The extension of neutralization theory to the academic dishonesty of college students

MISSING IMAGE

Material Information

Title:
The extension of neutralization theory to the academic dishonesty of college students
Physical Description:
ix, 243 leaves : ill. ; 29 cm.
Language:
English
Creator:
Polding, Brian E., 1962-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 230-241).
Statement of Responsibility:
by Brian E. Polding.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 34381809
ocm34381809
System ID:
AA00012911:00001


This item is only available as the following downloads:


Full Text








THE EXTENSION OF NEUTRALIZATION THEORY
TO THE ACADEMIC DISHONESTY OF COLLEGE STUDENTS











By

BRIAN E. HOLDING




















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1995


UNIVERSITY OF FLORIDA LIRAPS


































1995 Brian E. Folding














ACKNOWLEDGMENTS


There are many people to whom I owe a debt of gratitude

for the completion of this dissertation. While space does

not permit me to name all of them, several persons are

deserving of special recognition.

First and foremost, my deepest gratitude goes to my wife,

Brenda. She has given me her support, encouragement,

patience, and love, and I am forever grateful to her.

I extend my appreciation to two of my former grade school

teachers, Dr. Harry Senheiser and Ms. Sandy Morris, for

encouraging me to read and inspiring me to learn. I also

wish to acknowledge several of my former undergraduate

professors; the late Dr. Robert Macoskey for allowing me to

work with him and share his creative vision, and Dr. William

Oman, Dr. Bernard Freyberg and Dr. Allen Larsen for their

inspiration.

I offer my sincere appreciation for completing this

dissertation to Dr. James Wattenbarger, chairperson of my

committee, not only for his fulfillment of this formal role,

but for being my teacher, friend, and mentor. I also extend

a special thanks to Dr. Art Sandeen, for his support on this

project, his cogent advice regarding my career in student

affairs, and for serving as a professional mentor. I also








extend my warmest thanks to Dr. Michael Rollo who believed in

my potential as a student affairs professional enough to

allow me to serve in the Student Judicial Affairs Office for

three years. I learned much from him, and I thank him for

his friendship and support. My deepest appreciation is

extended to Dr. David Honeyman, cochair of my committee for

his support and encouragement; to Dr. David Miller for his

support, sense of humor, and assistance with the analysis of

the data; and to Dr. Richard Hollinger for his advice and

assistance.

I wish to express my appreciation to the staff of the

Student Services Office at the University of Florida and the

Dean of Students Office at the University of New Mexico for

encouraging me through this process and for displaying the

highest level of competence and commitment to their

profession. I thank them for allowing me to serve with them.

I thank my supervisor and friend at the University of New

Mexico, Karen Glaser, for her friendship and constant

encouragement throughout this process. I thank the

Association for Student Judicial Affairs for its financial

assistance in conducting the study. Finally, I wish to thank

my parents, Dana and Gaylene Polding, for encouraging me to

read and learn as a child and for encouraging me through this

process.















TABLE OF CONTENTS








ACKNOWLEDGMENTS ......................................... iii.

ABSTRACT ....... ............. .......................... vii.

CHAPTER I-DESCRIPTION OF THE STUDY 1
Introduction .............................................. 1
Theoretical Background ..................................8
Purpose of the Study ...................................11
Limitations ...................................... ........15
Definition of Terms ......................................16
Overview of the Methodology ...........................17
Selection of the Population and Sample ...........18
Self-Report .......................................18
Instrumentation .................................. 19
Data Collection .................................. 21
Analysis of the Data .................................. 21
Organization of the Study ........................23

CHAPTER II-REVIEW OF LITERATURE ............................25
Introduction .......................................... 25
Frequency of Academic Dishonesty ......................28
Individual Variables .................................. 34
Gender........................................... 34
Age .............................................. 47
Academic Achievement ............................. 48
Ability to Delay Gratification .................. 53
Internal/External Control Measurements ........... 53
Moral Development ................................ 55
Personality ...................................... 61
Situational Variables ................................. 65
Pressure to Succeed...............................66
Perceived Risk ................................... 69
Resentment ....................................... 73
Learning and Socialization ............................74
Modeling ......................................... 74
Group Identification .............................75
Parental Factors .................................79
Peer Reaction ....................................81
Reward and Punishment ............................ 82


v








Neutralization .......... ............................... 84
Conclusion ........................................... 101

CHAPTER III-RESEARCH METHODOLOGY .......................... 109
Purpose of the Research .............................. 109
Population and Sample ................................ 110
Instrumentation ...................................... 110
Development of Neutralization Statements ........ 110
Assessment of Frequency and Types of Cheating ... ll
Direction of Hypothesized Relationship ..........113
Self-Report ...................................... 114
Pilot Study .................................... 116
Administration of the Instrument ..................... 116
Treatment of the Obtained Data ....................... 116
Summary .................................................118

CHAPTER IV-PRESENTATION AND ANALYSIS OF THE DATA .......... 119
Purpose of the Research .................................119
Research Sample ...................................... 120
Analysis of the Data ..................................123
Independent Variable ............................123
Dependent Variable ................................130
Gender Differences in the Use of Neutralization.135
Differences in Frequency of Cheating by Gender .. 136
Relationship Between Neutralization and Cheating 137
Relationship Between Neutralization and
Cheating by Gender ....... .......................138
Summary ...................................... ..........145

CHAPTER V-SUMMARY AND CONCLUSIONS .........................190
Major Findings and Discussion ........................190
Implications for Practice .................. ...........198
Implications for Research ............................202
Conclusion ...................................... ........203

APPENDIX A-PANEL OF EXPERTS ............................204

APPENDIX B-STATEMENT ..................................... 205

APPENDIX C-NEUTRALIZATION STATEMENTS ...................... 206

APPENDIX D-ACADEMIC DISHONESTY STATEMENTS ... ...................208

APPENDIX E-PART I CONFIDENTIAL STUDENT SURVEY ............ 209

APPENDIX F-PART II CONFIDENTIAL STUDENT SURVEY ........... 222

REFERENCES ................... ............................ 230

BIOGRAPHICAL SKETCH ........................................ 242











Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE EXTENSION OF NEUTRALIZATION THEORY
TO THE ACADEMIC DISHONESTY OF COLLEGE STUDENTS

By

Brian E. Polding

December, 1995




Chairman: James L. Wattenbarger
Cochairman: David Honeyman
Major Department: Educational Leadership


The purpose of this study was to determine if the

neutralization theory of juvenile delinquency developed by

Gresham Sykes and David Matza and extended by Klockars and

Minor could be used to describe the reasons for cheating, and

whether gender differences existed in the relationship

between neutralization and cheating. Seven neutralization

techniques were examined: denial of responsibility, denial of

injury, denial of victim, condemnation of the condemners,

appeal to higher loyalties, defense of necessity, and

metaphor of the ledger.

The population for the study was 25,073 undergraduate

students at a large southeastern graduate research

institution. The sample consisted of 410 students in ten

different courses. The study was conducted by a two-part

questionnaire. Part I, administered at the beginning of the








semester, assessed the rate at which these students

subscribed to seven types of neutralization. Part II,

administered at the end of the semester, asked the

respondents to self-report how many times they had engaged in

ten types of cheating during the semester.

Responses to the 19 neutralization statements in Part I

were examined in relation to assignment to the seven types of

neutralization. Internal consistency coefficients were

calculated for each scale using Cronbach's alpha. A

confirmatory factor model corresponding to the assignment of

statements to techniques was estimated. Fit statistics and

internal measures of reliability were then examined. Factor

analysis was performed on the responses to the ten academic

dishonesty questions in Part II in order to classify

responses into three general types of cheating; lazy,

impulsive and premeditated. Gender differences in the use of

neutralization was examined through a MANOVA. Poisson

regression was then performed on models including gender,

neutralization technique and gender by neutralization

technique for each of the types of cheating. Additionally, a

model for total number of cheating acts was estimated.

The findings of this research provided some support for

the view that students who subscribe to neutralization cheat

more than those who do not. Results showed that males and

females subscribed to different types of neutralization at

different levels, males engaged in more cheating than females

due to higher neutralization scores, the relationship between


viii








neutralization types and the types of cheating in which

students engaged differed by gender, and a strong interaction

effect existed between gender and neutralization.














CHAPTER I
DESCRIPTION OF THE STUDY

Introduction

Several recent reports on higher education have focused

on how well higher education is fulfilling its mission of

educating students to become responsible citizens. Reports

from the National Institute of Education (1984), The

Association of American Colleges (1985), and The Carnegie

Foundation's Newman Report (1985) have raised important

questions regarding the quality of undergraduate education in

America, and specifically have addressed the failure of

higher education to provide students with a framework for

making ethical decisions. In addition, several recent best-

sellers, including Allan Bloom's (1987) The Closing Of The

American Mind. have indicted American higher education for

its failure to provide a focused curriculum which provides

moral guidance to students.

The concern for higher education's effectiveness in

fulfilling its mission of providing for a responsible,

enlightened citizenry has been heightened in response to

empirical data regarding the level of academic dishonesty

occurring on college campuses. A recent national study

indicated a 6.2% increase from 1987 to 1988 in the number of

freshmen who reported cheating at colleges and universities








(American Council on Education & the Higher Education

Research Institute, University of California, 1988).

Graduate and professional students have also not been immune

to academic dishonesty. Seib (1980) reported that 88% of the

medical students questioned admitted to cheating while in

college, with 58% reporting that they had cheated while in

medical school. Wynne (1979), after conducting two surveys

at American research universities, reported that the

proportion of students who had admitted cheating rose by 87%

from 1969 to 1976. Recent research by McCabe and Bowers

(1994) has questioned the popular belief that college student

cheating has increased substantially in the eighties and

nineties. However, regardless of the change in cheating

frequency, over time, the majority of students in this study

reported at least one incident of cheating while in college.

Also, research by McCabe and Trevino (1993) has provided some

evidence that the rate of increase for cheating may be

greater at large public institutions than at small private

ones.

Cheating on a school examinations results in an

inaccurate evaluation of a student's knowledge and a lack of

future commitment by the student to master the subject matter

being taught (Fischer, 1970). In addition, educators may

contribute to the development of lowered moral standards of

students by allowing students to discover the efficacy of

cheating (Harris, 1965; Mills, 1958; Weldon, 1966).








Concern about the moral and ethical development of young

people is not only of concern to the campus community, but to

the larger public as well. Research by Hersh (1979) and

Butts (1977) shows that the public ranks moral instruction as

one of the major responsibilities of education.

Academic dishonesty of college students has received

recent attention in the popular press, including People

magazine (Garred, Bacon, Matsumota & Skolnik, 1991) and the

USA TODAY newspaper (Kelly, 1991). Academic dishonesty is a

continuing problem for higher education, especially in the

current social climate where higher education is experiencing

a crisis of confidence from both inside and outside academe.

Higher education must be able to renew the confidence of its

constituents by preparing students who are "capable of

recognizing and dealing with ethical dilemmas that arise

among or from organizational pressures, conflicts of

interest, competition among multiple goods and consequences,

and ordinary human weakness" (Andrews, 1985, p.l as cited in

Kibler, Nuss, Patterson & Pavela, 1988).

If higher education is to accomplish the task of

equipping students with the skills for resolving ethical

dilemmas, it is necessary to determine where the primary

responsibility for this task resides within the institution.

An historical perspective is helpful in answering this

question.

In the early American colonial college, this character-

developing task was shared by faculty and administrators who








were concerned with the intellectual, religious and moral

development of students with "the pervasive impact of

Christian piety as the unifying aim of college education"

(Fenske, 1980, p. 5). Faculty and tutors utilized a

stringent disciplinary process to assist with the

accomplishment of this goal. At Harvard "the nurture of the

intellect required the standard means of discipline, which

until 1718, was flogging. Flogging was displaced in that

year by the perhaps somewhat more humane practice of boxing,

in which a bad boy was made to kneel at the feet of his

tutor, who proceeded to smack him sharply on the ear"

(Rudolph, 1962, p. 27).

In the mid-19th century higher education began to respond

to secular and technological trends by broadening its purpose

to include vocational training. This move was greatly

encouraged by the Morrill Land Grant Act of 1862 which

expanded the number of public institutions and further

liberated the American university "from the class-bound,

classical-bound traditions which for so long had defined the

American collegiate experience" (Rudolph, 1962, p. 263). The

liberalization of the role of the university was accompanied

by a relaxation of strict disciplinary codes which had

governed student life in most of the colonial colleges

(Rudolf, 1962, p. 106-107).

While the public sector of higher education was growing,

the mission of the university becoming more technical and

secular and the student body becoming more diverse, important








changes were also occurring in the American professoriate.

In early colonial colleges, faculty were responsible for

three main functions: teaching, research, and the development

of the character and values of students (Fenske, 1980).

However, as Knapp indicates, "the evolving role of the

college professor in America has been characterized by a

progressive decline of his character-developing function

along with a strong tendency for the research and the

informational functions to part company and form two separate

callings" (Knapp, 1962, p. 292).

The demise of the character-developing function of the

professoriate can be attributed to at least three different

factors: public institutions were inherently committed to a

greater pluralism of religious and moral thought than was

found in earlier colleges, thereby casting doubt on what

moral code should be taught; growing numbers of faculty

pursued graduate study at Germanic institutions where they

were introduced to scientific scholarly research, and a

disinterest toward students' activities beyond the classroom;

faculty reward systems preeminently rewarded faculty research

efforts (Fenske, 1980).

Although American institutions of higher education were

influenced in varying degrees by changes in the

professoriate, the largest and most complex research

institutions were affected the most. It was at these

institutions that the student services profession first

emerged and later blossomed (American Council on Education of








Student Personnel Administrators, 1987, p. 3). In the late

19th century many university presidents appointed deans of

men and/or deans of women to regulate student life, resolve

student problems, and oversee student social life including

Greek letter societies, intercollegiate athletics, drama and

debate teams, and student publications.

The student services profession received assistance in

its development by the mental testing and counseling

techniques developed by the American military during World

War I, and by the application of John Dewey's progressive

developmental theories to higher education. In 1937 a

comprehensive statement of beliefs concerning the student

services profession, The Student Personnel Point of View, was

published by the American Council on Education. The document

defined student personnel work, produced a statement of

purpose, and listed twenty-three specific services. The

philosophy of this statement was that of educating "the whole

student" which included "his intellectual capacity and

achievement, his emotional make-up, his physical condition,

his social relationships, his vocational aptitudes and

skills, his moral and religious values, his economic

resource, and his esthetic appreciations" (American Council

on Education, 1937, p. 1). Later statements by the ACE

(American Council on Education, 1949; American Council on

Education and National Association of Student Personnel

Administrators, 1987) reviewed this initial document, and

while they added some new specific goals for higher








education, the student services profession was still defined

as expressing a concern for the development of students as

whole people.

It is clear that the student services profession has an

historical and philosophical commitment to developing the

ability of students to address ethical dilemmas. Recent

research by Kibler (1993) has highlighted the importance of

student services administrators in addressing the problem of

academic dishonesty from the philosophical perspective of

developing the "whole student." Research by Ostroth,

Armstrong and Campbell (1978) and Steel, Johnson and Rickard

(1984) indicates that the role of administering student

discipline, including resolving academic dishonesty

complaints, resides with a student services administrator at

most public institutions of higher education. Kibler (1994)

has pointed out that disciplinary policies are the primary

source for guiding how institutions address academic

dishonesty. However, Kibler (1994) and Aaron (1992) point

out that systematic, comprehensive programs to promote

academic integrity, active communication with faculty

regarding academic dishonesty and active student involvement

in developing and enforcing academic dishonesty policies and

programs are not prevalent in institutions of higher

education.

The ambivalence of higher education towards the issue of

academic dishonesty may be related to the lack of a clear

understanding of the reasons that students cheat, the








relationship between student attitudes towards cheating and

cheating behavior, and the types of attitudes that need to be

addressed in creating a campus ethos that discourages

academic dishonesty. Numerous studies have been conducted to

identify those demographic, personality, or situational

factors that influence students to cheat. However, most of

these studies have been descriptive in nature, and suffer

from a lack of developed theory base. The purpose of this

study was to determine if the neutralization theory of

juvenile delinquency developed by Gresham Sykes and David

Matza (1957), and extended by Klockars (1974) and Minor

(1981) can be used to describe the attitudes related to

cheating behavior, and whether gender differences exist in

the operation of neutralization in the realm of academic

dishonesty.

Theoretical Backaround

Sykes and Matza (1957) developed a social learning model

of criminal deviancy which suggests that people learn deviant

behavior much as they learn conventional behavior. Sykes and

Matza maintained that most delinquents and criminals hold

conventional values and attitudes and participate in

conventional social activities. However, they learn

techniques that help them to neutralize these values and

drift back and forth between conventional and deviant

behavior. Sykes and Matza maintained that juveniles develop

a distinct set of justifications or rationalizations for

their behavior when this behavior violates societal norms.








These neutralization techniques allow youths to drift away

from socially accepted behavior and participate in

subterranean behaviors (Matza, 1961).

Sykes and Matza based their theoretical model on several

observations:

1. Juvenile offenders sometimes have a sense of guilt

over their illegal acts. If they had a stable

delinquent value system it would be unlikely that they

would feel remorse other than for being caught.

2. Juvenile offenders frequently respect and admire

honest, law-abiding people.

3. Juvenile offenders make a distinction between those

who they can victimize and those who they cannot.

Members of similar, social, ethnic or religious groups

are many times off limits.

4. Juvenile offenders frequently participate in the same

social functions as law-abiding youths, such as school

and family activities. (Siegel, 1986 p. 197)

Sykes and Matza offered this evidence to substantiate

their claim that delinquents drift from normative cultural

values to delinquency. This drift is made possible by their

learning of rationalizations that temporarily neutralize

their conventional value system. Sykes and Matza have

divided these techniques into five major types. Klockars

(1974) and Minor (1981) have proposed two additional

techniques of neutralization that are included in numbers six

and seven.








1. Denial of Responsibility--Offenders claim that their

actions were accidental or resulted from forces

outside their control.

2. Denial of Injury--Offenders claim that nobody is hurt

by their actions.

3. Denial of Victim--Offenders claim that although

someone may have been hurt by their actions, it is a

form of rightful retaliation or punishment.

4. Condemnation of the Condemners--Offenders claim that

those who criticize his/her actions are hypocrites.

5. Appeal to Higher Loyalties--Offenders claim that they

are caught in a dilemma of being loyal to their own

peer group while attempting to abide by norms of the

larger society. The demands of the localized peer

group take precedence because these demands are

immediate (Sykes & Matza, 1957).

6. Defense of Necessity--Offenders claim that an act is

necessary so they need not feel guilty about its

commission. (Minor, 1981).

7. Metaphor of the Ledger--Offenders claim that they have

done sufficient good to allow them an indulgence in a

dishonest act. Acceptable behavior is viewed as

accruing good credits which can be "traded in" for

participation in deviant acts (Klockars, 1974; Minor,

1981).

In more common terms, the first five neutralization

techniques describe the rationalizations of "I didn't mean








it", "I didn't really hurt anybody", "They had it coming to

them", "Everybody's picking on me" and "I didn't do it for

myself" (Sykes & Matza, 1957, p. 669). The last two

techniques describe the rationalizations of "I had no choice"

and "I've earned it".

The neutralization theory has been applied to the

academic dishonesty of college students in only a few cases

where an experimental method has been employed, and where the

interaction of neutralization with other factors has been the

focus of inquiry. The extension of this theory to the

problem of academic dishonesty will help to clarify the way

in which the theory describes deviant behavior in general,

and academic dishonesty in particular.

Purpose of the Study

Academic dishonesty has always been a concern for higher

education. However, higher education currently faces a

crisis of public confidence which may be further exacerbated

by reports of widespread academic dishonesty, and the

inability to produce students who are able to recognize and

resolve such ethical dilemmas. It is clear from recent

research that while the responsibility for resolving academic

dishonesty is shared between faculty, academic administrators

and student services administrators, the main responsibility

for transmitting ethical values to students resides in the

student services profession due to historical developments

and the profession's own philosophy. The academic integrity

of students is vital to the mission of higher education.








Yet, no clear theory base now exists for understanding

academic dishonesty.

The extension of neutralization theory (Sykes and Matza,

1957) to the problem of academic dishonesty is an attempt to

clarify both the problem of cheating, and the research

regarding the theory itself. The extension of this theory to

the problem of academic dishonesty will contribute a needed

addition to the research regarding cheating by offering a

developed theoretical approach to the problem.

In addition, the extension of neutralization theory to

the problem of academic dishonesty should help to clarify

several important theoretical problems evident in the

literature regarding neutralization. First, many studies of

neutralization theory have been confounded by the inability

to sort out the direction of the hypothesized relationship

between neutralizing attitudes and participation in

delinquent behavior (Ball, 1966; Minor, 1980). In order to

clarify an important assumption of neutralization theory,

that neutralizing attitudes precede delinquent behavior, this

study utilizes a two part questionnaire in which the

assessment of neutralizing attitudes precedes any

participation in the act of cheating.

Also, much of the research conducted regarding

neutralization theory has used incarcerated males as a

research sample. The use of subjects who have already been

incarcerated complicates the need to assess an important

assumption of neutralization theory; that most delinquent








activity is not a result of a thoroughly unconventional value

system, but rather a result of techniques of neutralization

that allow delinquents to "drift" back and forth between

conventional and unconventional behavior (Matza, 1964; Matza

& Sykes, 1961). In comparison to subjects who have already

been incarcerated, college students provide a population that

is fairly conventional and have not experienced the same

confounding "hardening experiences" that may interact with

neutralization to increase delinquent behavior (Minor, 1984).

Finally, the literature regarding academic dishonesty

reveals that gender is strongly related to cheating behavior.

As an independent variable, gender has been demonstrated to

be related to the frequency of cheating (Erickson & Smith,

1974; Haines, Diekhoff, LaBeff & Clark, 1986; Newhouse, 1982;

Parr, 1936;), and to the type of cheating (Burton, 1976;

Canning, 1956). In addition, gender differences have been

evident in a variety of the other factors that have been

shown to influence cheating, including the factors of risk

(Leming, 1980; Tittle & Rowe, 1973), the anticipated level of

academic difficulty in college (Roskens, 1976), concern

regarding cheating, (Roskens, 1976), the effect of an honor

code on cheating (Canning, 1956), personality factors

(Johnson & Gormly, 1972; Kelly & Worell, 1978; Ward, 1976),

parental factors (Kelly & Worell, 1978), the effect of

temptation and the risk of getting caught cheating (Jacobson,

Berger & Millham, 1970; Leming, 1980) and, academic standing

(Erickson & Smith, 1974).








However, because the vast majority of research regarding

neutralization theory has been conducted with young males, it

is unclear whether gender differences regarding the operation

of neutralization theory. An important contribution of this

study to the literature regarding neutralization is the

determination of whether gender differences exist in the

operation of neutralization theory in an area (academic

dishonesty) where important gender differences have already

been shown for a number of other variables.

By assessing the relationship between the seven types of

neutralization and ten types of academic dishonesty, this

study will also help to clarify the question of whether a

general neutralization trait exists, or whether, as suggested

by Mitchell and Dodder (1980), neutralization may be more

adequately understood as specific techniques related to

certain types of specific acts.

Finally, this study offers an important contribution to

the literature regarding neutralization theory because,

rather than relying on responses to an inventory of

hypothetical deviant acts (a method used in much of the

neutralization research), it offers self-report data of

actual involvement in deviant behavior. Through the use of

this self-report method, important factors regarding the

reliability of various neutralization data collection methods

may be assessed.

In addition to this contribution to the literature

regarding neutralization theory, educational administrators








and faculty may benefit from research regarding the

appropriateness of using criminal deviancy models to explore

the deviant behavior of college students. The neutralization

theory has had useful applications in explaining a variety of

deviant behaviors. If the appropriateness of this model is

confirmed by this study, the effectiveness of student

services administrators and faculty in understanding and

preventing academic dishonesty may be improved.

Limitations

1. The study was confined to one large public

southeastern graduate research institution. However,

findings from the study regarding the extent and types of

academic dishonesty reported were compared with studies from

several other similar institutions and with the findings of

several national studies regarding academic dishonesty to

help establish the validity of the findings of this study.

2. The sample was selected from a variety of

undergraduate classes where the instrument was administered

in-class. However, demographic data on the entire student

population in the semester in which the study was conducted

indicated that the selected sample was demographically

representative of the institution's student population.

3. The statements used to comprise the theory testing

independent variables were specific to the area of academic

dishonesty. Findings of this study may not be generalizable

to other forms of deviant behavior.








4. The study utilized a self-report survey to gather

data. There is evidence that self-reports may underestimate

cheating (Scheers and Mitchell, 1987). However, Hindelang,

Hirschi and Weis (1981) have conducted a thorough analysis of

the self-report methodologies utilized in the study of

juvenile criminal activity and have concluded that the

problems of accuracy in self-reports are surmountable and

that the self-reporting of criminal activity is for the most

part reliable.

5. The study utilized a self-report rather than an

experimental method. Care should be exercised in comparing

the results of this study to those that employ an

experimental method. The choice of a self-report method was

made for two reasons: (a) self-reports continue to be used as

a standard method of delinquency research, especially in the

analyses of neutralization theory, and (b) the study was

designed to assess participation in a wider range of

dishonest behavior than what could have been adequately

assessed using an experimental method.

6. Academic dishonesty has been operationally defined by

ten different items. When comparing the results of this

study to other research on the topic, attention should be

given to the ways that academic dishonesty has been

operationally defined in each study.

Definition of Terms

The following terms are defined for the purpose of this

study and to assist the reader.








The terms academic dishonesty and cheating are used

interchangeably in the study, as they are in the literature

on the problem.

Academic dishonesty is defined as engaging in the

following behaviors:

1. Copied someone else's homework or lab assignments.

2. Knowingly allowed someone to copy your answers to an

examination.

3. Copied answers from someone else taking the

examination.

4. Copied a few sentences of material from a reference

without footnotes or giving credit to the original

author in a paper.

5. Copied published material and turned it in as your

own.

6. Turned in another student's work as your own.

7. Made up a false excuse to delay or put off taking an

examination or turning in an assignment.

8. Secretly got access to examination questions or

answers prior to an examination.

9. Studied from a "hot" copy of an exam before it was

given to the rest of your class.

10. Used notes, books, "cheat/crib sheets", or other

unallowed materials during an examination.

Overview of Research Methodolovg

The purpose of this study was to determine if the

neutralization theory of juvenile delinquency developed by







Gresham Sykes and David Matza (1957) and extended by Klockars

(1974) and Minor (1981), can be used to describe the problem

of academic dishonesty, and whether gender differences exist

in the relationship between the operationalized measures of

neutralization and cheating.

Selection of the Population and Sample

The population of this study included 25,073 students at

a large southeastern graduate research institution of higher

education. The sample consisted of students in courses that

were identified by the dean of each college as fulfilling

general education requirements. The sample consisted of 410

students in 10 different courses. The sample included

students majoring in 19 of the university's 20 colleges. An

analyses of the demographic variables of gender, age, race,

class and college was conducted. This demographic breakdown

was compared to that for the entire population in the period

in which the study was conducted in order to help establish

the validity of the sample.

Self-Report

The literature regarding academic dishonesty reveals two

dominant methodologies that have been employed to study this

problem, an experimental design where students were given the

opportunity to cheat and a self-report questionnaire which

asks respondents to indicate their involvement in cheating

behavior. Each of these methods has its particular strengths

and weaknesses.








The self-report method was used in this study for several

reasons. First, the few previous studies which address the

relationship between academic dishonesty and neutralization

theory have employed such a methodology. Second, because the

objective of this study was the assessment of the strength of

a specific sociological theory in explaining academic

dishonesty rather than inference to a larger population, data

collection procedures emphasized achieving a high return rate

through the use of an anonymous self-report study.

Third, experimental designs can assess cheating only by

employing a vary narrow operationalized definition of

cheating. This study operationalized cheating in a manner

which included several different types of academic

dishonesty, rather than constraining the definition to one

type of behavior.

In general, the literature concerning juvenile

delinquency and the literature regarding academic dishonesty

suggest that self-report studies may under-report some forms

of deviancy. However, the reliability of self-report studies

is greatly increased by strongly establishing the anonymity

of the respondent. This study included a verbal and written

guarantee of anonymity, in order to increase the return rate

and the reliability of the data.

Instrumentation

The study was conducted by means of a two part

questionnaire. Part I of the questionnaire, which was

administered at the beginning of the semester, included a set








of 19 statements which assessed the neutralization model's

attributes. These statements were developed by the

researcher along with a committee of experts from the fields

of criminology and college student services work (Appendix

A). These statements represented the seven domains of the

neutralization model within a context of rationalizations

that could be given for engaging in cheating behavior. Using

these statements, an instrument was developed (Appendix B)

that included each of the statements with a five point Likert

type scale given to respond to each statement. The

respondents were asked to read each statement and indicate

how much they agreed or disagreed with the statement by

choosing one of five responses; "strongly agree," "agree,"

"neutral," "disagree," and "strongly disagree."

Part II of the questionnaire, which was administered at

the end of the semester, included ten items which had been

defined by the researcher as constituting academic

dishonesty. Respondents were asked to indicate how many times

they had engaged in each of these activities during the

preceding Fall semester. The response set included answers

from 0 to 9 or more times.

A pilot study of the instrument was conducted to

determine the clarity of instructions and statements given

and to determine the approximate amount of time needed to

complete the instrument. After completing the instrument,

the participants in the pilot study were questioned regarding

aspects of the instrument and the study.








Data Collection

The questionnaire was administered in-class by this

researcher and a group of students who had been trained to

administer this questionnaire. Before handing out the

questionnaire the facilitators read a prepared statement

(Appendix C) to the class which explained the purpose of the

study and insured the respondents that their answers were

completely anonymous and their participation completely

voluntary. Because an anonymous questionnaire was utilized,

the subject responses on the independent variables on Part I

were matched with responses on Part II by means of a coding

device found in the first three questions of each

questionnaire. Respondents were asked to indicate the month

that their mother was born, how many brothers and sisters

they had at the beginning of the semester, and on what day of

the month they were born.

Analysis of the Data

Of the 1,141 questionnaires collected, 410 were

determined to be completed and had a high level of assurance

in matching Parts I and II. These 410 subjects were used as

the sample for the study. The research hypotheses developed

for testing to achieve the purposes of this research were as

follows:

Hypothesis 1. The use of neutralization techniques

differs by gender.

Hypothesis 2. Males report a higher degree of cheating

than do females.








Hypothesis 3. A positive relationship exists between

neutralization and cheating.

Hypothesis 4. The relationship neutralization and

cheating differs by gender.

The level for statistical significance was set at .05 for

the factorial analysis and regression analysis procedures and

resulting probability values calculated for each hypothesis

tested.

Demographic data regarding the respondents were compiled.

The demographic data of the research sample were examined

using percentages for gender, race and year in school and

college. These percentages were compare to the total

undergraduate student population of the research sample in

order to help establish the validity of the sample.

The responses of the research sample to the 19

neutralization statements contained in Part I of the

questionnaire were examined in relation to their assignment

to the seven types of neutralization techniques. First,

internal consistency coefficients were calculated for each

scale using a Cronbach's alpha. Second, a confirmatory

factor model corresponding to the assignment of statements to

techniques was estimated to further ascertain the

appropriateness of the assignment. Fit statistics and

internal measures of reliability were examined.

A factor analysis was then performed on the responses of

the research sample to the 10 academic dishonesty questions

contained in Part II of the questionnaire in order to








classify the data into a reduced number of types of cheating.

Principal component analysis and cluster analysis was used to

accomplish this classification. The 10 academic dishonesty

questions were classified into three types of academic

dishonesty: lazy cheating, impulsive cheating and

premeditated cheating.

In order to determine whether gender differences existed

in the use of various neutralization techniques, a

multivariate analysis of variance was performed. Poisson

regression analyses were then performed on models including

gender, neutralization technique, and gender by

neutralization interactions for each of the academic

dishonesty types found in the prior analysis. Additionally,

a model for total number of cheating acts was estimated.

These analyses were used to test whether gender differences

existed in academic dishonesty, whether neutralization theory

could be used to describe the occurrence of academic

dishonesty and whether gender differences existed in the

relationship between neutralization and academic dishonesty.

A level of .05 was used to test for significance in all

statistical operations performed.

Organization of the Study

Chapter I has included an introduction, theoretical

background, purpose of the study, limitations, definitions

and overview of methodology. In Chapter II, relevant

research related to academic dishonesty and neutralization

theory is presented. A detailed description of the





24


methodology and design used to study the model is included in

Chapter III. The results and analysis of the data are

described in Chapter IV. Chapter V concludes the report with

a discussion of the results and implications for student

services practice and research.














CHAPTER II
REVIEW OF LITERATURE

Introduction

The purpose of this chapter was to review available

literature regarding academic dishonesty, and to review

available literature regarding Gresham Sykes and David

Matza's neutralization theory. The topic of academic

dishonesty has been studied by using two general types of

research methodology. One research methodology is an

experimental method whereby contrived situations provide

students with the opportunity to cheat. In most cases this

contrived situation involves returning to students for self-

scoring examinations which have been already been scored by

the instructor, and then comparing the instructor's scoring

to that of the students' self-scoring. This method has been

employed in a number of studies, and has yielded varying

results concerning the prevalence of cheating. Parr (1936)

reported a cheating prevalence of 42% among students given

the opportunity to self-score their examinations. Using the

same self-scoring methodology, Fakouri (1936) reported a

cheating rate of 15.6% among undergraduates on a test,

Zastrow (1970) reported a rate of 40% of graduate students on

a test, and Gardner, Roper, Gonzalez and Simpson (1988)








reported a cheating rate of 50.8% of undergraduates on a

study guide assignment.

The other methodology used in the study of academic

dishonesty is to have students admit their cheating behavior

on an anonymous and confidential survey. The extent of

cheating found through the employment of this self-report

method also varies. Baird (1980) reported over 75% of college

students admitted to at least one "cheating item" in a list

of various kinds of cheating. Eve and Bromley (1980) found

that 63% of students admitted to having been involved in one

of the given forms of cheating, and that 22% admitted to

cheating five or more times.

The difference in cheating frequency varies in both

experimental and self-report studies, in large extent,

because of the differences in the way that cheating behavior

is operationally defined in different studies. In both types

of study, experimental and self-report, the frequency of

cheating tends to increase as different kinds of behavior are

included in the operationalized definition. When comparing

various studies concerning academic dishonesty attention

should be given to whether the study employed an experimental

or self-report study, and the way that cheating has been

operationally defined for the study. As such, the following

review of literature identifies whether an experimental or

self report method has been employed in each of the studies

reviewed, and also states the way that cheating has been

operationalized in each of these studies.








For the most part, the research regarding academic

dishonesty has been descriptive in nature, with the analyses

of results focusing upon the frequency of cheating evident in

the study, and variables that are related to cheating as

operationalized. Some studies have focused on specific

traits of individuals including demographic and personality

variables, as contributing to cheating behavior. Other

explanations have focused on some aspect of a situation (such

as the risk of getting caught, or the threat of penalty) as

influencing cheating behavior. Lastly, some research

regarding cheating, focuses upon offering a social-learning

explanation for engaging in cheating (i.e. the viewpoint that

deviant dishonest behavior is learned through association

with certain persons or groups).

The following review of literature contains sections

regarding each of the descriptive elements that have been

addressed in the existing literature concerning academic

dishonesty:

1. Frequency of Academic Dishonesty

2. Individual Variables that are related to academic

dishonesty (such as race and gender),

3. Situational Variables that are related to academic

dishonesty (such as the risk of getting caught and the

threat of a penalty) and,

4. Socialization Variables that related to academic

dishonesty (such as modeling, group identification,

and parental factors).








The last section of the review of literature also

contains a description of the existing research regarding the

theory that is explored and extended in this study, that of

neutralization theory. The conclusion section of the

literature review synthesizes the academic dishonesty

literature and delineates its relationship to the research

regarding neutralization theory.

Freauencv of Academic Dishonesty

Determining the extent of academic dishonesty has been

the objective of many studies. However, the measures

reported in the literature vary greatly, due in large part to

the different definitions given to cheating in various

studies, and to differences in the method of data collection

which include self-report surveys and experimental methods.

Several studies have utilized experimental situations to

measure the extent of academic dishonesty. Specifically,

these studies have utilized an experimental situation in

which students were given the opportunity to grade their own

examinations, which unknown to them, had already been scored.

Cheating was operationally defined as the discrepancy between

the students' scoring and the instructor's secret scoring.

In a study conducted by Parr (1936) 42% of the subjects

cheated. Kelly and Worell (1978) used a similar self-scoring

method to assess the level of cheating among 629 students in

an introductory psychology course. Cheating occurred at a

rate of 19.5%. The same technique was utilized by

Hetherington and Feldman (1964), yielding a 59% cheating








rate; by Shelton and Hill (1969), yielding a 53% cheating

rate; by Zastrow (1970) on graduate students, yielding a 40%

cheating rate; by Vitro (1971) on female junior college

students, yielding a 45.6% cheating rate, by Fakouri (1972)

yielding a 15.6% cheating rate, and by Bronzaft, Stuart and

Blum (1973) yielding a 56% cheating rate.

In a study by Erickson and Smith (1974), college students

were given the opportunity to grade their own examinations,

yielding a cheating rate of 43%. This study also compared

the experimentally obtained cheating data with data obtained

from asking the same students to respond to a subsequent

self-report survey of cheating behavior. The difference

between the experimentally obtained cheating rates and the

self-reported cheating rates was found to be significant.

One of the other important findings of this study was that no

one who had not cheated reported that he or she had done so.

Also, for both males and females, the highest rates of self-

reporting were among those who had cheated the most.

Gardner, Roper, Gonzalez and Simpson (1988) conducted an

experimental study of 129 students enrolled in an

introductory psychology course at a regional state

university. A study guide designed to detect cheating was

used in each of the class sections. Chapters were assigned

simultaneously in the textbook and study guide, and students

were required to complete each study guide and turn it in.

Twenty percent of the course grade depended on the study

guide assignments. The first page of the study guide








indicated that students were allowed to check their answers

with the correct answers in the back of the study guide

before turning in the assignment, but they were advised not

to change any of their answers. Each study guide chapter

contained three to six questions in which wording of the

study guide's answer list differed in some insignificant way

from the wording in the textbook. These keyed questions were

used to identify unauthorized use of the answer list. The

mean cheating rate was 50.8%.

Karlins, Michaels and Podlogar (1988) conducted a study

of 1,374 undergraduates enrolled in an upper-division mass

lecture business course in order to examine the extent of

plagiarism on a library research assignment. This library

assignment required the students to write a three page paper

describing five published articles and/or books about

specific management topics. The study was conducted during

two semesters. All Fall 1987 and Spring 1988 papers were

classified by article topic and author. Any papers that

reviewed the same articles were compared to establish their

originality. All papers that were clearly copied or

plagiarized (as assessed by two independent raters) were

categorized as incidents of cheating. Only 3% were found to

have cheated on the assignment. The authors suggested that

this small percentage is due to the fact that cheating

behavior was very narrowly defined in this study, and the

study was conducted during only two semesters. The authors

compared the results of their experimental study with








substantially higher frequencies that have been obtained

through self-report studies (Baird, 1980; Haines, Diekhoff,

LaBeff and Clark, 1986; Liska, 1978; Singhal, 1982;), and

questioned whether self-report studies may overestimate the

frequency of cheating.

One of the self-report studies cited by Karlins, Michaels

and Podlogar (1988) is a study by Baird (1980), in which a

questionnaire was administered to 200 college students. The

questionnaire measured seven subject variables (gender, age,

class status, grade-point average, major, fraternity-sorority

membership, and extracurricular activities) and 33 behavioral

areas pertaining to the frequency, type, causes, consequences

and perceptions of cheating. The authors of this study

indicated that 75.5% of those students surveyed admitted to

cheating at some time in their college work. However,

cheating in college was operationally defined in a broad way

to include twelve different behaviors (obtaining test

information from other students, allowing someone to copy

your work, copying someone else's assignments, plagiarism,

copying someone's test work, concealing professor's errors,

using crib sheets, stealing or copying a test, changing a

test paper, taking a test for someone else, having someone

else take your test, and bribery or blackmail).

In addition, students reported whether they had engaged

in these activities at any time during their college work.

The broad operational definition employed and the large time-

frame studied could be expected to yield a far greater








frequency of cheating than that found in the previously cited

experimental study by Karlins, Michaels and Podlogar (1988)

in which cheating was operationally defined very narrowly and

studied during a narrow time-frame of two semesters.

Eve and Bromley (1981) conducted a self-report study of

681 undergraduate students, and reported a cheating frequency

similar to Baird (1981). Eve and Bromley defined cheating

with nine different behaviors; gave another student answers

during an exam, wrote a paper for another student, developed

a relationship with an instructor to get test information,

used notes or books during a test when prohibited, sold a

paper to another student, looked at a stolen copy of test

questions, copied answers from another student during an

exam, purchased a paper from another student and submitted a

paper written by another student (p. 8).

Respondents were asked to indicate how may times they had

engaged in any of these activities during their college

career, and during the previous semester. The authors

indicated that 63% of the respondents admitted to engaging in

at least one of these forms of cheating during their college

career; and 22% of the respondents reported cheating on five

or more occasions during their college years.

Singhal (1982) conducted a survey of 364 students in

Technology, Engineering and Agriculture, asking respondents

for a yes/no answer to the question of whether they have ever

cheated in any form on schoolwork at the college level.

Fifty-six percent of the respondents answered in the








affirmative; and only 3% reported getting caught cheating.

Haines, Diekhoff, LaBeff and Clark (1986) reported a similar

difference between reported cheating frequency and getting

caught cheating. In their survey study of 380 college

students, respondents were asked for a yes/no answer to

whether they had engaged in three different cheating

behaviors; cheated on major exams, cheated on daily/weekly

quizzes and cheated on assignments during their college

careers (p. 345). The authors reported that 54.1% admitted to

engaging in at least one type of cheating, while only 1.3%

reported being caught cheating.

Scheers and Dayton (1987) conducted a study of 378

students in order to address the accuracy of self-reports in

assessing the frequency of cheating. The study utilized a

randomized response technique (RRT) developed by Warner

(1965) that allowed sensitive information to be collected

without associating an individual with any particular

response. This data was then compared with data from an

anonymous questionnaire.

Cheating was defined by five different behaviors; "I have

lied to a teacher to avoid taking an exam", "I have lied to a

teacher to avoid handing in a term paper on time", "I have

turned in a term paper which was purchased from someone

else", "I have cheated on an exam by obtaining a copy of the

exam before taking it", "I have cheated on an examination by

copying the answers from someone sitting near me" (p.65).

The RRT resulted in higher percentages in all forms of








cheating. The differences between the RRT and the self-

report ranged from 39% to 83%, indicating that self-report

studies may significantly under-report the frequency of

cheating.

As these studies have indicated, the frequency of

academic dishonesty varies depending upon whether an

experimental or self-report method is utilized. In general,

experimental studies have tended to indicate a lesser

frequency of cheating because they can only assess a limited

number of forms of academic dishonesty, and because the time

span of the study is limited by the time span of the

experiment. In contrast, self-report studies have tended to

report a higher degree of frequency due to the broader

definition of cheating utilized, and because they tend to ask

respondents to indicate behavior over a longer time span.

Individual Variables

A number of studies have identified individual variables

which are correlated with cheating behavior. These variables

include; gender, age, academic achievement and IQ scores, the

ability to delay gratification, internal-external locus of

control, level of moral development and personality

variables.

Gender

Differences between males and females regarding cheating

have been analyzed in many of the academic dishonesty studies

conducted. Parr (1936) utilized a self-scoring technique to

conduct a study of 409 students over a two-year period. He








indicated a cheating frequency of 45% for males and 38% for

females. Parr noted higher psychological ratings for females

as a factor that may have contributed to the difference in

cheating between males and females, although the exact

psychological test administered to the students was not

identified in the study. In the previously cited study by

Fakouri's (1972), where undergraduates were given the

opportunity to self-score their examination, males cheated at

a rate of 24%, and females at a rate of 11%.

Tittle and Rowe (1973) conducted a study of students in

three sociology classes in which the subjects self-scored

their own examinations. In the class that received threats

of sanction for cheating behavior, the number of females who

cheated was reduced more than the number of males who

cheated. The authors suggested that females are more likely

than males to be deterred by threat of punishment.

Leming (1980) conducted a study of 153 college students

who were divided into an experimental and a control group and

given Hartshorne and May's (1928) circles test in high and

low risk settings. Cheating was ascertained by comparing the

scores of the experimental group that was given the

opportunity to cheat, with the scores of the control group

who were not given the opportunity to cheat. The

experimental group was further divided into sub-groups and

testing variables were manipulated to create situations with

both high risk and a low risk of being caught engaging in

cheating. Females cheated more than males overall, but the








margin between males and females was less in the high risk

than in the low risk situation. In the low risk condition

women cheated significantly more than men. Threats of

sanction for cheating (high risk condition) were found to

influence cheating only for women. This finding is

consistent with the study conducted by Tittle and Rowe (1973)

cited above.

The previously cited self-scoring study of 118 college

students conducted by Erickson and Smith (1974) yielded a

cheating rate of 35% for females and 54% for males. In the

previously cited study by Kelly (1978), 591 undergraduate

students self-scored their own examinations. Males had a

cheating rate of 24.4%, while females had a rate of 16.3%.

In a study conducted by Newhouse (1982) on 120 college

freshman, cheating was measured by an instrument in which

subjects interpreted 19 items as cheating or non-cheating.

The males were less likely to define situations as

constituting cheating than were females.

In the previously cited self-report study by Haines,

Diekhoff, LaBeff and Clark (1986), significant difference was

found between males and females when respondents were asked

to indicate whether they had ever engaged in three types of

cheating behavior; cheated on major exams, cheated on

daily/weekly quizzes and cheated on assignments during their

college careers (p. 345).

Several researchers have found no relationship between

gender and the frequency of cheating. Houston's (1977)







experimental study of 130 undergraduates compared the

examination answers of subjects with the answers given by

their neighbors in order to assess the frequency of cheating.

The difference between the mean amount of cheating by males

and females was not significant. Karabenick and Srull (1978)

conducted an experimental study in which 32 male and 32

female introductory psychology students were given the

opportunity to falsely report success on a series of

unsolvable achievement tasks. No significant difference in

cheating was found between males and females.

Burton (1976) analyzed gender as a predictor of cheating,

finding that gender is likely to interact with other factors,

and that gender itself may not be a significant predictor.

Several other researchers have analyzed the relationship

between gender and the type of cheating behavior engaged in

by subjects, and the interaction between gender and

personality or psychological variables. A self-reported

survey conducted by Roskens (1966) assessed attitudes and

behavior regarding cheating of 487 recent college graduates

from the Spring of 1961 and 2,384 college freshmen from Fall

of 1961 at a medium-sized Northeastern public institution.

Respondents were asked to indicate if they had ever engaged

in any of six specific types of cheating; cribbing, copying,

illegally obtaining examinations, plagiarism, ghost-writing

and cooperatively organizing cheating (p. 231).

Roskens found that the types of cheating in which the

respondents reportedly engaged were significantly related to








their gender. "Males reported greater than 'expected'

cribbing but less that 'expected' plagiarism and ghost

writing. Conversely, the responses of the females indicated

less than 'expected' cribbing, but more than 'expected'

plagiarism and ghost writing." (1966, p. 232). Also, females

expressed a greater concern about cheating than did males.

Roskens also found that students who had fathers with

higher status occupations had a greater concern about

cheating than did students who had fathers with lower status

occupations. The relationship between expressed concern

about cheating and the extent of self-cheating was

significant for both males and females. Little concern was

positively related to a high degree of self-reported cheating

and greater concern about cheating was associated with little

self-reported cheating. Roskens used a Chi-square analyses

and found a higher proportion of reported cheating among

males than among females, and less expressed concern

regarding cheating on the part of males than on the part of

females.

Roskens' study also examined the extent to which the

anticipated level of difficulty of college work was related

to the extent to which cheating in high-school was self

reported. Roskens found that females who admitted having

cheated expected greater academic difficulty in the future

than females who did not admit cheating. No significant

correlation was found for males regarding the relationship








between the anticipated level of difficulty of college work

and self-reported cheating.

Canning (1956) conducted an experimental self-scoring

study to assess the frequency and types of cheating found

among five lower-division sociology classes at Brigham Young

University in three different time periods; before, during,

and after the establishment of an honor system. The results

of this study showed that in the "Before" group the men

cheated significantly more than the women. In the "During"

group the proportions by gender were exactly the same. In

the "After" group the women cheated significantly more than

the men.

Canning also assessed the relationship between gender and

four methods of cheating: "(1) wrong answers were erased or

crossed through and correct answers inserted,(2) previously

left blanks were filled with correct answers, (3) answers

which were incorrect were not marked as such, and (4)

arithmetical 'mistakes' favorable to the students were made."

(p. 292)

Gender differences were present in the most popular forms

of cheating in two of the three periods studied. Males and

females used techniques (1)"Changing Wrong Answers" and (2)

"Filling Blanks" to the same degree in all three periods.

However, the male students participated more in (3) "Not

Checking Wrong Answers" than did females in all three periods

and females participated more in (4) "Making Favorable

Arithmetic Mistakes" than did males in all three periods.








Vitro (1971) conducted a study which employed a similar

procedure to that used by Canning (1956). Examinations which

had been secretly scored by the instructor were returned to

the students for self-scoring and any discrepancies between

the scoring of the instructor and the scoring of the student

were noted. Vitro's study, conducted on 70 female first year

junior college students enrolled in the same course,

attempted to determine the relationship of cheating

propensity to subjects' perceptions of parental disciplinary

methods employed in the child's early years. The parental

disciplinary method was determined by the Parental Punitive

Scale (PPS) developed by Epstein and Komorita (1965).

Vitro found that of the 70 subjects, 34 were cheaters

(45.6%) and 36 (51.4%) were non-cheaters. Cheaters reported

parental disciplinary techniques characterized by excess in

either punitiveness or laxity. Non-cheaters reported

moderate disciplinary techniques. Vitro concluded that if a

female child was subjected to excessively punitive

discipline, she probably learned to restrict her activity

because of the risk of being detected and the thought of

facing severe punishment. As a result, moral judgments were

more likely to be based on expediency and external factors

than on internal control. At the other extreme, female

children who had suffered little if any discipline retained

strong infantile impulses. No external control of these

impulses had been established by the parent, and the child

had not been able to develop internal control of these








impulses. Thus, the female child acted on these impulses

with little reluctance.

Vitro and Schoer (1972) utilized an experimental design

to assess the interaction of cheating and four other

variables; task importance, risk of detection, probability of

success, and gender. The gender of the subject did not

significantly interact with any of these variables.

Kelly and Worell (1978) explored the relationship between

cheating, personality factors, parent behaviors and the

gender of the subject. The sample for this study consisted

of 259 male and 370 female undergraduate students in four

different classes, who were enrolled in their first

introductory psychology course. Cheating behavior was

assessed by administering a self-scored reasoning test. The

self-scoring was compared to the actual student responses

that the students had discarded in the waste can.

Kelly and Worell reasoned that cheating represents a

failure of the subject to resist temptation and indicates a

weak internalized behavioral control mechanism. Based upon

this assumption, the researchers hypothesized that cheaters

would score higher than non-cheaters in the psychological

domain of impulsivity, and would score lower than non-

cheaters in the domains of harm avoidance, perseverance, and

orderly approach to tasks. Lastly, Kelly and Worell

hypothesized that the cheater is more likely to be motivated

by external tangible rewards than by intrinsic satisfaction.








Besides these psychological domains, the researchers also

examined the relationship to cheating of several parental

behaviors as measured by the Parent Behavior Form (Worell &

Worell, 1975). The PBF assessed the domains of Parental

Warmth, Active Involvement (with the child) Lax Control,

Conformity, Achievement Control, Strict Control, Punitive

Control, Hostile Control, and Rejection. Lastly, Kelly and

Worell assessed cognitive skills by data obtained from the

students' ACT scores, and assessed several personality

factors by administering the Personality Research Form

(Jackson, 1967).

The results of the study were as follows: One hundred

fifteen of the 592 subjects (19.5%) cheated; 57 of the 236

males (24.2%) cheated compared to 58 of 355 females (16.3%).

A chi-square analysis showed that males cheated more than

females. In addition, males who provided deviant responses

on the personality form were significantly more likely to

have cheated than males who did not give deviant responses.

This comparison was nonsignificant for females.

Univariate and multivariate analyses of variance were

performed on the personality and parent data in order to

compare cheaters with non-cheaters by gender. The following

significant differences were found; Male cheaters scored

higher in the domains of Aggression, Social Recognition,

Exhibition, Harmavoidance and Dominance than did male non-

cheaters. Male non-cheaters scored higher on the PRF Autonomy








Scale. Male cheaters also had lower ACT scores than male

non-cheaters.

Female cheaters scored lower in the domains of

Harmavoidance and higher in Impulsivity, Autonomy, and

Exhibition than female non-cheaters. Female cheaters also

had lower ACT scores than female non-cheaters.

Based upon the data from these personality scales, the

male cheater's personality appeared substantially different

than the male non-cheater:

The male cheater may be described as aggressive,
antagonistic and vindictive (PRF Aggression), interpersonally
domineering (Dominance), but at the same time highly
dependent upon other people's evaluations (Social Recognition
and Autonomy). The male transgressor is also vigilant and
concerned about the possibility of bodily harm
(Harmavoidance). (Kelly & Worell, 1978, p. 186)

These personality components suggest that male cheaters,

compared to male non-cheaters are overly reliant on external

sources of approval, and are attention-seeking and vindictive

rather than cooperative. As Kelly and Worell state, "the

personality characteristics somewhat parallel Kohlberg's

description of instrumental-relativistic moral reasoning" (p.

186).

According to Kelly and Worell, the female cheater is

characterized by an exaggerated thrill seeking and lack of

concern about physical harm. She is also less impulse

controlling (Impulsivity), and more likely to seek attention

through conspicuous, demonstrative behavior (Exhibition) than

the female non-cheater. Given her apparent rebelliousness

and nonconformity to social norms (Autonomy), the female








cheater appears to be an impulsive, nontimorous person who,

one might speculate, is also relatively alienated (p. 186).

Kelly and Worell also found that males and females

differed in their scoring on the parental scales. When male

cheaters were compared on the parental scales with male non-

cheaters, no significance was found for any of the father

behavior scales. The only parental scale which differed

significantly between male cheaters and male non-cheaters was

the Mother Lax Control scale. Male cheaters reported their

mothers as more lax in administering discipline than did male

non-cheaters.

When female cheaters were compared on the parental scale

with female non-cheaters, significant differences were noted

for both the mother and father scales. Females who cheated,

relative to females who did not, reported less Father Warmth,

greater Father Achievement Control, fewer Equalitarian

interactions with the mother, greater Mother Hostile Control,

and greater Maternal Rejection. In addition, female cheaters

reported their mothers to be more Strictly Controlling, less

Actively Involved, and reported less Maternal Warmth than did

female non-cheaters.

The authors note that the parent profile for males is

surprising considering the personality differences between

male cheaters and non-cheaters. They suggest that the lack

of statistically significant differences between male

cheaters and male non-cheaters on the parent scale may

indicate the weakness in assessing children's perceptions of








parental behaviors rather than assessing these behaviors by

some form of observation.

The parent profile for females indicated that female

cheaters have grown up in an atmosphere where parents were

both uninvolved verbally with their daughter and deficient in

expressing warmth. In addition, the father is reported to

exert pressure on the daughter to achieve, and the mother is

perceived by the daughter as imposing harsh, arbitrary

discipline. The lack of affection and inability to

communicate with the parents may have left the daughter no

recourse but to cheat in order to meet the achievement goals

of the father and avoid harsh discipline by the mother.

Johnson & Gormly (1972) conducted a self-scoring study

and showed gender differences among several personality and

situational variables as related to cheating. The study was

conducted with 113 fifth graders in four mathematics classes,

utilizing a self-scoring method. The researchers indicated

that cheaters had significantly lower course grades among

both males and females. It was also found that females were

more influenced by variables relative to self-devaluation

than were males. Female cheaters had a significantly lower

Achievement Motivation score, and a significantly higher

External Locus of Control score than did female non-cheaters.

Male cheaters had a significantly higher Achievement

Motivation score than did male non-cheaters, and no

significant differences between cheaters and non-cheaters was

found for the Locus of Control score.








Jacobson, Berger & Millham (1970) conducted a study which

investigated the effects of several personality variables and

gender differences on the tendency to cheat in a temptation

situation when confronted with failure. The subjects were

276 undergraduate volunteers (121 men and 155 women) enrolled

in an introductory psychology course at the University of

Miami. Cheating was defined as continuing to work at a task

after the allotted time had expired when explicitly

instructed not to do so. Experimental and control situations

were designed to assess what effect temptation and the threat

of getting caught would have on the cheating behavior of the

participants.

The relevant findings of the study were that;

(a) Women were found to cheat in the temptation situation

but men did not cheat, (b) The cheating effect found for

women was due primarily to the significant number of women

who categorized themselves as "self-satisfied," (c) Women

with a high need for approval showed a significant tendency

to cheat, (d) Both male and female subjects with a high

expectancy of success cheated but subjects with a low

expectancy of success did not cheat.

In general, the research indicates that males may be more

likely than females to cheat. However, gender as a predictor

is likely to interact with other variables including

personality variables, psychological variables, level of

moral thought, and parental factors.










The relationship of age and academic dishonesty has also

been examined in numerous studies. In the previously cited

study, Kanfer and Duerfeldt (1968) used an experimental

method to study the cheating behavior of 832 second through

fifth graders. Cheating was defined as fraudulently

reporting success at a series of unsolvable puzzles. The

results indicated that cheating was most prevalent among

second graders, decreasing through the fifth graders. In the

previously cited study by Parr (1936), in which 409 students

were given the chance to self-score a vocabulary examination,

cheating increased with age in the groups of seventeen

through twenty-one year olds. The twenty-two year olds

cheated less than any other age group, excepting the

seventeen year old.

In the previously cited research Haines, Diekhoff, LaBeff

and Clark (1986), who utilized a self-report survey of 380

college students, showed that those who admitted to cheating

had a mean age of 20.3 years and those who indicated that

they had not cheated had a mean age of 25.6 years.

In the previously cited self-report survey of 200 college

students, Baird (1980) showed that age was not correlated

with reported cheating. Analysis by Burton (1976) indicated

that the correlation between age and cheating in many

preceding studies of academic dishonesty could be explained

as an interaction between age and other variables including








stage of moral development, risk, and the ability to perform

a task without cheating.

Academic Achievement

Several researchers have also examined the relationship

of academic ability to cheating, including the relationship

of cheating to IQ scores, SAT scores, and overall grade point

average. In the previously cited study of second through

fifth graders conducted by Kanfer and Duerfeldt (1968)

cheating was most prevalent among those with lower grade

point averages. In a study conducted by Fischer (1970) 135

fourth through sixth grade boys and girls were given the

opportunity to cheat from the back of a test booklet during

an achievement test. No significant relationship was found

between IQ and cheating. However, among cheaters a lower IQ

was associated with a higher degree of cheating.

Several researchers have found that college students of

low academic standing cheat more than students of high

academic standing (Hetherington & Feldman, 1964, Kelly &

Worrell, 1978; Vitro, 1971). The previously cited self-

scoring study by Fakouri (1972) also involved giving the

Achievement Imagery Scale of Iowa Picture Interpretation Test

(Hurley, 1955) to 154 undergraduates students in a psychology

class. Fakouri indicated that those who cheated had lower

grades than those who did not. Of the 24 students who

cheated, 21 received lower than a "C+" in the course. In a

study of 245 college students conducted by Gardner, Roper,

Gonzalez and Simpson (1988) subjects were given the chance to








cheat on study guide assignments. Higher course grades were

found to be correlated with lower levels of cheating.

In a study by Bronzaft, Stuart and Blum (1973) 117

college students were given the Alpert-Haber Achievement

Anxiety Test (Alpert & Haber, 1960). After the first

examination, administered a week later, subjects were given

the opportunity to self-score their examinations. A

correlation was found between low grades and both test

anxiety and degree of cheating.

Drake (1941) conducted a study of 126 students in a

women's college who were given the opportunity to grade their

own examinations. Thirty of the students cheated. Of these

30, none were in the upper quarter based upon intelligence

test scores, 9 were in the second quarter, 6 in the third

quarter and 15 in the fourth quarter. The grades of those

who cheated were also analyzed showing that none of the "A"

students cheated, 4% of the "B" students cheated, 23% of the

"C" students cheated, 75% of the "D" students cheated and 67%

of the "F" students cheated.

The previously cited experimental study conducted by

Leming (1980), in which 153 undergraduate psychology students

were given the opportunity to cheat on an in-class

examination revealed that subjects above the mean grade point

average cheated significantly less under conditions of high

risk of being caught than under conditions of low risk of

being caught. Johnson and Gormly (1972) also conducted an

experimental study of cheating under high and low risk








conditions, finding that both males and females who cheated

had significantly lower grades than those who did not cheat.

Cheaters also had significantly lower intelligence scores

than non-cheaters in the high risk situation. These two

studies suggest that there is a point at which above average

students judge the risk of cheating to outweigh the benefit.

Below average students may not be as sensitive to the risk

level in their cheating behaviors.

Harp and Taietz (1966) conducted an analyses of self-

reported levels of cheating on term papers and Standard

Aptitude Test (SAT) scores among student in three colleges;

Arts and Sciences, Agriculture and Engineering. For each

college except Engineering, the level of self-reported

cheating on term papers was higher among those with lower SAT

scores. Kelly & Worrell (1978) conducted an experimental

study of 591 undergraduate students, comparing the level of

cheating on a problem-solving test with ACT scores. The

researchers indicated that the ACT scores of both males and

females who cheated were lower than those who did not cheat.

In the previously cited study by Baird (1980) an

anonymous questionnaire was given to 200 college students.

Baird indicated that grade point averages were negatively

correlated to self-reported involvement in 12 different

cheating behaviors; obtaining test information from other

students, allowing someone to copy your work, copying someone

else's assignments, plagiarism, copying someone's test work,

concealing professor's errors, using crib sheets, stealing or








copying a test, changing a test paper, taking a test for

someone else, having someone else take your test, and bribery

or blackmail (p. 517).

In the previously cited survey of 384 college students

from Technology Engineering and Agriculture schools conducted

by Singhal (1982), 62% of students with grade point averages

below 2.5 responded in the affirmative when asked whether

they have ever cheated in any form on schoolwork at the

college level.

A questionnaire was administered to 380 undergraduates by

Haines, Diekhoff, LaBeff and Clark (1986). The researchers

indicated that those who reported cheating behavior tended to

have lower grade-point averages than those who did not report

cheating. Scheers and Dayton's (1987) previously cited study

used a randomized response technique and an anonymous

questionnaire administered to 378 students. The researchers

indicated that the lower the grade-point average of the

respondent, the greater was the amount of admitted cheating.

In the previously cited study conducted by Erickson and

Smith (1974) 118 college students were given the opportunity

to self-score their own examinations. The subjects were also

given a questionnaire asking them to report their involvement

in various types of academic dishonesty. Students with

higher test scores cheated less frequently than those with

low test scores. Females with lower high school grade point

averages tended to cheat to a greater extent than females

with higher high school grade point averages. Also, self-








reported cheating was substantially lower than the actual

cheating measure obtained in the experimental portion of the

study.

Burton (1956) discussed the relationship between

intelligence and academic dishonesty and several studies in

which a correlation had been found between these variables;

including studies conducted by Hartshorne and May (1928),

Howells (1938) and Smith (1974). Burton concluded that the

relationship between academic ability and academic dishonesty

is situationally dependent. In tests dependent upon the

subject's ability to perform a task and associated with the

previous success or failure to perform the task, cheating was

correlated with academic achievement. In situations where

the subject had no history of success or failure, the

relationship between academic ability and cheating was

lessened.

The results of these studies indicated that a

relationship exists between academic achievement, academic

ability and academic honesty. There are several explanations

for this relationships, including the theory that the greater

academic ability of a student, the greater the perception of

risk involved in cheating. Therefore, those of higher

academic ability may be dissuaded from cheating because of

the perceived risk. The relationship between academic

ability and cheating may be explained by the theory that

academic success in the past has demonstrated to the student

that academic goals are obtainable by legitimate means. It








may also be true that in self report studies, those of lower

academic ability are more trusting of the study's intent

resulting in a higher incidence of self reported cheating.

Ability to Delay Gratification

Several studies have examined the relationship between

cheating and the ability to delay gratification. Johnson and

Gormly's (1972) conducted a self-scoring study 113 fifth

graders, finding that those children who chose to delay the

receipt of a reward in return for receiving a larger reward

at a later time, were no less likely to cheat than those who

chose an immediate smaller reward. The researchers also

found that the ability to delay a reward was correlated with

higher course grades and achievement measures, and with lower

intelligence scores.

Burton (1976) analyzed previously conducted studies by

Gilligan and Mischel (1964) Brock and DelGiudice (1963),

Aoyama (1970) and Johnson and Gormly (1972). Burton suggested

that in these studies the ability to delay gratification may

be related to a particular dimension of self-control, but

that honesty and reward delay may not necessarily have a

significant correlation.

Internal/External Control Measurements

Several studies have addressed the relationship between

cheating and whether one believes that his or her situation

is a result of internal controllable factors or external

factors beyond his or her influence. Johnson and Gormly

(1972) conducted a self-scoring study of female college








students, and indicated that those who cheated felt that

their academic achievements were the result of external

forces beyond their control.

Houston's (1977) analyzed the relationship between

cheating and four dimensions of the Rotter Internal-External

Scale (Rotter, 1966) as established by Collins (1974). The

four dimensions refer to beliefs about the world with respect

to a difficult-easy dimension, a just-unjust dimension, a

predictable-unpredictable dimension, and a politically

responsive-unresponsive dimension. Houston indicated a

significant positive relationship between perception of the

world as a difficult environment and cheating by copying from

a neighbor's paper. None of the other three belief

dimensions were correlated with cheating.

Leming (1980) replicated Houston's study by administering

Hartshorne and May's (1928) circles test and the Rotter

Internal-External scale (Rotter, 1966) to a group of 153

college undergraduates. However, the circles test was not a

measure of academic ability, as was the test used in

Houston's study. Leming compared the answers of the

experimental group, who were given the opportunity to cheat,

with the answers of the control group who were not given the

opportunity to cheat. Leming's method controlled for the

variables of the pressure to cheat and the threat of

detection. A positive correlation was found between the

Difficult-Easy Subscale and cheating behavior. The results

of these studies indicate that if a student sees the world as








difficult and complex then cheating may be seen as a means to

master this complexity.

Moral Development

Several researchers have explored the relationship

between cheating and the Level of Moral Thought as explained

by Lawrence Kohlberg's theory of moral development (Kohlberg,

1967). Kohlberg's Level of Moral Thought refers to six

developmental types of moral value orientation through which

the individual passes sequentially due to cognitive

development and social experience. The definition of each of

these six stages is as follows:

I. Preconventional Level

Stage 1. Punishment and obedience orientation.

The physical consequences of action determine its

goodness or badness regardless of the human meaning or value

of these consequences.

Stage 2. Instrumental relativist orientation.

Right action consists of that which instrumentally

satisfies one's own needs and occasionally the needs of

others.

II. Conventional Level

Stage 3. Interpersonal concordance, "good boy-nice girl"

orientation. Good behavior is that which pleases or helps

others and is approved by them. There is much conformity to

stereotypical images of what is majority or "natural"

behavior.

Stage 4. "Law and Order" orientation.








This orientation is toward authority, fixed rules, and

the maintenance of the social order. Right behavior consists

of doing one's duty, showing respect for authority, and

maintaining the given social order for its own sake.

III. Postconventional. Autonomous, or Principled Level

Stage 5. Social-contract legalistic orientation.

Right action tends to be defined in terms of general

individual rights, and standards which have been critically

examined and agreed upon by the whole society (generally with

utilitarian overtones). This is the "official" morality of

the American government and constitution.

Stage 6. The universal ethical principle orientation.

Right is defined by the decision of conscience in accord

with self chosen ethical principles appealing to logical

comprehensiveness, universality, and consistency. At heart,

these are universal principles of justice, of the reciprocity

and equality of human rights, and of respect for the dignity

of human beings as individual persons. (Leming, 1978, p. 215)

Schwartz, Feldman, Brown and Heingartner (1968) conducted

a study of 35 male freshmen recruited from introductory

English and psychology courses at the University of Michigan

in order to assess the relationships between cheating,

Kohlberg's Level of Moral Thought and the psychological

constructs of Achievement, Affiliation and Helpfulness. The

subjects in this study were provided with an opportunity to

cheat on an in class examination that was self-scored by the

student. The construct of Helpfulness was measured by








observation through a two-way mirror of subjects in groups of

two's helping each other with a puzzle exercise.

The Level of Moral Thought was assessed by administering

a moral dilemma test created by Kohlberg (1976). Achievement

was assessed with the Achievement Risk Preference Scale

devised by Atkinson and O'Connor (1963). Affiliation was

assessed with an eight-item scale (Brown, 1964). It measured

several related traits: desire to work with others,

preference for affiliation over being alone, desire to

maintain relationships. The Achievement and Affiliation

scales were administered by follow-up mail-back

questionnaire.

The authors indicated that significant relationships

existed between several variables. Subjects high in Level of

Moral Thought were less likely to cheat than those low in

Moral Thought. Subjects high in Achievement were less likely

to cheat than those low in Achievement. Subjects were more

likely to give help on the puzzle exercise if they were low

in Achievement and high in Affiliation.

Leming (1978) also conducted research which assessed the

relationship between cheating and Kohlberg's Level of Moral

Thought. The subjects for the study were 152 junior and

senior undergraduate psychology students at State University

of New York at Stony Brook. The Moral thought of the

subjects was assessed using James Rest's Defining Issues Test

(Rest, 1976). The subjects were then presented with a test

and an opportunity to cheat under two conditions; high threat








[of being caught], high supervision (HTHS) and low threat,

low supervision (LTLS). It was found that cheating behavior

was situationally specific for all three levels of moral

development (low, medium and high). Students in all three

categories cheated more in the LTLS than in the HTHS

situation. Students low in moral thought cheated

significantly more than those who scored medium or high in

moral thought.

A relationship was found between principled moral

reasoning and non-cheating behavior only in the HTHS

situation, where none of the 10 students who ranked High in

moral thought cheated. In the LTLS situation the Highs were

not significantly different from the Mediums or Lows with

respect to frequency of cheating behavior. Leming indicates

that preconventional and principled moral reasoners were

equally influenced by threat of detection when making moral

judgments.

Leming's findings seem to be inconsistent with Kohlberg's

theory of moral development, suggesting that morality may be

situationally specific. It is possible that differences

between moral stages depends on the type of dilemma being

addressed and that in the case of examinations, there is

little difference between preconventional and conventional

moral thinkers.

A self-scoring study conducted by Zastrow (1970) focused

on the incidence of cheating, reasons for cheating, the kinds

of behaviors considered to constitute cheating and the








differences in personality characteristics between cheaters

and non-cheaters. The study was conducted on 45 first year

social work graduate students. Zastrow had hypothesized that

if personality type was positively correlated with cheating,

it would be likely that one who cheated would have a higher

score than a non-cheater on both the Lie Scale and

Psychopathic Deviate Scale of the Minnesota Multiphasic

Personality Inventory (MMPI). However, no significant

differences between cheaters and non-cheaters were found on

any of the eleven MMPI scales.

This finding has implications for the specificity-

generality of moral behavior topic. The generality of moral

behavior hypothesis is that a general trait of honesty exists

and that a person will be consistent in his behavior

regarding honesty over many kinds of situations. If such a

general trait exists, we would expect to find that students

who cheated have higher scores on the Lie (L) scale and the

Psychopathic Deviate (Pd) scale. The MMPI results, however,

seem to support the theory of specificity of moral behavior

which states that a person acts in each situation according

to the way that he has learned to act in similar situations

in the past.

The predictability of behavior does not depend on a

general trait of honesty, but rather on the number of

identical elements which two situations share. If no general

trait of honesty exists than it is more difficult to state

that cheating on an examination indicates that a person will








be dishonest in other situations. However, it is also

possible that the MMPI is not an appropriate instrument for

assessing the personality of non-deviant traditional age

college students.

In a study conducted by Lanza-Kaduce and Klug (1986), an

anonymous survey was given to 196 undergraduate students,

asking respondents to report their cheating behavior and that

of their best friends. Respondents were also given James

Rests's Defining Issues Test (1976). A differential

association was obtained between the level of cheating

engaged in by the subjects, and the extent to which the

subjects reported best friends were involved in cheating.

Moral development was compared to differential association

scores. Moral development was found to interact with

differential association among those with low and medium

level moral development. The authors suggested that this

interaction indicates that those with lower levels of moral

development are more influenced by peers than are those with

higher levels of moral development.

In general, the literature offers some support to the

view that those who have a higher level of moral thought may

be dissuaded from participating in cheating behavior that is

extremely high-risk in nature, while those of a lower level

of moral thought are less affected by risk factors involved

in cheating. Part of this difference may be attributed to

the fact that a higher level of cognitive ability is a

prerequisite to higher level moral reasoning. Thus, those of








higher levels of moral thought may not have the need to cheat

in order to do well, and they may be better at recognizing

situations that pose a high level of risk. Also, those with

higher levels of moral development may be less influenced by

peer associations than those of a lower levels. However,

questions remain regarding the ability of even those of

higher level moral reasoning to generalize moral behavior

across a variety of situations which pose moral dilemmas.

Personality

The relationship of cheating and personality

characteristics has also been explored by several

researchers. In Hetherington and Feldmans's (1964)

previously cited self-scoring study of 78 college students,

subjects were given the opportunity to self-score their

examinations in three different situations. They were also

administered the Minnesota Multiphasic Personality Inventory

(MMPI), the California Personality Inventory (CPI), and the

Edwards Personal Preference Schedule (EPPS) to assess

personality types. Significant differences were found

between those who cheated and those who did not on the

Repression Scale of the MMPI; the Sociability, Achievement,

Socialization, Responsibility and Intellectual Efficiency

Scales of the CPI; and the Autonomy, Intraception and

Deference Scales of the EPPS.

The previously cited self-scoring study conducted by

Zastrow (1970) focused on the incidence of cheating, reasons

for cheating, the kinds of behaviors considered to constitute








cheating and the differences in personality characteristics

between cheaters and non-cheaters. After conducting this

study on 45 first year social work graduate students, Zastrow

found no significant differences between cheaters and non-

cheaters on any of the eleven personality variables studied.

In a study by Kelly & Worrell (1978), 629 students were

given the opportunity to self-score their own problem-solving

test and were also given the Personality Research Form

(Jackson, 1967). The researchers indicated that female

cheaters scored higher on the scales of Impulsivity,

Autonomy, Exhibition and scored lower on Harmavoidance scales

than did female non-cheaters. Males who did not cheat scored

higher than male cheaters on the Autonomy scales. Male

cheaters scored higher on scales on Aggression, Social

Recognition, Dominance and Harm-Avoidance than did male non-

cheaters.

In a study conducted by Bronzaft, Stuart and Blum (1973),

117 subjects self-scored their own examinations after being

given the Alpert-Haber Achievement Anxiety Test (1960). No

correlation was found between test anxiety and cheating.

Shelton and Hill (1969) studied 49 male and 62 females in

five tenth and eleventh grade college preparatory classes in

order to assess the relationships between cheating,

achievement anxiety and knowledge of peer performance of an

examination. The students were given the opportunity to

cheat on a word construction task. Subjects were also given

the Alpert-Haber Achievement Anxiety Test (1960) and assigned








to one of three levels of anxiety; high, middle and low.

Fifty-three percent of the students cheated on the word

construction task given. There were no differences between

males and females in anxiety scores or incidence of cheating.

Anxiety scores proved to be associated positively with

cheating in those subjects who were provided knowledge that

their peers had performed well on the assignment. Level of

anxiety was not significantly related to cheating among

subjects who received no information concerning peer

performance.

The effects of the peer knowledge manipulations on

cheating differed at each of the three anxiety levels. Among

subjects with high anxiety, more cheating occurred in the

conditions of success or failure on the exercise than in the

control condition. Among moderately anxious subjects, only

the failure/control condition was significant. There were no

significant differences between any of the conditions among

subjects with low anxiety. The findings of this study

suggest that knowledge of peer success may interact with

anxiety as a determining factor in cheating.

Campbell (1933) conducted a study of 173 students at a

large southern state universities in order to identify the

relationship between cheating and several personality

variables; neurotic tendency, dominance-submission,

introversion, extroversion and self-sufficiency. Cheating

was assessed by use of an observation "spy" system for in-

class examinations. The personality variables were assessed








by Bernreuter Personality Inventory (Campbell, 1933). It was

found that the cheaters were more neurotic, less self-

sufficient, more introverted and more dominant than non-

cheaters. An interesting finding was that for cheaters, the

normal grouping of these variables was not found. Usually,

those who are neurotic and introverted are not dominant, and

those who are dominant are not also submissive.

In a study of 357 college students, conducted by

Eisenberger and Shank (1985), the personal work ethic was

measured with a work values survey and analyzed for possible

correlation with cheating. Cheating was defined as falsely

stating a solution for an unsolvable anagram task. It was

found that students who had a high work-ethic as measured by

the Survey of Work Values (Wollack, Goodale, Wijting, &

Smith, 1971) worked twice as long at the assigned task before

cheating.

Ward (1976) analyzed cheating and self-esteem by

administering to 165 college undergraduates, a questionnaire

which assessed self-concept, and then giving these subjects a

chance to grade their own examinations. For males, no

correlation was found between cheating and self-esteem. For

women, a significant positive correlation existed between

self-esteem and cheating. Thirty-three percent of the women

with low self-esteem cheated, while only 13% of the women

with high self-esteem did so.

Aronson and Mettee (1968) conducted a study of 45 females

in an introductory psychology class at the University of








Texas in order to assess the relationship between cheating

and induced levels of self-esteem. After taking a

personality test, the subjects were given feedback aimed at

temporarily inducing either an increase or decrease or no

change in self-esteem. The subjects were then allowed to

participate in a game of cards in which opportunities to

cheat were provided. It was found that cheating was

influenced by the level of self-esteem. Significantly more

people cheated in the low self-esteem condition than in the

high self-esteem condition. It is important to realize that

the subjects were not told anything about themselves

regarding their morality or honesty during the manipulation

of the self-esteem variable. Rather, their general level of

self esteem was manipulated. These findings suggest that

women with lower levels of self-esteem may be more prone to

cheat than women with higher levels of self esteem.

Situational Variables

Other researchers have focused upon situational factors

to explain or predict cheating behavior. These situational

explanations have focused on the circumstances presented to

the subject, rather than traits possessed by the individual

as the causal structure in cheating behavior. The

situational factors that have been studied include; pressure

to succeed, appeal to honest behavior, perceived risk, and

resentment.








Pressure to succeed

Several studies have examined the relationship between

cheating and the pressure that a student feels to succeed.

Zastrow's (1970) self-scoring study of 45 graduate social

work students included a questionnaire which asked the

subjects what their reasons were for cheating. The reason

for cheating that was given most often was the pressure to

achieve good grades. The second most given reason was

feeling unprepared for the test.

Vitro and Schoer (1972) studied 611 fifth and sixth

graders in order to analyze the relationship between cheating

and the factors of risk, probability of test success and test

importance. Subjects were given the opportunity to cheat by

looking at the answer sheet while taking a fifty item

multiple choice vocabulary test. Probability of test success

was manipulated by telling subjects that they performed well

or poorly on a pre-test. Test importance was manipulated by

telling subjects that the test was either very important to

their grade or not important at all. Risk was manipulated by

either having a teacher present and walking around the room,

having the teacher present and seated at the front of the

class, or having the teacher leave the room. The researchers

indicated that the probability of test success was positively

correlated with cheating in the high importance, low risk

situation, and negatively correlated in the low importance,

high risk situation.








Gilligan & Mischell (1964) studied 49 sixth-grade boys

from two public schools in the suburbs of Boston who were

given the opportunity to cheat on a game in which a ray gun

was fired at a target. It was found that fear of failure had

a significant positive relationship to cheating.

A questionnaire was administered to 384 college students

by Singhal (1982). Singhal indicated that 68% of the

students believed that pressure to get good grades influenced

cheating behavior. Nuss (1984) surveyed 146 undergraduates

and 169 faculty, finding that 45% of the students and 37% of

the faculty stated that avoiding class failure was the

primary reason for cheating. Schab (1971) administered an

anonymous questionnaire to 855 honor students in Georgia, 155

in the province of Quebec, Canada, and 132 in Scotland. The

questionnaire consisted of 50 items concerning cheating

behavior, and attitudes about dishonesty in society. All

three groups indicated that fear of failure is the primary

motivation behind cheating.

Appeal to honest behavior

Several researchers have examined whether cheating may be

influenced by informing the subject of the undesirability of

dishonest behavior and appealing to the subject's honesty.

In the previously cited study by Kanfer and Duerfeldt (1968)

832 second through fifth graders took part in a guessing game

in which guessing the correct number was highly improbable.

The subjects were given the opportunity to reward themselves

for correct guesses. One group was told that cheating was








unacceptable and one group was not. The two groups did not

differ significantly in their cheating scores.

Fischer (1970) conducted an experimental study of 135

students in fourth, fifth and sixth grades in order to

analyze the relationship between cheating and five variations

in practical classroom conditions. Cheating was defined as

the subjects' use of an answer page rather than his or her

own knowledge in answering a 60 item multiple choice test.

The classroom conditions were as follows:

Informative Appeal to Honesty Condition- These groups

were told the importance of obtaining true measures of their

knowledge in order for the teacher to evaluate her own

teaching techniques and materials effectively. The tone of

the explanation was matter-of-fact appeal and information-

giving.

Public Affirmation of Value Condition- While waiting for

the testing time to begin the examiner casually asked what

the students thought about cheating. After a brief

discussion among the students, the examiner asked each

subject to give his of her name and reasons why he or she

would not cheat on a test like the one being given.

Value-Relevant Threat of Punishment Condition-The

following statement was made to the subjects:

"Children in your grade have tended to cheat
on this test. They get answers to questions
they don't know from the answer page. If I
should find that any of you has cheated, he
or she will have, as a punishment, to write a
sentence 50 times. The sentence would be,
'Although I do not believe in cheating, I
cheated on this test.'"








Non-Value-Relevant Threat of Punishment Condition-The

following statement was made to the subjects:

"Children in your grade have tended to cheat
on this test. They get answers to the
questions they don't know from the answer
page. If I should find that any of you has
cheated, he or she will have as a punishment,
to write numbers from one to 100, 25 times."
(p.13).

The author indicated that 23 subjects (65%) cheated under

the Control conditions. The proportion of subjects who

cheated under the Informative Appeal condition was not

significant. The proportion of cheaters in the Public

Affirmation of Value, Value-Relevant Threat of Punishment,

and Non-Value-Relevant Threat of Punishment conditions were

all significantly lower than in the Informative Appeal

condition. But they did not differ significantly from each

other.

In the previously cited study by Tittle and Rowe (1973),

college students were given the opportunity to self-score

their own examinations. Half of the subjects were also told

that they had a moral obligation to be honest and that they

were being trusted to grade their quizzes accurately. No

significant difference was found between the frequency of

cheating of the two groups.

Perceived Risk

Attention has also been given to the relationship between

cheating and risk in a particular situation. Steininger,

Johnson and Kirts (1964) administered a questionnaire to 46

college students, asking the subjects to comment concerning








the likelihood and reasons for cheating in 32 different

classroom situations. The subjects stated that they would be

more likely to cheat if the teacher left the room, although

they would feel guilty about cheating in that situation.

In a study of 153 college undergraduates conducted by

Leming (1980), subjects were administered the Hartshorne and

May (1928) circles test in which they were asked to close

their eyes and attempt to write the numbers one through ten

in circles of varying size on a single sheet of paper.

Subjects then scored their own papers and their scoring was

compared to the performance of a control group that had been

blindfolded in order to determine the degree of cheating in

the experimental group. One group of students was given a

verbal warning concerning cheating; and the professor walked

through the aisle during the course of the examination. The

other group was given no warning and no instructor was

present. The researchers indicated that those subjects in

the low risk setting cheated more than those in the high risk

setting. The data were also analyzed with respect to the

relationships between gender, academic ability and cheating

behavior. In the low risk condition women cheated

significantly more than men. The high risks condition was

found to reduce the incidence of cheating only for women.

Cheating was not related to academic ability; but under the

high risk condition high ability students cheated

significantly less than under the low risk condition.








Hill and Kochendorfer (1968) conducted a study which

analyzed the relationship between cheating and risk of

detection and knowledge of peer performance. The subjects

were 60 sixth-grade boys drawn from seven classrooms in seven

different schools. The subjects were assigned to one of four

cells of experimental conditions; (a) High Risk of Detection

with No Knowledge of Peer Performance (HR-NK), (b) Low Risk

of Detection with No Knowledge of Peer Performance (LR-NK),

(c) High Risk of Detection with Knowledge of Peer Performance

(HR-K), and, (d) Low Risk of Detection with Knowledge of Peer

Performance (LR-K).

Subjects played a ray-gun game in which their scores were

secretly manipulated. Students were responsible for keeping

their own score. Conditions associated with the potential

for being caught cheating on one's score were also

manipulated. Students who scored well were given a badge to

wear upon returning to the classroom.

More Low Risk subjects cheated than did High Risk

subjects. Fifty-seven percent of the subjects who had

knowledge of peer scores cheated and only 33% of those

subjects without such knowledge cheated. The greatest number

of cheaters was found in the Low Risk- Knowledge group and

the least number was found in the High-Risk-No Knowledge

group. Knowledge of peer scores appeared to increase

cheating, regardless of the risk involved, and the risk of

detection appeared to inhibit cheating to some degree

regardless or whether the subject knew the scores of peers.







The authors suggested that peer pressure in the sense of

attempting to "measure-up" to one's peers, is an important

element in the determination of cheating behavior and that

this factor is sometimes not outweighed by risk.

Vitro and Schoer (1972) also assessed the relationship

between cheating and risk of detection. In addition, they

also examined the effects of probability of test success and

test importance. In the study, 24 fifth and sixth grade

classes (611 students total) were randomly assigned to

treatment conditions formed by all possible combinations of

the three variables being studied: high or low probability of

test success, high or low risk of detection, and high or low

test importance. The incidence of cheating was measured by

secretly scoring student responses on an examination before

it was returned to the student to be self-scored. The

results of this study were as follows:

(a) Probability of success had a significant influence on

cheating when combined with high importance and low risk

detection and with low importance and high risk, but not when

combined with high importance and high risk and low

importance and low risk,(b) Risk had a significant influence

on cheating only when combined with high importance and low

probability of success, and,(c) Importance had a significant

effect on cheating when combined with high risk and low

probability of success and with low risk and low probability

of success (Vitro & Schoer, 1972, p. 274).








It is important to note that probability of test success,

as manipulated by comments made to students concerning an

earlier pretest, was determined to be the most important of

the four variables investigated in terms of its effect on

cheating. Those with high probability of success (based on

the feedback that they received from the pretest) cheated far

less than those with a low probability of success. This

finding suggests that educators can deter cheating by paying

more attention to manipulating how a student feels in terms

of ability to succeed.

Burton (1976) also identified and analyzed several

studies that support the hypothesis that risk is related to

cheating, including studies conducted by Hartshorne and May

(1928), Canning (1956), and Howells (1938). He concludes

that risk of getting caught is an important factor in

cheating.

Resentment

Several researchers have examined the hypothesis that the

greater resentment or hostility a student feels about an

assignment or examination, the greater will be the incidence

of cheating. Steininger, Johnson and Kirts (1964)

administered a questionnaire to 49 college students, asking

the subjects to comment on the likelihood and reasons for

cheating in 32 different classroom situations. The

respondents indicated that they would be most likely to cheat

in situations where an examination was based on senseless








detail, when the teacher was inadequate, and when the test

was perceived as difficult.

Reeves (1987) conducted an experimental study in which

subjects were asked to ingest a capsule that could create

side effects prior to taking an examination. The results

indicated that varying levels of resentment were present in

the subjects, but that these resentment levels were not

accompanied by varying degrees of cheating.

Learning and Socialization

Other researchers have focused on the theory that

cheating is learned through socialization and is not

necessarily trait or situationally specific. Aspects of

socialization that have been examined regarding their

relationship to cheating include; modeling, neutralizing

attitudes, group identification, parental factors, peer

reactions, and reward and punishment.

Modeling

Several researchers have addressed the question of

whether dishonest behavior is learned through observing the

dishonest behavior of others. Stein (1967), Ross (1971), and

Rosenkoetter (1973), indicate that dishonest behavioral

models can increase dishonesty in children. In a study by

McQueen (1957) cheating behavior in a classroom defined as

the failure of a student to report that he or she had

erroneously been given an excess number of points on an

examination, decreased when a peer reported an overscoring of

his grade, and was further decreased when this person was








praised by the teacher, indicating that honesty models can

influence the behavior of children. However, research by

Wolf and Cheyne (1972) indicated that honesty models have not

been as clearly shown to have an effect on children, as

dishonesty models.

In a study by Wilson (1988) 84 first-grade and second-

grade girls were studied as to the effect that conforming and

nonconforming models would have on their behavior. The

nonconforming model was a videotape of a peer opening a toy

box that the child was specifically told not to open. The

subjects exposed to this model took a shorter time to engage

in this nonconforming behavior than did those children who

were not exposed to this model. However, subjects exposed to

two nonconforming models did not take a shorter time to

engage in nonconforming behavior than did those children

exposed to only one nonconforming model.

Group Identification

Several researchers have examined the hypothesis that

deviant behavior is learned and maintained by one's

association with certain groups and by identifying with the

normative beliefs of these groups. In a study by Parr (1936)

409 students were given the chance to self-score a vocabulary

test. Cheating was positively correlated with sorority

involvement for female subjects. Cheating was also

positively correlated with fraternity involvement for males.

Overall, fraternity or sorority involvement resulted in 6%

more cheating behavior than for independent students.








In the previously cited self-scoring study of 126

students in a women's college conducted by Drake (1941),

sorority members had a 36% involvement in cheating and non-

fraternity members had only a 16% involvement. The authors

indicated that there was no significant difference between

the two groups in intelligence or academic achievement, and

postulated that the difference in cheating frequencies was

due to fraternity pressure to maintain good grades.

Hetherington and Feldman (1964) conducted a study of 78

college students who were given the opportunity to cheat in

three different situations, finding no significant

correlation between cheating and sorority or fraternity

involvement. A positive correlation did exist between self-

reported church attendance and cheating. However, the

authors pointed out that church attendance is frequently

over-reported in survey studies.

Harp and Taietz (1966) conducted a study of term paper

cheating among students in three colleges. Cheating was

found to be more prevalent among fraternity members, and

engineering students who belonged to a fraternity had the

highest level of cheating. The authors suggested that

cheating among these students is increased by the

availability of opportunity afforded by the testing methods

used in the college.

In a study by Baird (1980) 200 college students answered

self report questionnaires. Fraternity and sorority members

admitted to more cheating than did independent students. The








students who admitted cheating in college also admitted a

higher incidence of cheating during high school.

In the previously cited study by Haines, Diekhoff, LaBeff

and Clark (1986) a self report questionnaire was administered

to 380 college students. Respondents were asked to indicate

their past involvement in twelve different types of cheating

behavior. Admitted cheaters were more involved in sorority

and fraternity activities, as well as sports activities.

They were also less likely to be employed than were those who

did not report cheating. The authors suggested that the

cheater has less of an economic stake in his or her

education, and thus fears sanction less than students who

have to support themselves financially. Also, greater

involvement in outside activities takes away from study time

and may lead to an increased reliance on cheating to maintain

grades.

Liska (1980) administered a questionnaire to 359 college

students in which they were asked to report involvement in

seven different cheating behaviors; plagiarism, lying to an

instructor to take an examination late; seeking and receiving

help on a take-home examination, using the same term paper in

two or more classes, securing a copy of an old examination to

prepare for a forthcoming examination, receiving information

about the contents of an examination from students who had

already taken the examination (p. 77-78). The researchers

found that both the subjects' attitude and association with








others involved in deviant behavior were found to be

correlated with cheating behavior.

In a study of 120 college undergraduates, Newhouse (1982)

measured cheating with an instrument developed by Lewis

(1965), in which subjects interpreted 19 items as cheating or

non-cheating situations. The subjects were also given a

modified version of Scrole's Scale of Anomia (Clinard, 1964)

to assess the degree of alienation that the subject felt

towards the faculty, student body, services, administration,

and acquaintances. Students with high levels of alienation

had higher incidence of cheating than those with low levels

of alienation. The authors suggested that those divorced

from the social fabric of the college may have a greater

tendency to cheat than those who are committed socially to

the institution.

In a study conducted by Lanza-Kaduce and Klug (1986) an

anonymous survey was given to 196 undergraduates asking them

to report their cheating behavior and that of their best

friends. A differential association was obtained between the

level of cheating engaged in by the subject, and the extent

to which the subject reported best friends were involved in

cheating. Moral development was also assessed by Rests

Defining Issues Test (1976) and moral development was

compared to differential association scores. Moral

development was found to interact with differential

association among those with low and medium level moral

development. The authors suggested that this interaction








indicates that those with lower levels of moral development

are more influenced by peers than are those with higher

levels of moral development.

Parental Factors

Several researchers have addressed the relationship of

cheating to parental factors. Parr's (1936) used a self

scoring study, finding that cheating was more prevalent among

students who had parents in vocational occupations (64%) than

among those who indicated that their parents were in

professional occupations (30%).

Vitro (1971) conducted a study of 70 female junior

college students enrolled in a human anatomy and physiology

course. Subjects were given the chance to self-score their

own examinations. The study analyzed the relationship of

cheating propensity to subjects' perceptions of parental

disciplinary methods employed in the child's early years.

The parental disciplinary method was determined by the

Parental Punitive Scale (PPS) developed by Epstein and

Komorita (1965).

Vitro found that of the 70 subjects, 34 were cheaters

(45.6%) and 36 (51.4%) were non-cheaters. The measure between

mother punitiveness and cheating was not significant. The

relationship between father punitiveness and cheating was

significant. Cheaters reported parental disciplinary

techniques characterized by excess in either punitiveness or

laxity. Non-cheaters reported moderate disciplinary

techniques. The authors concluded that if a female child was








subjected to excessively punitive discipline, she probably

learned to restrict her activity because of the risk of being

detected and the thought of facing severe punishment. As a

result, moral judgments were more likely to be based on

expediency and external factors than on internal control. At

the other extreme, female children who had suffered little if

any discipline retained strong infantile impulses. No

external control of these impulses had been established by

the parent and the child had not been able to develop

internal control of these impulses. Thus, she acted on these

impulses with little reluctance.

In a study by Kelly (1978) 592 undergraduate students

were given the opportunity to self-score their examinations.

The Parent Behavior Form (Worell & Worell, 1975) was given to

measure parental qualities. Kelly found that 115 of the 592

subjects (19.5%) cheated. Fifty seven of the 236 males

(24.2%) cheated compared to 58 of 355 females (16.3%). The

father behavior scales were not correlated with cheating for

male subjects. The behavior scale of Mother Lax Control was

slightly higher for male cheaters than for male non-cheaters.

Female cheaters reported less Father Warmth, and greater

Father Achievement Control, Mother Rejection and Mother

Hostile Control than female non-cheaters.

Research by Yarrow, Campbell and Burton (1968) and

Hartshorn and May (1928), supports the relationship between

parental warmth and honesty. Research by Mussen (1970), and

Sears, Rose and Alpert (1965) on nursery school students








suggested that parental warmth is not a necessary requirement

for honest behavior. In fact, both studies showed no

correlation between resistance to temptation and parental

warmth. Instead, the high standards of both parents and, in

particular, the father's belief in the importance of teaching

the meaning of right and wrong, appeared to be the strongest

indicators of honest behavior of the nursery school students

under the four experimental situations studied.

Hart (1988) conducted a study of the moral development of

male subjects over a 20 year period, and indicated that

higher levels of involvement with the father while the

subject was a child correlated positively with higher levels

of moral development throughout early adulthood. Maternal

involvement showed no significant relationship to moral

development. In a study of 120 families, Steinberg, Elmen

and Mounts (1989) analyzed the relationship between parenting

and academic success. All three aspects of authoritative

parenting that were measured; curfew, leisure activities and

decision making, were found to be positively related to

academic success.

Peer Reaction

Several researchers have examined the relationship of

cheating and the reaction of one's peers. Liska (1978)

conducted a study of 359 college students which involved a

self-report questionnaire regarding cheating behavior and

what the subjects thought there parents and peers reaction

would be to cheating. Liska indicated that the perceptions








of peers may slightly decrease cheating behavior. In a

survey study of 1500 students at Iowa State University

conducted by Barnett and Dalton (1981), a majority of

students surveyed indicated that their peers only slightly

disapproved of cheating. Only 19% of the subjects indicated

that a close friend would strongly disapprove of their

cheating behavior.

In a study by Lanza-Kaduce and Klug (1986) an anonymous

survey was given to 196 undergraduates. Subjects were asked

to report their past involvement in cheating behavior and to

indicate how their best friends would feel about their

cheating. The relationship between peer reactions and self-

reporting of the frequency of cheating on tests within the

semester during which the survey was given, was found to be

significant in the group with lower levels of moral

development as measured by Rest's Defining Issues Test

(1976).

Reward and Punishment

Several researchers have analyzed the relationship

between cheating and the punishment that would be incurred

for cheating, or the reward that would be gained from

refraining from cheating. In the previously cited study by

Fischer (1970) 135 fourth through sixth graders were given

the opportunity to cheat during an achievement test.

Cheating was defined as the subjects' use of an answer page

rather than his or her own knowledge in answering a 60 item

multiple choice test. Fischer indicated that punishments








announced before the examination resulted in lower levels of

cheating.

In a study by Tittle and Rowe (1973) students were given

the opportunity to grade their own quizzes. Treatment groups

were notified of sanctions that would be imposed if they were

caught cheating. Cheating behavior was significantly lower

for the groups that received the threats than for the groups

who did not.

Winston (1978) conducted a study of three grade school

boys in which the subjects were given the opportunity to

cheat on arithmetic problems. The subjects received various

rewards and punishments for cheating and for admitting to

cheating. When praise and reward were given for admitting to

cheating, the rate of admission increased to 87.5%, but did

not affect the frequency of cheating behavior. When praise

and punishment were given for admitting cheating, the

admission of cheating decreased to 0%. The frequency of

cheating did not change.

In a study of 245 college students conducted by Gardner,

Roper, Gonzalez and Simpson (1988) students were given the

opportunity to cheat on a study guide assignment by referring

to an answer sheet in the back of the book. Praise and

punishment was manipulated as result of admitting cheating.

Results were consistent with those of Winston (1978).

Although the admission of cheating was influenced by rewards

and punishment, the frequency of cheating was not.








Burton's (1976) discussion of incentive and dishonesty

analyzes several other studies including; Mills (1958),

Brodsky and Jacobson (1970), Vitro (1969), and Merrit and

Fowler (1948), finding that incentive and risk may interact

in dishonest behavior. Burton suggested that if risk is held

constant, the likelihood of one cheating is related to the

reward that is expected for success.

Neutralization

Neutralization is one of several types of social learning

theories that are offered to explain deviant behavior. The

fundamental premise of the three types of social learning

theory (differential association, differential reinforcement

and neutralization) is that deviant behavior is a product of

learning the norms and values associated with deviant

behavior. Social learning can involve the techniques of

crime or deviancy (breaking into a house, for example) as

well as the psychological techniques to rationalize or

neutralize the guilt associated with deviant behavior.

Sykes and Matza (1957) questioned Cohen's (1955)

subcultural theory of juvenile delinquency and sought to

clarify elements of Sutherland's (1947) differential

association theory by arguing that much of the delinquent

behavior among juveniles occurs because adolescents, who

experience pressure to conform with the larger society and

pressure to conform with a subculture of differing values,

devise a set of psychological defenses which allow them to

justify participation in deviant behavior. Sykes and Matza








maintained that most delinquents hold conventional values and

attitudes but master "techniques of neutralization" that

allow them to neutralize these conventional values and

"drift" back and forth between conventional and "subterranean

behaviors" (Matza, 1964; Matza & Sykes, 1961). Sykes and

Matza maintained that juveniles develop a distinct set of

justifications or rationalizations for their behavior when

this behavior violates societal norms. Sykes and Matza based

their theoretical model on several observations:

1. Juvenile offenders sometimes have a sense of guilt

over their illegal acts. If they had a stable

delinquent value system it would be unlikely that they

would feel remorse other than for being caught.

2. Juvenile offenders frequently respect and admire

honest, law-abiding people.

3. Juvenile offenders make a distinction between those

who they can victimize and those who they cannot.

Members of similar, social, ethnic or religious groups

are many times off limits.

4. Juvenile offenders frequently participate in the same

social functions as law-abiding youths, such as school

and family activities. (Siegel, 1986 p. 197)

Sykes and Matza have divided these techniques which

temporarily neutralize conventional value systems into five

major types. Klockars 91974) and Minor (1981) have proposed

two additional techniques of neutralization that are included

in numbers six and seven.








1. Denial of Responsibility- Offenders claim that their

actions were accidental or resulted from forces

outside their control.

2. Denial of Injury- Offenders claim that nobody is hurt

by their actions.

3. Denial of Victim- Offenders claim that although

someone may have been hurt by their actions, it is a

form of rightful retaliation or punishment.

4. Condemnation of the Condemners- Offenders claim that

those who criticize his/her actions are hypocrites.

5. Appeal to Higher Loyalties- Offenders claim that they

are caught in a dilemma of being loyal to their own

peer group while attempting to abide by norms of the

larger society. The demands of the localized peer

group take precedence because these demands are

immediate. (Sykes & Matza, 1957)

6. Defense of Necessity- Offenders claim that an act is

necessary so they need not feel guilty about its

commission. (Minor, 1981).

7. Metaphor of the Ledger- Offenders claim that they have

done sufficient good to allow them an indulgence in a

dishonest act. Acceptable behavior is viewed as

accruing good credits which can be "traded in" for

participation in deviant acts. (Klockars, 1974; Minor,

1981).

In more common terms, the first five neutralization

techniques describe the rationalizations of "I didn't mean







it", "I didn't really hurt anybody", "They had it coming to

them", "Everybody's picking on me" and "I didn't do it for

myself" (Sykes & Matza, 1957, p. 669). The last two

techniques describe the rationalizations of "I had no choice"

and "I've earned it". According to neutralization theory,

the juveniles most likely to continually violate societal

norms are those who can most successfully rationalize their

behavior through the techniques of neutralization.

Neutralization theory has been used to explain a variety

of deviant behaviors including; having an abortion (Brennan,

1980), dishonesty in selling shoes and selling automobiles

(Friedman, 1974), smuggling (Paulus and Simpson, 1981),

becoming a taxi-dancer (Hong & Duff, 1977), pedophilia (De

Young, 1988), homosexual prostitution (Salamon, 1989), heroin

use (Covington, 1984) and becoming a hit man (Levi, 1981).

While each of these studies have shown some support for the

neutralization theory of deviancy, they have employed

differing methodologies in exploring the application of the

neutralization theory.

Brennan's (1980) paper presented a philosophical analysis

of how several forces including abortion legalization,

advancing abortion technology, the establishment of abortion

as a valid medical procedure, population pressures, scarcity

of resources and quality of life considerations have provided

authenticity to neutralization rationalizations for abortion

(p. 358). Friedman (1974) presented a case studies of two

sales people to show that neutralization techniques








consistent with Sykes and Matza's theory were utilized by

these sales people in justifying their dishonest practices.

Paulus and Simpson (1981) used a survey of 700 persons

crossing between Canada and the United States to show that

neutralization techniques affect the incidence of smuggling

low cost materials between the two countries.

Kong and Duff (1977) conducted interviews with more than

70 dances hostesses in Los Angeles over a four year period to

determine the applicability of neutralization to the choice

of this occupation. Kong and Duff found that taxi dancers

who persist in their occupation engaged in neutralization

techniques with the effect of facilitating commitment to the

job, and that the type of justification engaged in by these

subjects changes over time.

De Young (1988) analyzed printed material from three

American pedophile organizations in order to find what

neutralization techniques might be engaged in by these

organizations to minimize or disavow their deviance. She

found that all of the material studied utilized the

techniques of denial of injury, denial of the victim,

condemnation of the condemners and appeal to higher

loyalties. Salamon (1986) conducted interviews with one

homosexual escort agency owner and twenty-eight male escorts

and found that the moral questionability of this service was

neutralized by the organization of names, space and

structure. While Salamon does not discuss Sykes and Matza's








theory specifically, his findings regarding a general theory

of neutralization are consistent with Sykes and Matza.

Covington (1984) conducted interviews with 72 heroin

users in metropolitan Detroit public treatment programs to

determine if their acceptance or rejection of labels would be

affected by the defense mechanisms that they employed, the

degree of labeling that they experienced and their background

characteristics. Neutralization in the form of

externalization of blame (denial of responsibility) was found

to be an effective defense in minimizing respondent self-

rejection, thus fending off societally imposed labels. Those

respondents who adopted a denial of responsibility technique

of neutralization experienced a greater resistance to

rehabilitation programs than those who did not engage in this

technique. The defense of condemning the condemners was not

significant in insulating the heroin user from self-

rejection.

Levi (1981) conducted interviews with one incarcerated

"hit man" in metropolitan Detroit. Levi found that this

subject engaged in several neutralization techniques

including the denial of victim, the denial of injury, and the

denial of responsibility. The study also chronicled a

progression of participation in this deviant career which is

presented as a "process of reframing, [where] the experience

of victim-as-target emerged as the 'main story line,' and the

experience of victim-as-person was downgraded from the main

track to the disattend track to the overlay track" (p. 61).








The reframing hypothesis presented by Levi supports the

hypothesis of Sykes and Matza that individuals are able to

psychologically "drift" between socially acceptable and

deviant acts.

Despite the existence of these colorful studies, the area

in which neutralization has been most extensively explored

and debated is the area for which it was first proposed, that

of juvenile delinquency. Mitchell and Doder (1980)

administered questionnaires to 694 social science students at

Pan American University and at Oklahoma University in order

to explore the relationships between father's occupation,

delinquent peer association, tendency to neutralize, and

self-reported delinquency. Using path analysis, on the

tendency to neutralize delinquent behavior, and self-reported

delinquency, the researchers revealed that the strongest

relationships existed between delinquency and (a) the

tendency to neutralize and; (b) delinquent peer association

(p. 247). It was also found that as the seriousness of self-

reported delinquent acts increased, the tendency to

neutralize decreased. Neutralization was also found to be a

more viable explanation for delinquency (especially less

serious delinquency) among Anglos than among Mexican

Americans. However, the relationship between delinquent

peers and self-reported delinquency was stronger among

Mexican Americans than among Anglos.

The authors postulated that "Mexican Americans may be

more likely to fit into Sutherland and Cressy's (1976)








differential association [another type of social learning

theory of criminal deviancy], in which peers are relied upon

for support in committing and justifying delinquent acts

somewhat independently of internalizing a normative

structure" (p. 247). The findings of Mitchell and Dodder

raise the question whether neutralization may be more

adequately understood as specific techniques related to

certain types of specific acts, and suggest that racial

differences may exist in the use of neutralization

techniques.

Mitchell and Dodder (1983) conducted a questionnaire

study of 298 public high school students and 53

institutionalized male delinquents in Oklahoma. Delinquent

acts were measured by a modified version of the Nye-Short

(1958) delinquency scale. Subjects' tendency to neutralize

were measured using a revised version of Ball's (1966)

inventory of hypothetical deviants acts. Results of the

study included the following:

1. The finding that "each technique of neutralization was

accompanied by varying degrees of agreement as to its

efficacy as a solution to the hypothetical situation

illustrating a deviant act. This suggest that

attitudes toward neutralization vary according to the

particular type of neutralization" (p. 317);

2. The finding that there is a consistent pattern of

agreement among the five neutralization techniques

across the subsamples. In each case acceptance of the




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ESCEI4UMV_0GU0L3 INGEST_TIME 2012-12-07T21:09:15Z PACKAGE AA00012911_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES