcation At Time Of Parole
5) Age Upon Parole From Present Prison Commitment

The matching effort was directed at establishing two (2) Groups of subjects who
were. in a collective sense. as alike as possible in regard to certain important factors.
The matching variables selected related to conditions (factors) which had a well-
established linkwith recidivism. were widely suspected as having an important tie with
recidivistic behavior. or assisted in the control of the study.

Because of distinct cultural differences associated with the factor of “race.” it was
included as a matching variable. To have two (2) Groups with dramatic diﬁ'erences in
regard to racial make-up would have been to disregard the strong inﬂuence of cultural
heritage and the variation that brings to language use. social views and practices. and
value system formation.

 

127

128

Marital and employment status at the time the subjects committed the offense for
which they were incarcerated. according to the literature. reﬂected on the condition
of social stability. The literature strongly suggested that a person without
community andfamily ties is more likely to be involved in criminal activity.
Although this is an area which is very much in need of further research. ignoring these
conditional elements would have opened the study ﬁndings to serious question.

The MDOC Assaultive Risk Classiﬁcation code relates directly to the degree of
supervision administered to an inmate once he is paroled to the free community.
Namely. those with high assaultive risk classiﬁcation codes are placed under more
rigid supervision. not the least of which is the frequency with which they must
personally report to their parole ofﬁcer. Thus. in the interest of comparing subject
Groups who had similar parole experiences (the lack of similar parole experiences is
why outstate parolees were eliminated from the study). this factor was included as a
matching variable.

The factor of age upon parole was chosen as a matching variable because the
literature clearly established this conditional element as having a direct relationship
with recidivism. As indicated in Table 4. l on page 79. the two (2) Groups were without
signiﬁcant diﬁ'erences in regard to the matching variables.

The two (2) Groups were also without signiﬁcant diﬁ'erences in regard to demo-
graphic and background variables. as reported in Table 4.2 on page 8 l . The literature
gave mild support to a relationship between certain background and demographic
elements (such as substance abuse history. upbringing environment. and family
emotional support system) with that of recidivism. Thus. the lack of signiﬁcant
diﬁ‘erences between the two (2) Groups in this area gave added support to the eﬁ'ort
put forth to formulate two (2) similar study Groups.

The criminal history comparisons between the two (2) Groups reﬂected the fact that
in a collective sense they diﬁ'ered signiﬁcantly in this area. Namely. members of Group
II had a more extensive history of property oﬂ'enses both as juveniles and as adults.
than was true of those subjects in Group I (see Table 4.3. p. 83).

 

Chapter V: Summary. Conclusions and Recommendations

128

In an examination of the “type of Offense(s)“ for which these study subjects were
incarcerated (present criminal history). it became apparent that Group I subjects had
a much greater number of violent oﬁenders (see Table 4.4. p. 85). A fair summary in
regard to the criminal history and oﬁense type make-up of the two (2) Groups was that
Group I consisted mainly of violent offenders. whereas Group II had a higher
percentage of property oﬁ'enders. This circumstance must be considered in drawing
comparisons between the two (2) Groups.

The area of institutional history reﬂected major differences in regard to the two (2)
Groups (see Table 4. 5. p. 88). The members of Group I received sigriﬁcantly fewer non-
bondable misconduct reports while in prison. Speciﬁcally. forty-nine (49) (42.296)
members of Group I received one (I) or more such reports. while sixty-eight (68) (58.696)
members of Group 11 received one (1) or more such reports. To add practical meaning
to these Group diﬁerences. we must also give notice to the fact that members of Group
I served a mean time of 76.08 months in prison. whereas members of Group II served
a mean time of 49.27 months (see “Length Qfﬂme Served“. p. 1 15). Thus. despite
the fact that members of Group I served on average considerably more time in prison.
they received signiﬁcantly fewer non-bondable misconduct reports.

The two (2) Groups also signiﬁcantly differed in regard to their age at the time they
entered prison (see Table 4.5. p. 88). The mean age for members of Group I upon entry
into prison was 20.7 years of age. whereas it was 22. 1 years of age for members of
Group 11. Thus. we saw that members of Group I were on average a year-and-a-half
younger than their counterparts in Group 11 upon prison entry.

The two (2) Groups also reﬂected sigriﬁcant diﬁ'erences in regard to the area of
educational history (see Table 4. 6. p. 90). Seventy (70) members of Group I had a
completed high school education or above when they entered prison. whereas only
forty-nine (49) members of Group II were in that category. Group I members had a
mean of l 1.52 years of academic education upon prison entry. whereas Group II
members had a mean of 10.88 years of academic education upon entry into prison.
The variable of “academic educational level at time of instant oﬂ’ense(s)“ was

 

Chapter V: Summary. Conclusions and Recommendations

130

originally selected as a matching variable. but the limited size of the Comparison Group
Pool (from which Group II subjects were selected) prevented this eﬁ’ort from material-
izing. Thus. we found that members of Group I were on average better educated upon
prison entry. and because they successfully completed the requirements for an
associate degree while in prison. were better educated upon parole to the ”free
community. "

Recidivism

The overall rate of recidivism for members of Group I was thirty point two percent
(30.296). and for those in Group II it was thirty-nine point seven percent (39. 7%). In a
statistical sense. these diﬁ'erences proved not to be signiﬁcant. but a nine point ﬁve
percent (9.5%) diﬁ'erence between the two (2) Groups was certainly noteworthy.

The study ﬁndings clearly pointed out the inﬂuence of post-secondary academic
educational experiences on recidivism could not be examined in isolation. Rather.
academic educational experiences had to be examined as a complete and integrated
collection of inﬂuence factors. For example. of the forty-six (46) Group II subjects who
recidivated. thirty- six (36) (78.3%) of them were not high school graduates when they
were paroled. While the thirty-six (36) recidivists who did not have a completed high
school education upon parole represented only thirty-one percent (3196) of the Group 11
population. they represented seventy-eight point three percent (78.3%) of the total
recidivists in thatgrOup (see Figure 5.1. nextpage).

The study ﬁndings in regard to academic educational inﬂuence targeted those who
entered prison with less than a completed high school education. and left prison for
parole status without having completed the requirements of a high school education
or its equivalent (a GED Certiﬁcate). This study revealed that seventy-eight point three
percent (78.3%) of the people in this educational category recidivated. and supports
the ﬁndings of other researchers (Waldron. 1974: Thomas. 1957) that attainment qf
acertain minimal academieeducational level appears to lowerthelikelihood
d recidivism

 

Chapter V: Summary. Conclusions and Recommendations

131

RECIDIVISTS — COMPARISON GROUP
High School Education Upon Parole

/— Graduates
)0 (21.7%)

  

/_

Non-Graduates
36 (78.3%)

Flgure 5.1

 

In an eﬁ'ort to develop predictive/causative factors of recidivism. a series of
crosstabulation analyses revealed that twenty-ﬁve (25) independent variables needed
to be tested in regard to the statistical relationship they shared with the outcome
variable (recidivism). Of the twenty-ﬁve (25) independent variables tested. ﬁve (5)
provided strong statistical evidence they were signiﬁcantly related to recidivistic
behavior. Those ﬁve (5) variables (in a priortized listing of strength relative to the

outcome variable) were:

1) Age Upon Parole From Present Prison Commitment

2) History Of Substance Abuse

3) Type 01’ Instant Offense(s)

4) Academic Educational Level At Time 01’ Instant Offense(s)
5) Prior Adult Felony Conviction(s)

These ﬁve (5) variables. along with the “given” variable Group Membership (1 or II).
were buﬂt into a discriminant junction analysis with a forward selection (stepwise)

 

Chapter V: Summary, Conclusions and Recommendations

132

(Wilks' Lambda) algorithm to help determine (individually and in cluster form) the
major predictive/ causative factors of recidivism. as related to the variables examined

in this study.

Four (4) of the six (6) variables proved to be signiﬁcantly related to the outcome
variable (recidivism). They were:

1) Age Upon Parole Prom Present Prison Commitment

2) History Of Substance Abuse

3) Type Of Instant Oﬂ’ense(s)

4) Academic Educational Level At Time Of Instant Oﬂensem

These four (4) variables. along with the “given” variable Group (I or II). were then
built into the conﬁrmatory model (a hierarchical log-linear analysis with a backward
elimination algorithm) in a further effort to measure the signiﬁcance of their relation-
ship relative to the outcome variable (RECIDALL). The initial generating class as
determined by the conﬁrmatory model (a hierarchical log-linear analysis) in their
order of statistical strength (signiﬁcance levels) relative to the outcome (recidivism)

was:

1) Group Membership

2) Age Upon Parole From Present Prison Commitment

3) History 01‘ Substance Abuse

4) Type Of Instant Oﬁensem

5) Academic Educational Level At Time 01’ Instant Oﬂense(s)

The ﬁnal goup of effects (variables) which remained after this elimination process
was completed were determined to be the major predictive] causative factors associ-
ated with the outcome (recidivism). They were:

1) Age Upon Parole Prom Present Prison Commitment

2) History Of Substance Abuse

3) Type 0! Instant Oﬁenseis)

4) Academic Educational Level At Time Of Instant Offense(s)

 

Chapter V: Summary. Conclusions and Recommendations

133

Results of the study indicated no overall statistically signiﬁcant relation-
shipbetweencornpletioantheC.O.P.E.Prograranstudyand reducedrates (J
recidivismfor participants. There was. however. some limited basis to suggest that
completion of the C. O.P.E. Program may be predictive of parole success for relatively
small sub-groups of the population. speciﬁcally: those inmates who are less than
twenty-six (26) years of age upon parole. who do not have a history of substance abuse.
who are incarcerated for a violent type offense. and/ or those who do not have a high
school diploma or GED CertUlcate upon entry into prison. However. the results which
suggest a limited impact on reduced recidivism rates must be viewed with caution in
light of the small study sample as well as the other study limitations (see “Limita-
tions" on pgs. 13-14 of Chapter I).

Survey Results

Forty-four (44) (3 7. 9%) of the one-hundred sixteen (n=1 1 6) subjects in Group I
(C. OPE.) responded to the survey (see pgs. 1 1 9-123 of Chapter IV). ‘

This survey instrument was not part of the primary data analyses. and was
included to provide readers of the study a “subjective feel“ for the reasons members
of the C.O.P.E. Group (I) decided to attend the C. O.P.E. Program and how they felt the
experience impacted on them once they were paroled from prison.

While a high percentage of participants initially entered the Programto impress the
Parole Board and / or to “kill time.” ninety-eight percent (9896) of the respondents
indicated their interests in time turned toward self-improvement. As for attitudinal
changes regarding the possibility of personally becoming involved in further criminal
activity. the survey results were mixed. Clearly. most of the respondents indicated they
had no desired to return to criminal activities once they were paroled from prison.
However. the point at which that attitudinal change took place is indistinguishable
from the survey results. The results did. however. attribute much of that attitudinal
change to the inﬂuence of the C.O.P.E. Program.

 

Chapter V: Summary. Conclusions and Recommendations

134

The respondents felt that graduation from the C. O.P.E. Program helped them to
better adjust to parole (95%). and to get a job once they were paroled from prison (73%).
They strongly indicated via the survey results that they felt the C. O.P.E. Program
provided them with valuable experiences which better prepared them for their re- entry
into the “free community.“

Testlng Oi Hypothesls

The following primary research hypothesis was formulated: people who success-

fully complete the C.O.P.E. Program of study will have signiﬁcantly lower rates
qf recidivism than other like groups qf inmates (generalized).

The primary tests of the hypothesis were accomplished through a series of
crosstabular analyses designed to assess the bivariate relationship between recidi-
vism and a wide range of independent variables. the major one being completion of the
C. O.P.E. Program.

These and other statistical analyses failed to reveal a statistically signiﬁcant
relationship between completion of the C. O.P.E. Program and reduced rates of
recidivism. There was revealed. however. evidence linking certain sub-groups within
the C.O.P.E. Group (Group I) to an amrmative answer relative to the research
hypothesis. Speciﬁcally. those sub-groups were: those inmates who were under the
age of twenty-six (26) upon parole. those who did not have a history of substance
abuse. those who were incarcerated for a violent oﬂense. and/ or those inmates who
entered prison without a high school diploma or GED Certiﬁcate.

Group membership (C.O.P.E. orComparisonldid not emerge as asigniﬂcant
predictorqfrecidivism. Thismayhavebeenattributableeithertoamlack

ofauociation between completion qftheC.O.P.E.Programqfstudyandreduced
ratesqfrecidivism, orbecausethebendicialeﬂ'ectsqftheC.O.P.E.Progr-am
repoﬂedbythisstudywerelhnitedwawmparaﬂvelysmallsubsetoftheamup
I (C.O.P.E.) population.

Chapter V: Summary. Conclusions and Recommendations

185
Conclusions

An analysis of the basic relationship between completion of the C.O.P.E.
Programqfstudy andreducedratesqfrecidivism indicated therewasnooverall
statistically signmcant dw‘erence between C.O.P.E. and non-C.O.P.E. groups
regarding the probability of being arrested for a newfelonyfollowing parole.
Theﬂnding was continued by a multivariate analysis. When several indepen-
dent variables were simultaneously used to predict recidivism, C.O.P.E. Pro-
gram completion did not emerge as a signiﬁcant predictor of parole success.
However. it should be noted that C.O.P.E. Program graduates through a survey
instrument (subjective) overwhelmingly felt it was a beneﬁcial experience. one
which better prepared them for their readjustment to the "free community."

There was some evidence to suggest that graduation from the C. O.P.E. Program
may be predictive of parole success for selected sub-groups of the prison population.
Use of control variables indicated that prisoners who were less than twenty-six (26)
years of age upon parole from prison. did not have a history of substance abuse. were
sent to prison for a violent oﬁ'ense. and / or entered prison without than a high school
diploma or GED Certiﬁcate showed lower rates of recidivism than did prisoners with
similar backgrounds who did not complete the C.O.P.E. Program of study.

The ﬁndings mu st be viewed with caution in light of the small sample size. and the
other limitations of the study (see p. 13 of Chapter I). Results obtained under these
conditions are highly volatile. particularly in the instance of any sort of multivariate
analysis. Further. it must be remembered that the study was limited to the Ionia
(Michigan) institutions. which were not representative of the entire correctional
system. A ﬁnal more important limitation has to do with the impact of self-selection.
As noted earlier. the design of the study did not allow for an assessment of the effects
of prisoner motivation on the overall outcomes. or the possible impact of diﬁ’erences
between the study Groups on oﬁ'ense type and prior education. Absent such an

assessment. it was very possible that the same motivation which lead a prisoner to

 

Chapter V: Summary. Conclusions and Recommendations

136

enroll in and complete the C. O.P.E. Program played a major role in reducing the
likelihood of his recidivating following parole.

' Insummary, results indicated theC.O.P.E.Program cannotbesaid to have
a measurable qﬂ‘ect on recidivism rates. While there was some limited basisfor
inferring an qﬂ'ectfor speciﬁed sub-groups of the C.O.P.E. population, further
continuation is necessary before drawing any ddinitive conclusions.

Recommendations

The study ﬁndings support and offer four (4) major recommendations:

1)

2)

3)

4)

The Legislature. Parole Board. corrections ofﬁcials. and general public should
recognize academic education oﬁered to prison inmates as a worthwhile
expenditure of public funds. The subjective merits alone seem to justify
the continuation of such programs.

The inmates who fail to complete the requirements for a high school diploma
(GED) prior to parole should be viewed as the at risk group by the Parole Board.
These are the people who according to the ﬁndings of this study are most likely
to recidivate (78.3%).

In a limited budget situation. prison inmates who apply for entry into a post-
secondary academic program should be screened to determine if they win
be under the age of twenty-six (26) upon parole from prison. do not have
a history of substance abuse. were sent to prison for a violent oﬂense. and]
or entered prison without a high school dipoma or G. Certiﬁcate.
because the study ﬁndings indicate inmates who fall into one or more of these
categories would beneﬁt most from such education.

The Parole Board should develop and utilize a recidivistic proﬁle assessment
tool that will rate (predict) the chances of an inmate being involved in
recidivistic behavior upon release from prison. The results of this assessment

 

Chapter V: Summary. Conclusions and Recommendations

137

should be used in determining whether an inmate should be paroled from
prison prior to completion of his / her maximum sentence. The ﬁndings of this
and other like studies could be used in developing the assessment instrument.

Suggestion For Future Research

The ﬁndings and implications of this study lead to the following recommendations
for further research:

I) The relationship between post-secondary education programs within other U.S.
prison settings. and rates of recidivism.

2) The affect of the attainment of a high school diploma or GED Certiﬁcate while in
prison on rates of recidivism.

3) The qtfect of academic education programs within US. prison settings on
institutional control.

4) In a limited budget situation. the need for giving priorityﬁlndtng to the high
school] GED Certiﬁcate programs.

5) The beneﬁts of occupational training programs within US. prisons.

Concluding Observations

The expansion of prisoner rights and opportunities in the 50's. 60's. and 70's came
from a forgiving generation. Until very recent days. we as a nation have been willing
to forgive convicted felons by encouraging them to “rehabilitate” themselves. and
releasing them when they convinced us they were “well.“

The national mood today is swinging away from the “liberal” approach. which by
the evaluation of its own supporters has not been very effective. Instead. we are
currently determined to stiffen prison sentences. reduce and / or eliminate “rehabili-
tative“ programs within our prisons. and greatly broaden the use of stiﬁ'er penalties

 

Chapter V: Summary. Conclusions and Recommendations

138
including that of death for certain crimes. The opponents of these harsher approaches.
however. are still to be found in large numbers in this nation. At this point in our
national history. we ﬁnd attitudes regarding the treatment of criminals to be highly
polarized. People at both ends of the spectrum feel strongly about their positions.

Regardless of where one stands on these issues. it is clear that American
corrections constitutes a profession in search of a philosophy: for currently the
“philosophy“ is one that swings erratically between deterrence and rehabilitation. It
is a “philosophy” oﬁen born in an atmosphere of crisis. one which is essentially reactive
as opposed to proactive in nature.

As the corrections profession enters perhaps the most critical period in its history.
the correctional education process is itself undergoing a searching examination. Most
penologists openly admit they know little about what causes crime. However. an
examination 0f oﬁ‘ender histories clearly reveals that today's prison papulation suﬁers
from undereducation. unemployment. drug and/ or alcohol use. and in general poor
societal adjustment. Education of inmates. addressing as it does the ﬁrst two (2)
problems. claims signiﬁcant support among many people— both inside and outside
the criminal justice system (Becker. 1983).

As pointed out by this and other like studies. research concerning the impact of
correctional education on paroled inmates has produced highly mixed results.
Nevertheless. one cannot and should not ignore the deﬁnable themes which have
emerged from the broad collection of criminal justice research. such as: 1) academic]
occupational training which prepares tnrnates for marketable jobs have proven suc-
cessﬁdﬂrmanyhstarwesandshouldbecondrmedandexpandedandmthe
sulyecttveandnon-measw'ablebeneﬂtsqfacademic/occupational training needto be
considered in funding such programs.

 

Chapter V: Summary. Conclusions and Recommendations

 

APPENDICES

 

 

 

 

 

 

 

 

 

A. LETTER REQUESTING PRELIMINARY STUDY APPROVAL FROM MCC ........... 139
B. LETTER FROM MCC GRANTING PRELIMINARY STUDY APPROVAL .............. 140
C. RESPONSE TO LETTER GRANTING

PRELIMINARY MCC STUDY APPROVAL 14)
D. LETTERS FROM MCC TRANSMITTING PROPOSAL TO THE MDOC ................ 143
E. LETTER REQUESTING STUDY APPROVAL FROM MDOC ................................ I45
F. LETTER OF STUDY APPROVAL FROM MDOC 147
G. LETTERS REQUESTING STUDY APPROVAL FROM UCIRHS ............................. I48
H. LETTERS OF STUDY APPROVAL FROM UCRIHS 153
I. LETTER OF FINAL STUDY APPROVAL FROM MCC 155
J. MDOC DATA PROCESSING SERVICE REQUEST 156
K. VARIABLES DICTIONARY 158
L. CODING SHEET I77
M. SPSS PROGRAMS I79
N. ATTITUDINAL SURVEY INSTRUMENT 181

 

 

Appendices

APPENDIX A

Letter Requesting Preliminary Study Approval From MCC

 

Appendices

139
January 6. 1986

Mr. Dan Herman

Director. C.O.P.E. Program
Montcalm Community College
Sidney. Michigan 48885

Dear Mr. Herman:

I am a doctoral student at Michigan State University in the area of College and
University Administration. with a cognate in Criminal Justice. Adult Corrections.
My major advisor is Dr. Eldon Nonnamaker.

I am at the dissertation stage of my Ph.D. Program and am currently seeking a
research topic which will bridge my two areas of concentration: higher education
and adult corrections. The C.O.P.E. Program which you administer in the Ionia.
Michigan prisons is of great interest to me. Judging from the preliminary data I
have collected. the C.O.P.E. Program is extremely well run and most effective.
Therefore. I am respectfuny requesting permission to conduct a descriptive study
of the program. I hope that such a study would be of beneﬁt to those of you at
MCC who administer the prison program. as well as the larger community of
peOple who are interested in and concerned with correctional education.

I fully realize that approval to conduct such a study must come from a number
of parties. My ﬁrst step. however. is to receive your preliminary approval. If you
grant that approval. I will then write a formal proposal and submit it to all
concerned parties.

Thank you for the time and attention I know you will give to this request. I
await your answer....

Sincerely.

WW

Arthur Kirk

1842 Hamilton. Apt. Bl
Okemos. Michigan 48864
(517) 349-6941

 

Appendices

APPENDIX B

Letter From MCC Granting Preliminary Study Approval

 

Appendices

140

m

mo ntcalm community college

SIDNEY MICHIGAN 48885

14 January 1986

Arthur Kirk
1842 Hamilton, Apt. 81
Okcmos, HI 48864

Dear Mr. Kirk:

Thank you very much for your letter concerning the possibility of
doing a descriptive study of the COPE program. I took the liberty of
sharing that letter with the vice President and President of Montcalm
Community College both of whom were very supportive. Since I share
their enthusiasm the preliminary approval you requested is granted.

It will be necessary, of course, to also secure the support of the
Department of Corrections. 1 think that it would be best if they were
asked to respond to a formal proposal. They will have questions
concerning purpose, design, confidentiality of information, specific
areas to be investigated, etc, which would be most effectively dealt
with in a formal proposal.

Once completed the proposal should be sent to me for endorsement by
the college after which I will gladly forward it to the proper
authorities .

if I can be of assistance please feel free to contact me at (517)
328-21ii or (616) 527-2500 extension 319.

Sincerely,

  

- tcalm Community College

 

 

 

TRUSTEES: Chairperson Beatrice Dear. ViccCharrpcrson Omllc Trsoran; Treasurer Paul Wamsnurs: David Hayes:

 

Appendices

APPENDIX C

Response To Letter Granting Preliminary MCC Study Approval

 

Appendices

141
January 18. 1986

Mr. Danny G. Herman
C.O.P.E. Program Director
Montcalm Community College
Sidney. Michigan 48885

Dear Mr. Herman:

Thank you very much for your letter of January 14. 1986. in which you grant
me preliminary approval to study the C.O.P.E. Program. People in Michigan know
it to be a well-run, highly effective program. and to say I am pleased over this
opportunity is a real understatement.

My job now is to write a proposal and submit it to your ofﬁce for what I hope
will be ﬁnal study approval. It will be awhile before I can get this proposal in hard
copy form. I want to be certain to have a well-designed proposal which will then
lead to a meaningful and comprehensive study. I want the parameters of the
study to be well-deﬁned so that all concerned parties know exactly where this
thing will take us.

Up to this point I have been referring to the study as a descriptive type. In
actuality I envision a hybrid design approach. or more formally: an observational
design. In descriptive fashion I do want to look at the C.O.P.E. Program in terms
of its history. organizational structure. and ﬁnancial support base. but ifI stopped
at that point the bottom line question would be: So what? I feel the C.O.P.E.
Program affects in a positive way. the quality of life of its participants. It really
goes without saying that such a belief is very difﬁcult to support through formal
research techniques. To my nation. the single most interesting element to examine
is that of recidivism rates. I intend to include in this research design. a statistical
model which will examine that variable as an outcome criterion. As part of the
model I plan to include other factors which previous research indicates might well
be signiﬁcantly related to recidivism. By introducing and controlling these other
variables. we should be able to get a “clean” look at the recidivism rates of those
who graduated from the C.O.P.E. Program. versus the recidivism rates of other like
inmates who have less than a completed post-secondary education. The end
result I hope will support my intuitive notion that the C.O.P.E. Program has a
positive (reducing) inﬂuence on the rates of recidivism of its graduates.

In a broader sense. I hope the study meets the following criteria:
1) Should relate to the ﬁelds of higher education and adult corrections.

2) Should make a positive contribution to the ﬁelds of higher education and
criminal justice. '

 

Appendices

142

3) Should be limited enough in scope to be thoroughly researched.
4) Should have signiﬁcance and be of interest to the author and his audience.

5) Should help promote the concept of prisoner rehabilitation through aca-
demic education.

6) Should make a contribution toward the improved welfare of prison inmates.

In conclusion let me assure you that if granted ﬁnal approval to conduct the
study. I fully intend to protect the privacy rights of the study subjects. Reporting
will be done in goup form. and the individual identities of the study participants
will go unreported and unavailable to members of the general public.

Thanks again for granting me preliminary study approval. I will keep you fully
appraised of activities in regard to the conduct of this study.

Sincerely.

WES...

Arthur Kirk

1842 Hamilton. Apt. B1
Okemos. Michigan 48864
(517) 349-6941

 

Appendices

APPENDIX D

Letters From MCC - Transmitting Proposal To The MDOC

 

Appendices

143

m

montcolm community college

SIDNEY MICHIGAN MS

May 7, 1986

Terry Murphy

Acting Chief of Research

Michigan Department of Corrections
P.O. Box 30003

Lansing, MI 48909

Dear Mr. Murphy:

In January, I was contacted by Mr. Arthur Kirk, a Michigan
State University Doctoral student, who proposed doing a study
of the COPE (College Opportunity - Prison Extension) Program
conducted by Montcalm Community College within the Ionia
Complex. After much discussion the attached proposal was
agreed upon as a viable study, the design of which would
yield information valuable to Mr. Kirk, Montcalm Community
College and the Michigan Department of Corrections.

The Montcalm Community College Administration supports the
concept of the study and will assist Mr. Kirk's efforts in
any way possible.

While the study will not involve inmates directly, access to
their records will be required. To that end, we request your
support and authorization to conduct the research described
in the attached proposal.

Mr. Kirk wishes to have information gathering completed by
August of this year. Therefore, time is of some importance.

If you have any questions please feel free to contact me at
(517) 328-2111, ext. 266 or (616) 527-2500, ext. 319.

Thank you for your assistance.

Sincerely,

 

Di
DGH/jel iii;

If

mmewm‘ . .a .. .i. . Y .. - Au-“ -

 

 

ErieNet-eenn: IeeenUnseen: Orville‘l'rsbini

 

Appendices

144

m

montcolm community college

SIDNEY. MICHIGAN 48885 5” 128211

November 7, 1986

Mr. Terry Murphy

Acting Chief of Research

Michigan Department of Corrections
P.O. Box 30003

Lansing, MI 48909

Dear Terry:

I want to thank you for taking time from your busy schedule
to talk with me about Arthur Kirk's research proposal.

I was a little anxious about the amount of time that had
passed since the proposal was submitted (May 7, 1986), but
Vou have put my mind at ease. Your commitment to having a
resolution to our request within the next two weeks was
greatly appreciated. We are slightly behind schedule on our
time line, but if we have the approval as soon as you have
indicated. I think Art can get back on schedule in short
order.

Again. I appreciate all you've done in regard to this
proposal; I know it was not an easy task.

Thank you for your support.

Sincerely,

   

. Herman, Director
Col ge Opportunity Prison Extension

cc: Arthur Kirk
File

DGH/lel

 

mug - - - ~__ -— .. Trebran: 3.....- r... “
Robert Palmer: Karen Carbonem: Eric Haivoraen

 

 

Appendices

APPENDIX E

Letter Requesting Study Approval From MDOC

 

Appendices

145
August 23. 1986

Mr. Terry Murphy
Acting Chief of Research

Michigan Department of Corrections
P.O. Box 30003
Lansing. Michigan 48909

Dear Mr. Murphy:

Back on May 7, 1986. Dan Herman sent you ,a copy of a proposal in regard to a
study I hope to conduct. Because this study calls for gathering information from
certain inmate ﬁles held by the MDOC. approval by your department is being
sought.

I spoke with you about a month ago regarding this study. and at that time you
made mention of the need for some additional information regarding the study
subjects. With the strong hope of gaining study approval. I have gone ahead and
developed Chapter I. which is simply an expansion of the proposal. I have
enclosed a copy of this document in the hopes it will satisfy your questions
regarding subject recruitment. subject “treatment.” the subject information
(variables or data points) being sought from Montcalm Community College and
your department (MDOC). particulars regarding design. and general ideas

regarding privacy protection.

I call your attention to the last paragraph on page I l. in which mention is
made of a survey instrument. A copy of both the cover letter and the instrument
itself is enclosed so you can determine exactly what kind of information I am
seeking via t hat instrument. The instrument will be returned to PD. Box 6756.
East Lansing. Michigan 48823. This post omce box is located in the MSU Student
Union. and I am the only person who has access to it. I will be the only person to
view the returns. As you can see by the survey instrument. it is designed so the
person returning it mnnot be identiﬁed. This will protect identities and encourage
openness. Further. I will be the only person to tabulate the retums. and the
resulting information will be reported in collective form only.

Almost all of the information I seek in regard to the study subjects must come
from their MDOC ﬁles located in Lansing. Michigan. Some additional information
regarding members of Group I (C.O.P.E. graduates) will have to come from records
maintained by Montcalm Community College. If granted study approval. I will
treat all the data extracted from these records with utmost caution and concern in
regard to the privacy rights of these subjects. I am well aware of my legal and
ethical responsibilities.

 

Appendices

148

I do hope the information contained in this packet serves to answer your legal
and ethical concerns in regard to this study. If any questions remain. please
contact me and I will see to it that you receive a prompt response. Thank you for
the time and effort you are putting into this study approval request. Like you. I
too want to be certain that all the proper procedures are being followed.

Sincerely.

WW

Arthur Kirk

1842 Hamilton. Apt. Bl
Okemos. Michigan 48864
(517) 349—6941

cc: Mr. Dan Herman. C.O.P.E. Program Director
Mr. Zolton Ferency. Dissertation Committee Member

encl.: Chapter I - Dissertation
Attitudinal Survey Instrument - C.O.P.E. Graduates

 

Appendices

APPENDIX F

Letter Of Study Approval From MDOC

 

Appendices

14ar

MICHIGAN DEPARTHEIT'OP'CORRECTIONS
RESEARCH DIVISION

H.E H10 R A N D U."

'10: Art Kirk

m: Terry Murph

DATE: October 6, 1987

SUBJECT: .HontcaLe Can-unity College Evaluation

I am pleased to inform you that the Research Division has formally
approved your request to conduct an evaluation on the Montcalm Community
College Program. This approval is subject to the conditions that we
discussed and agreed upon earlier. The Research Division will provide
reasonable assistance in locating and retrieving files. consult in
variable specifications. and consult on the analysis phase. The
Research Division will also provide temporary coding space to assist you
with this project. Good luck on your project and please feel free to
contact me at any time. The contact person for this project will be R.
Douglas Kosinski.

TM/rw

 

Appendices

APPENDIX G

Letters Requesting Study Approval From UCRIHS

 

Appendices

148
July 17. 1986

University Committee on Research Involving
Human Subjects (U CRIHS)

238 Administration Building

Michigan State University

East Lansing. Michigan 48824- 1046

Dear Sirs:

I am a doctoral student at Michigan State University in the area of College and
University Administration. My major advisor is Dr. Eldon Nonnamaker. I am
currently at the dissertation stage of my studies and have written my study
proposal. a copy of which is enclosed. Also enclosed is a copy of my signed
Proposal Approval Form.

I am applying to this Committee for study approval in that my study involves
human subjects. I request exemption from full Committee review bemuse my
study directs itself to the observation of public behavior (1D). At no time will the
identity of the subjects be revealed to members of the general public. All subjects
will be identiﬁed in the study by numeric code. the key to which will only be
lmown to me.

In anticipation of exemption from full Committee review. I have enclosed 9n:
copy of the information required for full Committee review.

Thank you for the time and attention I know you will give this request.

Sincerely.

WM...

Arthur Kirk

1842 Hamilton. Apt. B 1
Okemos. Michigan 48864
(5 17) 349-6941

 

Appendices

149
August 23. 1986

University Committee on Research Involving
Human Subjects (U CRIHS)

238 Administration Building

Michigan State University

East Lansing. Michigan 48824- 1046

Dear UCRIHS Member:

I received a letter from Dr. Bredeck dated August 21. 1986. in which he
requested additional information on my proposed study. Much of the information
requested is covered in Chapter I. which is now complete: a copy is attached. This
document covers questions you have expressed regarding subject recruitment.
subject “treatment.” the subject information (variables or data points) being sought
from Montealm Community College and the Michigan Department of Corrections.
particulars regarding design. and some general comments regarding privacy
protection. Much of this information was not developed. and hence not available
at the time I ﬁrst applied for study approval from this Committee. I do hope the
information contained in this instrument serves to satisfy your questions in the
above listed areas. If not. I will be happy to supply additional information.

I eall your attention to the last paragaph on page 1 1. in which mention is
made of a survey instrument. A copy of both the cover letter and the instrument
itself is enclosed so that you can determine exactly what kind of information I am
seeking via that instrument. Montcalm Community College and/ or the Michigan
Department of Corrections will supply me with the last Imown address of all
subjects in Group I. The instrument will be mailed to all members of Group I. and
their replies will be mailed to Northern Research Associates. P.0. Box 6756. East
Lansing. Michigan 48823. Northern Research Associates is a one-person. non-
proﬁt corporation incorporated by me in the State of Michigan for the purpose of
assisting me in the conduct of this and other research. The PD. Box is located in
the MSU Student Union. and I am the only person who has access to it. I will be
the only person to see the returns. As you can see by the survey instrument. it is
designed so the person returning it cannot be identiﬁed. This will protect
identities and encourage openness. I will be the only person to tabulate the
returns. and the resulting information will be reported in collective form only.

Another concern expressed by some UCRIHS Committee Members has to do
with my access to Michigan Department of Corrections data. I call to your
attention the ﬁrst paragaph on page 9. I will be the only person having access to
personal identities by name. address. institutional number. social security
number. etc. Almost all the information I seek regarding the study subjects must
come from their MDOC institutional ﬁles maintained in Lansing. Michigan. Some
additional information regarding members of Group I (C.O.P.E. gaduates) will

 

Appendices

150

have to come from records maintained by Montcalm Community College. In short.
there is no way the information can be obtained without access to these two sets of
ﬁles. These records are not public information. but under ageement with the
parties mentioned. can be made available to researchers involved in properly
sanctioned studies such as this dissertation. I further call your attention to the
letter from Dan Herman of Montcalm Community College. dated January 14.

1986. As you can see by that letter. I have study support from the administrators
at Montcalm Community College. I also eall your attention to the letter from Dan
Herman dated May 7. 1986. In it he refers to a conversation with William Kime.
Deputy Director of Progams and Planning for the MDOC. It appears from that
letter the MDOC is more than willing and interested in having the study
conducted. You will also see attached. a letter of proposal transmittal from Dan
Herman to Terry Murphy. dated May 7. 1986. Mr. Murphy works under Deputy
Kime in Lansing. I personally talked with Terry Murphy about this study and he
indicated that because the study is not sanctioned by the Legislature. it is
considered low priority. However. he did indicate that he would send me a letter
requesting some additional information. and upon receipt of that information.
study approval would be no problem. He said the MDOC would require me to sigr
a statement of conﬁdentiality. I told him the additional information and the signed
statement would be submitted to him upon request. I have not yet heard from
him. but have taken the liberty of submitting to his emce. the information I
suspect he needs. Attached is a copy of the cover letter used in transmitting that
information to his ofﬁce.

I do hope the information contained in this packet serves to answer your legal
and ethical concerns regarding the conduct of this study. If any questions remain.
please contact me and I will see to it that you receive an immediate response.
Thank you for the time and effort you are putting into this review. Like you. I too
want to be certain that proper procedures are being followed.

Sincerely.

we‘d-L—

Arthur Kirk

1842 Hamilton, Apt. Bl
Okemos. Michigan 48864
(517) 349—6941

cc: Dr. Eldon R. N onnamaker

encl.: Chapter I - Dissertation
Attitudinal Survey Instrument - C.O.P.E. Graduates
Letter from Dan Herman dated January 14. 1986
Letter from Dan Herman dated May 7. 1986
Letter to Terry Murphy dated August 23. 1986

 

Appendices

151
October 24. 1987

University Committee on Research Involving
Human Subjects (UCRIHS)

238 Administration Building

Michigan State University

East Lansing. Michigan 48824-1046

Dear Sirs:

Back in August of 1986. I applied to your committee for approval to conduct a
dissertation project involving human subjects. Following review. I was informed
that approval would be ganted after the following conditions were met:

1) Study approval by the Michigan Department of Corrections.

2) Revision of a cover letter on an attitudinal survey form to be sent to sub-
jects. so they would be absolutely clear on the point that their participation
is voluntary.

l have now met those stipulations. Enclosed you will ﬁnd a letter of study
approval from Mr. Terry Murphy. Chief of Research for the Michigan Department
of Corrections. Also enclosed. a copy of the cover letter I intend to send to the
subjects in question. Lastly. I have enclosed a copy of the study proposal and
variables dictionary which were approved by the MDOC.

I sincerely hope I have met the committee’s requirements for approval. If not.
please let me know and I will take whatever action you deem necessary.

Respectfully.

WW

Mr. Arthur Kirk
41 1 North Cedar #106

Lansing. Michigan 48912

 

Appendices

152
March 30. 1989

Dr. John Hudzik. Chairman

University Committee on Research Involving Human Subjects (UCRIHS)
206 Berkey Hall

Michigan State University

East Lansing. Michigan 48824- 1046

Dear Dr. Hudzik:

I am a doctoral student in the College of Education. Department of College and
University Administration. My committee chairman is Eldon Nonnamaker.

My dissertation is about 98% complete. and my hopes are to meet all the
requirements for my degee this Spring Term.

My dissertation does on a very limited basis involve the participation of human
subj ects. in the form of their responding to a questionnaire. On November 3. 1987
I received written permission from the UCRIHS to conduct my research. However.
this permission was extended for a period limited to one ealendar year.

I am now seeking renewed permission to conduct my study. Certain
procedural changes in regard to statistical applications have been made since
UCRIHS permission was last ganted. There have not. however. been any changes
in the procedures involving the participation of the human subjects. Nevertheless.
I have enclosed a current copy of the proposal so the changes of which I speak can
be noted. Further. I have enclosed a copy of the attitudinal survey instrument
which will be sent to the subjects. along with the cover letter which will
accompany it. The latter has not changed since original permission was ganted.

Please let me know if there is anything further I must do in order to be ganted
renewed permission to conduct my study.

Respectfully.
Wm...

Mr. Arthur Kirk
411 North Cedar. #106

Lansing. Michigan 48912

 

Appendices

APPENDIX H

Letters of Study Approval From UCRIHS

 

Appendices

18E!

MICHIGAN STATE UNIVERSITY

 

UNIVERSITY couurrm-z 0N eesmcu INVOLVING EAST LANSING ° MICHIGAN ' 43'3“.“
HUMAN susyrcrs IucanISI

m ADMINISTRATION BUILDING

(SI?) 355-2106

November 3, 1987

Mr. Arthur Kirk
411 North Cedar #106
Lansing, Michigan 48912

Dear Mr. Kirk:

Subject: Proposal Entitled, "Inmate Graduation from an Academic
College Program: How It Affects Their Rates of Criminal
Recidivism Upon Release From Prison"

UCRIHS' review of the above referenced project has now been completed. I am
pleased to advise that the rights and welfare of the human subjects appear to
be adequately protected and the Committee, therefore. approved this project
at its meeting on November 2, 1987.

You are reminded that UCRIHS approval is valid for one calendar year. If you
plan to continue this project beyond one year. please make provisions for
obtaining appropriate UCRIHS approval prior to November 2, 1988.

Any changes in procedures involving human subjects must be reviewed by the
UCRIHS prior to initiation of the change. UCRIHS must also be notified
promptly of any problems (unexpected side effects, complaints. etc.)
involving human subjects during the course of the work.

Thank you for bringing this project to our attention. If we can be of any
future help. please do not hesitate to let us know.

Sincerely.

Henry E. Bredeck, Ph.D.
Chairman , UCRIHS

HEB/jms

cc: Dr. Eldon Nonnamaker

MSU is an III/Moe Action/Equal Oppon-euv Immense

 

Appendices

APPENDIX I

Letter Of Final Study Approval From MCC

 

Appendices

155

(CTCY

montcolm community college

SIDNEY, MICMIGAN ‘8885 517 325-211

OFFICE OF THE PRESIDENT
October 30, 1987

Mr. Art Kirk
411 North Cedar #106
Lansing, MI 48912

Dear Mr. Kirk:

We, at Montcalm Community College, are pleased that you are studying
the relationship between criminal recidivism and inmate participation
in academic college programs.

It is my understanding that you have gathered most of the necessary
information needed from, or about. inmates in our C.O.P.E. (College
Opportunity -- Prison Extension) program.

This letter is intended to authorize final approval for your study.
If you need further assistance from MCC, please contact Mr. Dan
Herman.

Sincerely:

ﬂwc Aw

onald C. Burns, Ph.D.
President

DCB/pam

 

Appendices

APPENDIX J

MDOC Data Processing Service Request

 

Appendices

156

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MICHIGA CAM”
N DATA PROCESSING SERVICE REQUEST
DEPARTMENT
OF
DATA PROCESSING
CORRECTIONS PROJECT NUMBER
USER CONTROL NUMBER:
PROJECT TITLE: HQNIQAI'M ggmmm ”:1 QQMJGE EEALUATION
Requester Mama—Titwrphv . RN00“ 0..., 3/02/_87
Requester Wort: Unit: Research D1V1310“ Phone Number: 334-7857
(Division/Oureeu/lnstltutlon)
5mm: om summer [I] Yes E No Date enema “3908811919— 59 New Application
Reason tor Required om: C] cum. To Existing Appllcatlon,
D Discrepancy
Detailed Description oi services required. or change to be made: (it specillc lntormatlon is requested in report torm. indicate what
Items are needed on the report).
Print-out and tape of all persons who received a parole from January 1, 1980 and December 31.

1984 and who:
I. Served all or part of their incarceration at an institution in Ionia.

a. MTU Michigan Training Unit
b. RCF Riverside Correctional Facility
c. RHI Ionia Reformatory
2. Did not have an academic educational level equal to two years of college or above

upon parole.

Specifically

D 500 Allsthments

 

 

"rstlﬂcation. Benefits or reason tor request:
Non-priority Doctoral dissertation to evaluate Montcalm Community Program of

Ionia complex.

B ,ee Attachments

 

 

 

“Warm I0 was mu ‘1 pnepouie'

 

 

 

 

 

 

 

 

 

A
( questor tenet I/. I neuron». Authority) 1;
5 To Be Completed By Data Processing
; PRELIMINARY nevuew
:Asslsned To: Due Date:
ORecommended Action: PR Completion Dell:
PROJECT TIME ESTIHATES
Assigned To: Due Data:
Analyst Hours: Estimated Protect Completion Date:

 

Programmer Hours:

 

 

Computer Hours:

PROJECT COMPLETION

 

 

I Date Completed Authorized Signature

HOD” III leilCItOI Stink Otllhll Ill. 00. CD” to D.“ W Olvlsbll. ROM” 000 cm '0! A”! III“.

Appendices

11H?

MICHIGAN DEPARTMENT OF CORRECTIONS
RESEARCH DIVISION
M E M 0 R A N D U M

'10: Larry Walker
FROM: Terry Murphf, Chief DATE: July 17, 1987
SUBJECT: Service Request for Montcalm Community College Evaluation

Thank you for contacting me regarding further clarification on the
service request for the Montcalm Community College Evaluation. As we
discussed on the telephone on July 15, 1987, the study requires a print-
out and tape of all persons who received a parole from January 1, 1980

through December 31, 1984 (inclusive) and who:

I. Served all or part of their incarceration at an institution in
Ionia. Specifically,

a . M'I‘U
b. RCF
c. RMI

2. Did not have an academic educational level equal to 2 years of
college or above at commitment.

During our conversation you indicated that the following information
could be included on the tape and print-out without a significant delay

in obtaining the request:

a. Prisoner I.D.

b. I.D. Prefix

c. Ionia institution placement history (locations and dates)

d. Highest grade at entry

e. Date of Birth

f. Date Received

3. Race

h. Marital Status at Arrest

1. Occupation at Arrest

j. Parole Date (If multiple paroles in applicable term, include
all of the dates)

k. Assaultive Risk Classification

Finally, we discussed the possibility of including the subroutine that
generates a crime category field, based upon the MCL number field. I am
referring to the same crime category field that you produced on the CMIS
tapes for the Population Projection project.

I would like to thank you again for contacting me concerning our request
and appreciate your assistance in this matter. Please feel free to
contact me if further clarification is required. Thanks.

 

Appendices

APPENDIX K

Varlables chtionary

 

Appendices

158

VARIABLES DICTIONARY

The College Opporlunily Prison Extension (COPE) Study

by

Arthur Kirk

 

Appendices

159
GENERAL INSTRUCTIONS

Accuracy of the collected data is of utmost importance. If a coding question arises
which is not discussed verbally and / or addressed in this variables dictionary. write a note
in the space provided on the coding sheet explaining the problem and set that record aside.
The researcher will be available on a daily basis to resolve these problems.

Many inmates have more than one MDOC ﬁle because they served or are serving more
than one prison term in Michigan. When coding be absolutely certain you are extracting
data from the correct ﬁle (more commonly called the instant oﬂ'ense ﬁle). Following are
comments which will help you identify the correct ﬁle:

For purposes of this study, instant oﬂ'ense(s) is a reference to the oﬁ'ense(s) for which
the subjects were incarcerated and, in turn, paroled during the years 1980 through 1984.
The subjects could very well have been incarcerated prior to 1980, but the parole must
have occurred within that time frame. Some subjects will have been paroled more than
one time during these years. This study is designed to examine the ﬁrst parole in the time
frame for those subjects who comprise Group II. As for Group I subjects. the ﬁrst parole
in the time frame refers to the ﬁrst parole following graduation from the COPE Program.
The present prison commitment refers to the period of incarceration the subjects served
for the instant oﬂ'ense.

In the case of thou subjects who have more than one departmental ﬁle, you are
requested to check all of them for general background information. You are to use these
secondary sources simply to clarify and/or support data from the primary source. Just
be certain the data related to variables concerned with the instant oﬂ'ense are extracted
from the correct ﬁle.

In the case of subjects who escaped ﬁ'om community supervision, it is possﬁile the
record pulled for your inspection will be the escape ﬁle by itself. and the instant Oﬁ'ense
ﬁle will not be included. If that happens write a note to that eﬂ'ect and put the ﬁle aside.

You will note that missing data is accounted for in two ways: 1) 88=unlrnown, and 2)
99=unavailable. Unknown means the data cannot be retrieved because it does not exist.
Unavailable means the information does exists somewhere, but cannot be attracted from
the available records). and the time and expense involved in gathering the data makes it
impractical in regard to this study.

When you begin work on a ﬁle please follow these procedures:

1) Place a check next to the corresponding prison number shown on the master list

2) Put your initials in the upper right-hand corner of the coding sheet

3) Put the current date in the upper left-hand corner of the coding sheet

4) Place all completed ﬁles in one stack so they may be returned to the records
department

5) Placeallﬁleswhichpresmtacodingprobleminanotherstackso theresearcher
can inspect them daily

6) Always keep clearlylnmind the responslbﬂltyyou have tomsintainndmm
mum.

 

Appendices

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

DataSource:

Variable:

DataSource:

Variable:

Commentary:

DataSource:

160
PERSONAL SUBJECT DATA

Name

Provide the subject’s full name. Display the last name, then the ﬁrst name.
followed by the middle name or middle initial.

File jacket (box)/PSI/BIR

Last name, ﬁrst name, middle name or initial

Prison Number

This is the number assigned to the subject by the Michigan Department of
Corrections upon original entry into the Michigan prison system. Be certain to
include the letter prefix such as A, B, C, etc. The correct letter preﬁx for
purposes of this study is the one which relates to the subject’s instant Offense.
See the commentary associated with Variable #04 for a deﬁnition of instant
Offense.

File jacket (tab)/BIS

Letter designation followed by six digits

Social Security Number

Give subject’s social security number(s).
PSI/BIR

Nine digit code

Address

listsubject’slastlmownaddress. Iftherecordsdonotprovidethisinformation
give the address of his wife, parent(s), brother(s), sister(s). etc. In most instances
driswillbetheaddresstowhichthesubjectparoled. Thcaccuracyofthis
informationis especially criﬁcalinthecaseofGroupImembers because an
attitudinalsunreyistobcmailedtothosesubjectsaspanoftbesmdyplan.
PERissuedpriortOﬁrstpamie

Listthesubject’slastknownmailingaddress. Tlﬁscouldbetheaddressofhis
parents,brotherorsister,etc.

Note: The data related to these four variables will not be contained in the
computer ﬁles.

 

Appendices

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

Data Source:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

161
STUDY VARIABLES

01 - Subject Number

This numberwill be assigned by the researcher. A l preﬁx will relate tothe
subjects in Group I, a 2 preﬁx to members of Group II. The next three digits
will identify the individual subject in that group. For example, the subject
number 1127 will relate to the 127th subject in Group I. This code will serve as
the means of accessing individual records from the computerized data base
(primary key). Names, institutional numbers. social security numbers. and
addresses will not be contained within the general computer ﬁles.

Assigned by researcher

A four digit code

02 - Date of Birth

Based on subject’s date of birth. The reporting format is: mm/cki/yy.

File jacket (box)/PSI/BIR

1-12=month of birth. l-31=day of birth, 00-75=year of birth. 88=unknown.
99=unavailable

03 - Race

The categories are: black, white. hispanic, and other.

ﬁle jacket (box)/PSI/BIR

1=black. 2=white, 3=hispanic. 4=other, 88=unknown. 99=unavailable

04 - Date Of Prison Entry For Instant Offense(s)

This is the date of actual prison entry for the instant offense(s). It does um
include jail time, or time spent on bond while awaiting trial and/or case disposi-
tion. The instant Offense(s) is/are the criminal Offense(s) for which the subject
was incarceratedinprison. andfrom whichhe was grantedhis ﬁrstparoleduring
the years 1980 through 1984. For members of Group I this is the ﬁrst parole
following graduation from the COPE Program. For members of Group II it is
the ﬁrst parole during the time period 1980 through 1984.

File jacket/BIS

1-12=month of entry, l-31=day of entry, 00-86=year of entry (mm/ddlyy)

05 - Academic Educational Level At Time 0! Instant Offense(s)

Based on the number of years of academic education the subject actually com-
pleted at the time he committed the instant offense(s). A code of 14 for this
item would mean the subject had completed two years of college. If a subject
completed his GED, a grade level completion will not be expressed in the
record(s). In that case, list the functional level of the subject next to the 77 code
(e.g. - 77 9th). Also, in the case ofa special education subject, list in the com-
ments section of the coding sheet the impairment (in words) which required him
to be placed in that track. .

PSI/BIR

0=never attended school, 1-12=grades 1 through 12. 13-25=years of college
completed, 66=Sp.Ed.. 77=GED. 88=unknown. 99=unavailable

 

Appendices

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

DataSour'ce:

Variable:

Commentary:

Data Source:
Coding:

162

06 - Type 01' Instant Offense(s)

Based upon the following crime-type categories the subject’s instant Offense(s)
falls into: property, drug, violent. property and drug property and violent, drug
and violent, property, drug, and violent. See Appendix A for a lisu'ng of prop-
erty offenses, drug offenses, and violent offenses as deﬁned by Michigan stat-
utes.

PSI/BIR/Appendix A

1=property. kdrug, 3=violent, 4=property and drug, 5=property and violent.
6=drug and violent, 7=property, drug and violent, 88=unknown, 99=unavailable

07 - Date Of First Arrest
Thisisthedatethesubjectwasﬁrstanestedeitherasajuvenileoranadult. If
the date of ﬁrst arrest is not given, code this variable by giving subject’s age at
first arrest. Ifyou have toexercise this last option, putanoteinthe comments
section of the coding sheet to that effect.

PSI

1-12=month of arrest, l-31=day of arrest, 00-86=year of arrest. 88=unknown,
99=unavailable

08 - Prior Adult Felony Conviction(s)

Based on a listing of all felony convictions as an adult. It excludes the felony
conviction(s) related to the subject’s instant offense and any which might have
followed the instant offense. Often times the record will list the offense for
which the subject was convicted. but not specify whether it was a felony offense.
AlistingofthemostcommontypesofmisdemeanorswillbefoundinAppendix
B. Ifthe offense falls withinoneofthese categoriesitis amisdemeanoroffmse
and must not be included in this variable.

PSI/Appendix B

0-75=number of prior adult felony conviction(s), 77=yes, but count tmknown.
88=unknown, 99=unavailable

09 - Criminal Profile - Juvenile Property Offense(s)

This relates to study subjects who have one or more prior felony arrests for a
property offense as a juvenile. It excludes the felony Offense(s) related to the
subject’s instant Offense(s) and any which might have followed the instant
offense. SeeAppendifooralistingofpropertyoffenses, andAppendixB to
be certain you are dealing with a felony offense.

PSI/Appendix A/Appendix B

0-75=number of felony arrests for a property offense as a juvenile, 77=yes, but
count unknown, 88=unknown, 99=unavailable

10 - Criminal Profile - Juvenile Drug Offense(s)

This relatestostudy subjectswhohave one ormore prior felonyar'restsfora
drug offense as a juvenile. It excludes the felony offense(s) related to the
subject’s instant Offense(s) and any which might have followed the instant
offense. SeeAppendifooralistingofdrugoffenses, andAppendixB tobe
certain you are dealing with a felony offense.

PSI/Appendix AlAppendix B
0-75=number of felony anests for a drug offense as a juvenile, 77=yes, but count
wn wn ' 1

 

 

Appendices

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

163

ll - Criminal Profile - Juvenile Violent Offense(s)

This relates to study subjects who have one or more prior felony arrests for a
violent offense as a juvenile. It excludes the felony offense(s) related to the
subject’s instant offense(s) and any which might have followed the instant
offense. SeeAppendix Aforalistingofviolentoffenses, andAppendixB tobe
certain you are dealing with a felony offense.

PSI/Appendix A/Appendix B

0—75=number of felony anests for a violent offense as a juvenile. 77=yes, but
count unknown, 88=unknown, 99=unavailable

12 - Criminal Profile - Adult Property Offense(s)

This relates to study subjects who have one or more prior felony arrests for a
property offense as an adult. It excludes the felony offense(s) related to the
subject’s instant offense(s) and any which might have followed the instant
offense. See Appendix A for a listing of property offenses, and Appendix B to
be certain you are dealing with a felony offense.

PSI/Appendix A/Appendix B

0-75=number of felony arrests for a property offense as an adult. 77=yes, but
count unknown, 88=unknown, 99=unavailable

l3 - Criminal Profile - Adult Drug Offense(s)

This relates to study subjects who have one or more prior felony arrests for a
drug offense as an adult. It excludes the felony offense(s) related to the subject’s
instant offense(s) and any which might have followed the instant offense. See
Appendix Aforalisting ofdrugoffenses, andAppendixB tobecertainyou are
dealing with a felony offense.

PSI/Appendix A/Appendix B

O-75=number of felony anests for a drug offense as an adult. 77=yes, but count
unknown. 88=unknown, 99=tmavailable

l4 - Criminal Profile - Adult Violent Offense(s)

This relates to study subjects who have one or more prior felony arrests for a
violent oﬁense as an adult. it excludes the felony offense(s) related to the
subject’s instant offense(s) and any which might have followed the instant
offense. SeeAppendix Aforalistingofviolentoffenses, andAppendix B tobe
certain you are dealing with a felony offense.

PSI/Appendix AlAppendix B

0-75=number of felony anests for a violent offense as an adult, 77=yes, but
count unknown. 88=unknown, 99=unavailable

15 - In The Community At Least Three Years Prior To Prison Commitment
For Instant Offense(s)

Based upon whether ornot the subject was free ofajail or prison term for at
leastthree years priortohis presentprisoncommitrnent. It does not include
short-term jail detentions of ninety (90) days or less for minor offenses.

PSI

l=yes, 2=no, 88=unknown, 99=unavailable

 

Appendices

Variable:

Commentary:

DataSource:

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

164

16 - Marital Status At Time Of Instant Offense(s)

This pertains to the marital status of the subject at the time he committed the
instant offense(s).

PSI/BIR

1=single, 2=manied. 3=separated. 4=divorced, S=widowed, 88=unknown,
99=unavailable

l7 - Employment Status At Time Of Instant Offense(s)

This relates to the subject’s employment status at the time he committed the
instant offense(s). A coding reﬂecting no work history (0) means the subject
never held a job. whereas intermittently (3) means the subject worked from time
to time on a non-pattemed basis. A coding for laid-off (4) means the person had
a job and was awaiting recall, whereas unemployed (5) means the person was
without a job.

PSI/81R

0=no work history, 1=full-time. 2=part-time.3=interrnittently. 4=laid-off,
5=unemployed, 6=student, 88=unknown, 99=unavailable

18 - History Of Substance Abuse

This variable considers only these substances: cocaine. heroin, and alcohol.
Because of the questionable nature of addiction associated with the use of
marijuana. that substance is not to be considered in addressing this variable. We
are not necessarily looking at addiction in this item, rather for evidence the
subject used one or more of these substances in an abusive way. where abusive
means the use of the substance interfered with the subject’s social and/or occupa-
tional life one or more times within a thirty-day period of time. Also, the fact
the subject may have been under the inﬂuence of one or more of these substances
at the time of the instant offense(s) does not by itself call for a yes (1) response.
PSI/Psychological report

1=yes. 2=no, 88=unknown, 99=unavailable

19 - Evidence Of A Serious Physical Illness Or Disability At Time Of
Instant Offense(s)

For purposes of this study. a serious physical illness or disability is one which
interfered with the subject’s social and/or occupational life at least once every
thirty-days for a period of one or more years immediately prior to the commis-
sion of the instant offense(s). Please be certain your coding response relates to
the subject at the time he committed the instant offense(s)

PSI (Family Backgroundyl‘ranscase

1=yes. 2=no. 88=unknown, 99=unavailable

20 - Evidence Of A Serious Emotional Or Psychological Problem At Time

Of Instant Offense(s)

For purposes of this study, a serious emotional or psychological problem is one

which interfered with the subject’s social and/or occupational life at least once

every thirty-days for a period of one or more years immediately prior to the

commission of the instant offense(s). Please be certain your coding response

relates to the subject at the time he committed the instant offense(s).

PSI (Family Background)/I‘ranscase/Psychological report
MW

Appendices

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

165

21 - Upbringing

This variable attempts to identify the person or perstms who raised the subject
from childhood (below 18 years of age) to adult status. If none of the coding
categorieslistedforthis variable ﬁts the subject. make anotetothateffectinthe
comments section of the coding sheet and express the coding in words.

PSI

1=natural parent(s), 2=step-parent(s). 3=a natural parent and a step-parent.
4=relative(s). S=guardian(s). 6=foster parent(s). 7=members of an institution
such as an orphanage. 88=unknown, 99=unavailable

22 - Financial Status Of Upbringing Environment

Based onthe ﬁnancialincome ofthe people whoraisedthe subjectuptoadult
status (18 years old). Ifthe subject left the upbringing environment prior to 18
yearsofage,codethisitemuptothepoimofhisdeparnue.Ifnoneofthecoding
categories ﬁt the subject, the condition did not remain constant throughout the
upbringing period, orifhe was raisedin aninstitution. make anoteto thateffect
onthecodingslwetandexpressthecodinginwords.

PSI

l=wealthy (income of $50,000 per year or above), hﬁnancially stable (yearly
income between $15,000 and $50,000). 3=poor (yearly income below $15.00).
88=unknown, 99=unavailable

23 - Academic Educational Level Of Mother At Time Of Subject’s Instant
Offense(s)

Based on the number of years of academic education the subject’s biological
mother actually completed at the time of the subject’s instant offense(s). If the
subject’s mother completed her GED. list her functional level next to the 77 code
(e.g. - 77 10th) if it is provided in the record(s). If she was in special education.
list the impairment (in words) whichrequiredhertobeplacedinthattrack (if
given).

PSI (Family Background)

0=never attended school. 1-12=grades 1 through 12. l3-25=years in college.
66=Sp.Bd., 77=GED, 88=unknown, 99=unavailab1e

24 - Academic Educational Level Of Father At Time Of Subject’s Instant
Offense(s)

Based on the mrmber of years of academic education the subject’s biological
father actually completed at the time of the subject’s instant offense(s). If the
subject’s father completed his GED. list his functional level next to the 77 code
(e.g. -10th)ifitis providedinthe record(s). Ifhe wasin special education. list
the impairmentanwords) whichrequiredhim tobeplacedinthattrackﬁf
given).

PSI (Family Background)

0=never attended school. 1-12=grades 1 through 12. 13-25=years in college.
66=Sp.Ed.. 77=GED, 88=rmknown. 99=unavailable

25 - Family Emotional Support System At Time Of Subject’s Instant
Offense(s)
Indnsmstancethereferenceistothesubject simmediatefamily: wife. mother

 

 

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

168

Emotional support simply means one or more of these parties was obviously
concerned about the subject’s welfare. as reﬂected by a display of care and
concern. Thus. the subjecthadoneormore parties onwhom he couldcall for
help and assistance when one or more personal and/or ﬁnancial problems arose.
PSI (Family Background)

l=strong support from one or more parties. 2=some support ﬁom one or more
parties, 3=no support from any of the parties, 88=unknown, 99=unavailable

26 - Number of Non-Bondable Major Misconduct ReportsIn Prison For
Which The Subject Was Found Guilty During His Present Prison Conunit-
ment

Based only on non-bondable major misconduct reports in prison. for which the
subject was found guilty. The variable does not include bondable major
misconduct reports, or minor misconduct reports. If a major misconduct report
involved multiple charges, it is to be counted only one time. Time spent at a
correction center is considered part of the present prison commitment. but time
spent on parole is not part of the present prison commitment. See Appendix C
for a listing of non-bondable offenses. Please recall. the present prison commit-
ment is that which relates to the incarceration period the subject served for the
instant offense(s).

Major Misconduct Reports/PER (Institutional Adjusnnent)/Appendix C
0-85=number of non-bondable major misconduct reports. 88=unknown, 99=un-
available

27 - Date Of First Parole For Instant Offense(s)

Date subject was paroled to the free community from his present prison commit-
ment. IndrecaseofGroupIsubjectsthe ﬁrstparole fortheinstantoffenseisthe
ﬁrst parole following graduation from the COPE Program. For Group 11 sub-
jects. ﬁrst parole refers to the ﬁrst parole during the time period 1980 to 1984.
The reporting format is: mm/ddlyy.

Parole Board Action Notice/Parole Board Order For Parole/CMIS

1-12=month of parole. l-31=day of parole, 00-86=year of parole. 88=unknown,
99=unavailable

28 - Place Rom Which Paroled
TheMDOCfacihtyﬁomwhichdresubjectwasgrantedhisﬁrstparolefromids
present prison commitment. See Variable #27 for a deﬁnition of ﬁrst parole, and
Variable #26 for a deﬁnition of present prison commitment.

Parole Board Action Notice/CMIS

Give thename ofthe facility from which subject was grantedhis first parole
following present prison commitment. This will be reduced to a numeric code
by the researcher.

29 - Parole Placement
Based on the person(s) to whom the subject was initially paroled from his
presentprisoncommitment. Makeanoteonthecodingsheetifacombinationof
the coding categories ﬁt the subject.
PER issued prior to ﬁrst parole

1=wife, 2=parent(s) (biological or step). 3=brother(s) or sister(s) (biological or

 

 

 

Variable:

Commentary:

DataSource:

Variable:

Commentary:

Data Source:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:
Coding:

167

30 - Academic Educational Level Upon Parole From Present Prison Com-
mitment

Based on the number of years of academic education the subject had actually
completed at the time he was paroled from his present prison commitment. In
the case of a subject who worked toward or completed his GED during his
present prison commitment, a grade level of completion will not be expressed in
the record(s): rather. his progress will expressed as a functional level. Therefore.
ifyoucodethesubjectasa77(GED)onthisvariable.pleaselisthisfunctimal
level next to the 77 code (e.g. - 8th). If the subject was not involved in the
education program during his present prison commitment. then his academic
educational level for this variable will be identical to what it was at the time of
the instant offense(s) (Variable #05). Any problems coding this variable must
be brought to the attention of the researcher via a note on the coding sheet.

PER issued prior to parole

0=never attended school. l-12=grades 1 through 12. l3-25=years in college.
66=Sp.Ed., 77=GED, 88=unknown, 99=unavailable

31 - MDOC Assaultive Risk Classification At The Time Of Parole

This relatestotheMDOC assignedriskclassiﬁcationineffectatthetimethe
subject was paroled ﬁom his present prison commitment.

SCRS/PER issued prior to parole

1=very low. 2-—-low, 3=middle. 4=high. 5=very high. 88=unknown, 99=unavail-.
able

32 - Evidence Of A Serious Physical Illness Or Disability At Time Of Parole
From Present Prison Commitment

For purposes of this study, a serious physical illness or disability is one which
interfered with the subject’s social and/or occupational life in prison at least once
every thirty days for a period of one or more years immediately prior to parole
from hispresentprison commitment. The conditionmusthavebeenin effectat
the timeofparolefromhispresent prisoncommitmentinordertocodethis as 1
(yes).

PER issued prior to parole

1=yes. 2=no, 88=unknown, 99=unavai1able

33 - Evidence Of A Serious Emotional Or Psychological Problem At Time
Of Parole From Present Prison Comrnltment

For purposes of this study. a serious emotional or psychological problem is one
which interfered with the subject’s social and/or occupational life in prison at
least once every thirty-days for a period of one or more years immediately prior
to parole from his present prison commitment. The condition must have been in
effectatthe time ofparole from his presentprisoncommitrnentinordertocode
this as a 1 (yes).

PER issued prior to parole/Psychological Report issued prior to parole

1=yes. 2=no, 88=unknown, 99=unavailable

Note: Coders. do not go beyond this point. The balance of the variables will be
calculated or determined by the researcher.

 

Appendices

Variable:

Commentary:

DataSource:

Variable:

Commentary:

Data Source:
Coding:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

168

34 - Age At First Arrest
Basedontheactualageinyearsofthesubjectwhenhewas firstarrested. The
arrestneednothave resultedinaconviction. Also.thisarrestcouldverywell
have taken place when the subject was a juvenile.

Variable 07 (converted to subject’s age at the time) through the use of Variable
02

0-85=age in years, 88=unknown, 99=unavailable

3S - Age At Time Of Present Prison Commitment

Based on the subject’s age in years at the time he actually entered prison to
begin his present prison commitment. Jail time credits are not usedin orderto
make this determination.

Variable 04 (converted to subject’s age at the time) through the use of Variable
02

0-85=age in years, 88=unknown, 99=unavailable

36 - Year Of Graduation From COPE

This will be furnished by Montcalm Community College for Group 1 members.
Montcalm Community College records

0=Members of Group II, 0-86=Year of graduation. 88=unknown, 99=unavailable

37 - Age Upon Parole From Present Prison Commitment
Basedonthesubject’sageinyearsatthetimehewasparoledfromhispresent
prison commitment.

Variable 27 (converted to subject’s age at the time) through the use of Variable
02

0—85=age in years. 88=unknown, 99=unavailable

38 - Length Of Time Served For Instant Offense(s)

Based onthe actual numberofmonthsthe subjectservedinprisonduringhis
present prison commitment. It does not include jail time or time spent on
parole. but it does include time spent in a correction center.

Variable 27 (converted to a term of months) through the use of Variable 04
00-999=months of incarceration

39 - Academic Educational Attainment Level Increase During Present
Prison Commitment

This is a transformed variable obtained by subtracting the subject’s academic
educational level at the time of present prison commitment from his academic
educational level upon parole from present prison commitment. In the case of
subjects who worked toward or completed the GED. functional levels may be
used in place of years of education completed.

Variable 30 (converted to a term of years expressing academic attainment)
through the use of Variable 05

0-25=years of education. 66=Sp.Ed., 77=GED, 88=unknown, 99=unavai1able

40 - Criminal Recidivism - Property Offense(s)
The data for this variable will be obtained from the Law Enforcement Informa-
tion Network (LEIN) maintained by the Michigan State Police. and rs based

 

 

Data Source:
Coding:

Variable:

Commentary:

DataSource:

Variable:

Commentary:

DataSource:

169

offense during the two-year period of time following the ﬁrst parole from his
present prison commitment. The arrest need not have resulted in a conviction.
See the General Instructions and/or Variable #04 for a deﬁnition of ﬁrst parole
and present prison commitment. A distinction between a juvenile and adult
offense is not addressed by this variable.

LEIN/Appendix A/Appendix B

0-75=number of felony arrests for a property offense, 77=yes, but count un-
known, 88=unknown, 99=unavailable

41 - Criminal Recidivism - Drug Offense(s)

The data for this variable will be obtained from the Law Enforcement Informa-
tion Network (LEIN) maintained by the Michigan State Police, and is based
upon whether or not the subject was arrested for a felony class drug offense
during the two-year period of time following the ﬁrst parole from his present
prison commitment. The arrest need not have resulted in a conviction. See the
General Instructions and/or Variable #04 for a deﬁnition of ﬁrst parole and
present prison commitment. A distinction between a juvenile and adult offense
is not addressed by this variable.

LEIN/Appendix NAppendix B

0-75=number of felony arrests for a drug offense. 77=yes, but count unknown.
88=unknown, 99=unavailable

42 - Criminal Recidivism - Violent Offense(s)

The data for this variable will be obtained from the Law Enforcement Informa-
tion Network (LEIN) maintained by the Michigan State Police. and is based
upon whether or not the subject was arrested for a felony class violent offense
during the two-year period of time following the ﬁrst parole from his present
prison commitment. The arrest need not have resulted in a conviction. See the
General Instructions and/or Variable #04 for a deﬁnition of ﬁrst parole and
present prison commitment. A distinction between a juvenile and adult offense
is not addressed by this variable.
LEIN/Appendix A/Appendix B
0-75=number of felony arrests for a violent offense, 77=yes, but count unknown.
88=unknown, 99=unavailable

 

Appendices

170

APPENDICES
to the Variables Dictionary

 

Appendices

171

APPENDIX A
Listing Of Property/Drug/Vloleni Offenses

PROPERTY OFFENSES

Arson
(All except dwelling)

Burslm

Breaking and entering

Entering without breaking

Breaking and entering: or entering without breaking: buildings. tents. boats. railroad cars: entering
public buildings when expressly denied

Burglar’s tools, possession

Larceny

Larceny

Larceny from motor vehicle or trailer

Breaking and entering coin operated telephone

Larceny from vacant dwelling

Larceny from building

Larceny by conversion

Larceny by false personation

Larceny from libraries

Receiving or concealing stolen property (may be referred to as RCSP)

Auto Theft
U.D.A.A. (unlawfully driving away an automobile)
U.D.A.A. (without intent to steal)

Forgery - Uttering and Publishing (may be referred to as U&P)
Forgery of records and other instruments

Uttering and publishing

Forgery of notes

Forgery of bank bills and notes

Possession of counterfeit notes. etc.. with intent to utter same
Uttering counterfeit notes. etc.

Possession of counterfeit bank. state. or municipal bills or notes
Afﬁxing ﬁctitious signature

Counterfeiting and possession of coins

Certifying checks/insufﬁcient funds

Checks without accounts or insufﬁcient funds, usually over a certain amount

Embeulement
Allforms exceptwhennotedasunderacertain amount

 

Appendices

Fraud

172

Building connector funds-fraud, use
False pretenses with intent to defraud
Personal property. fraudulent disposition

Malicious Destruction
All forms except when noted as under a certain amount

Weapons

Carrying concealed weapons

Carrying weapon with unlawful intent
Weapons manufacture

Drugs

DRUG OFFENSES

Because of the State Police reporting format, it is sometimes difﬁcult to distinguish between misde-
meanors and felonies. The general rule is that illegal use or possession with intent to use is a misde-
meanor and the sale of. or possession with intent to sell is a felony. Unfortunately. the State Police
may only list Dangerous Drugs or Violation of Drug Law (VDL). The following procedures
should minimize any coding difﬁculties.

1)

2)

3)

Dangerous Drugs or Violation of Drug Law with the designation of use is
considered a misdemeanor. Illegal Use. Possession of Drug Paraphernalia are
also misdemeanors.

Dangerous Drugs or Violation of Drug Law with the designation of sale or
manufacture is considered a felony.

When the only information available is Dangerous Drugs. use the disposition
(if listed) to determine seriousness. A disposition of greater than 1 year is
considered a felony (e.g. . 2 years probation. 6 months jail and 5 years probation
are two examples of felony dispositions). All prison sentences are felonies (e.g. .
6 months to 2 years. 10 to 20 years). Sentences of jail terms only are
misdemeanors (e.g.. 6 months jail time. 30 days jail time).

4) When no disposition is available and you cannot determine use or sale. then

assume a felony when only designated as Dangerous Drugs or VDL.

 

Appendices

 

173

VIOLENT OFFENSES

Homicide

First degree murder
Second degree murder
Manslaughter
Attempted murder

Rape/CSC

Rape, forcible (do not include statutory)
Assault with intent to rape

Criminal sexual conduct lst. 2nd, and 3rd
Attempt or assault to commit CSC

Kidnapping
Kidnapping (all forms)

Assault

Felonious assault

Assault with intent to commit murder

Assault with intent of great bodily harm less than murder
Assault with intent to maim

Assault with intent to commit felony

Extortion

Robbery

Robbery armed - any weapon or indication thereof
Robbery unarmed

Bank. safe. and vault robbery

Assault to commit robbery armed

Assault to commit robbery unarmed

Attempted robbery

Larceny from person

Children

Child exposing with intent to injure
Cruelty to children

Torturing of children

Sex

Sodomy

Gross indecency between males

Gross indecency between females

Males under 15. debauching by females

Males under 15, debauching by males

Ravish abuse of female patient in an institution for the insane
Carnal knowledge - female ward

 

Appendices

174

Other Vlolent Offenses

Arson of a dwelling

Placement of explosives to damage or injure
Explosive device

Careless use of ﬁrearms to kill

 

Appendices

175

APPENDIX 3
Listing Of Common Misdemeanor Offenses

MOST COMMON MISDEMEANOR OFFENSES

In certain instances. a misdemeanor and felony are distinguished by a certain
dollar amount (e.g.. $50. $100).

1) Assault and battery (A&B)
2) Aggravated assault
3) Resisting ofﬁcer
4) Larceny under $50, $100
5) Anything under vs. over is a misdemeanor
6) Shoplifting
7) Petty theft
8) Petty larceny
9) Simple larceny

10) Jay riding

ll) Disorderly

12) Illegal entry

13) Checks NSF under $50

14) Motor vehicle. tampering

 

Appendices

1)

2)

3)

4)

8)

9)

10)

176

APPENDIX C
Ustlng Of Non-Bendable Offenses In Prlson

NON-BONDABLE MISCONDUCT OFFENSES IN PRISON

Escape or attempt to escape
Leaving or failing to return to lawful custody without authorization. Failure to return
within two hours after designated time from furlough or pass.

Felony
Committing any act which would be a felony if prosecuted under Michigan law.

Homicide
Causing the death of another person by any means.

Assault
Physical confrontation where one party is the victim and the other is the assailant. Injury
is not necessary. but contact is necessary.

Intimidating or threatening behavior
Words, actions, or other behavior expressing an intent to injure, which places another in
fear of being physically harmed or assaulted This includes attempted assault.

Sexual assault
Physical confrontation for sexual purposes, where one party is the victim and the other is
the assailant. Non-consensual physical contact for sexual purposes.

Fighting
Mutual physical confrontation, including a swing and miss. even when not done in anger.

Incite to riot or strike (includes participation)

Advocating or instigating actions which are intended to seriously endanger the physical
safety of the facility. persons, or property. or to disrupt operation of the facility by group
cessation of normal activity.

Dangerous contraband
Possession of weapon(s). explosives. acids, caustics. materials for incendiary devices,
escape materials. or “critical” tools.

Drug offenses
Use. possession, or sale of narcotics or amphetamines.

 

Appendices

APPENDIX L

Coding Sheet

 

Appendices

Name

177
Coding Sheet

 

Prison Number 8.8. Number

Address

#1

assesses

3
e

#11
#12
#13
#14
#15
#16
#17
#18

#19

#21

 

 

Subject Number

 

Date Of Birth _.Z_L_

Race

 

Date Of Prison Entry _L_l__
Academic Educational Level (Instant Offense(s))

Type Of Instant Offense(s)

 

DateOfFIrstArrest _L_L_

Prior Adult Felony Conviction(s)

 

 

Criminal Profile - J uvenile Property

 

Criminal Profile - Juvenlle Drug

Crlmlnal Profile - Juvenile Vlolent

 

 

Criminal Profile - Adult Property

 

Criminal Profile - Adult Drug

 

Criminal Profile - Adult Violent

 

In Community Three Years

 

Marital Status (Instant Offense)

 

Employment Status (Instant Offense)

 

History Of Substance Abuse
Serious Physical Problem (Instant Offense) —
Serious Emotional Problem (Instant Offense) ' '

Upbringing

 

Appendices

 

3

§

§ 2?

#31

#32.

§§§

#39
#40
#41

#42

Notes:

178
Financial Status

 

Mother’s Educational Level

 

Father’s Educational Level

 

Emotional Support System (Family)

 

Non-Bondable Misconduct Reports (Guilty)
Date Of First Parole _L_L

Place Of Parole

 

 

Parole Placement

Academic Level At Parole

 

MDOC Risk Classification

 

Serious Physical Probiem.(Parole)

 

 

Serious Emotional Problem (Parole)

 

Age At First Arrest

 

Age At Time Of Prison Commitment

Year Of Graduation From C.O.P.E. ——

 

Age Upon Parole

Length Of Time Served (Instant Offense) —

 

Academic Level Increase While In Prison

Criminal Recidivism - Property Offense(s) ‘-

 

Criminal Recidivism - Drug Offense(s)

Criminal Recidivism - Violent Offense(s) ‘—

 

 

 

 

 

 

Appendices

APPENDIX M

SPSS Programs

 

Appendices

179

0010FIIE TYPE NESTED FILE=MACOCOLL RECORD=RECIYPE 5-6
0020CASE=PRSNRID 7- 13(A) DUPLICATE=CASE

0030RECORD TYPE 01 / ‘MASTER RECORD
0030DATA LIST / RECLEN 1-4

0040 BIRTHYR 86-87

0050 BIRTHMO 82-83

0060 BIRTHDA 84-85 .

0070 RECEIVYR 111-112

0080 RECEIVMO l 07- l 08

0090 RECEIVDA. 109- l l 0

0 100 SEX 1 14(A)
0 l 10 RACE 1 15(A)
0120RECORD TYPE 02 /"ID RECORD

0130DATA LIST I‘MARRYARR 92(A)

0 140 HIGRADE 95-96

0 l 50 OCCARR l 05- l 07(A)

0 l 60 ASSLTRSK 2 1 5

0 1 70RECORD TYPE 14 I‘TRANSIT RECORD
0180DATA LIST /TRANFROM 36-38(A)

0190 TRANTO 39-41(A)

0200 TRANYR 50-51

02 l 0 TRANMO 46-47

0220 TRANDA 48-49

0230 MOVECODE 52-53

0240EN'D FILE TYPE

0250RECODE RECLEN (SYSMIS=0)

026OSELECT IF (RECLEN GT 0)

0270WRITE OUI'FILE=MACOCOL2 TABLE
0280FREQUENCIES VARIABLES=RECEIVYR HIGRADE T'RANFROM TRANTO
0290TRANYR MOVECODE

 

Appendices

160

FILE HANDLE MACOCOL2 / NAME = ’MACOCOL2 DATA A‘
FILE HANDLE MCOLGRP2 / NAME = ’MACOCOL2 GROUP n'
/ 1 A 1 (A)
PRSNRID 2-7
BIRTHMO 12- 13
BIRTT-IDA 1 4- 1 5
BIRTHYR 18- 19
RACE 23 (A)
MARRYARR 25 (A)
HIGRADE 27-28
OCCARR 30-32 (A)
ASSLTRSK 34
TRANFROM 36—38 (A)
TRANTO 40-42 (A)
TRANMO 44-45
TRANDA 47-48
TRANYR 50-5 1
MOVECODE 53-54
SELECT IF (MOVECODE E9 61 AND RANGE(TRANYR.80.84))
COMPUTE X=(TRUNC((YRMODA(BIRIT-IYR.BIRI‘I-IMO.BIRI‘HDA) -
YRMODAI'I'RANYRTRANMOJ'RANDAD/36525D
SORT CASES BY PRSNRID A TRANYR
AGGREAGATE OUTFILE=‘
/BREAK=PRSNRID A
/PAROLYR=FIRST(TRANYR)
/PAROLMO=FIRSTT1‘RANMO)
/PAROLDA=FIRST(rRANDA)
/MARRIED=FIRST(MARRYARR)
/EMPLOYED=FIRST(OCCARR)
/PAROLAGE=FIRST(X)
/RACE2=FIRST(RACE)
/ARISK=FIRST(ASSLTRSK)
/EDLEVEL=FIRST(HIGRADE)
/PREF1x=FIRST(A)
/IDNUMBER=FIRST(PRSNRID)
FORMATS PAROLYRTO PAROLDA (F2.0) MARRIED (A1) EMPLOYED (A3) PAROLAGE
(F2.0) RACE2 (A1) ARISK (F1.0) EDLEVEL (F2.0) PREFIX (A1) IDNUMBER (F6.0)
SORT CASES BY PAROLAGE EMPLOYED EDLEVEL RACE ARISK MARRIED
WRITE OUTFILE=MCOLGRP2 TABLE
LISTVARIABLES=PREFIX IDNUMBERPAROLAGE EMPOYED EDLEVEL RACEARISK
MARRIED
FREQUENCIES VARIABLES=PAROLAGE TO MARRIED

 

Appendices

APPENDIX N

Aiiitudinal Survey Instrument

 

Appendices

181

Mr. Robert Subject
12 12 Street
Lansing. Michigan 48912

Dear Mr. Subject:

I am a doctoral candidate at Michigan State University in the area of College
and University Administration. To satisfy my dissertation requirements I am doing
a study of the C.O.P.E. Program at Montealm Community College. I have the
approval of the administrators at Michigan State University. Montcalm Community
College. and the Michigan Department of Corrections to conduct this study.

The speciﬁc purpose of the study is to determine if graduation from the
C.O.P.E. Program provides a positive (reducing) inﬂuence on the criminal
recidivism (repeat crime) rates of its participants.

Inasmuch as you are a graduate of the C.O.P.E. Program. I wish to enlist your
help in one part of the study. That part serves to examine the reason(s)
individuals like you decided to enter and. in turn. graduate from the C.O.P.E.
Program. It alsoseeks to examine any attitudinal change that may have taken
place in regard to personal behavior. once you were released from prison.

Your participation in this study is WI If you decide to help
me in my research. I can assure you that your responses to the enclosed
questionnaire will be seen only by me. and reporting will be done in collective
(group) form only. No one other than I will view your answers to the survey
questions. and once the survey responses are put into group form. the
instruments will be destroyed.

I urge you to help me. and perhaps give a hand to those who are still
incarcerated. Please use the stamped. self-addressed envelope to return your
survey form. 1204101 put any identiﬁer such as your name on this form. and do
not place your name or return address on the envelope provided. That way. even I
will not know the identity of the person who returned the questionnaire. Thank
you!

Respectfully.
WW

Arthur Kirk

P.O. Box 637 1

Michigan State University

East Lansing. Michigan 48826

 

Appendices

182
ATTITUDINAL SURVEY INSTRUMENT - C.O.P.E. GRADUATES

Mm: Beside each of the statements. please indicate with an X whether you
Strongly Agree (SA) . Agree (A). Disagree (D). Strongly Disagree (SD). or are Undecided
(U).

SA A D SD U
l. Ientered the C.O.P.E. Program to impress the Parole Board ........... 0 0 0 0 0
2. I entered the C.O.P.E. Program to kill time ..................................... () () 0 () ()
3. I entered the C.O.P.E. Program to improve myself .......................... () 0 () O ()
4. I ﬁrst entered the C.O.P.E. Program to impress the Parole Board and/or
to kill time. but my interests later turned to self-improvement ......... O 0 0 O O
5. My C.O.P.E. studies helped me to better understand myself ............ () () 0 0 ()
6. My C.O.P.E. studies helped me to better deal with
my incarceration ............................................................................. () () 0 O O

7. My C.O.P.E. studies helped me to better understand other people 0 0 0 () O

8. Participation in the C.O.P.E. Program helped me to
more clearly deﬁne my personal goals ............................................ () 0 0 () 0

9. The C.O.P.E. Program helped me become a better person ............... () 0 0 0 O

10. I decided not to become involved in further criminal activity
prior to entering the C.O.P.E. Program ........................................... () 0 0 0 0

11. I decided not to become involved in further criminal activity
while in the C.O.P.E. Program ........................................................ () () 0 0 0

12. I decided not to become involved in further criminal activity
after I graduated from the C.O.P.E. Program .................................. 0 0 0 0 O

13. I never decided against being further involved in criminal activity .. () 0 0 O O

14. My graduation from the C.O.P.E. Program helped me
gain release from prison .................................................................. () () 0 () 0

15. My graduation from the C.O.P.E. Program helped me
better adjust to parole and/or release from prison ............................ 0 0 0 O 0

 

Appendices

16.

1‘7.

18.

19.

21.

22.

23.

25.

183

SA A
My graduation from the C.O.P.E. Program helped me
get work once I was released from prison ........................................ () ()
Graduation from the C.O.P.E. Program has
caused me to view myself in a more positive way ........................... () ()
Graduation from the C.O.P.E. Program has
given me more self-conﬁdence ........................................................ () 0
I have or will continue my college education even further ............... () ()
I think all prisons should have a college program like C.O.P.E. ...... () ()
If all prisons had a program like C.O.P.E., there would be
less trouble among the inmates ........................................................ () ()
If all prisons had a program like C.O.P.E., many inmates would be
able to improve their chance of successfully completing parole ....... 0 ()
If all prisons had a program like C.O.P.E., there would be less
repeat crime .................................................................................... () ()
Without the beneﬁts of the C.O.P.E. Program. I probably would
have continued with my same old life patterns ................................ () ()

Without the beneﬁts of the C.O.P.E. Program, I probably would
be back in prison ............................................................................. 0

DSDU

0

()

O

()

0

()

()

O

()

()
0

0

 

Appendices

 

BIBLIOGRAPHY

 

Adams. 8. (1968). Winners. Los Angles. California:
University of California.

Adams. 8. (1970). I
Washington. D. C.: Department of Corrections.

Adams. 8. (1973. November). Higher learning behind bars. Change. 5(9). 45-50. (ERIC
Document Reproduction Service No. EJ 086 300).

Adams. 8. (1973). The P100 project. In A.R. Roberts (Ed.). Beadinemmucation
(pp. 279-299). Springfield. Illinois: Charles C. Thomas.

Adams. 8. (1975). W. Washington. D.C.:
American University Law Institute.

Adams. 8.. Connolly. J.J. (1971. March). Role of junior colleges in the prison commu-
nity- WW 11(6). 92- 96.

Adwell. S.’l‘.. Wolford. 8.1. (1983). The development and growth of standards for correctional
education. W.Mﬁh 123- 125.

Allen. R.A. (1974. October). Inmates go to college. W.
ﬂ(2).146-149.

Ambrussi. W.A.. Jr. (1984). Compulsive maladaptive behavior as a cause for criminal
recidMsm. WW. 55(06). SEGA. PP1640. (University
Microﬁlms No. ADG84-18845. 0000).

American Correctional Association. (1983). WWW...
College Park. Maryland: Author.

American Psychological Association. (1983). W
W. (3rd ed.). Washington. D.C.: American Psychological

 

Association.

Anderson. 3.16.. Field. T.G. (1979). Study of selected demographic-variables and recidivism
with adult male probationers. WWW. 16(2). 11 1414.

Anderson. DB. (1981). The relationship between correctional education and parole
success. WEED/4). 13-25.

Arizona Department of Corrections. (1976. January). W (Research
Report No. 76-1). Arizona: Author.

Artist. 11.. Oolton. RR. (1970). W. New York. New York: Barnes & Nobel
Books.

 

Bibliography 134

185
atone. N.A. (1982). Survey of research ﬁndings related to oﬂ'ender reintegration. Indian
WW 19(2). 91-99.
Astonc. N.A. (1982). What helps rehabilitation? A survey of research findings. W
: 1.2§(2).109-120.
Atwood. 8.1!. (1970). Some other institutions. In R.M. Smith. G.F. Aker. 81 J .R. Kidd
(Eda-I. Handhmkntadnluducatian (PP- I. New York Macmillan.

Austin. ILL. (1976-77. Winter). Oﬁ’ense history and recidivism. W3.
1(2). 209'226.

Avio. 8.1.. (1975). Recidivism in economic model of crime. W. 13(3) . 450-456.

Ayers. J.D.(1976. August). 0 ...» : .
Wes. Unpublished manuscript, University of Victoria.

Ayers. 1).. and others. (1990). on ...; _- =
m. (ERIC Document Reproduction Service No. ED 201 798).

Babbie.E . (1986). Wed). Belmont. California: Wadsworth
Publishing Company.

Bailey. W.C. (1966). Correctional outcomes: An evaluation of 100 reports. M
We. 51 157-160 Baker J E (1973)
Social education in a penitentiary. In A.R. Roberts (Ed) W
Mn (pp. 240-250). Springﬁeld. Illinois: Charles C. Thomas.

Baker.K.. Irwin.J..Leonard. 1).. Haberfeld.9.. &.Seashore. 1141973). Sum

 

 

 

 

Marshall Kaplan. Carts. and Kahn.

Baker. W. (1972. Spring. From the community college a response. W
W. 5(4). 96-98.

Beck. JJ... Boﬁnan. P.B. (1978). Time served and release performance. W
W. New York. New York: Wiley. 745-749.

Becker. D.C. (1989). Corrections challenge educators. WW. 11(4).
336-340.

Bell. IL. ct IL (1979). Watts. Washington. 110-:
National Institute of Law Enforcement and Criminal Justice.

Bends. B.B. (1979). Criminal recidivism: horn adolescence to adulthood. Dissertation

WW. 59(07). SEGA. PP4234. (University Microﬁlms No. ADG79-
27156. 0000).

Bcrccochu. J.E.. Jaman. 13.11.. at Jones. WA. (1973). W
W. Sacaramento. California: Department of Corrections.

Berry. ILL. (1985). Shape-up: The eﬁ'ects of a prison aversion program on recidivism and

family dynamics (deterrence model). WWW. 46(08).
SEGA. PP2449. (University Microﬁlms No. ADG85-23372. 8602).

Bishop. TX. (1981). The eﬂ'ects of the Recorder’s Court Restitution Program on the inmate
population. criminal recidivism and educational and sociological status of oﬂ'enders.

. 52(11). SECA. PP4942. (University Microﬁlrns
No. ADG82-09270.0000).

 

Bibliography

 

Black. LR. (1975. April). '
Yellow Springs. Ohio: Union for Experimenting Colleges and Universities. (ERIC
Document Reproduction Service No. ED 107 603).

Blackburn. LS. (1979). The relationship between recidivismand participation in a
community college associate of arts degree program for incarcerated offenders.

, 49(10). SECA. PP5294. (University Microﬂlrns
No. ADG80-O7936. 0000).

Blackburn. 1".S. (1981). The relationship between recidivism and participation in a

community college program for incarcerated offenders. W};
m. 32(3). 24'30-

Blackwell. P.H.. Jr. (1973). Higher education in prison: A study of the impact of college
education upon selected inmates of Draper Correctional Center. Elmore. Alabama.

WWW. 34(06). SECA. PP3080. (University Microﬁlms
No. ADG73-31615. 0000).

Blake. S. (1972. Spring. In the community college. potentially. W
W. 5(4). 103-105.

Blalock. 8.1L. Jr. (1970). WWW. Englewood Cliﬁ's. New
Jersey: Prentice-Hall.

Blurnstein. A.. Cohen. J. (1974)-
mm Pittsburg. Pennsylvania: Carnegie-Mellon University.

Blunrstein. A.. Larson. RC. (1971. July). Problems in measuring and modeling recidi-
vism. WWW. 6(2). 124-132.

Boas. me. (1976). An evaluative study of Project Outreach to inmates: A higher education
progam oﬂ'ered by the Universtiy of Virginia at three of Virginia's correctional institu-

tions. WW. 31(08). SEGA. PP4887. mmversity M1010-
ﬁlms No. ADG77-02760. 0000).

Books behind bars. (1970. December). W. 5(10). 28-29.

Bonifaeio. P. (1975). W. (ERIC Document Reproduction
Service No. ED 147 734).

Bosarge. B.B. (1993). Representative Conyers urges support for $25 million per year
Correctional Education Act. com. “(22). 1-4.

Boudouris. J. (1984). Recidivism as a process. WWW
mo 5(3). 41-51-

Boyd. ND. (1973. October). The prison and educational possibility. W.
22(4). 132-136.

Britton. 0.11. a Glass. J.C. (1974). Adult education behind bars: A new perspective.
WWW. 26(2). 6-7.

Buikhuis. W.. Boekstra. ILA. (1974). Factors related to recidivism. W
melon. 14(1). 63-69.

Bureau of Prisons. (1981. November 12). W
mm. Washington. D.C.: U.S. Government Printing omce.

Business is booming. (1987. January 11). W. p. 15.

 

Bibliography

187
Buttrm. J4... Men. ILA-(1977) WWW

Document Reproduction Service No. ED 159 216). Philadelphia. Pennsylvania: Re-
search for Better Schools. Inc.

Campbell. D.‘l‘.. Stanley. J.C. (1966). Mm ~ . -
magnum. Chicago. Illinois. Rand McNalley & Company.

Campbell. J.. Jr. (1973). An experience in group counseling. InA.R. Roberts (Ed.). Readings
WW. (pp. 271-278). Springﬁeld. Illinois: Charles C. Thomas.

Carr. T.S. (1980). The effects of crowding on recidivism. cardiovascular deaths. and infraction

rates in a large prison system. WW. .4200). SECB.
PP3931. (University Microﬁlms No. ADG81-08348. 0000).

Carty. J.W.. Jr. (1976). Education goes to prison in West Virginia. Wm.
39(4). 14-15. (ERIC Document Reproduction Service No. E1 139 160).

Casselman. J .. Blake. S. (1972. Spring. Community college of Philadelphia at Holmesburg
prison- rionmalmammnnexeienmentinmileaﬁen. 5(4). 98-102.

Castro. D.A. (1977). From the joint to the campus: Ex-oﬁ'enders in transition. mm
W. 1(3). 251-255. (ERIC Document Reproduction Service No. El 161
841).

Cerullo. 03. (1978). A project for community-based and community-sponsored higher
education of inmates who have completed two years of college since their incarcera-

tion. WWW. 3.902). SECB. PP61 12.
(University Microﬁlms N0. ADG79-l 1951. 0000).

Catillas. ILL. Russo. J .A. (1976). The funding of prison education programs: An approach.

W63). 52-56. (ERIC Document Reproduction Service No.
El 164 599).

ChamberofCommerce. (1972). 0.: ~.= .
Washington. D. C.: Chamber of Commerce of the United States.

Chappell. A.D. (1977). The family home evening program at the Utah State Prison: Its

history and relationship to recidivism. W. 380 l).
SECA. PP6528. (University Microﬁlms No. ADG78-07084. 0000).

Ciom. 1". (1981). Teaching college humanities courses in prison. W
W. 5(1). 49- 59. (ERIC Document

 

 

 

Reproduction Service No. EJ 252 444).

Clements. C.. Jr. Self. P. (1980). Credit-free programs in correctional centers.
W. 8(2). 50-53. (ERIC Document Reproduction Service No. EJ 225 942).

Ciendenen. ILJ.. Ellington. J.R.. a Severson. ILJ. (1979). Project Newgate: The ﬁrst
ﬁve years- Want. 25(1). 55-64.

Community Corrections Resource Programs. Inc. (1976. August 12). 99W
WWI: AnnArbor Michigan: CCRP Inc

Coombs. “(1965. June). . u :

(Research Review No.20). State of
Washington: Department of Institutions.

 

 

Bibliography

188

Corcoran. 1". (1985). Pedagoy in prison: Teaching in maximum security institutions.
WWII. 35(1). 49-58. (ERIC Document Reproduction Service No.
EJ 31 l 336).

Correctional Education Association. (1978). W
W. Vienna. Illinois: Author.

Cortwrlght. R.W. (1973. January). Guidelines for adult correctional education. Adult
leadership. 21(7). 224-226.

Coman. J .W.. Forster. W. (1977). Higher-education of prisoners. 99mm. 19(4) . 83-84.

Craig. J .M. (1983). A study of inmate participation in college-level academic programs and
recidivism

.Dnsertatiormsiraetsinimtienal. 43(05). SECA. PP1587. (University
Microﬁlms No. moss-22191. 0000).

Cullen. l".'l‘. 8: Gilbert . LE. (1982). W. Cincinnati. Ohio: Ander-
son Publishing Company.

Curry. E.A. (1979). Recidivism in the mentally ill criminal oﬁ'ender: A comparative study

in forensic psychiatry. W. 49(06). SECab. PP2607.
(University Microﬁlms N0. ADG79-271 13. 0000).

Curry. WJ. (1974). Academic and motivational characteristics of prison-inmates enrolled
in the community college program at North Carolina's correctional instititutions: An
crploratory study. WW. 3502). SEGA. PP7593.
(University Microﬁlms No. ADG75- 13138. 0000).

David. K. D. (1978). Offender education in the American correctional system: An
historical perSpective. W. 5.4. 30-38.

Davis. RA. (1973). Editorial. new. 53(2).

DeJoie. C.H. (1979). The university and social education for prisoners.
Rm. 30(4). 242-252. (ERIC Document Reproduction Service No. EJ 215 281).

Dell’Apa.1|‘.(1973). . - : . . .
Boulder. Colorado. Western Interstate Commission for Higher Education.

Department of Housing and Urban Development (HUD). (1976). W. Washington.
D.C.: U.S.Government Printing Oﬁce.

Devine. R.P.. Falk. L.L. (1972). .. = ' - =
W. Morristown. New Jersey: General Learning Press.
Donnelly. P.A. (1980). The psychological and sociological factors that predict recidivism of

criminal oiienders. WWW. AMOS). SECB.PP1970.
(University Microﬁlms No. ADG80-24091. 0000).

Drlsoon. EM. (1971). . ~

W. (ERIC Document Reproduction Service No. ED 060
801).

Education Commission of the States. Correctional Education Project. (1975). 11;:
WW. Denver. Colorado: Author.

Education Commission of the States. Correctional Education Project. (1976).
WWW (Report No. 76). Denver. Colorado:
Author.

 

 

 

 

Bibliography

189
Education Commission of the States. Correctional Education Project. (1976). W

”IQI.‘ 3Q. 'O‘Hlﬂl’QCo‘O. 0 0'3. |v£3 6‘ C"3.C.3 0.3 09“.. ‘A‘Oﬂ

WW (Report No.8 1). Denver. Colorado: Author.

Emmert. 3.3. (1976). .15 i c a = ‘

W. Washington D C American Association of
Community and Junior Colleges.

Enocksson. K. (1980). Correctional programs: A review of the value of education and

training in penal institutions. WWW
lation 5(1). 5-18.

mm Is. NM. To (1977’s \f ‘ V 7 3 .
W. New York. New York: Columbia University Press.

Farrington. D.P.. Nuttall. C.P. (1983). ‘Overcrowding and recidivism” : A response to
Gaee's comment. MW. 11(3). 269-271.

Feldman. Sun (1975). mmmeﬂmdeumeanonalandedueaﬁenalmamsm
- ~ - - - - - -~ ...:Washington.DC American

 

 

 

Association of Community and Junior Colleges.

miller. RE. new“. J.. a Smith. RC. (1974).mmateassindem;anmzesﬁaaﬁenetformerht
WWW (Report No. 21).

Columbia. South Carolina: University of South Carolina.

Focal. D. (1979). WWW Cincinnati.
Ohio: Anderson Publishing Company.

Forst. B.. Rhodes. 17.. Dimm. J.. Gelman. A.. & Mullin. B. (1983). Targeting Federal
resources on recidivists: An empirical view. W. 46(2). 10-20

Foust. J.M.. Reagen. M.V.. a Stoughton. D11. (1978). School behind bars - descriptive
overview of correctional education in American prison system. W.
n‘1)e 225-225.

Franklin. B.B. (1979). Rehabilitating prison education. Change. 11(8). 18-21. (ERIC
Document Reproduction Service No. El 2 14 520).

Friedman. A.Ii. (1978). Community college faculty members’ attitudes toward correctional
instruction at correctional institutions. WWW. 39(08).
SECA. PP4679. (University Microﬁlms No. ADG79-04277. 0000).

Friedman. C.J.. Mann. F. (1976). Recidivism- fallacy of prediction. W
. ‘ i = . Own In 0-4 ”29(2) 153-164.

hits. B.. Lewis. MN. (1975). Prison education and rehabilitation - illusion or reality.
WW. 21(3). 291-292.

Furtado. A.. J ohnson. D. (1980). Education and rehabilitation in a prison setting. Journal
WW. 4(3). 247-273.

Gees. 0.6. (1983). Farrington and Nuttal’s “overcrowding and recidivism“. W
Wee. 1.1. 265-267.

Gaither. C.C. (1983). TDC’s Junior College Program: Promising results. Cumming
We 45(4)s 58‘84.

George. P.S.. and others. (1980). The college program in the Georgia State Prison.
. 8(2). 21-25. (ERIC Document Reproduction Service No.
El 225 938).

 

 

Bibliography

190

Gerts. H.G.. a Talarico. 8.1L (1977). Problems of reliability and validity in criminal
Justice research. WW. 5. 217-224.

Gest. T. (1984. April 23). Bulging prisons curbing crirne-or wasting lives? W

 
 

m: .tm . New York: Bobbs-Merrill.

Glaser. D. (1966. March/April). The eﬁ‘ectiveness of correctional education. American
JellmaleLCenection. 28(2). 4—9.

Glaser. D. (1974). Remedies for the key deﬁciency in criminal justice evaluation research.
W. 1.1. 144-154.

Glaser. D. (1975). Achieving better questions: A half-century’s progression in correctional
research. Eeileralmllation. 3.9. 3-9.

Click. m. a New. VN. (1976. October). WWW-
Sacramento: California Youth Authority.

Goldfarb. RJ... Singer. LR. (1977). W. New York. New York: Simon and
Schuster.

Good. C.V. (Ed.) (1973). W (3rd ed.). New York: McGraw-Hill.

Greenﬁeld. 11.x. (1970. April). The college goes to prison. W. 42(7).
17-20.

Griswold. D.B. (1977). Perception of legitimate opportunities. legal self-concept.
adherence to focal concerns. identiﬁcation with criminal/non-crirninal referfence
maps. and recidivism A multivariate analysis. Disseﬂaﬁenhbeiraetslniemaﬁenal.
38(09). SEGA. PP5734. (University Microfilms No. ADG78-01480.0000).

Griswold. D.B. (1978). A comparison of recidivism measures. W.
ﬁe 247’252.

Guild. LP.T. (1977). The role of Michigan’s community colleges in the education of

incarcerated adults. W. 3.80 l). SECA. PP6467.
(University Microﬁlms No. ADG78-04715. 0000).

Haber. GM. (1983). The realization of potential by Lorton. D.C. inmates with UDC college
education compared to those without UDC education. W
. 2(3-4). 37-55. (ERIC Document Reproduction
Service No. EJ 289 363).

Harris. CM” Moitra. S.D. (1978. July). Improved techniques for the measurement of
recidivism. National Council on Crime and Delinquency. W

Cnmeandnelimuem.15(2). 194-213.

Bessel]. N.B. (1978). The perceptions of selected groups of incarcerated and
non-incarcerated junior college students regarding certain teacher traits and teaching

effectiveness. W. 3.90 1). SEGA. PP6488. (University
Microﬁlms No. ADG79-09662. 0000).

Bavilan. J .J . (1962). A study of the diﬁ'erences between prison college graduates and the
total released inmate population on recidivism by risk category. Dissertation

Abstractslntematienal. 43(04). SEGA. PP1304. (University Microﬁlms No. ADG82-
21732. 0000).

Hawkins. G. (1976). W. Chicago and London: University of
Chicago Press.

 

Bibliography

 
 

Herron. R.F.H.. and others. (1973). \ = z . - ~
WW. Oﬁlce of Economic opportunity. Washington. D. C. (Report No.
FGK 65995). (ERIC Document Reproduction Service No. ED 138 751).

Herron. R... Muir. J. (1974).

W. Oﬁice of Economic Opportunity. Washington. D.C.
(Report No. EGK659995) (ERIC Document Reproduction Service No. ED 137 574).

Herschler. E. (1976. Spring. Education: Weapon against crime. Compact. 1Q(2). 4-6.
Higher education behind bars. (1972. April). Appalachia. 5(5). 32-34.

Hlnck. J .H. (1975). Diﬁ'erences in selected variables of prison inmates who are completers

and non-completers of a three-year college program. Dissenaﬁeaohsimetsintema:
um. 3.603). SECA. PP1836. (University Microﬁlms N0. ADG75-19447. 0000).

MB. (Ed..) (1981). MW. (ERIC Document
Reproduction Service No. ED 212 199).

Hoﬁman. P.B. (1983). Screening for risk: A revised salient factor score. W
neutralise. 11(6). 539-547.

Hoﬂman. P.B.. & Beck. J.L (1984). Burnout--Age at release from prison and recidivism.
Jeumaletcnnnnalilustieedz. 617-623.

Hoffman. P.B.. & Beck. J.L. (1985). Recidivism among released federal prisoners.
Wham. 12(4). 501-507.

Hoﬁman. P.B.. & Stone-Heierhoefer. B. (1980). Reporting recidivism rates: The criterion
and follow-up issues. W. a. 53-60.

Hopkins. A. (1978). Imprisonment and recidivism. W. New York.
New York: Wiley. 750-762.

Horton. A.M.. Medley. DM. (1978). Prediction of recidivism in a criminal population by
birth-order and family-size. W. .420). 27-3

Horton. J .H. (1978). The inﬂuence of selected vocational and academic educational
experiences on self concepts of college students in prison.
International. 39(1 1). SEGA. PP6569. (University Microfilms N0. ADG79-09662.
0000).

Hutchinson. R.L. (1978). Attitudes of selected college administrators and correctional
facility personnel toward the development of post-secondary prisoner education

programs in New York state. WM. 3900). SEGA.
PP5873. (University Microﬁlms N0. ADG79-07219. 0000).

Ineiardi. J .A. (1973). Parole prediction: A ﬁfty year fantasy. W. (5). 42-50

Ingallis. G.R. (I973). The relationship between educational programs and the rate of
recidivism among medium security prison parolees and mandatory supervision cases
from Drumheller Institution in the province of Alberta. WW
national. 32(02). SEGA. PP609. (University Microﬁlms N0. ADG78-l 1897. 0000).

Irwin. J. (1980). W. Boston & Toronto: Little. Brown and Company.

Jacobs. J.B. (1977). W. Chicago and London:
University of Chicago Press.

 

Bibliography

192

Jenkins. W.O.. Witherspoon. A.D. (1972). -
mm. Elmore. Alabama: Rehabilitation Research Foundation.

Johnson. CM. (1984). The effects of prison labor programs on post-release employment
and recidivism (criticaltheory. vocational education. work release). W

W. aims). SECA. PP2660. University Microfilms No. ADGS4-
24609. 0000).

Johnson. D.C.. Shearon. R.W.. a Britton. GAE. (1974). Correctional education and
recidivism in a woman's correctional center. W. 25(2). 121-129.

Johnson. E.H. (1965). Prerequisites to extension of prisoner education. W
W. .12. 1 7- 19.

Johnston. N.. Savlta. LB. (1978). Operationalizing recidivism. W.
New York. New York: Wiley. 763-766.

Hachigln. 8.x. (1986). ~ - ~.
Wetlands New York: Radius Press

Ksssebaum. 0.. Ward. D.A.. & Wilner. p.111. (1971). W.
New York: John Wiley and Sons. Incorporated.

Kendall. 6.1!. (1973). What makes correctional education correctional? In A.R. Roberts

(Ed.). Wanna (pp. 95-103)- Sprlnsﬂeld. Illinois Charles C-
Thomas.

Eerie. H. (1973). Penal education: United States and Europe. W. 53(2).
4-25.

Eerie. x. (1974. May) WWW Paper presented at
the meeting of the Institute on Action Research and Justice Management. The

American University.

Kirk. RE. (1982). W- (2nd ed.). Monterey. California: Brooks/Cole
Publishing Company.

 

Klein. R.J.. and others. (1982). The continum of criminal oﬂ'em instrument: mother
development and modiﬁcation of Sellin and Wolfgang: s original criminal index.

 

ment Reproduction Service No. EJ 280 761) I
minds. EH. Babst. PM (1964. November 2t!) W

 

0‘ .09.. .s _‘II|‘! : A ‘ so .3 ‘ do so o.’ 0.10’10. or

(Progress Report No. 6). State of Wisconsin:

Department of Public Welfare.
Lanna. WI. (1935. September). Parole prediction as science. W
W.

Lavin. J.H. (1977). Changing lives. Change. 9(12). 49-50. (ERIC Document Reproduction
Service No. EJ 170 197).

Law Enforcement Assistance Administration (v.8. Department of Justice). (1978).

W. Washington. DC: U.S- Government
Printing Oﬁlce.

 

Bibliography

193

Lee. A. (1973). Evaluation of adult basic education in correctional institutions. In A.R.

Roberts (Ed.). We (pp- 377-388i- Springﬁeld. lllinois=
Charles C. Thomas.

Lee. A. (1990). - - - ~
Report). Lansing. Michigan: Ofﬁce of the Auditor General.

m. JsPs. ”a“. JsGo (1976). at 3

 

 

.(ERIC Document Rte-production

Service No. ED 128 650).
Lewis. MN. (1973). The humanities in prison. W. 53(2). 26-35.

Lewis. NW. (1973. June). WUMNW
Park. Pennsylvania: Institute for Research on Human Resources. Pennsylvania State
University.

11nd. 9.0. (1985). Comparison of prison inmate college students with campus students:
Emphasis on self-concept (incarcerated adult males). WWW
Initial. 55(09). SECA. PP2584. (University Micromms No. ADG85-25341. 8603).

Linden. 11.. Perry. L. (1982). The eﬁ'ectiveness of prison education programs. mm
W. 6(4). 43-57.

Lindsey. B.A. (1982). Correctional education in North Carolina as a function of locality.

community college. and prison characteristics. WW.
53(03). SECA. PP639. (University Microﬁlms N0. ADGB2-17045. 0000).

lalptorr. n.. Martinson. B.. awmu. J. W
Writes. New York. New York: Praegcr-

Littleﬂeld. J.l".. Woiford. B.L (1982). A survey of higher education in us. correctional
institutions. WWW. 31(4). 14-18.

lace. J .L. (1978). The eﬂ'ect of correctional institutions' education programs on inmates
societal adj ustrnent as measured by post-release recidivism. W
W. 39(03). SECA. PPl296. (University Microﬁlms No. ADG78-16974.
0000).

Maciekowich. 2.1). (1976). Academic education/vocational training and recidivism of

adult prisoners. W. 31(08). SECA. PP5080.
(University Microﬁlms N0. ADG77-03087. 0000).

Hahoney. J. (1976/1977. December/January). Keeping them out of jail. 99mm
Mammal. 511(4). 46-48.

Mulroney. J.R.. Emmett. 19.3.. 0: Russell. WJ. (1976). Qﬁendenaesistancuhmgh

W. Washington. DC: American Association of
Community and Junior Colleges.

Halts. I.D.. Cleary. R. (1977). The mathematics of behavioral change: Recidivism and
construct validity- W. 1. 421-438.

Mandel. N.G.. Collins. B.B.. Moran. M.R.. Barron. A.J.. Felbrnann. Ii‘.G.. Gadbois. C.B..
& Heminstein. P. (1965). Recidivism studied and defined. WEE.
We. 5.6. 59-66.

Ilarken. R.N.G. (1974). All the most valuable things are useless. WM. 2(3).
49-54. (ERIC Document Reproduction Service No. El 1 10 665).

 

Bibliography

194

Marsh. J.J. (1979). Higher education in American prisons. W
11112:. 5(1). 139-155. (ERIC Document Reproduction Service No. EJ 077 61 l).

Marsh. J .J . (1978). Philosophical considerations of prison education: “Pro and con. ”In
M.V. Reagen. D..M Stoughton. T.E. Smith. &JC Davis. W

 

(Abridged ed. ). Syracuse: Syracuse University Research Corporation. Policy Institute.

Marsh. J.. 8: Adams. 8.N. ( 1973). Prison education tomorrow. In M.V. Reagen. D.M.

Stoughton. T.E. Smith. & J.C. Davis . --
WWW (Abridged ed.). Syracuse:
Syracuse University Research Corporation. Policy Institute.

Mahatma... mm. (1973. April). W
W SanPrancisco: Author

ﬂush-ﬂnpiumcunu. undKuhn. (1975. March). WWW
- -. -.:Sanﬁancisco Author.

Martin. RT. (1979). A brief history of prisoner education. In M.V. Reagen. D.M.

Stoughton. T.E. Smith. & J.C. Davis. WWW
WWW (Abridged ed.). Syracuse:
Syracuse University Research Corporation. Policy Institute.

Martinson. R. (1974. Spring. What works? Questions and answers about prison reform.
Wain-.51. No. 35.

 

 

m. Hackensack. New Jersey: National Council on Crime and Delinquency.
Msslow. A. (1948). A theory of human motivation. W. 5.0. 370-396.

McCsbe. M.P.. Driscoll. B. (1971).
(ERIC Document Reproduction Service No. ED 061 906).

McCabe. P.. Driscoll. B. (1971). - .-
W. Ashland. Kentucky: Federal Correctional Institution.

IlcCleary. R. (1978). W. Beverly Hills & London:
Sage Publications. Incorporated.

McCollrnn. 8.6. (1973. June). New designs for correctional education and training pro-
grams. W. 31. 6- 1 1.

McCollurn. 8.6. (1976). College for prisoners. W11. 39.
98-106.

McConum. 8.0. (1976). WWW: (ERIC
Document Reproduction Service No. ED 104 282).

McCollurn. 8.0. (1975). ' .~ -. .
Reproduction Service No. ED 126 830).

McCollurn. 8.0. (1977. June). What works! A look at eﬂ'ective correctional education and
training experiences. We pp. 32-35.

HcOonm. 8.0. (1984. April 3). WW Paper presented to
annual conference of the American Association of Community and Junior. Colleges.

Washington. D.C.

 

 

 

Bibliography

195
McDougall. E.C. (1976). Corrections has not been tried. W. 1. 63-76.

McFadden. J.. McFadden. G.J. (1976). Education as a reform mechanism. W
Wanna. 21(4). 4-6.

McKinlay. TM (1977). Prison inmates: Institutional adjustment. educational levels.
recidivism. and escapism related to 16 personality factor scores. mm
mm. 16(04). PP247. (University Microﬁlms No. ADGl3-l 1570. 0000).

McNamara. C.E. (1976). Insight into corrections education. W
No 38(3). 10-1 1-

Mchlliams. J.P. (1971. March). Rehabilitation versus recidivism. MW.
51(6). 88-90.

Mendez. G.A.. Jr. (1979). The relationship of prison higher education to institutional

adjushnent.Dissrnatim1Ahstracis.lnieman0nai.40(10). SECA. PP536. (University
Microﬂlrns No. ADG80-07787. 0000).

Ilichigan Departurent of Corrections. (1985). W
W. Lansing. Michigan: Author.

Moberg. 13.0.. Erickson. 11.0. (1972). A new recidivism outcome index. W.
35.. 50-57.

Hogilnicki. 11.1.. (1972). Continuing education - in prison..10nmal.0LQ0ntinningEsiucancn
mm.1(4). 251-257. (ERIC Document Reproduction Service No. E1 058 737).

note. P.. similarity. J. (1986). Wine
W. Wilmington. Ohio. Wilmington College. Research and Development
-Project Talents.

Morris. E.C. (1972). A national survey of college programs in corrections. W
Wm. mm. 26-28.

Morris. N. (1974). W. Chicago and London: University of Chicago
Press.

Morse. 8.1.. (1976. March)- WWW (Report No
82). Denver. Colorado: Education Commission of the States. Correctional Education
Project.

Moseley. W.H. (1976). Parole: How it is working. WW. 5. pp. 185-203.

Mullen. J.. Smith. a. (1980. October).
MW. National Institute of Justice.

Murphy. T3. (1978). WWW (Final Report AP-O)
Lansing. Michigan: Michigan Department of Corrections. Program Bureau.

Nacci. P.L. (1978). The importance of recidivism research in understanding criminal
behavior. Writer. 5. 253-260.

National 1nstitute of Law Enforcement and Crimninal Jrntice. (1979. June). W

 

(Series A. Number 22). Washington. D. C.: U. S. Government Printing Ofﬁce.

National Parole 1nstitutcs.(1964). .~ -
New York: National Council on Crime and Delinquency.

Nelson. N. (1976. June). Prisons and colleges. W. 2302). 372-383.

 

 

Bibliography

196
Nettles. G. (1974). W. New York: McGraw-Hill.

Newmark. J. (1983). W. New York. New York: Saunders
College Publishing.

New York State Divison of Parole. (1964. September). W

W WWI. New
York: Author.

1988 saw record prison population. (1989. April 24). W. p. 9A.

Norneet. ML. (1972). W
W. (ERIC Document Reproduction
Service No. ED 063 524).

Ofﬁce of Economic Opportunity. (1973). WWW
. Oﬁce of Economic Opportunity. Washington. D.C. (Report No.
FGK 65995). (ERIC Document Reproduction Service No. ED 096 496).

O'Hlyre. B.B.. Coffey. 0.1). (1982). W
W. Washington. D.C.: U-S- Depart-

ment of Education.
Ohlin. L. (1975). Harvard recidivism study. W. 11. 21-23.

Oldroyd. R.J.. 8tapley. M. (1976-77). Some correctional programs do reduce recidivism.
WW. 1. 132-141.

O’Leary. V. (1976). Programs of correctional study in higher education. mm
mm. 22(1). 52-66.

O'Huircheartaigh. 0.. Francis. D.P. (1981). WWW.
London. England: Arrow Books.

Palmer. J.. Carbon. P. (1976). Problems with the use of regression analysis in prediction
studies. Wm. 13. 64-81.

Parlett. T.A.A. (1975). A rationale for advanced education in prison. MW
Education. 21. 10-12.

Parson “-3-- 991107. J.P. (1976. October 28) WW
- . = - = , .Paper presented to 1976
National Convention of the Community College Social Science Association. Kansas
City. Missouri.

Peak. K. (1984). Postsecondary correctional education. Contemporary program nature and
delivery systems in the U S W. 35(2). 58—62.

 

 

Washington. D. C. Education Commission of the States.

Pipkin. 8.1!. (1972). WWW. Huntsville. Texas: Texas
Department of Corrections.

     

Polk. K. (1969). lo . me .
Washington. D. C.: Joint Commission on Correctional Manpower and Training.

Pritchard. CJ. (1980). W. (ERIC Document Reproduction
Service No. ED 192 316).

 

Bibliography

197
Pritchard. D.A. (1979). Stable predictors of recidivism - summary. mm. 11(1). 15-21 .

Petersilia. J. (1979). Which inmates participate in prison treatment programs? M
William. 4(2). 121-135.

Quinn. J.R. (1973). Predicating recidivism and type of crime from the early recollections of

prison inmates. W. 35(01). SECA. PP197. (University
Microfilms N0. ADG74-16186. 0000).

Rahming. E. ( 198 1). Participation in a prison vocational / educational program and recidivism

among parolees in Missouri W. .42( 12). SECA. PP5250.
(University Microfilms No. ADGB2-12332. 0000).

Rattenbury. 1".R. (1986). The outcomes of hospitalized and incarcerated sex oﬂ'cnders: A
study of offender types. recidivism rates. and identifying characteristics of the repeat
oﬂ’ender (longitudinal. incest. rape. pedophilia). WWW.
AZ(01). SECB. PP387. (University Microﬁlms No. ADG86-0555. 8607).

Rcagen. my" Stoughton. D.M. (Eds) (1976). WWW
‘ . . ~. am am . Metuchen. New Jersey: The

 

Scarecrow Press. Incorporated.

Reagen. M.V.. Stoughton. D.M.. Smith. T.E. & Davis. J. C. (1978) W

 

(Abridged ed. ). Syracuse. Syhracuse University Research Corporation. Policy Insti-
tute.

Reed. W.E. (1982). Motivation of inmates for college enrollment and the eﬂ’ect of higher
education and vocational training upon inmate discipline. W W
Iniematipnal. 43(06). SECA. PP21 1 1. (University Microfilsm No. AD682-25402.
0000).

Rice. W.E. (1975). Recidivism: A multivariate explanation of the length of time recidivists
remained free before returning to prison incarcerations. W
mm. 3.600). SECA. PP6990. (University Microﬁlms No. ADG76-09285.
0000).

Roberts. A.R. (1971). . .
nlinois: Charles C. Thomas.

Roberts. A.R. mm). (1973). W. Springﬁeld. Illinois: Charles C.
Thomas.

Roberts. AR. 0: Carley. 0.1). (1976. September). W
W College Park. Maryland: American Correctional Association.

Rose. C.. Nyre. G.F. (1979). . .... » : ' :
W. (ERIC Document Reproduction Service No. ED 175 347).

Rothman. D.J. (1980). . : :
W. Boston & Toronto. Little. Brown and Company.

Rutledge. J.‘I‘. (1980). Post release inmate employment: An experimental study concerning
the eﬂ‘ects of job placement counseling on the post release employment and recidivism

of inmates. WW. 43(01). SECA. PP152. (University
Microﬁlms No. ADG82-12703. 0000).

m T.A. (1973). Model components. In A.R. Roberts (Ed.). W
(pp. 55-62). Springﬁeld. Illinois: Charles C. Thomas.

 

e' . Springﬁeld.

 

 

 

Bibliography

19S

Saden. 8.J. (1962. October). Correctional research at Jackson prison. W
Washer). 15.

Barbe. 117.3. (1974. April 11). W. Washington. D.C.:
Attorney General's Office.

Schneider. A.L. (1984). Deinstitutionaliztion of status oﬂ'enders: The impact on recidivism
and secure confinement. Wm. 16(3). 410-432.

Schnm'r. A.C. (1948. September). The educational treatment of prisoners and recidivism.
MW. 54(2). 142-147.

Schore. G.R. (1978). Mom of the press behind barts. WW.
5(4). 17-18. (ERIC Document Reproduction Service No. E1 188 561).

Schwartz. 8. (1975). Teaching at Walkill Prison (Opps. I mean correctional facility). W
W. 6(2). 101-108. (ERIC Document Reproduction Service No. E.) l 15 749).

SEARCH Group. Incorporated (1976). WWW. lst
ed. Washington. D.C.: U.S.Department of Justice.

Seashore. M.J.. Haberfeld. 8.. Irwin. J.. tBaker. K. (1976). W
We. New Yorle New York“ Praeger Publishers

Sherman. n. a Hawkins. G. (1981). W. Chicago
and London: University of Chicago Press.

Shinbaum. M.G. (1977). Development of a model for prediction of inmate interest in

prison-sponsored academic and vocational education. WW
tional. 38(05). SECA. PP2470. (University Microﬁlms No. ADG77-24505. 0000).

Shuman. C.C. (1976). The eﬂ'ects of vocational education on recidivism of formerly

incarcerated individuals. WW. 38(02). SECA. PP753.
(University Microﬁlms No. ADG77-13290. 0000).

Smith. D.B.. Reagen. M.V.. & Stouﬂiton. D.M. (1979). School behind bars - descriptive
overview of correctional education in the American-prison system. W.
59(2). 217-219.

Social: Educational. Research and Development. Incorporated. (1974. August 12).An

Washington. D. C.: SERD. incorporated.

Staﬂord. RM (1977. Spring. Project Soledad: Community based higher education for
the incarcerated. W. 2(1). 4-9.

Stollmac. 8. (1973). Predicating inmate populations from arrest. court disposition. and
recidivism rates. WWW. 19(2).14l-162.

Stouﬂiton. D.M. a Reagen. M.V. (1973). Prisoner education today. In M. V. Reagen. D. M.
Stoughton. THE Smith. &J..C Davis WM

MW (Abridged ed. ). Syracuse:
Syracuse University Research Corporation. Policy Institute.

Study ﬁnds crime awaits many of us. (1987. March 9). W. p. 3.

Sykes. G.M. (1974). I. :.. .
Jersey: Princeton University Press.

   

 

 

Bibliography

199
Syracme University Research Corporation. (1973). WWW

 

York: Ford Foundation.

‘nask Force on Corrections and the Higher Education Syhstenr in Wisconsin. (1973).

 

Document Reproduction Service N0. ED 094 677).

Taylor. A. (1974). Beyond rehabilitation: The Federal City College Lorton Project--A model

prison highear education program. WW1}. 43(2). 172-178.
(ERIC Document Reproduction Service No. EJ 106 803).

Thomas. D.B. (1978) A descriptive study of correctional education for long-term inmates
at the Maryland House of Correction at J essup through a community college

associate of arts degree program. WW. 39(04). SECA.
PP1982. (University Microfilms No. ADG78-18561. 0000).

Thomas. E.L. (1987). -

WWW Unpublished master’ 8
thesis. Brigham Young University.

Thomas. R.L. (1981). Role of the community-college in continuing-education for the
correctional inmate. W. 43(4). 41-44.

Thompson. J .A. (1979). An analytic model designed to predict inmate success for
management of a prison-based college degree program. WW
national. 59(06). SECA. PP3076. (University Microﬁlms No. ADG79-27599. 0000).

Thorpe. T.. MacDonald. B.. a Bala. G. (1984). Follow-up study of oﬁenders who earn

college degrees while incarcerated. WW. 35(3). 86-
88.

Tibbits. C. (1931). Success or failure on parole can be predicted. W
andsrirninom. 22. 1 1-50.

W. J.(1984). o in oiloot '
Reproduction Service No. ED 252 136).

Trent. C.. Ragsdale. J .F. (1976. February). Community college programs for prisoners.
WW. 4(2). 43-47.
Tucker. H.C. (1977). A comparison of verbal behavior of teachers teaching both regular

college classes and college classes for incarcerated students. W
W. 38(05). SECA. PP2474. (University Microfilms No. ADG77-24847.
0000).

 

 

Tulardilok. A. (1977). Assessment of the college level educational program at the State

Prison of Southern Michigan. WW. 39(01). SECA. PP
159. (University Micromms No. ADG78-10127.0000).

11.8. Federal Bureau of Prisons. Education Branch. (1978). WWW
Wm (Draft). Unpublished manuscript.

0.8. Federal 3mm of Prisons. (1976). Warn.
Washington. D.C.: Author.

0.8. President's Task Force on Prlsoner Rehabilitation. (1971). The criminal oﬂ'ender--
What should be done? Tire report of The President’s Task Force on Prisoner Rehabili-

tation--April 1970. W1. 242-266.

 

Bibliopaphy

Valletutti. B.. & Mopsik. 8.1. ( 1973). Aconceptual model for correctional education programs:

Aspecial education perspective. InA.R. Roberts (Ed.). Wlpp.
1 15-127). Springfield. Illinois: Charles C. Thomas.

Vukevich. 0.8. (1973). An attempt at self-direction through social reeducation. In A.R.

Roberts (Ed.). 82W (pp. 251-257). szringt‘iield.
nlinois:Charles C. Thomas.

Wagner. P.A.. Jr. (1976. April). Adult education and the prison. W. 24(9).

263-264.

Waldron. 114141974. September). W
W. (Technical Note No. 5). State of Texas: Texas Department
of Corrections.

Wallack. WJL. Kendall. C.E.. a Briggs. EL. (1939). W.
Kingsport. Tennessee: Kingsport Press.

Wallerstedt. J.1i‘. (1984). W (Bureau of Justice Statistics Special Report No.
NCJ-95700). Justice Statistics Clearinghouse. National Criminal J ustiee Reference
Service. Box 6000. Rockville. Maryland 20850.

Washington State Board For Community College Education. Vocational Education

Special Praieetr Division. (1972 June). W
W. Olymma. Washington. Author.

Washington State Board for Community College Education. (1973. J1me). W

- . ~ .(Proceedings of a
conference at Yakima Valley College). Olympia. Washington: Author. (ERIC Document
Reproduction ServiceNo. ED 094 835).

Weeks. B.B.. Jr. Coltharp. J .C. (1985). College prison education in Georgia: A proﬁle of post-
seeondaiy education institutions. WWW. 36(3). 94-97.
(ERIC Document Reproduction Service No. EJ 327 516).

Weiser. MA. (1979). A career proﬁle of a sample of male prison inmates enrolled in

courses in community colleges in the state of Maryland. W
International. 40(07). SECA. PP3993. (University Microﬁlms No. mesa-00767.
0000).

 

Weiss. N.. Bassett. M. (1982). W. Reading. Massachusetts: Addison-
Wesley Publishing Company.

 

Center for Vocational Education. -

Whyte. C.B. a Whyte. W.R. (1982). Accelerated prison programs-behind prison walls.
W. 15(1). 70-72.

Willetts. D.A. (1971). The college behind bars. W (Trenton). 22(3). 19-21.
Willis. RM" and others. (1978). W

 

Document Reproduction Service No. ED 159 428).

Wilmington College of Ohio - Ofﬁce of Continuing Education. (1986). W
W. (1986. September). Wilmington. Ohio. Author.

 

Bibliopaphy

201

Wilson. J.C. (Ed.). (1983). W. San Francisco. California: Institute
For Contemporary Studies.

Wold. J.C. & Sykes. D. (1981). mmmmmmmmmm

10$ '9‘... '51. 9'9 0 0 ‘.3'\fo!.o"' ‘ ."!.‘6 '1'" ..ol 0 03'." 9.1..

W. (ERIC Document Reproduction Service No. ED 226 073).

Wolford. B.1.. Littleﬂeld. J.F. (1983. June 19-22). -

W. Paper presented at the
38th International Confernece of the Correctional Education Association. Houston.
Texas.

Wooldridge. 8. (1976. October). College for prisoners: Ohio's open door. Change. 8(9). 17-20.
(ERIC Document Reproduction Service No. E.) 146 830).

Yarborouﬂl. T.B. (1980). Motivational demographic and academic characteristics of prison
inmates enrolled in community college programs in the state of Maryland. Dissertation

W. .4108). SECA. PP3409. (University Microfilms No. ADG81-
03898. 0000).

Yarborouﬂl. T.B. (1983). An inmate viewpoint of teaching and curriculum in community
college programs. W. 11(4). 407-409.

Yarborough. T.B. (1985). Some inmate viewpoints on teaching and curriculum in community
college programs. WWW. 36(3). 92-93.

 

Bibliography