Cheating and Honor: Lessons from a Long-Term Research Project

Abstract

This chapter reviews key findings from a research project into student academic dishonesty conducted over a period of approximately 15 years. The project replicated and extended a large-scale seminal study which was conducted across 99 US campuses in the 1960s Bowers (1964). Over the life of the project, thousands of students have provided self-report data about their own dishonest academic behaviors including those involving various forms of copying, cheating on tests and exams, and fabricating data. Twelve of the 28 behaviors measured in the project were replicated from the Bowers study, enabling comparison of results over approximately half a century. Interestingly, a consistent reduction in reported engagement in dishonest behaviors is seen over time in most of the domains measured. The chapter also provides an overview of the role that honor codes play in many of the participating institutions and the effects of these codes on cheating behavior, as witnessed over the lifetime of the project.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The author’s research agenda has focused on issues of academic integrity and student cheating for roughly 25 years. His introduction to the concepts around academic honesty, and dishonesty, came over 50 years ago, when he received his admissions package in the mail from Princeton University, including information about their honor code. As the son of an alumnus, the author was aware of the basic code, but this was his first explicit, intense exposure to it. He wondered if his new classmates, who had been described as the top-ranked students from the best high schools around the country, had made it to the top through cheating. Would he witness a rampant lack of academic integrity (which was not even a common term in 1961) among his peers? If so, would he have the courage to report it? Thankfully, neither happened. His worries were for naught as he never observed a single individual cheating on a test or exam.

After some vacillation, the research has led the author to remain a strong proponent of honor codes as one of the most effective strategies to reduce cheating in academia and the larger society that is read about daily in the newspaper or seen on the television news.

A Long-Term Research Project into Academic Dishonesty

At around the same time, seminal research by Bowers (1964, discussed in more detail below) reported less cheating by students from honor code schools. Almost 30 years later, McCabe and Trevino – both of whom had been undergraduates in schools with honor codes – undertook a research project designed to test the proposition suggested first by Bowers (1964) that cheating was lower in institutions that had honor codes. The study was replicated and somewhat extended 5 years later. The results of the two studies (McCabe and Trevino 1993; McCabe et al. 1999) bore out Bowers’ findings that students’ self-reported engagement in nine different dishonest academic behaviors was consistently lower in schools with honor codes than in those without codes.

About the Project

The research project continued and evolved from that time with data being collected from thousands of students over a period of more than 20 years. The goal of this chapter is to review the findings from data collected from college and university students and faculty over a particular 15-year period of that longer-term research into academic dishonesty. The research was primarily conducted to aid in understanding the level of cheating in which students engage.

The benchmark data for the project was collected in the seminal study conducted by Bowers (1964) in a survey of almost 6,000 students across 99 campuses. For the first time on this scale, Bowers asked the surveyed students to self-report their own dishonest academic behaviors. Most of the data in the McCabe project were gathered over the period from the fall semester of 2002 to the spring semester in 2013 for all types of schools – 2-year and 4-year, large and small, and geographically dispersed (although all in the USA) – probably the most robust sample collected in terms of school characteristics. Although the students in these studies were located in the USA and Canada, surveys were also administered to students from Mexico, Egypt, the UK, Australia, the United Arab Emirates, and Hong Kong.

Although additional data was gathered in 1991 and 1996 by McCabe and colleagues (McCabe and Trevino 1993, 1997; McCabe et al. 2001), this additional data is not reviewed here in detail, although reference is made to it on occasion.

As noted in McCabe (in press), the student and faculty surveys that were employed in this project “…remained relatively consistent over the period in question with only minor changes being made and most of these changes have involved additions to the survey…Most of these changes have taken place in the section of the survey which attempts to ask both students and faculty about the frequency with which they have either engaged in (in the case of students) or observed (in the faculty survey) selected behaviors which some might consider cheating.”

With the exception of some additions, the survey instrument involved in this project remained relatively constant over time. The same basic survey was used for both students and faculty with only minor necessary changes. The student survey can be found at https://honesty.rutgers.edu/rutgers.asp, and the faculty survey can be found at https://honesty.rutgers.edu/rutgersfac.asp.

The primary emphasis of this chapter is on the specific behavior section of the survey which details behaviors which one might consider cheating. Twelve of the 28 behaviors used are directly from Bowers (1964), while two represent electronic types of plagiarism not possible in Bowers’ day. The other 14 include such behaviors as homework copying, fabricating laboratory or research data, and copying someone else’s computer program.

Some Key Findings Over the Years

Plagiarism

Only cursory attention is paid to plagiarism in this chapter, leaving a broader discussion of deliberate plagiarism to Jude Carroll in the next chapter. Plagiarism of various types has been researched in each of the surveys reviewed here. In each case, of course, it is made clear to survey respondents that if the (student) author has cited the text in question, there is no violation of the rules of plagiarism.

Typically, lesser forms of plagiarism are described as “cut and paste” plagiarism, where selected sentences are woven together to construct the answer to a particular question or are woven together throughout an entire essay. Large-scale plagiarism, either from written sources or the Internet, is where essentially the entire paper is being taken from another source. Not surprisingly, surveyed students readily admit to the “cut and paste” category of offenses which they do not generally view as a “big deal” and either are more reluctant to admit to more significant plagiarism or are defining larger-scale plagiarism differently than faculty might.

Review of the data collected from ten large state universities in 1995 (McCabe et al. 2012), some Canadian universities collected in 1994, and follow-ups to an honor code study conducted in 1996 and 2006 (McCabe and Trevino 1997; McCabe et al. 2001, 2012) reveal some interesting findings. As shown in Table 1, for self-reported plagiarism a somewhat surprising pattern can be seen – a reported decrease in engagement in these activities by students over the period studied. This is despite ongoing media reports to the contrary and is especially surprising in the case of Internet plagiarism where the media would have us believe there has been a dramatic increase in cheating.

Table 1 Reported student engagement in plagiarism

It is likely that these results can be influenced by a number of factors, including the following:

• 1. The manner in which the statistics have been calculated – e.g., in the case of smaller quantities of text, if we go back as far as the Bowers data in the 1960s, it can be seen that Internet plagiarism was zero at this time as it was nonexistent.

• 2. The reliance on self-report data which may result in the data being influenced by self-presentation bias. Although it might be expected that such bias is relatively constant over time, we do not know this for a fact.

• 3. The manner in which the surveys have been conducted with a switch to Internet surveys rather than paper-based. This could result in an underreporting of actual engagement in dishonest behaviors where respondents feel their identity may be traced.

Views on the Seriousness of Plagiarism

The research has also looked at evaluations of the seriousness of cheating. In the 2002/2003 survey, 43 % of respondents rated small Internet plagiarism as not cheating or merely trivial cheating. Ten years later in 2012/2013, this number had fallen quite significantly to only 33 % of respondents who viewed the behavior as trivial (McCabe et al. 2012). Similar increases were also seen in perceived seriousness of plagiarism from written sources (McCabe et al. 2012, p. 59). The shift suggests that over time respondents have come to view Internet cheating as a more serious offense.

Other Forms of Cheating on Written Work

Data were collected on three other forms of cheating on written work beyond just plagiarism – submitting work done by someone else, working collaboratively when not permitted, and fabricating or falsifying a bibliography. From the data collected in the period from 2002 to 2013, a general pattern emerges of a decrease in cheating between the Bowers study in 1963 and the data collected in 2012/2013. This is the case for all types of schools, and the pattern is repeated for many of the different cheating behaviors investigated.

As shown in Table 2, we observe roughly the same pattern of engagement for self-reported (non-plagiarism) cheating on written work that can be seen in Table 1 for plagiarism. While some minor differences to the overall trend are noted for fabricating a bibliography in 2002/2003 and working collaboratively early in Table 2, both of these variations are very minor and have little effect on the observable trend.

Table 2 Self-reported engagement in cheating on written work other than plagiarism

Cheating by Copying on Tests

When looking at behaviors involving copying from others on tests and exams, as shown in Table 3, again we see the same basic pattern for the self-reported engagement described above for plagiarism and other forms of cheating on written work (other than plagiarism).

Table 3 Percentage engagement of copying on tests

As expected, the percentage of engagement is consistently lower in the case of those copying with other’s knowledge than for those copying without the other’s knowledge. This is not surprising since copying with knowledge technically involves both people cheating, and we would expect that in most cases the source of the material is an individual who would prefer not to be involved.

If we look at the seriousness of these two offenses that are shown in Table 4, we see that in every case respondents consider copying without the other person’s knowledge is slightly more of an issue. Although the difference is consistent, the difference is indeed minor suggesting that students do not see much of a difference between the two offenses.

Table 4 Perceived seriousness of copying on tests

Some Notes About the Data

The reader should be aware that the first column of results represents a 1-year period, and all the others cover a 2-year period. However, the first column has more data than the last three. The differences in sample size and time frame simply reflect how many schools decided to participate in the survey and when. In this sense researcher control was sacrificed for greater number of respondents – perhaps a good trade-off, but one that requires further examination.

In particular, both undergraduate and graduate students’ numbers are “lumped” together in the sample even though it has been shown previously (e.g., McCabe et al. 2012) that graduate students cheat less often than undergraduates or at least report less cheating. However, closer examination of the sample suggests this is not a major factor in the large samples. Examination of the composition of each annual segment in the sample reveals a total range of only 6 % from a low of 10 % graduate students in 2005/2006 to high of only 16 % in 2007/2008. Therefore, it is unlikely that these results are skewed because of different graduate student representation in the different samples.

Some Observations on Motivation to Cheat

Across the length of the research project, the contextual influences of peer behavior, campus culture of integrity, and the perceived opportunity to cheat are the three that bear consistently strong relationships with cheating (McCabe et al. 2012). Of these, students’ perceptions of peer behavior were the strongest influence on decisions to engage in dishonest behavior themselves.

At the individual characteristic level, the research project reported here has shown that cheaters tend to be males majoring in business or science, and these respondents suggest that faculty have not discussed the rules as much as the faculty of those who do not cheat (McCabe and Trevino 1995). This latter relationship is not surprising, as noted earlier, cheaters often tend to place some blame on faculty for their cheating.

Indeed, when students are asked about motivations to cheat, two reasons have been repeatedly offered more than any others: firstly, professors have not made the rules clear, and secondly, the student must get an A grade. While these observations are based on student comments, the frequency with which they are offered adds to the researcher’s confidence that they are highly relevant in any discussion of motivation to cheat.

Unclear Rules and Unrealistic Expectations

In relation to making the rules clear, there are equal arguments to be made that faculty instructions are not as complete as they should be and that students have a certain obligation to take responsibility to educate themselves. This is not meant to free either party of all blame as there is certainly a shared responsibility here and students should have already internalized these “rules” to a certain degree. For example, the comments of the following students from a large public university in Canada touch on these points:

When there is an evaluation specifically assignments and essays, it would be great if professors could go over plagiarism and academic misconduct because when I first came, it took a long time for me to properly cite references.

Although it may seem like students know how to properly cite, profs should go over their expectations for every assignment. I just had a prof let us know while she handed back the assignment that using our textbook as a reference was wrong. It was considered a secondary source and she wanted us to go to the primary sources and read/cite those. No one in our class had a clue that we were supposed to do that. Clarity works.

There would not be cheating if students valued learning more than their grades but that’s not the case and multiple professors stack the amount of work due and between our life and everything else cheating is easiest.

An A Grade: “The Coin of the Realm”

In general, it appears that “cheaters” are sometimes people who do not feel the rules apply to them for some reason and very often people who either feel the need to cheat to “survive” or to compete.

The following student from a major public university in the southeastern USA seems concerned about the competition for good grades:

The cheating is extraordinarily rampant to the extent that if you do not at least a little [cheating] it is nearly impossible to compete.

Another disturbing response was from a junior at an honor code university in the East who suggested that although he did not cheat in his first 2 years at university, he felt he might have to cheat from now on due to the fact that the administration had been putting pressure on the faculty to reduce the number of A’s given in their courses. This student believed that a straight A average was more important than any concept of honor.

Also concerning grades, although not necessarily A’s in this case, is a quote from a faculty member at a large, public university in the southeastern USA concerning admission into selective programs on campus:

I explain the use of test banks in my class this way – “these are in place in order that someone who cheats can’t get as high of a grade as you and get into the competitive nursing … program before you.”

Perhaps the most disturbing quote noted over the years on the topic of grades was from a young gentleman attending a very prestigious school in the mid-Atlantic region. Since he was not at Harvard or Yale, he was concerned about competing against students from those universities when he graduated and looked for a job on Wall Street. In his mind, this required that he cheat to maintain the straight A’s needed, and if he had to cheat to get them, so be it.

Faculty Views

There are also some data available from the only other group intimately involved with student plagiarism – teaching faculty. Generally, faculty feel there is a need to improve the education of both students and faculty in relation to academic integrity . For example, the following comments from a faculty member at a medium-sized private university in the US Midwest support this need:

Make the academic integrity policy more readily available to everyone. For instance, I checked the University website on Academic dishonesty and it refers people to [a page], which is nowhere to be found on the web. The process needs to be as streamlined as possible to reduce the cost to faculty of using the system. Savvy students know faculty don’t want to deal with it…and that lowers the expected costs of cheating.

Since teacher pay is tied to student evaluations of teaching, teachers are legitimately afraid of alienating students. Take this out of the equation and teachers will be more bold about confronting cheaters.

Such comments capture faculty dissatisfaction with the policies currently in place in their schools. They also suggest reasons why the research reports more students cheating than might be expected. The reference by the faculty member quoted above to teacher pay being tied to student evaluations is particularly relevant for adjunct faculty whose livelihood depends on their continued employment at the institution and hence to student ratings of their performance. The bottom line is that students today feel that they can get away with cheating which helps in part to explain the high incidence of engagement in dishonest behaviors.

Some Observations on Honor Codes

There is no universal agreement on what constitutes an honor code, and not every code nor system which an individual school classifies as an honor code has all four of the characteristics normally associated with such codes. As suggested by Melendez (1985), these four characteristics are the following:

• 1. The signing of some kind of pledge, ranging from one designed to be signed once, typically upon matriculation, to one designed to be signed with the submission of any work governed by the code (tests, examinations, or a written assignment);

• 2. Unproctored tests and examination;

• 3. A student-controlled hearing process; and

• 4. A degree of obligation on students to report any cheating which they learn about or observe.

Melendez (1985) argued that any one of these elements indicated the school used an honor code, but a more rigorous definition of a minimum of two of these characteristics with some elements of a third is used for inclusion in the honor code category in the project discussed in this chapter. Most codes, but not all, have some form of student governance although the exact form may vary widely in practice. We see variation ranging from total student control over the sanctioning process through to the involvement of administration at some point in the process, often in the assignment of sanctions.

The success of honor codes like the 1960s Princeton code mentioned at the beginning of the chapter can, in the author’s view, be attributed to three things: (1) it was simply a different time with students more concerned about Vietnam and the draft and less concerned about their class rank, including some who were more concerned with simply graduating; (2) fear of being asked to leave school, which occurred on occasion, was probably the primary motivator for many, including this author (who could not imagine explaining expulsion to his father, who was so proud of his son’s admission to Princeton); and (3) the fact that Princeton had a “partial” code that only covered in-class tests and not plagiarism on written work. This latter feature was undisputedly a point of contention as students, who felt they were not being completely trusted, pushed for full control of the code, while school administration resisted. In spite of all the rhetoric associated with this debate, the typical Princeton student surely had to feel less trusted than a student governed by a full honor code – like the one at Washington and Lee University. Although it was rarely stated aloud, one of the factors keeping the focus of the code on tests and exams was some lack of trust by administration, and possibly faculty, that students would not do their written work honestly. Not surprisingly, most students disputed this “fact.”

Nonetheless, in conversations the author held with about one hundred classmates after he started this project a quarter century ago, he only encountered alone classmate who knew of a single attempted incident of cheating. These conversations also seemed to confirm the fact – at least among the Princeton students of the early 1960s – that the fear of being dismissed was a primary motivator for students to comply with the very concept of honor itself.

Conclusions

Probably the most surprising conclusion of this work is the finding that students are suggesting cheating is decreasing while they rate the behaviors we would consider cheating more seriously of late. This holds for every relationship studied.

Over time the research revealed an increasing reluctance on the part of students to report cheaters for fear of being ostracized (summarized in McCabe et al. 2012). Students typically associated terms such as “narc” and “tattle” with reporting cheating. In addition, they seem to have no intention of reporting any cheating they might see. For example, the following are quotes from students at a major public university in the South:

Students are very unlikely to report another student, especially a friend, but possibly creating an anonymous system where people could report incidents could help.

I don’t feel students should be held responsible for other students’ cheating habits unless they are directly involved.

A student at a 2-year school in the Midwest seems to agree:

Students should play no part in academic integrity because it pertains to tattle tailing. What if a student tells on their ex friend for cheating when they really didn’t. There are a lot of immature students and allowing students to tell on each other will increase the caddishness in the future work field.

Finally, the body of research also revealed a tendency for students to seek out like-minded students in regard to cheating. The research has not yet identified a school where all the students surveyed felt there was absolutely no cheating at all. On every campus studied over the years (now approaching 200 campuses), there was a cluster of students who felt there is no cheating, while at the same time another cluster felt that cheating is rampant.

The difference evident between the campuses is the respective size of the two groups. Campuses with an honor code generally have a large segment of non-cheaters who seem to be attracted by the code, whereas those students with strong pro-cheating attitudes seem to seek out non-code schools. These two comments from students at the university in the South cited above seem to offer excellent summary comments:

Honestly, it is going to happen regardless. Students are smart enough to get around anything and everything that the University could do to regulate cheating. It is nearly impossible to regulate each and every single student. Realistically, no one is going to report someone cheating because no one cares. It is a serious issue but honestly all college students are in college for themselves. Someone can cheat on an exam but when they go to take the MCAT or LSAT, it is more difficult to cheat and they will just suffer there. I do the best that I can and that’s all I care about. This comment might sound harsh but it’s reality. I’m not going to take the time to report another student. I have myself to worry about and my grades. [This university] would probably be wasting time, money, and energy by trying to more highly regulate cheating. Students will just develop smarter ways to cheat.

Students know which classes and which professors tend to keep the same tests year after year, and it just supports the circular idea of cheating. If I know that a friend of mine has answers to all of an online class’ tests and assignments that have been passed down for a few semesters that haven’t changed, I’m much more likely to take that class, especially if it counts for a requirement that doesn’t fall within my major or interests.