Academic and Personal Dishonesty in College Students

Abstract

Three studies used self-report measures to examine predictors of dishonesty across the personal and the academic domains of college students' lives. Dominance analysis methods (Azen & Budescu, 2003) determined the relative importance of predictors from a range of variables. High levels of predictability were found to be the result of a small set of dominant variables: dishonesty in the other domain, the impression management subscale of Paulhus's (1988) Balanced Inventory of Desirable Responding, and partying. This replicated pattern of results adds to a growing body of work that indicates dishonesty is related across domains. This behavioral consistency in our college student samples is discussed in terms of the substantive nature of the Paulhus (1988) scale and social-cognitive approaches to personality. ********** Dishonesty in both academic and personal domains appears to be quite common. Whitley's (1988) review of 48 studies in the academic domain reported a mean prevalence rate for total cheating of 70%. While less research has investigated the extent of dishonesty in personal relationships, there appears to be correspondingly high prevalence rates for a variety of behaviors. For example, Drigotas, Safstrom and Gentilia (1999) found that 72% of college-aged students involved in exclusive dating relationships admitted to some degree of emotional intimacy with another partner, while 48% reported physical intimacy with another. Similarly, Kashy and DePaulo (1996) estimated that about one in four social interactions involve lying. Is there a relationship between academic and personal dishonesty for college students and if so, can the relationship be accounted for by a common set of variables? Our strategy was to evaluate the relative contribution of situational and individual difference predictors, a frequent categorization of predictors in the academic dishonesty literature. These goals were furthered by using a statistical methodology, dominance analysis (Azen & Budescu, 2003), designed to assess the relative importance of an individual predictor in the context of other predictors. Dominance analysis improves upon standard multiple regression techniques in that it does not assume that the effects of other predictors are held constant, a tenuous assumption when predictors are correlated. Our definition of dishonesty was broad, encompassing a variety of concrete behaviors. What all the behaviors have in common is that they violate established standards of behavior. The research in academic dishonesty largely reflects this broad definition and we have adopted that approach for both domains considered here. In the academic domain, for example, Newstead, Franklyn-Stokes and Armstead (1996) emphasized that cheating behaviors consist of a variety of behaviors with some being more common than others. An extensive body of research has explored the predictors of academic dishonesty (Whitley, 1998). Until recently, this work often categorized predictors as either situational or dispositional in nature. Hartshorne and May's (1928) classic studies of deceit indicated that schoolchildren showed surprisingly low cross-situational consistency in their dishonest behaviors in the academic domain. Their findings were interpreted as demonstrating the dominance of situational factors in dishonesty. Subsequent research supported this position. McCabe and Trevino (1997) examined individual and situational factors of student cheating in a study of approximately 1,800 students at 9 different universities and concluded that situational factors such as peer acceptance, peer disapproval, and severity of penalties were the predominant predictors of academic cheating. Pulvers and Diekhoff (1999) found that perceptions of the classroom environment predicted academic dishonesty. Whitley's (1998) review of factors associated with academic cheating across 107 studies from 1970 to 1996 similarly found large effects for the perception of social norms that condone cheating and moderate effects for class size, perceived work load, and aspects of the testing environment. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.