Approach, Honor Code, Mediation, Organizational Context, Peer Environment ... (1) copied a few sentences of material without footnoting it in a paper; (2) padded a .... why young students entering college are less inclined to cheat, but they are then easily swayed by their ..... Of these students, 1578 (63.0%) were White,.
Why Students Cheat: A Conceptual Framework of Personal, Contextual and Situational Factors Hongwei Yu Baylor University, USA Perry L. Glanzer Baylor University, USA Byron Johnson Baylor University, USA ABSTRACT The authors presented an extensive summary of significant factors associated with college student cheating. The authors then compared these findings to a synthesis of the authors’ own research findings from empirical studies based upon a large national sample of college-aged students. Overall, the authors found student characteristics and precollege experiences (e.g. gender, age, family financial background, self-control, life purpose), individual college experiences and peer environment (extracurricular involvement, favorable perception of cheating environment), organizational context (student perception of faculty’s actions towards academic cheating) are all significant factors associated with academic cheating. More importantly, student academic preparation, extracurricular activities, attitude toward cheating, and perceived opportunities all served as important mediating variables between lack of selfcontrol and academic misconduct. Implications about research and practice and directions for future research were presented at the end of the chapter. Keywords: Academic Misconduct, College Impact Model, Faculty, Four-Frame Model, Holistic Approach, Honor Code, Mediation, Organizational Context, Peer Environment INTRODUCTION There is little doubt that academic misconduct among college students is a pervasive and serious problem on college campuses in the United States (Beasley, 2014, 2016; Bernardi, Baca, Landers, & Witek, 2008; Bertram Gallant & Drinan, 2006; Jurdi, Hage, & Chow, 2011; Levy & Rakovski, 2006; McCabe, Butterfield, & Treviño, 2012; Schmelkin, Gilbert, Spencer, Pincus, & Silva, 2008; Wowra, 2007). According to a review of various national studies over the past five decades, scholars found that more than two-thirds of college students consistently self-report being involved in incidents of academic dishonesty (McCabe, Butterfield, & Treviño, 2012, p. 71). To better understand this problem, scholars in the United States (US) undertook numerous studies of correlated factors of academic misconduct from educational, managerial, psychosocial,
1|Page
and sociological perspectives (e.g., Bowers, 1964; Davis, 1993; Haines, Diekhoff, LaBeff, & Clark, 1986; McCabe et al., 2012; Tibbetts & Myers, 1999; Vowell & Chen, 2004; Whitley, 1998). With rare exceptions, prior studies took a segmented rather than a holistic approach when studying student academic misconduct. In other words, they selected a few important factors that are hypothesized to be linked with academic misconduct (Whitley, 1998). While the first part of this chapter provided an overview of these findings, the authors agree with other scholars who acknowledged piece-meal strategies or initiatives derived from this segmented approach was ‘not the most effective way to manage the problem’ (Bertram Gallant & Drinan, 2006, p. 840). Given that a paucity of studies exist that took a systematic approach to study academic cheating, the authors introduced a conceptual framework that can help readers holistically study college students’ academic misconduct. The authors then reviewed the findings from prior studies using this framework. The authors further utilized this framework to summarize multivariate analysis of a large sample of college students at both two-year and four-year institutions based on the Gallup Organization’s daily tracking survey (n = 2,503). In particular, the authors examined how some of these important factors work together to explain or predict student academic misconduct/cheating among college students. In the end, the authors offered various suggestions for institutional initiatives and interventions that might be most effective in light of the research findings. BACKGROUND Defining Academic Dishonesty One important challenge when studying academic misconduct or cheating involved appropriately defining the term or terms used i.e., the authors used academic misconduct and academic cheating interchangeably throughout this chapter (Eve & Bromley, 1981; Haines, Diekhoff, LaBeff, & Clark, 1986; McCabe et al., 2012; Schmelkin, Gilbert, Spencer, & Silva, 2008). Previously, researchers such as Brown and Emmett (2001) suspected that studies reporting a high cheating rate among college students simply defined the term too broadly. Indeed, the variation in some findings among cheating studies often derived from a definition of academic cheating that focuses only upon one particular behavior such as plagiarism (e.g., Karlins, Michaels, & Podlogar, 1988) versus a whole range of behaviors (e.g., McCabe et al., 2012). Determining how best to define the range of behaviors may also prove challenging. One study simply left this task to the university and thus defined cheating as ‘behaviors that undermine academic integrity because they do not comply with [faculty or university] rules, norms, or expectations’ (Bertram Gallant, 2008, p. 10). The authors contended that a more helpful approach is the one adopted by McCabe et al. (2012) who defined academic misconduct/cheating as student engagement in a list of specific behaviors generally understood as cheating. Therefore, in the second half of the chapter the authors adopted this approach by defining academic misconduct/cheating as occurring when a student engages in one of the following nine behaviors largely adopted from the McCabe et al.’s (2012) definition: (1) copied a few sentences of material without footnoting it in a paper; (2) padded a bibliography or included sources in a bibliography that you did not use in the paper or project; (3) plagiarized from public materials on papers; (4) gotten questions or answers from someone who had already taken the exam; (5) worked on the same homework with several other students when the teacher
2|Page
did not allow it; (6) turned in papers done entirely or in part by other students; (7) given answers to other students during an exam; (8) copied off of another student during an exam; (9) taken unauthorized material, such as notes, into an examination. The authors should note that the limit of this approach is that it still assumes that students can recognize, for instance, when they ‘plagiarized from public materials on papers’. Yet, researchers knew that students and even faculty may not always agree upon what constitutes plagiarism (Roig, 1997, 2001). Nevertheless, the authors believed the behavioral approach to establishing a definition provides the best way to identify what cheating or academic misconduct means. CONCEPTUAL FRAMEWORK To help organize the presentation of the disparate research findings, they drew upon a college impact model/framework proposed by Terenzini and Reason (2005). The model was built using three decades of research on student learning and persistence at institutions of higher education (HEIs) (Reason, 2009) as well as the extended student learning and persistence models proposed by Astin (1985, 1993), Tinto (1975, 1993), Pascarella (1985), and Berger and Milem (2000). It identified four groups of variables reported to have an influence on student learning and behavior: (1) student precollege characteristics and experiences—includes socio-demographic traits, academic preparation and performance, and personal and social experiences; (2) organizational context-includes internal structures, policies and practices, academic and cocurricular programs, policies, and practices, and faculty culture; (3) peer environment; and (4) individual student college experiences-includes curricular and co-curricular experiences. In the overview below, the authors used this framework to discuss what factors are reported to be associated with academic misconduct among college-aged students (Figure 1). Insert Figure 1 Here LITERATURE REVIEW Student Precollege Characteristics and Experiences Individual students’ characteristics and experiences can exert a significant influence on subsequent student learning, change, and development (Terenzini & Reason, 2005). Prior studies of cheating suggested that a wide range of student characteristics are associated with academic misconduct, although in some cases the literature painted an equivocal picture. Some of the major individual factors considered important were gender, age, and the academic level of one’s parents. While these factors may be beyond an individual’s control, studies have found that certain character qualities and religious activities within one’s control were also related to academic cheating. In general, studies found that male students are more likely to report academic misconduct (Bowers, 1964; Davis, Grover, Becker, & McGregor, 1992; Tibbetts, 1999; Hensley, Kirkpatrick, & Burgoon, 2013) although there were a couple of exceptions that found women cheat more than men (Anton & Michael, 1983; Graham, Monday, O’ Brien, & Steffen, 1994; Leming, 1980). One of the explanations for this difference was that women were socialized to obey rules more than men (e.g., Ward & Beck, 1990). If this theory is correct, changing approaches to socialization might alter this pattern. In fact, a few studies indicated the gender
3|Page
gap may be disappearing or perhaps never existed in the first place (Lipson & McGavern, 1993; Ward & Beck, 1990). Another explanation could be that women simply self-report cheating less than men, an explanation supported by Harding et al.’s study (2007) which revealed that female students were less likely to report cheating on tests. Within the same major, the gender difference is ‘often very small’ (McCabe, Trevino, & Butterfield, 2001, p. 228), suggesting the interaction (moderation) effect be investigated when exploring the relationship between gender and academic misconduct. Researchers also found that gender plays a significant role in cheating intentions. While female students’ cheating intentions were reported to be primarily associated with anticipated shame of failing, male students’ cheating intentions were reported to be mostly linked with lack of self-control (Tibbetts, 1999). When it comes to the characteristic of age or class level, scholars consistently reported that younger college students cheat more than their older peers (Klein, Levenburg, McKendall, & Mothersell, 2007; McCabe & Treviño, 1997; Mustaine & Tewksbury, 2005; Olafson, Schraw, L. Nadelson, S. Nadelson, & Kehrwald, 2013). Yet, there was some lack of clarity regarding whether age or class (e.g., 1st, 2nd, 3rd or 4th Year) is the key factor since the two were highly correlated. One explanation for this finding was that first year or second year students are more likely to be at early stages of cognitive and moral development where they are swayed by peer influences and are therefore less likely to develop their own ideas and take appropriate actions (Baxter Magolda, 1999; Evans, Forney, & Guido-DiBrito, 1998). This theory would help explain why young students entering college are less inclined to cheat, but they are then easily swayed by their peer culture (McCabe et al., 2012). In addition, the educational attainment of one’s parents appears to be negatively related to students’ engagement in academic misconduct (Bowers, 1964; Kerkvliet, 1994). One suggested explanation offered for this difference is that students’ parents’ educational attainment is positively linked with student socioeconomic status, which means these students were more likely to have the resources to prepare academically and were therefore less likely to be engaged in academic cheating (McCabe et al., 2012). The authors would add that parents with more educational experience are more likely to know how to help their children study and navigate the educational experience. While the above individual student factors were less under students’ control, there is an increasing body of research that examines the relationship between students’ attitudes, habits, and character and cheating. In general, scholars found an important relationship between the two. For instance, studies suggested that the habit of cheating clearly starts earlier than college (Whitley, 1998). Researchers reported that students who are engaged in cheating at the high school level are more likely to report cheating at the college level (Harding, Carpenter, Finelli, & Passow, 2004, 2007; Rabi, Patton, Fjortoft, & Zgarrick, 2006). Research also suggested that students’ intentions to be engaged in academic cheating mediate the relationship between past cheating behaviors in high school and college cheating behavior (Harding et al., 2007). Using Ajzen’s (1991) ‘Theory of Planned Behavior’ (TPB), researchers examined the linkage between attitudes and behavior and reported that there is a significant positive relationship between attitude toward cheating and actual cheating behaviors (Bolin, 2004; Haines et al., 1986; Rettinger & Kramer, 2009).
4|Page
More recently, scholars begun to pay more attention to the association between academic honesty and ethically desirable traits. For instance, scholars found an inverse relationship between higher levels of self-control (Bolin, 2004), moral reasoning (Harding, Mayhew, Finelli, & Carpenter, 2007; Mayhew, Hubbard, Finelli, & Harding, 2009), aversion to cheating (Vandehey, Diekhoff, & LaBeff, 2007), and student academic misconduct. Other researchers have investigated the relationship between religiosity/religion and student academic cheating and found that while religious identity makes little difference in student academic misconduct, religious services attendance is found to be inversely related to academic cheating (Bloodgood, Turnley, & Mudrack, 2008; Huelsman, 2006; Storch & Storch, 2002). At least among students who are religious, those who attend religious services more frequently were associated with less engagement in academic misconduct in all the courses (Rettinger & Jordan, 2005). Individual Student Experiences Students’ own college experience in their academic and nonacademic lives also had a significant influence on student learning and development (Terenzini & Reason, 2005). In particular, researchers examined the influence of student college experience on academic misconduct and have found that both students’ curricular and co-curricular involvement are strong predictors of academic misconduct (McCabe et al., 2012; Whitley, 1998). With regard to curricular-related factors, three in particular was found important. First, students’ GPA was negatively associated with academic dishonesty. In other words, students with a high GPA cheat less and students with a low GPA cheat more (Aniton & Michael, 1983; Baird, 1980; Bowers, 1964; Haines et al., 1986; Klein et al., 2007; Lipson & McGavern, 1993; Michaels & Miethe, 1989; Olafson, et al., 2013; Roig & Caso, 2005; Singhal, 1982). Although, it should be noted that McCabe et al. (2012) found evidence that students in the top of the rankings who appeared extremely competitive also demonstrated a higher propensity to cheat. The second factor that appeared to be important, not surprisingly, is the level of students’ academic engagement or preparation. Scholars found that the level of student academic engagement/preparation is negatively associated with student academic misconduct (Davis, 1993; Whitley, 1998). Finally, researchers determined that students with certain majors, such as engineering and business are more likely to cheat than students in other majors (Clement, 2001; Harding, Finelli, & Carpenter, 2004, 2006; Levy & Rakovski, 2006; McCabe et al., 2012; Rettinger & Jordan, 2005; Smyth & Davis, 2004). A survey of 15,000 students of engineering, business, science, and humanities indicated that business major students are the least academically honest student population; a total of 87% of business major students self-reported academic misconduct (Caruana, Ramaseshan, & Ewing, 2000). Students majoring in business were reported to have permissive attitudes toward cheating and greater grade orientation, both of which can lead to more academic cheating (Rettinger & Jordan, 2005). Engineering major students also reported a higher cheating rate than natural science and social sciences major students (McCabe, 1997). Previous research estimated that 96% of engineering students committed academic cheating at least once during their college studies (Harding et al., 2004). When it comes to the co-curricular realm, two types of factors played an important role. First, increased levels of extracurricular involvement in general tended to be linked to cheating
5|Page
(McCabe & Treviño, 1993, 1997; Mustaine & Tewksbury, 2005). Logically, frequent involvement in extra-curricular activities often means students have little time to study, which increases students’ pressure to be engaged in academic misconduct (Ma, McCabe, & Liu, 2013). For example, Davis (1993) found one of the primary reasons for academic misconduct is that students often admit they ‘usually do not study’ (p. 26). Second, students who were heavily involved in particular co-curricular activities such as athletics and fraternity or sorority organizations are even more likely to report academic misconduct (McCabe & Treviño, 1993, 1997; Mustaine &Tewksbury, 2005; Whitley, 1998). Scholars claimed the relationship between fraternity/sorority membership and student cheating behaviors can perhaps be partially explained by social learning theory, which posited students learn knowledge and behaviors by observing other people’s behaviors and its consequences for them (McCabe & Treviño, 1993; Vowell & Chen, 2004). Informal Social Norms: Peer Environment The influence of peers has long been known to ‘exert normative power over student behavior’ (Reason, 2009, p. 671). Indeed, peer characteristics influenced every aspect of student development (cognitive, affective, psychological, and behavioral) (Astin, 1993). For instance, seminal studies conducted by William Bowers (1964), Donald McCabe and colleagues (1993, 1997, 2001, 2012) all reported peer environment as one of the most influential factors associated with student academic dishonesty. For instance, Bowers (1964) discovered a strong positive correlation exists between student self-reported cheating and their perception of peers’ attitude toward cheating. College students who perceived their peers permissive of cheating are more likely to report academic cheating. McCabe and Treviño (1993) also reported that student perception of their peers’ cheating behavior is the ‘most powerful influence on self-reported cheating’ (p. 533). In fact, the influence of peer environment on academic misconduct was even stronger than that of organizational norms (Bertram Gallant, 2008). Organizational Context Researchers acknowledged the role of organizational context in either inhibiting or promoting student academic dishonesty behaviors. As Lang (2013) claimed, the level of academic misconduct may ‘very well depend on the structures of the learning environment’ (p. 37). Numerous studies investigated the relationship between organizational context and student academic cheating (e.g., Bertram Gallant & Driana, 2006; Bertram Gallant, 2008; McCabe & Treviño, 1993, 1997). McCabe and various colleagues continually found that traditional honor codes can play an important role in decreasing academic misconduct (e.g., McCabe, 2003; McCabe & Treviño, 1993, 1997; McCabe, Treviño, & Butterfield, 1999, 2001; McCabe, Butterfield, & Treviño, 2012). These traditional codes usually required students to report themselves and others when a violation has occurred. Faculty must also report any violations and adjudication often heavily involves the students in the judicial process. One study suggested only 54% of students at honor code institutions reported engaging in one or more incidents of cheating compared to 71% of students on non-honor code institutions (McCabe, Treviño, & Butterfield, 2001). In addition, 53% of students at private institutions with honor codes, 58% of students at large public institutions with honor codes, and 68% of students at institutions with no honor codes admitted one or more instances of serious cheating (McCabe & Pavela, 2000). In particular,
6|Page
honor codes that are longer, require faculty and student reporting, and contain clearly stated consequences, often adjudicated by students, are associated with a lower level of academic misconduct (Gurung, Wilhelm, & Filz, 2012). The qualifiers are important since one group of researchers found that modified honor codes, where the faculty have the option of adjudicating violations on their own and students are less involved in the adjudication process, were not as effective as the traditional honor system described above (Schwartz, Tatum, & Hageman, 2013). McCabe and colleagues (1993, 1997, 1999, 2001) identified several reasons why an honor code may play a role in reducing student academic cheating behaviors. First, in these institutions, academic cheating was more clearly defined and students were less likely to rationalize their academic dishonesty behaviors (McCabe et al., 2012). Second, students have the responsibility of addressing peer academic cheating in honor code institutions. Students in honor code institutions were often given privileges such as unproctored exams, and they were more likely to preserve such rights by abiding by the academic honesty policies and rules (McCabe & Treviño, 1993; McCabe et al., 2012). Yet, the influence of honor codes on student academic cheating was not ‘sufficient in and of itself’ (Bertram Gallant & Drinan, 2006, p. 850). Even McCabe et al. (2012) acknowledged that the mere existence of honor codes is not effective in addressing academic dishonesty; rather, an effective honor code should be ‘well implemented and strongly embedded in the campus culture’ (p. 106). Bertram Gallant and Drinan (2006) investigated student academic cheating using organizational behavior theories. Integrating Bolman and Deal’s (2008) approach to manage change with Huntington’s (1968) theory of institutionalization, the authors offered specific strategies that help educational leaders manage academic misconduct on college campuses. For instance, they argued that educational leaders should acknowledge academic cheating is systematic and harmful to the values of academia, and they should communicate a ‘compelling version’ of curbing academic dishonesty and raise ‘standards and expectations of the organization’ in regard to academic cheating (Bertram Gallant & Drinan, p. 854). Lang (2013) argued how faculty construct the learning environment can have a great influence on student academic dishonesty and student learning on college campuses. He contended faculty who construct a learning environment that helps students develop an interest and fascination in the subject matter and see the relevance of the subject matter for their lives or futures reduces the environmental factors that lead to cheating. Moreover, classroom climate research suggested that student academic cheating can be significantly eliminated if faculty devote their attention to cheating and clearly communicate what constitutes academic cheating in their classroom (Beasley, 2014; Bertram Gallant, 2008). On the other hand, if students perceive faculty members are lenient toward cheating or show little interest in academic integrity, student academic cheating is more likely to occur (Bertram Gallant, 2008; Broeckelman-Post, 2008; Genereux & McLeod, 1995; Lang, 2013; Liebler, 2015). In reality, faculty members reacted to student academic misconduct quite differently, and a considerable percentage of faculty (40%) reported ignoring at least one incident of student academic cheating when responding to the surveys administered by McCabe and his colleagues between the year 2002 and the year 2010. Research indicated that many faculty ‘do little or nothing’ to respond to such behavior (Schmelkin, Gilbert, Spencer, Pincus, & Silva, 2008, p.
7|Page
588). Many faculty believed that the time and effort to report student academic cheating through ‘established channels is just not worth it’ (McCabe et al., 2012, p. 139). As can be ascertained from the overview, factors in all four broad categories outlined in the conceptual framework have been found to be associated with academic cheating/misconduct among college students: (1) student precollege characteristics and experiences—includes sociodemographic traits, academic preparation and performance, and personal and social experiences; (2) organizational context—includes internal structures, policies and practices, academic and cocurricular programs, policies, and practices, and faculty culture; (3) peer environment; and (4) individual student college experiences—includes curricular and co-curricular experiences. In the following section, the authors attempted to formulate a broader understanding of the relationship between these factors by summarizing research findings using this framework and compare research findings from prior studies with those from the analysis of a large national sample of college-aged students. THE FINDINGS: THE NATIONAL STUDY Despite the significant amount of research examining the factors mentioned above, McCabe et al. (2012) noted with regard to these areas, ‘we often don’t have enough high-quality research to make strong claims about such relationships. Clearly, more studies are needed’ (p. 89). The national study of American college students sought to address this need. The guiding research questions for the study were: (1) Whether and to what extent are the four groups of variables discussed above associated with student academic cheating?; and (2) In what ways do these variables interact with each other to explain college students’ academic cheating, if indeed they do? Data and Research Methods The respondents for the study were drawn from the Gallup daily tracking sample that is a nationally representative sample of US adults aged 18 or older conducted by Gallup every day, 350 days per year1. The final sample included a total of 2,503 college students from both twoyear and four-year institutions. Among these students, 1,451 of them (58%) were male students, while 1,052 of them (42%) were female students. Of these students, 1578 (63.0%) were White, 402 (16.1%) were Hispanic Americans, 300 (12.0%) were African Americans, 171 (6.8%) were Asian Americans, 33 (1.3%) were others that include American Indian or Alaska Native and Native Hawaiian or Pacific Islander, and 19 were missing this information (Table 1). Insert Table 1 Here Respondents were asked to self-report the frequency of their engagement in nine types of academic dishonesty behaviors during their college studies on a four-point Likert scale (Very Often = 1, Often = 2, Rarely =3, Never = 4) in the spring of 2014. The reliability coefficient (Cronbach’s alpha) for academic misconduct/cheating was .77, indicating good reliability. Data were drawn from a large sample of 2,503 college-aged students from the nationally representative Gallup daily tracking survey. Then, the authors conducted a series of multivariate analyses (ANOVA, hierarchical multiple regression, as well as structural equation modeling) to answer the proposed research questions listed above. The authors reported the research findings using the conceptual framework proposed by Terenzini and Reason (2005) described earlier.
8|Page
Student Characteristics and Experiences Consistent with past large-scale, multiple-campus studies (Bowers, 1964; McCabe & Treviño, 1997) and one meta-analysis (Whitley, Nelson, & Jones, 1999), the authors found a gender difference in student academic misconduct at HEIs. Through rigorous multivariate analysis, the authors found that female students reported fewer incidents of academic misconduct than their male classmates (-.143, -.128, -.112, -.082 p
