 

 

 

\
“m LIBRARY
Michigan State
.... University

 

 

This is to certify that the
thesis entitled

SOCIAL NORMS AND ACADEMIC DISHONESTY

presented by

ERIC BEASLEY

has been accepted towards fulﬁllment
of the requirements for the

MASTER OF degree in SOCIOLOGY
ARTS

 

 

{Emma/t CW

Major Proféssor’s Signature

1 2-07-09

 

Date

MSU is an Aﬁirmative Action/Equal Opportunity Employer

u----------o-v--a--v---u-u--d--n--.c-o

-I-u-I-I-l-l-o-I-0-I-I-l-l-I-i-I---l-I--I90--o-I-n-l-n-u‘r--l-I-l-l-O-V.

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Aoc&PresIClRC/Daleoue.hdd

SOCIAL NORMS AND ACADEMIC DISHONESTY
By

Eric Beasley

A THESIS

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

MASTER OF ARTS
Sociology

2009

ABSTRACT
SOCIAL NORMS AND ACADEMIC DISHONESTY
By
Eric Beasley
This article aims to evaluate the salience of the perception of the beliefs and

behavior of peers as a predictor of students’ own beliefs and behaviors regarding
academic dishonesty. Particular attention is paid to discerning the relative predictive
power of different peers or peer groups (e. g. the average student, your best friend, etc.).
Speciﬁcally this study offers support for the following conclusions (especially in regards
to academic dishonesty): students over-perceive their peers’ delinquency; these
misperceptions increase as social distance increases; there is a positive correlation
between one’s perception of 1118 peers’ delinquent behavior and one’s own; and the
strength of this correlation generally lessens as the social distance of the peer group being

referenced increases.

ACKNOWLEDGMENTS

The author would like to thank Dr. Tom Conner for his assistance in the
completion of this project. An extension of thanks is also offered to Dr. Stan Kaplowitz,
Dr. Soma Chaudhuri, and Dr. Ralph Pyle for their insightful comments and suggestions.
The author would also like to thank his fellow graduate students for their words of advice
and encouragement along the way. Lastly, the author would like to express his gratitude

to Meaghan Carpenter for the motivation she provided and the loving support she gave.

iii

TABLE OF CONTENTS

LIST OF TABLES
INTRODUCTION ..................................................................................................... l
HYPOTHESES .......................................................................................................... 6
METHODS ................................................................................................................ 9
Participants .......................................................................................................... 9
Survey ................................................................................................................. 10
RESULTS .................................................................................................................. 1 1
DISCUSSION ............................................................................................................ 18
Limitations .......................................................................................................... 23
Implications ......................................................................................................... 25
REFERENCES .......................................................................................................... 27

LIST OF TABLES

TABLE 1 .................................................................................................................... 13
Comparing Student Self-Report Frequency of Misconduct with Their
Perceptions of Their Peers across Five Different Forms of Academic Dishonesty

TABLE 2 .................................................................................................................... l4
Self-Report and Perception of Various Peers’ Ethical Beliefs on Academic
Misconduct

TABLE 3 .................................................................................................................... 15
Mean Differences and Correlations of Self-Reports and Perception of Various
Peers’ Frequency of Academic Misconduct and Related Ethical Beliefs

TABLE 4 .................................................................................................................... 16
Self-Report and Perception of Various Peers’ Frequency of Academic
Misconduct

Introduction

It has been well documented that academic dishonesty regularly occurs on college
campuses with some reports indicating that almost 90% of college students cheat at least
once during their academic career (Brown and Choong 2005; Sims 1993). Although it
can be reasoned that academic dishonesty has been around as long as academic
institutions have, it seems that the frequency and intensity of the dishonest acts are on the
rise (Drake 1941; McCabe and Trevino 1997). In fact, cheating on college campuses has
become such a problem that some researchers have deemed it an epidemic (Lambert and
Hogan 2004). But what inﬂuences student cheating?

There is an abundance of literature that seeks to explain why students commit
academically dishonest acts, and a variety of explanations have been offered and tested
using various methodologies (Bolin 2004; McCabe 1999; Cochran et a1 1999; Michaels
and Miethe 1989; etc.). One of the largest studies was conducted by McCabe and Trevino
(1997), who ran a multi-campus survey of nine mediurn-to-large state universities
looking to identify contextual and individual factors that inﬂuenced academic dishonesty.
They discerned that of all the factors that were tested, peer disapproval provided the most
inﬂuence on academic dishonesty. Speciﬁcally, the authors identiﬁed that the stronger an
individual thought that her peers would disapprove of her cheating, the less often she
reported she cheated. This sort of ﬁnding corroborated an earlier ﬁnding of theirs, which
was that self-reports of academic dishonesty are more highly associated with perceptions
of peers’ behavior than many other factors (McCabe and Trevino 1993).

Many theories and models exist to help one understand the potential logic and

mechanisms that cause there to be such a salient relationship between deviant peer

behavior and one’s own deviant behavior. One of the most tested and used is social
learning theory (Akers and Sellers 2006). The cornerstone of social learning theory is
that all human behavior, whether it is deviant or conforming, is learned in the same way.
Inﬂuenced by symbolic interactionism, social learning theory posits that what sways
someone towards conformity or deviance is the type of things that are learned in
everyday interaction. Thus, those who associate with deviants are rewarded for being
deviant and attach positive meaning to deviance. This, in turn, makes them more likely
to commit deviant acts (Akers and Sellers 2004). For example, a person who spends his
time socializing with people who steal cars and is rewarded when he himself steals a car
will think positively about car stealing and be more likely to steal a car than a person
without any car stealing ﬁiends who has not been rewarded for being deviant.

A way to explore some of the mechanisms at play in social learning theory more
deeply is through the lens of social norms theory (Berkowitz 2004). Social norms theory
is harmonious with social learning theory in that it ﬁts under the umbrella of ideas that
social learning theory examines but speciﬁcally looks at only one element contained
under the totality of social learning theory. Social norms theory, like social learning
theory, is concerned with differential association. It asserts that the way individuals
perceive the behavior and beliefs of their peers inﬂuences their own behavior. Thus,
those who an individual associates with are a very powerful predictor of what that
person’s beliefs and actions will be. Unlike social learning theory, which concerns both
the normative and interactional dimensions of differential association and the subsequent
deﬁnitions, reinforcement, and direct imitation that may likely follow (Akers and Sellers

2004), social norms theory concentrates speciﬁcally on the inﬂuence of perceived peer

norms on an individual. Speciﬁcally, it posits that an erroneous overestimation by an
individual with regards to the frequency in which his peer group does a particular activity
or has a particular opinion will lead to an increased likelihood that he will do that activity
or subscribe to that opinion (Berkowitz 2004).

Similarly to social learning theory, there are numerous studies that empirically
validate social norms theory (Mattem and Neighbors 2004, Borsari and Carey 2001,
Perkins 2002, etc). Most of the empirical research on social norms theory, however, has
been done on one substantive area: collegiate alcohol use. Despite the difference in
dependent variables, the research that has been done with social norms theory and
collegiate alcohol use can possibly help us understand academic dishonesty among
college students.

Since the extant research on academic dishonesty points to peer behavior as one
of the most salient correlates of self-reported academic dishonesty (Whitley 1998,
McCabe and Trevino 1997), and since there is a rapidly expanding literature on social
norms theory (complete with validating empirical tests using multiple methods, see
Berkowitz 2004), it makes sense to examine collegiate academic dishonesty from the
perspective of social norms theory.

Hard, Conway, and Moran (2006) explicitly applied social norms theory to academic
dishonesty. Their ﬁndings indicate that students overestimated how often their peers
were academically dishonest, and that they found a statistically signiﬁcant positive
correlation between what the students thought of their peers and their own self-reported

academic misconduct. Following the logic of social norms theory, this misperception is

liable to lead towards a higher frequency of academically dishonest acts being committed
(Berkowitz 2004).

A shortcoming of Hard et al.’s (2006) paper is that it measures only one
conception of a peer group, speciﬁcally “other university students.” As Berkowitz (2004)
notes, individuals belong to a variety of groups. Each group that a person is a part of is
different; thus, an individual is inﬂuenced differentially, both in intensity and direction,
by the various groups of which he perceives he is a member. For instance, an
individual’s recreational softball teammates might compel him to spend a couple of hours
socializing at the local tavern with them, but that compulsion might be overridden by his
membership and allegiance to a religious organization that disapproves of alcohol
consumption. So, it is not that measuring self-reported differences between an
individual’s own behavior and that of “other university students” is not useful; it is rather
that a deeper understanding of how peers might inﬂuence behavior can be gained by
assessing the relative saliencies of the different peer groups to which one belongs.

Some research has been done on comparing the relative misperceptions and
correlations of several self-other relationships that differ by reference group. Differences
dependent upon the type of peer group referenced were found, underscoring the idea that
as social distance increases, so do the size of the misperceptions (Berkowitz 2004,
Borsari and Carey 2003). In other words, a person’s misperceptions about the thoughts
and actions of a distant acquaintance are likely to be greater than misperceptions about a
close friend (Borsari and Carey 2003). In addition, current research shows that perceived

norms for people who are the least socially distant are stronger predictors of an

individual’s beliefs and actions than are the perceived norms for people who are more
socially distant (Borsari and Carey 2003, Korcuska and Thombs 2003).

This ﬁnding validates the worth of research that distinguishes clearly and
speciﬁcally the reference group being used as the peer variable since it is clear that an
individual’s actions and beliefs are differentially related to different peer groups.
Furthermore, if one person was asked about his peers, he might think of his three closest
friends, while another individual might answer thinking of all others within a couple
years of his age who live in his city. These answers to the same question would likely
produce disparate results because they are measuring different things. Therefore, along
with the need to measure different peer groups in trying to ascertain the effect of peers on
individuals, there is also a need for researchers to standardize and use speciﬁc terms in
assessing peer group inﬂuence (e. g. using “The average undergraduate student” rather
than “your peers”) (Borsari and Carey 2003). This is because it is hard to deduce
relationships from a cross-study synthesis without the widespread use of common terms
as variables. It cannot be said with much analytic clarity that the predictive power of
peers on some phenomenon is X if there is not uniformity in the peer variable. When one
study asks students to rate their “peers” and another asks them to rate their “friends” it is
hard to compare the ﬁndings of the two studies. Also, since it has been illustrated that
different peer groups do have varying effects, understanding of the process will be
enhanced if there is identiﬁcation and more thorough understanding of the unique
differential effects of speciﬁc peer groups on individual behavior and thoughts. Once
again, however, most of the research that points towards varying saliency amongst

different peer reference groups is based on collegiate alcohol use. Thus, there is a need to

evaluate the relative predictive power of different peer groups on academically dishonest
behaviors and thoughts.

Although Jordan (2001) did measure the effect of a variety of peer groups on
academic dishonesty, there has not been enough cross-study uniformity of peer variables
or congruency of stem questions to make for a critical analysis of the varying predictive
power of different peer groups on academic dishonesty. Furthermore, Jordan’s study
included only three conceptions (all students, students in resident hall, and ﬁiends) of
peer groups. Thus, even within a singular study it is hard to make conclusions about the
varying effects of peer groups. The present study plans to address these needs through the
concurrent examination of the relative predictive power of different peer groups using
uniform stem questions. Moreover, in the interest of substantiating earlier results and
building a literature that can be viewed more as a totality, the study will employ much of

the same experimental design as Hard et al. (2006).

Hypotheses

For the following hypotheses the word “peers” will account for the six different
peer groups (“Your best friend,” “Your signiﬁcant other,” “All of your current and
former college roommates [on average],” “The average MSU undergraduate [in your
major],” “The average MSU undergraduate [all majors],” and “Your ﬁve closest ﬁ'iends
on campus [on average].”) that were tested and their aggregate. In addition, “social
distance” will be seen as a continuum with the ordering of peer groups from least socially
distant to most socially distant: “Your signiﬁcant other,” “Your best friend,” “Your ﬁve

closest ﬁ’iends on campus (on average),” “All of your current and former college

roommates (on average),” “The average MSU undergraduate (in your major),” “The
average MSU undergraduate (all majors)”

Hypothesis 1: Students over-estimate their peers’ ethical approval of academically
dishonest acts.

Hypothesis 2: Students over-estimate the frequency in which their peers take part in
academically dishonest acts.

Hypothesis 3: The size of the misperceptions students have about their peers’ degree of
ethical approval of academically dishonest acts increases as peer group social distance
increases.

Hypothesis 4: The size of the misperceptions students have about the frequency in which
their peers take part in academically dishonest acts increases as peer group social distance
increases.

Hypothesis 5: A positive correlation exists between one’s own behavior/beliefs
regarding academic dishonesty and the behavior/beliefs she perceives her peers
participating in and believing.

Hypothesis 6: The predictive power of the perception of peer beliefs and behaviors and
one’s own beliefs and behaviors regarding academic dishonesty decreases as peer group
social distance increases.

Hypothesis 7: Presenting a colloquial deﬁnition of what constitutes academic dishonesty
may mediate the disparity between students’ own beliefs and behavior regarding
academic dishonesty and the perception of their peers corresponding beliefs and

behaviors.

Besides adding to the literature, this study can serve as a guiding tool in setting up
a social norms intervention to reduce academic dishonesty. Social norms interventions
seek to correct inﬂuential misperceptions that people have about their peers by providing
factual information, which is hypothesized to change actual behavior and values
(Berkowitz 2004). Much empirical data has been gathered in support of the efﬁcacy of
this approach (Mattem and Neighbors 2004, Scher, Bartholow, and Nanda 2001,
Berkowitz 2004, etc.), and although social norms interventions have been targeted at
correcting a variety of self-other misperceptions, once again, most of the interventions
have targeted collegiate alcohol use.

Therefore, there is room for development and testing of a social norms
intervention aimed at lowering the documented pervasive over-perception of peers’
approval and acting out of academically dishonest acts, which, following social norms
theory (Berkowitz 2004), will lead to actual lower approval and misconduct overall. In
choosing how to most effectively administer an intervention, it would help to know the
relative predictive power of a variety of peer groups and the size of the self-other
misperception on each particular group. The knowledge would be even more beneﬁcial
if it were speciﬁcally based on the particular phenomenon that the intervention is
attempting to inﬂuence. Knowledge of these principles derived from alcohol use studies
is useful, but it is quite worthwhile in attempting to establish the framework of an
effective social norms intervention on academic dishonesty to seek information that is

solely based upon academic dishonesty.

Methods

In the summer of 2009 several instructors of in-person summer courses at a large
public Midwestern university were contacted via email to see if they would be willing to
allow their students an opportunity to participate in an anonymous and voluntary survey.
Ten instructors agreed to allow the survey to be administered either immediately
following or during their class. One instructor offered extra credit to his students for their
participation in the study. The courses surveyed tended to be mostly social science
classes as the instructors for those classes were both, disproportionately contacted out of
convenience, and disproportionally agreed to participate. The students were told by the
researcher that a survey was to be administered that seeks to help further understanding
of why people both approve of and commit academically dishonest acts. It was
explicitly stated that the students’ participation was completely voluntary and that their
names would in no way be attached to their answers. The great majority of students
chose to participate.
Participants

There were 184 surveys turned in. Although not every student answered every
question (it was an option not to answer if the question did not apply to the individual),
and a few students seemed to just hurriedly mark the same answer for every question, no
surveys were excluded from all analyses. The sample was 66.8 percent female and 33.2
percent male, with every respondent declaring their gender. Racially, the sample was
70.1 percent White, 17.9 percent Black, 5.4 percent Asian, 4.3 percent some other race,
1.1 percent two or more races, and .5 percent American Indian. One respondent left the

race category blank. As is quite characteristic of in-person summer courses, most of the

respondents were upperclassmen. The sample was 56.0 percent Seniors, 37.5 percent
Juniors, 3.3 percent Sophomores and 3.3 percent of the sample responded “Other.” Most,
62.0 percent, of the respondents reported having a major in the Social Sciences and 22.8
percent of students marked “Other” when asked about their academic major. All other
major areas were thoroughly underrepresented in this sample. The average age of the
respondents was 21.2 years old (note that the category of 24 years of age and older was
coded as “24”).
Survey

Respondents were asked to report their age, class standing, major area, sex and
race. Following these demographic questions, on roughly half of the surveys (58.2
percent), was a paragraph that described in colloquial terms what constitutes academic
dishonesty (the other surveys had a blank spot where the paragraph was). This paragraph
was adapted from Michigan State University’s ofﬁcial deﬁnition and read:

A person is being academically dishonest or “cheating” if he does any of the

following things: he claims that another’s work is his own; obtains an

examination or assignment without authorization to do so; he does an assignment

or examination for another person; he tampers with the academic work and (or)

research of another person, he lies about his data.

(Condensed from https://www.msu.edu/unit/ombud/RegsOrdsPolicies.html) *note

that the “he” pronoun was used on the survey but is not used in MSU’s official

deﬁnition.

In the next part of the survey students were asked, to the best of their knowledge,

how often they and others had performed certain academically dishonest acts. In all, the

10

students reported their own behavior and that of six peer groups (“Your best friend,”
“Your signiﬁcant other,” “All of your current and former college roommates (on
average),” “The average MSU undergraduate (in your major),” “The average MSU
undergraduate (all majors),” and “Your ﬁve closest friends on campus (on average).”) on
ﬁve different types of academic dishonesty. The ﬁve types of academic dishonesty were
planned and unplanned copying of another’s paper, allowing another to copy your paper
during an exam, submitting another person’s material as your own, and copying
information from Internet websites and submitting it as your own. The students were
asked to select answers on a ﬁve point scale: never, once or twice, three to ﬁve times, six
to ten times, or more than ten times. Additionally, they were asked to skip a question if it
did not apply to them. The types of academic dishonesty and the scale were adapted
from Hard et al. (2006) in the interest of comparing earlier results and adding to the
literature. Following the section that asked about the frequency of committing
academically dishonest acts was a section that inquired about how ethical it was to
commit the same acts. The responses in the scale used were as follows: very unethical,
unethical, neither ethical nor unethical, ethical, and very ethical. Once again, students
were also instructed to skip a question if it did not apply to them.

Results

In order to test the hypotheses put forth in this study more succinctly, some data
was aggregated. All peer group responses, in both the frequency and ethical belief
section, for each speciﬁc form of academic dishonesty were then averaged forming two
new variables used in Table 1: “Student Mean Frequency of Behavior Belief of
Aggregated Others” and “Student Mean Ethical Belief of Aggregated Others.” The

average response of all peer groups, across all ﬁve forms of academic dishonesty, for

11

both frequency and ethical belief questions respectively, are designated as the variable
“Aggregated Others” in Tables 2 and 3. In Table 4, the peer group data is based upon
answers for that particular peer group, in either the frequency or ethical belief categories,
across all ﬁve forms of academic dishonesty.

Hypothesis 1: Students over-perceive their peers’ degree of ethical approval of
academically dishonest acts.

Hypothesis 2: Students over-perceive the frequency in which their peers take part
in academically dishonest acts.

On average, the results show that students reported that they both committed less
often and were more ethically approving of each form of academic dishonesty than their
peers (see Table 1). The average difference between the Student Mean Self Frequency of
Behavior and the Student Mean Frequency of Behavior Belief of Aggregated Others was
.38 with the differences ranging from .36 to .41 depending on the type of academic
dishonesty. The average difference between the Student Mean Self Ethical Belief and the
Student Mean Ethical Belief of Aggregated Others was .28 with differences ranging from
.21 to .36 depending on the type of academic dishonesty.

The act that students reported being involved with most frequently was copying
from another’s paper. The mean score for students’ self report on copying from another’s
paper was 2.01 (on the response scale a 2 corresponded with having participated in the
behavior once or twice). Copying from another’s paper was also one of two acts that
students were most ethically accepting of receiving a mean score of 1.96 along with the
act of allowing another to copy on an exam (a 2 on the ethical response scale represented

thinking the behavior was “Unethical” whereas a I meant thinking it was “Very

12

unethical” and a 3 “Neither ethical nor unethical”). For the Aggregated Others scores,

answers were only included if the respondent provided an answer for each peer group on

that particular form of academic dishonesty.

Table 1

Comparing Student Self-Report Frequency of Misconduct with Their Perceptions of Their Peers across

Five Different Forms of Academic Dishonesty (Standard Deviations Appear Parenthetically)

 

 

 

 

 

 

 

Type of Student Student Mean Self-Other Student Student Self-Other

Academic Mean Self Frequency of Mean Mean Self Mean Mean

Dishonesty Frequency of Behavior Frequency Ethical Ethical Ethical

Behavior Belief of of Behavior Belief Belief of Belief

Aggregated Difference Aggregated Difference
Others Others

Planned in 1.86 (1.08) 2.21 (0.84) 0.36 1.78 (0.85) 2.09 (0.73) 0.31

advance and N=l 79 N=124 N=l 83 N=132

then copied

ﬁom another

Copied from 2.01 (1.07) 2.42 (0.95) 0.41 1.96 (0.83) 2.17 (0.65) 0.21

another N=l83 N=126 N=184 N=132

person’s

paper

Realized 1.66 (0.91) 1.99 (0.81) 0.33 1.96 (0.91) 2.17 (0.74) 0.21

during an N=182 N=130 N=184 N=132

exam that

another

student

wanted to

copy from his

or her paper,

and allowed

that student to

COPY

Submitted 1.24 (0.64) 1.65 (0.61) 0.41 1.39 (0.65) 1.75 (0.65) 0.36

another N=182 N=l3l N=184 N=135

person’s

material as

one’s own

Copied 1.52 (0.83) 1.93 (.076) 0.41 1.63 (0.72) 1.92 (0.67) 0.29

information N=l 83 N= 1 30 N=183 N=130

from Internet

websites and

submitted it

as his or her

own work

 

 

 

 

 

 

 

 

Hypothesis 3: The size of the misperceptions students have about their peers’
degree of ethical approval of academically dishonest acts increases as peer group
social distance increases.

The “You” mean ethical belief was lower than that of any of the peer groups,
although the difference between “You” and “Signiﬁcant Other” was negligible. The
mean perception of each peer group was unique with students perceiving that
undergraduates in all majors are the most ethically approving of cheating with a mean
score of 2.32 (falling between “Unethical” and “Neither ethical nor unethical” on the
scale). Table 2 lists Student Mean Ethical Belief responses for themselves and others on

the ﬁve different types of academic dishonesty.

Table 2
Self-Report and Perception of Various Peers’ Ethical Beliefs
on Academic Misconduct

 

 

 

 

 

 

 

 

 

 

 

 

 

Mean (Std. N
Deviation)
You 1.74 (.631) 182
_S_i_gniﬁcant Other 1.75 (.663) 128
Best Friend 1.80 (.628) 172
Five Closest Friends On Campus 1.96 (.654) 174
College Roommates 2.05 (.617) 166
Undergraduates 1n Major 2.20 (.627) 170
Undergraduates All Majors 2.32 (.643) 173
Aggegated Others 1.99 (.574) 122

 

Further testing yielded an analysis of means derived only from the particular cases
that had answers for both the “You” and the speciﬁc peer group category across all ﬁve
forms of academic dishonesty. Analyzing the data this way allowed for a Paired-Samples
T Test to be conducted on each self-other difference. The self-other mean ethical
difference for “Signiﬁcant Other” was not found to be statistically signiﬁcant at the.01
level. The difference between oneself and his or her best ﬁiend was found to be

signiﬁcant at the .01 level. The four remaining self-other peer group differences were

14

found to be signiﬁcant at the .001 level. The degree to which the ethical mean of each
peer group differed from the respondents self-report of his or her own ethics varied. The
mean perceived ethical belief of the “Best Friend” was just .080 more than that of “You,”
whereas the undergraduates in all majors mean was .592 (over a full half-step on the
response scale) more than the mean student ethical self-report. All in all, the hypothesis
was conﬁrmed as the self-other mean ethical difference increased as the social distance of
the peer group increased (note that the difference for “Signiﬁcant Other” was not found

to be statistically signiﬁcant) (see Table 3).

Table 3
Mean Differences and Correlations of Self-Reports and Perception of Various Peers’ Frequency of
Academic Misconduct and Related Ethical Beliefs

 

 

 

 

 

 

 

 

 

 

Pearson Self-Other Pearson Self-Other
Correlation For Mean Correlation For Mean Ethical
Aggregated Frequency Aggregated Difference
Frequencies Difference Ethical Beliefs
Signiﬁcant Other .617" .009 .879“ .050
N=116 N=116 N=127 N=127
Best Friend .865“ .110" .835” .080"
N=157 N=157 N=171 N=171
Five Closest .741 "”“ .345“ .779" .230"
Friends On N=163 N=163 N=173 N=1 73
Campus
College .637" .392“ .630" .345"
Roommates N=156 N=156 N=165 N=165
Undergraduates .504“ .751“ .538" .486”
In Major N=158 N=158 N=169 N=169
Undergraduates .405“ 1.01" .380“ .592"
All Majors N=159 N=159 N=172 N=172
Aggregated .751" .441 *"' .807" .297"
Others N=109 N=109 N=121 N=121

 

 

 

 

 

 

" denotes signiﬁcance at the .01 level “denotes signiﬁcance at the .001 level

Hypothesis 4: The size of the misperceptions students have about the frequency in
which their peers take part in academically dishonest acts increases as peer group

social distance increases.

 

Students mean frequency of behavior responses for themselves and others on the
ﬁve different types of academic dishonesty were calculated. Once again, the N values
differed by type due to the fact that not everyone answered every question. All peer
variables besides “Signiﬁcant Other” had mean scores higher than those that the
respondents self-reported. Also, as it was with the ethical belief questions, the mean
perception of each peer group was unique with students perceiving that undergraduates in
all majors are academically dishonest most often. The mean score on the frequency
component for undergraduates of all majors was 2.65 (falling between “Once or twice”

and “Three to ﬁve times” on the frequency of misconduct scale) (see Table 4).

Table 4 Self-Report and Perception of Various Peers’ Frequency
of Academic Misconduct

 

 

 

 

 

 

 

 

 

 

 

 

 

Mean (Std. N
Deviation)
You 1.66 (.664) 177
Significant Other 1.58 (.682) 120
Best Friend 1.74 (.680) 162
Five Closest Friends On Campus 2.00 (.733) 168
College Roommates 2.04 (.747) 161
Underggduates In Major 2.39 (.754) 163
Undergraduates All Majors 2.65 (.785) 164
Aggregated Others 2.01 (.618) 113

 

Once again, further analysis of means derived only from the particular cases that
had answers for both the “You” and the speciﬁc peer group category was performed. The
self-other frequency of misconduct difference for “Signiﬁcant Other” was not found to be
statistically signiﬁcant at the.01 level; however, each of the remaining peer group means
were signiﬁcantly different than the mean self-reported frequency of behavior at the .001
level. The degree to which the frequency of behavior mean of each peer group differed
from the respondents self-report of his or her own cheating behavior varied. The mean
perceived frequency of misconduct of the “Best Friend” was just .110 more than that of

“You,” whereas the undergraduates in all majors mean was 1.01 (over a full step on the

16

scale) more than the mean student self reported frequency of misconduct (see Table 3).
The hypothesis was conﬁrmed as the self-other mean frequency difference increased as
social distance increased (note that the difference for “Signiﬁcant Other” was not found
to be statistically signiﬁcant) (see Table 3)

Hypothesis 5: A positive correlation exists between one’s own behavior/beliefs
regarding academic dishonesty and the behavior/beliefs she perceives her peers
participating in and believing.

The aggregated ethical belief and frequency of misconduct mean peer variables
were correlated with the mean “You” variable. Results show a strong positive correlation
for both the frequency and ethical belief correlations, with a frequency correlation of .751
and an ethical belief correlation of .807. Each of these correlations is signiﬁcant at the
.001 level (see the bottom of Table 3).

Hypothesis 6: The predictive power of the perception of peer beliefs and behaviors
regarding academic dishonesty decreases as peer group social distance increases.

Each ethical belief and frequency of misconduct mean peer variable was
correlated with its respective mean “You” variable to see the strength of each
relationship. The Pearson correlation analysis (Table 3) shows that each relationship that
was tested is signiﬁcant at the .001 level. The strength of these correlations did vary by
peer group. For the ethical belief category, “Signiﬁcant Other” was most strongly
correlated with “You” with a correlation of .879 whereas undergraduates of all majors
was the least powerful predictor of self-reported ethical belief with a correlation of .380.
In the frequency of misconduct category, the weakest correlate was also undergraduates

of all majors with a correlation of .405, but the most powerful predictor was “Best

17

Friend” with a correlation of .865. In this category “Signiﬁcant Other” was only the
fourth strongest correlate with a correlation of .617. The predictive power of the
perception of ethical beliefs did decrease as peer group social distance increased. Also,
with the exception of the aforementioned “Signiﬁcant Other” being only the fourth
strongest correlate, the predictive power of the perceptions of peer behaviors decreased as
peer group social distance increased.

Hypothesis 7: Presenting a colloquial deﬁnition of what constitutes academic
dishonesty may mediate the disparity between students’ own beliefs and behavior
regarding academic dishonesty and the perception of their peers corresponding
beliefs and behaviors.

The analysis of the data failed to show any important differences dependent on
whether or not the respondent received a survey that asked them to read and presented
them with a colloquial deﬁnition of academic dishonesty.

Discussion:

In general, the analysis yielded the expected results. Students were found to over-
perceive both their peers degree of ethical approval of academically dishonest acts and
their peers’ frequency of committing these acts. The ﬁndings are consistent with Hard et
al. (2006), and Scanlon and Neumann (2002). Contrary to these ﬁndings, a study by
Jordan (2001) found that students underestimated the cheating behavior of their peers.
This difference in result could be the result of sample differences. The sample in
Jordan’s study included only members of a small liberal arts school with an honor code
that responded to a survey by mail, whereas this study and that of Hard et a1. were drawn

from surveys conducted in the classroom at a large and medium sized university,

l8

respectfully. The sample in the Scanlon and Neumann study was made up of students on
nine campuses. Previous research has conﬁrmed that self-other misperceptions are
present for other types of behaviors as well. These behaviors include alcohol use,
cigarette smoking, gambling, eating behaviors, and attitudes towards homophobia.
Berkowitz (2004:10) sums up the ubiquity of self-other misperceptions by stating that
they “have been consistently documented for a variety of behaviors and social contexts
and in a variety of student and adult populations and sub-populations in both individual
studies and in meta-analyses.”

As anticipated, the size of the misperceptions, in both the ethical belief and
frequency of misconduct categories, varied depending on which peer group was
referenced. This analysis of the distinctive characteristics of different self-other
misperceptions is somewhat unique for studies on academic dishonesty and seemingly
has never been done with as many as six different peer targets. In the study, the size of
the misperceptions grew larger as the social distance from the respondent increased,
which corroborates with earlier studies on collegiate alcohol use (Borsari and Carey
2001)

The self-other mean frequency differences were larger in the frequency of
behavioral misconduct than in ethical approval. This difference may be attributed to a
methodological ﬂaw in the study design. The frequency of misconduct questions
preceded the ethical belief questions on every survey. Seemingly, this order of questions
had some effect on the data because respondents were more likely to provide straight
answers (providing the same response for both the self and all peer group portions of any

or all forms of academic dishonesty) on the ethical belief questions. It is hypothesized

l9

that this trend is at least partly the result of students hurrying through the second portion
of the survey (which looked very similar to the ﬁrst) and that this difference would be
lessened if the order of the categories was randomized.

A student’s beliefs about his or her peers’ behavior and ethics regarding academic
dishonesty were shown to be strong correlates of that individual’s own academically
dishonest behavior and related ethical beliefs. This replicated many previous ﬁndings on
academic dishonesty (Burrus et a1. 2007, Hard et al. 2006, McCabe et al. 2001). The
strength of the correlations differed by peer group. With only one exception (self-
signiﬁcant other frequency), the strength of the correlations became weaker as social
distance increased. The ﬁnding that the strength of self-other correlations decreases as
social distance increases and self-other misperceptions increase is well documented in the
collegiate alcohol use literature (Borsari and Carey 2003). This study adds to that
literature in that it shows similar ﬁndings for academic dishonesty.

The one exception that was found to these general rules was the relationship
between self-frequency of misconduct and perception of the frequency of misconduct of
the respondent’s signiﬁcant other. For this self-other category the mean frequency
difference was lower than any other, which is to be expected if it is assumed that of all
the peer variables there is the least amount of social distance between a respondent and
her signiﬁcant other. What was not expected, however, was that of all six peer groups,
the self-signiﬁcant other correlation was only the fourth strongest. Previous literature
and theory would have predicted it much stronger.

There are many possible explanations for this. Perhaps the most parsimonious is

to consider the unique values of the signiﬁcant other as a gendered variable. It may be

20

that, although socially close, an individual’s perception of his or her signiﬁcant other is
relatively less powerful in predicting that individual’s behavior because the individual
does not see his or her signiﬁcant other’s actions as the norm for people like himself or
herself. In other words, a male might be aware of his girlﬁiend’s actions but would not,
deliberately or otherwise, in order to ﬁt the social norm, try and align his behavior with
hers. Instead he might conclude (consciously or unconsciously) that what she does is just
“what girls do” and need not inﬂuence his behavior. In this scenario it might be more
likely that this individual’s behavior would be more correlated with a same-sex best
ﬁiend who not only is socially close but also does “what guys do.” Of course, this
analysis is assuming that a majority of couples are heterosexual and best friends are
mostly of the same sex. It also doesn’t speak to why this difference was not present in
the ethical belief category. Further study is needed to indentify the unique inﬂuence and
potential gender differences in the predictive power of signiﬁcant others on academic
dishonesty.
Limitations

This study is certainly not without limitations. Firstly, the relationship between a
person’s behavior/beliefs and his perception of his peers’ behaviors/beliefs is
correlational; thus, causation should not be inferred. It could certainly be the case that
law abiding individuals tend to choose other law abiding individuals as friends,
signiﬁcant others, etc. Even if this is the case though, it certainly does not negate the
possibility that these peer groups inﬂuence an individual’s behavior and beliefs after

joining their ranks. It is not at all contradictory for a person to both select friends due to

21

similarities in activities and beliefs and to be inﬂuenced by these friends thereafter. More
rigorous empirical studies are needed to establish causality.

It is worthwhile to note that methods have been taken in other research in attempts
to ascertain the effect that change in a person’s perception of their peers has on his or her
own behavior and beliefs. Studies have been conducted on alcohol usage in which an
experimental group, but not the control group, is privy to information that shows actual
drinking norms (which are less than perceived norms). At the end of the intervention,
students in the experiment group reported that they personally consumed less alcohol and
that they also thought that their peers consumed less alcohol than they did pre-
intervention (Peeler et a1 2000, Schroeder & Prentice 1998).

Another limitation of this study is that it relies on self-report data. Therefore,
answers that are given are subject to social desirability bias (although the promise of
anonymity would seem to lessen this) and mediated by memory. Nonetheless, self-report
methods are readily used and are seen by many as a reliable way to study delinquent
behavior (Harris and Benson 1999, Lab & Allen 1984, Hindelang 1981). Moreover,
many of the self-report studies on collegiate alcohol use that have yielded positive
correlations between individuals and their peer groups have been validated by research
that uses methods that are not self-report. For instance, Foss et al. (2003. 2004) used
measures of blood alcohol concentration to measure change in behavior and concluded
that after students were shown actual statistics on drinking that they drink less.

A third limitation of this study is that it does not measure what people will do,
only what they feel and say they have done. It is possible that although a person

truthfully reports what they have done in the past, they may not continue to behave and

22

perceive the same way in the future. Nonetheless, studies point to past deviant behavior
as the best predictor of future deviant behavior (Akers and Sellers 2004). Moreover,
one’s attitudes about delinquent behavior have been shown to be good predictors of
actual behavior (Bolin and Heatherly 2001).

A fourth limitation is that time-order effects were not controlled. Thus, every
student was presented with the same set of questions in the exact same order. All of the
frequency questions came before the ethical belief questions, and questions asking about
individual behavior and beliefs always preceded the questions about one’s perception of
his various peer groups. Systematic variation of questions should be considered in future
research.

The last limitation to be discussed is the study’s lack of ecological validity. The
results found here may only be generalizable to undergraduates taking summer courses at
large public universities who are Juniors or Seniors who are mostly white and female.
Still the ﬁndings here on the misperception of the cheating behavior of one’s peer group
and the correlation between self-reports and perception of peers on academic dishonesty
corroborate with earlier ﬁndings with somewhat different samples. For instance, Hard et
al. (2006) also found that students overestimated how often their peers were academically
dishonest and that there was a signiﬁcant positive correlation between student behavior
and perception of peer behavior from analyzing data from 421 students in general
education courses at a medium sized public university. Conversely, the aforementioned
study by Jordan (2001) found that students underestimated their peers’ cheating behavior

while still ﬁnding a strong correlation between student behavior and that of their peers

23

The ﬁndings from this study heed the call of replication put forth by earlier
studies (Hard et al. 2006, Berkowitz 2004) and offer substantiation and increased
ecological validity of conclusions drawn by many. Speciﬁcally this study offers support
for the following conclusions (especially in regards to academic dishonesty): students
over-perceive their peers’ delinquency; these misperceptions increase as social distance
increases; there is a positive correlation between one’s perception of his peers’ delinquent
behavior and one’s own; and the strength of this correlation generally lessens as the
social distance of the peer group being referenced increases.

Implications

The particular ﬁndings of this research may be of use in designing a social norms
campaign aimed at ultimately lowering the prevalence and severity of academically
dishonest acts on college campuses. The theory behind a social norms campaign asserts
that people behave in ways that are consistent with their peers. Furthermore, it states that
an over—estimation of the behavior of peers may lead to a person increasing her behavior
in the same direction as the over-estimation (Whitley 1998). A social norms campaign
then seeks to correct these misperceptions by providing accurate information about the
norms of peer groups. The correction of the misperception is reasoned to result in the
lessening of the behavior (Berkowitz 2004). Thus, the person who believes that the
average undergraduate student at his institution cheats ﬁfty times each semester is more
likely to be academically dishonest himself than if he thought the average undergraduate
never cheated.

Some decisions need to be made when developing a social norms campaign.

Perhaps the most important is deciding which norm to seek to correct. This importance is

24

not lost on Alan Berkowitz, the co-creator of the social norms approach, who reasons,
“Selecting the most relevant and salient norms for a particular intervention and the
appropriate strategy for changing those norms is an important part of the planning
process of a social norms intervention” (Berkowitz 2003:13)

The extant literature shows that self-other misperceptions tend to be greater for
injunctive norms (those based on morals or ethics) than they are for descriptive norms
(those based upon behavior). The self-other correlations also tend to be stronger for
injunctive norms (Borsari and Carey 2003). Despite this, the majority of successful
campaigns have used descriptive norms (Berkowitz 2004). One can conclude that either
type of norm may be of value in a social norms campaign and thus both were included in
this study.

Another factor to consider when looking for the most useful norm to try and
correct is whether to use people or institution based norms. The extant literature points to
people based norms as being more inﬂuential than institution based norms (Borsari and
Carey 2003). Thus, this study looked exclusively at people based norms.

So then, what might be taken away from this study as far as knowledge about
which norm to attempt to correct in implementing a social norms campaign with the goal
of lessening academically dishonest behavior? Ideally, one would choose a norm that is
highly correlated with the individual behavior, misperceived to a large degree, easy to
correct, and whose correction yields a big change in the perceiver’s behavior.
Unfortunately, this study did not test for all of these factors, and even if it did it would be

unlikely that a single norm exists that ﬁts all of the criteria. What this study does offer is

25

the point that for academic dishonesty as social distance increases misperceptions

increase and the strength of correlations decrease.

26

References

Akers, Ronald and Christine S. Sellers. 2004. "Social Learning Theory." Pp. 81-110 in
Criminological Theories: Introduction, Evaluation and Application. Los
Angeles, CA: Roxbury Publishing Company.

Akers, Ronald and Gary F. Jensen. 2006. "The Empirical Status of Social Learning
Theory of Crime and Deviance: The Past, Present and Future." in Taking Stock:
The Status of Criminological Theory. Advance in Criminological Theory, vol. 15,
edited by J. P. W. Francis Cullen, and Kristie R. Blevins. New Brunswick, NJ:
Transaction Publishers.

Berkowitz, Alan D. 2004. "The Social Norms Approach: Theory, Research, and
Annotated Bibliography." Trumansburg, NY: www.alanberkowitz.com.

Bolin, Aaron U. 2004. "Self Control, Perceived Opportunity, and Attitudes as Predictors
of Academic Dishonesty." The Journal of Psychology 138:101-1 l4.

Bolin, Aaron and Linette Heatherly. 2001. “Predictors of Employee Deviance: The
Relationship Between Bad Attitudes and Bad Behavior.” Journal of Business and
Psychology 15 :405-41 8.

Borsari, B. and Kate. B. Carey. 2001. "Peer inﬂuences on college drinking: A review of
the research." Journal of Substance Abuse 13:391-424.

Borsari, B. and Kate. B. Carey. 2003. "Descriptive and injuctive norms in college
drinking: A meta-analytic integration." Journal of Studies on Alcohol 64:331.

Brown BS. and P. Choong. 2005. "An investigation of academic dishonesty among
business students at public and private United States universities." International
Journal of Management 22 :20 l -214.

Burrus, Robert T., KimMarie McGoldrick, and Peter W. Schuhmann. 2007. "Self-
Reports of Student Cheating: Does a Deﬁnition of Cheating Matter?" Journal of
Economic Education 38:3-16.

Choong, Brown BS. and P. 2005. "An investigation of academic dishonesty among
business students at public and private United States universities." International
Journal of Management 22:201-2 14.

Cochran, John K., Mitchell B. Charnlin, Peter B. Wood, and Christine S. Sellers. 1999.
"Shame, Embarrassment, and Formal Sanction Threats: Extending the
Deterrence/Rational Choice Model to Academic Dishonesty." Sociological
Inquiry 69:91 -105.

Drake, C. A. 1941. "Why Students Cheat." Journal of Higher Education 12:418-420

27

Foss, R., S. Deikkman., C. Bartley and A. Goodman. 2004. “Social Norms Program
Reduces Measures and Self-Reported Drinking at UNC-CH.” The Report on
Social Norms: Working Paper #14. Little Falls, NJ, PaperClip Communications.

Foss, R., S. Deikkman, A. Goodman, and C. Bartley. 2003. “Enhancing a Norms
Program to Reduce Hi gh-Risk Drinking Amongst First Year Students.” University
of North Carolina Chapel Hill.

Hard, Stephen F ., James M. Conway, and Antonia C. Moran. 2006. "Faculty and College
Student Beliefs about the Frequency of Student Academic Misconduct." The
Journal of Higher Education 77:1058-1080.

Harris, Diana and Michael Benson. 1999. “Theft in Nursing Homes: An Overlooked
Form of Elder Abuse.” Journal of Elder Abuse & Neglect 11:73-90

Hindelang, M., Hirschi T., and J. Weis. 1981. Measuring Delinquency. Beverly Hills,
CA: Sage Publications.

Jordan, A.E. 2001. "College Student Cheating: The Role of Motivation, Perceived
Norms, Attitudes, and Knowledge of Institutional Policy." Ethics & Behavior
11:233-247.

Lab, S. and R. Allen. 1984. “Self-report and ofﬁcial measures: A further examination of
the validity issue. Journal of Criminal Justice 12:445-455.

Lambert, E. G. and N. L. Hogan. 2004. "Academic Dishonesty among Criminal Justice
Majors." American Journal of Criminal Justice 29:1-20.

Korkuska, J. and D. Thombs. 2003. “Gender Role Conﬂicts and Sex-Speciﬁc Drinking
Norms: Relationships to Alcohol Use in Undergraduate Women and Men.”
Journal of College Student Development 44:204-215.

Mattem, Jody L. and Clayton Neighbors. 2004. "Social Norms Campaigns: Examining
the Relationship between Changes in Perceived Norms and Changes in Drinking
Levels." J. Stud. Alcohol 65:489-493.

McCabe, Donald L. 1999. "Academic Dishonesty Among High School Students."
Adolescence 34:681-687.

McCabe, Donald L., Linda K. Trevino, and Kenneth D. Butterﬁeld. 2001. "Cheating in
Academic Institutions: A Decade of Research." Ethics & Behavior 11:219-232.

McCabe, D. L. and L. K. Trevino. 1993. "Academic Dishonesty: Honor Codes and Other
Contextual Inﬂuences." The Journal of Higher Education 64:522-538.

28

McCabe, Donald L. and Linda K. Trevino. 1997. "Individual and Contextual Inﬂuences
on Academic Dishonesty: A Multicampus Investigation." Research in Higher
Education 38:379-396.

Michaels, James W. and Terance D. Miethe. 1989. "Applying Theories of Deviance to
Academic Cheating." Social Science Quarterly 70:870-885.

Neumann, Scanlon P. M. and D. R. 2002. "Internet Plagiarism Among College Students."
Journal of College Student Development 43:374-385.

Peeler, C., J. Far, J. Miller and T. Brigham. 2000. “An Analysis of the Effects of a
Program to Reduce Heavy Drinking Among College Students.” Journal of
Alcohol and Drug Education 45:39-54.

Perkins, H. W. 2002. "Social Norms and the Prevention of Alcohol Misuse in Collegiate
Contexts." Journal of Studies on Alcohol Supplement 14:164-172.

Scher, K., B.D. Bartholow and S. Nanda. 2001. "Short-and Long-Term Effects of
Fraternity and Sorority Membership on Heavy Drinking: A Social Norms
Perspective." Psychology of Addictive Behaviors 15:42-51.

Schroeder, C. and D. Prentice. 1998. “Exposing Pluralistic Ignorance to Reduce Alcohol

Use Among College Students.” Journal of Applied Social Psychology 28:2150-
2180.

Sims, R. 1993. "The relationship between academic dishonesty and unethical business
preactices." Journal of Education for Business March/Apri11207-21 l.

Whitley Jr., Bernard E. 1998. "Factors Associated With Cheating Among College
Students: A Review." Research in Higher Education 39:235-274.

29