Examining Cognitive Predictors Of Academic Cheating Among Urban Middle School Students: The Role of Home-School Dissonance Free

Academic cheating is a significant international educational quandary that (1) is pervasive from middle school throughout postsecondary education, (2) has many forms, including plagiarism, digital/online cheating, and contract cheating, and (3) varies among students with different academic capabilities (i.e., gifted and talented and low-performing) (Anderman & Midgley, 2004; Bernandi et al., 2004; Ma, Wan, & Lu, 2008; Murdock & Anderman, 2006; Walker & Townley, 2012). According to some research, including reports from the Josephson Institute of Ethics and the Center for Academic Integrity at Duke University, up to 90% of students have cheated at least once prior to completing secondary education (McCabe, 1999; McCabe & Trevino, 1993, 1997). Other works have demonstrated that academic cheating is often recurrent among many students rather than an isolated and singular incident; that is, students’ engagement in academic cheating tends to increase rather than decrease with time and transition to higher grade levels (Anderman & Midgley, 2004). The recurrence of academic cheating among students suggests that the decision to cheat may be informed by factors within the individual and his/her thinking about their ability to excel (Brunell, Staats, Barden, & Hupp, 2011; Eisenberg, 2004; Finn & Frone, 2004; Murdock & Anderman, 2006; Murdock, Hale, & Weber, 2001; Niiya, Ballantyne, North, & Crocker, 2008) or factors within the schooling environment that may influence his/ her thinking about the importance of doing well in school (i.e., contextual factors) (Evans & Craig, 1990; Murdock et al., 2001; Tas & Tekkaya, 2010).

To address the possible sources of academic cheating, some educational researchers have begun to examine the psychological and social/contextual correlates of academic cheating (Anderman, Griesinger, & Westerfield, 1998; Anderman & Midgley, 2004; Anderman & Murdock, 2007; Finn & Frone, 2004; Murdock & Anderman, 2006; Murdock et al., 2001). A consistent finding among these studies—which have sampled participants ranging from middle school to college students—is that students’ overall feelings about and attitudes toward school and their teachers are associated with academic cheating behaviors (Anderman & Midgley, 2004; Calabrese & Cochran, 1990; Finn & Frone, 2004; Murdock et al., 2001; Taylor, Pogrebin, & Dodge, 2006). Some studies have reported that students have lax or lenient attitudes toward academic cheating (Jensen, Arnett, Feldman, & Cauffman, 2002; Stephen & Wangaard, 2013).

Many of these feelings emerge from perceptions of the schooling context, particularly the achievement goals within students’ classrooms and/or learning environments. Niiya and colleagues (2008), for example, found that male college students engage in cheating behaviors in order to preserve their self-worth within a competitive context where they do not expect or have not performed well on a given academic task. Other research has shown that students who have a lower sense of belonging at their school and overall negative perceptions of school and their classroom teachers engage in higher academic cheating (Calabrese & Cochran, 1990; Finn & Frone, 2004; Murdock et al., 2001). Also, students’ perceptions of classroom performance goal structures (i.e., whether a given classroom and its goals for learning are reflective of greater comprehension and mastery or if such classroom goals reflect an exclusive orientation toward optimal performance among peers) are linked to academic cheating behaviors (Anderman et al., 1998; Murdock et al., 2001; Peklag, Kalin, Pecjak, Zuljan, & Levpuscek, 2012; Tas & Tekkaya, 2010). In essence, it is likely that students who cheat—particularly those in the middle and high school levels—may do so as a consequence of being in a formal learning environment that (a) stresses performance outcomes over comprehension and overall mastering of class material and/or (b) undervalues and alienates them (Anderman & Midgley, 2004; Evans & Craig, 1990).

The purpose of this study is to extend this line of research by examining middle school students’ perceptions of home-school dissonance and their association with academic cheating reports. Home-school dissonance is the perceived difference between the values and operations extant in students’ home and those values and operations salient within their formal schooling experiences (Arunkumar, Midgley, & Urdan, 1999; Kumar, 2006). Some empirical reports have uncovered the role of home-school dissonance on student performance and school-related behaviors and attitudes (Arunkumar et al., 1999; Kumar, 2006; Tyler et al., 2010). No study to date, however, has examined whether home-school dissonance is a significant predictor of academic cheating among middle school students. Examining the association between home-school dissonance and academic cheating with a middle school sample is particularly noteworthy, as academic cheating peaks during high school (Cizek, 1999; Davis, Grover, Becker, & McGregor, 1992; Evans & Craig, 1990; Jensen et al., 2002), but is likely to begin during the middle school years, where students endure significant transitional periods inclusive of matu- rational and developmental changes, as well as changes within the formal learning environment. Specifically, the middle school years have been found to be particularly challenging as students, likely for the first time in their academic careers, are exposed to increased class sizes, larger number of teachers and greater academic responsibilities, including press from teachers and administrators to do well on standardized exams, larger school buildings, and less nurturance/social interaction with classroom teachers as well as significant cognitive and physical developments (Bailey, Giles, & Rogers, 2015; Smart, 2014). Thus, it is likely that these changes and contextual pressures may provide a stimulus for academic cheating behaviors. In addition, this study will also explore the cognitive factors associated with academic cheating specifically by examining whether students’ reports of several academically related cognitive factors (i.e., academic efficacy and achievement goal orientations) mediate the relationship between home-school dissonance and academic cheating. A review of the literature on academic cheating, its sources, and correlates such as home-school dissonance is provided below.

Athanasou and Olasehinde (2002) describe academic cheating as students’ conscious involvement or participation in deception (i.e., lying, falsifying, misrepresenting, corruption, plagiarism, copying, or the unlawful assistance provided to someone else), typically for the purpose of performing well or giving the perception of performing well on an academic task. Cizek (1999) offers three classifications of cheating behaviors, which include (a) cheating by taking, giving, or receiving information from others (e.g., copying another student’s work with or without their permission), (b) cheating through the use of forbidden materials or information (e.g., plagiarism), or (c) cheating by circumventing the process of assessment (e.g., a student taking or having someone else take an examination for another).

Recently, Raskin (2013) indicated that the academic cheating phenomenon cannot be resolved with a focus on more tests, more test preparation and thus, more test anxiety. Rather, the author argues for a rethinking and redesigning of learning contexts. Raskin (2013) identifies the source of academic cheating being the formal learning context itself, while many educational and psychological research studies on academic cheating have attributed the phenomenon to multiple motivational influences within the individual as well as within his or her context. Some research studies in these fields have included individual factors such as gender (Fisher & Brunell, 2014; Whitley, 1998; Whitley, Nelson, & Jones, 1999), self-worth (Niiya et al., 2008), perfectionism (Krone, Rouse, & Bauer, 2012), narcissism (Brunell et al., 2011), morality (Eisenburg, 2004), fear of failure (Taylor et al., 2006), materialism (Koul, 2012), academic performance (e.g., grades and time management) (Jones, 2011), and school identification (Finn & Frone, 2004), while contextual factors have included parental support/parent-child interaction/peer interactions (Bong, 2008; Carrell, Malmstrom, & West, 2008), studentteacher interactions (MacGregor & Stuebs, 2012), low reinforcement of academic punishment for cheating (Ma et al., 2008; Ma, Lu, Turner, & Wan, 2007) and perceptions of both teacher and classroom learning goal orientation (Bong, 2008; Glazer, 2013; Peklaj et al., 2012).

Addressing the presence of both individual and contextual antecedents for academic cheating, Anderman and Murdock (2007) proposed a conceptual model that underscores the individual and contextual factors that precede the decision to engage in academic cheating. In their model, Murdock and Anderman introduce several microlevel, context-based factors (e.g., home and school) that influence students’ decision to engage in academic cheating. The second set of theories referred to in Murdock and Anderman’s model include person-centered, individual factors such as intrinsic motivation and achievement goal theories, which explain achievement by examining students’ cognitive rationales for engaging in school work (Anderman & Murdock, 2007; Murdock & Anderman, 2006). Murdock and Anderman (2006) argue that academic cheating is not typical among those students possessing a high intrinsic motivation and/or mastery goal orientation (i.e., learners who value understanding and deep cognitive processing). Rather, academic cheating is considered characteristic of students with high extrinsic motivation and performance goal orientations (i.e., learners more interested looking like they are doing well, Murdock & Anderman, 2006). Thus, in their model, Murdock and Anderman (2006) suggest that a performance goal orientation versus a mastery goal along with an extrinsic motivation versus an intrinsic motivation for achievement is directly linked to academic cheating behaviors.

Though Murdock and Anderman (2006) have constructed a conceptual model that describes both individual-level and contextbased factors that facilitate the decision to engage in academic cheating, their conceptual model does not exhaust the list of possible individual and contextual factors that may be predictive of academic cheating behaviors. To be sure, the literature is replete with studies that have shown how perceptions of self and classroom goal orientations are statistically associated with academic cheating (Anderman et al., 1998; Bong, 2008; Koul, 2012; Murdock & Anderman, 2006; Murdock et al., 2001; Niiya et al., 2008; Tas & Tekkaya, 2010), with each study indicating a positive association between performance goal orientations and academic cheating and negative associations between mastery orientations and academic cheating. Yet, in keeping with the focus on identifying contextual as well as individual factors associated with academic cheating, it is necessary to examine additional variables that may explain variance in academic cheating, particularly among middle school students who—given the transition from elementary to middle school and the accompanying academic and social pressures—may become more prone to cheating behaviors at this stage in their academic careers. Two variables to be considered in the current study include homeschool dissonance and academic efficacy as possible predictors of academic cheating reports among middle school students.

Kumar defines home-school dissonance as the perceived differences between the values and operations extant in students’ home or outof-school environment and those values and operations salient within their formal schooling experiences (Arunkumar et al., 1999; Kumar, 2006). These values and operations can include, but are not limited to preferred modes of learning and working (e.g., working with others versus working alone). According to Kumar, students from all grade levels experience dissonance when the cultural values, beliefs, and norms of their home are incongruent with those found in the school. In particular, Arunkumar and colleagues write “students from cultures outside the mainstream may experience a sense of dissonance when they encounter a devaluing of their beliefs and behaviors at schools that reflect the dominant White, middle-class ideology” (p. 442). The effects of exposure to a dissimilar or dissonant learning environment have proven, however, to be debilitating for many students, including White, middle-class students.

For example, Arunkumar et al. (1999) found no significant differences in reports of home-school dissonance between African American and European American middlegrade students. However, they showed that students reporting high levels of home-school dissonance also reported lower levels of future hopefulness, academic efficacy, self-esteem, and grade point average (GPA). Arunkumar and colleagues (1999) also reported a positive association between home-school dissonance reports and higher levels of anger and self-deprecation (Arunkumar et al., 1999). In a later study, Kumar (2006) used multilevel growth curve analysis to examine the associations between middle school students’ perceptions of classroom goal structures and teachers’ reported classroom practices and home-school dissonance. Analyses revealed that middle school students’ perceptions of classroom performance goals were predictive of home-school dissonance. In addition, teachers’ reported mastery goal instructional practices were significantly related to lower home-school dissonance reports. Additional studies have found similar results with home-school dissonance as a predictor of various academic outcomes and their psychological/cognitive antecedents among high school students (Brown-Wright et al., 2013; Tyler et al., 2010).

Given the predictive nature of home-school dissonance across these studies, it is expected that home-school dissonance reports will be predictive of academic cheating. In addition to examining this association, the data will also be examined for any possible mediation effects for the proposed home-school dissonance—academic cheating relationship. Similar to other research that has shown academic cheating to be the result of a cognitive/motivational process (Anderman et al., 1998; Murdock et al., 2001; Murdock et al., 2004), wherein students’ motivation to cheat is linked to contextual as well as personal factors, home-school dissonance may facilitate in students’ decisions to cheat specifically by being associated with an increase or decrease of some academically relevant cognitive factor. That is, homeschool dissonance “does” something to the middle school student, specifically through its association with cognitive factors germane to his or her school performance. Unlike Kumar’s (2006) study, where home-school dissonance was viewed as a criterion variable to be predicted by learning goal orientation factors, in the current study, home-school dissonance is used as a predictor variable. Here, it is believed that perceptions of middle school students’ formal learning context (i.e., the degree of dissonance between their home and school) may impact their decisions to engage in academic cheating, particularly by impacting their own achievement goal orientations and academic efficacy. It is hypothesized, therefore, that home-school dissonance reports will be associated with middle school students’ reports of academic efficacy and achievement goal orientation and in turn, these factors will be associated with academic cheating reports. The major research question driving this study asks, “Does the reported degree of dissonance between home and school have any association with middle school students’ academic cheating behaviors?” Below is a brief discussion of the variables serving as potential mediators in the current study.

Academic efficacy is an individual’s confidence that he or she can carry out academic tasks successfully (Mercer, Nellis, Martinez, & Kirk, 2011). Widely regarded in the educational research literature as the strongest predictor of academic performance, academic self-efficacy is derived from self-efficacy (Weiser & Riggio, 2010), a component of social cognitive theory, introduced in the psychological literature by Bandura (1986). Selfefficacy is considered the most significant of cognitive factors as all others are “rooted in the core belief that one has the power to produce effects by one’s actions” (p.187) (Bandura, Barnaranelli, Caprara, & Pastorelli, 2001). In academic settings, efficacy is operationalized as the degree of interest in and persistence with challenging and often, difficult tasks (Baird, Scott, Dearing, & Hamill, 2009). Students with high academic efficacy are viewed as working harder, participating more readily, highly pursuant of challenging tasks and achieving goals, having incremental views of intelligence, and persisting in the face of adversity and/or task difficulty (Weiser & Riggio, 2010). Regarding academic performance outcomes, higher academic efficacy has been associated with grade point average and higher grades (Caraway, Tucker, Reinke, & Hall, 2003; Finn & Frone, 2004; Mercer et al., 2011). Lower levels of academic efficacy have been associated with maladaptive schooling behaviors and attitudes such as poor academic performance, classroom disruptive behaviors and academic cheating (Finn & Frone, 2004).

Achievement goal orientation is a motivation-based set of cognitive variables in the educational research literature that articulate varying motives and purposes students may possess for engaging achievement-related behaviors (Anderman, Andrzejewski, & Allen, 2011; Midgley, 2002). Students possessing a mastery goal orientation toward their achievement-related behaviors strive to develop and refine their competence in either several or all of their academic domains. Specifically, they welcome the challenge of difficult concepts and lessons as these often provide the cognitive growth mastery oriented students seek while engaged with academic tasks. As a result, mastery oriented students tend to have positive academic and affective outcomes, including adaptive attitudes toward school, higher achievement, and thus, high academic efficacy (Anderman et al., 2011; Peklaj et al., 2012). In contrast, performance oriented students often used social comparison rather than individual cognitive growth as the motivation factors preceding their engagement in academic tasks (Anderman et al., 2011). Specifically, performance approach students often compare their performance with other students as they seek to demonstrate their knowledge of subject matter publicly, while performance avoidance students tend to limit their achievement-related behaviors in an effort not to appear incompetent (Peklaj et al., 2012). While the negative association between performance avoidance goal orientation and achievement outcomes has been stable (Peklaj et al., 2012), the literature has been mixed in reporting the effects of performance approach goal orientation on achievement outcomes and corresponding cognitive variables like academic efficacy (Linnenbrink, 2005; Midgley, 2002).

Given the aforementioned results, the purpose of this study is twofold: First, I will examine the statistical association between homeschool dissonance and academic cheating among middle grade students. Second, I will examine whether such a relationship is mediated by academic efficacy and achievement goal orientation reports. It is expected that academic efficacy and mastery goal orientation will mediate the relationship between homeschool dissonance and academic cheating. Academic efficacy and achievement goal orientation is a mediator variable for homeschool dissonance and academic cheating because it is expected that aspects of the formal learning context (e.g., home-school dissonance) often shape the cognitive factors (e.g., goal orientation and academic efficacy) deemed relevant to schooling outcomes (e.g., academic cheating).

Six hundred and sixty (N = 660) middle school students from two randomly selected middle schools in the Southeastern region of the country participated in the current study. Sixty eight percent of the participants attended one of the middle schools. Nearly 53% of the total sample was African American, while 47% of the sample was Caucasian. The sample was virtually even in gender with 49.7% male middle school students and 50.2% being female middle school students. One student did not report his/her/their gender. Additionally, 28% of the sample emerged from the sixth grade, nearly 40% from the seventh grade, and 31% from the eighth grade. Six students did not indicate their grade level. The age range for middle school participants was 10 years to 15 years of age with 13 years of age being the largest percentage (35.5%). Selfreported grades among the middle school participants yielded over 75% of students receiving As and Bs for language arts, 68% receiving As and Bs for math, 70% receiving As and Bs for science, and 72% receiving As and Bs for social studies. Average self-reported GPA for student participants was 3.25.

Demographic information. Several variables provided demographic information on the study sample. These included the categorical variables of ethnicity (1 = African American, 2 = White), grade (1 = sixth grade, 2 = seventh grade, 3 = eighth grade), and gender (1 = male, 2 = female) and a continuous GPA where students reported information about their current GPA when available. Middle grade student participants were also asked on the demographic page about their current grades and responded to the queries by indicating the letter grades they received in Language arts, math, social studies, and science. Scale responses for these ordinal variables included 1 = “F”, 2 = “D”, 3 = “C”, 4 = “B”, and 5 = “A”.

Patterns of Adaptive Learning Scales. The Patterns of Adaptive Learning scale (PALS; Midgley et al., 2000) was developed to examine the relationship between student motivation, affect, and behavior and the learning environment. Items on the PALS are scored on a 5-point Likert-type scale from 1 (not at all true) to 5 (very true). The PALS have been administered to ethnically diverse samples in elementary, middle, and high school. In the current study, the cheating behavior subscale (3 items, α = .86) is the criterion variable. A sample item from the cheating behavior subscale was, “I sometimes copy answers from other students during tests.” The dissonance between home and school subscale (5 items, α = .88) was the principal predictor variable in the current study. A sample item from the dissonance between home and school subscale was, “I feel troubled because my home life and my school life are like two different worlds.” Regarding the proposed mediator variables, hereby referred to as “cognitive variables,” a sample item from the academic efficacy subscale (5 items, α = .78) includes “Even if the work is hard, I can learn it.” A sample item from the mastery goal orientation subscale (5 items, α = .85) is “One of my goals in class is to learn as much as I can.” A sample item from the performance-approach goal orientation (5 items, α = .89) is “One of my goals is to show others that I’m good at my class work.” Finally, a sample item from the performanceavoidance goal orientation (4 items, α = .74) is “One of my goals in class is to avoid looking like I have trouble doing the work.”

Institutional review board approval was obtained from the institution hosting the research. The associate superintendent for research for the public school system also granted approval for research where the two participating middle schools were located. Subsequently, an initial meeting was arranged with each school’s administrative personnel to introduce the study and to coordinate data collection. Written informed consent was obtained from participants’ parents while student assent was obtained from all participants prior to survey completion. The survey packet was administered to participants during a single classroom session in one day, and students were given 45 minutes to complete the survey protocol. Students received $5 department store gift cards for their participation and completion of the surveys.

Several preliminary analyses were performed prior to analyses examining the initial research question. To begin, a MANCOVA was computed to determine whether scores on academic cheating, academic efficacy, mastery goal, performance approach, performance avoidance, and home-school dissonance vary as a function of ethnicity, gender, and/or class rank. GPA served as the covariate and was self-reported in the current study. In addition, a bivariate correlation matrix was computed to determine whether significant associations emerged among home-school dissonance, the cognitive variables, and academic cheating. Along with these correlation analyses, a series of linear regression models was computed to determine whether (a) home-school dissonance was predictive of the cognitive variables (i.e., academic efficacy, mastery goal, performance approach, and performance avoidance), (b) the cognitive variables were predictive of academic cheating, and (c) home-school dissonance was predictive of academic cheating. According to several researchers (Baron & Kenny, 1986; MacKinnon, Fairchild, & Fritz, 2007), these analyses are aligned with the criteria necessary to examine mediation effects. Finally, a Sobel test of mediation (Preacher & Hayes, 2004, 2008) was performed to determine whether each cognitive variable mediates the relationship between home-school dissonance and academic cheating.

A MANCOVA was computed to determine whether scores in home-school dissonance, cognitive variables, and academic cheating vary as a function of ethnicity, gender, school, and class rank, with student reported GPA as a covariate. MANCOVA is the appropriate analytical tool to use when the researcher seeks to minimize error variance—through statistical control of a given covariate—while examining the effects of multiple independent variables on multiple dependent variables (Mertler & Vannatta, 2010; Weinfurt, 1995). Significant F statistics emerged for the ethnicity [F(10, 461) = 4.54, p <. 01, η² = .09] and gender [F(10, 461) = 1.65, p <.01, η² = .03] variables. No other variables or interaction terms between them were statistically significant. Separate univariate analyses were computed to examine where statistically significant differences emerged for the ethnicity and gender main effects. For ethnicity, significant F statistics emerged for home-school dissonance [F(1, 470) = 5.15, p = .02, η² = .01], a [F(1, 470) = 10.65, p < .02, η² = .02], performance approach goal orientation [F(1, 470) = 4.99, p = .03, η² = .01], and GPA [F(1, 470) = 31.67, p < .02, η² = .06]. However, upon closer inspection of the effect sizes for significant F statistics, along with the means between the African American and Caucasian middle grade students across these factors, it was concluded that only the ethnicity main effect for GPA was practically meaningful. Specifically, African American middle school students, on average, reported very slightly higher scores on homeschool dissonance (2.38 versus 2.30) and performance approach (3.11 versus 3.01). Average GPA scores between the two groups yielded more interpretable significance; with African American middle grade students reporting significantly lower GPAs (2.80) than their Caucasian counterparts (3.46). Regarding gender, significant F statistics emerged for the performance approach [F(1, 470) = 4.63, p = .03, η² = .01], and performance avoidance [F(1, 470) = 6.63, p = .01, η² = .01] variables. Similar to the ethnicity main effect, mean scores between male and female middle school students for performance approach (3.11 versus 3.04) and performance avoidance (3.09 versus 3.02) had less than 1/10th of a point difference and thus, were not practically significant. Table 1 presents the overall descriptive statistics, including dependent variable means and bivariate correlations for the study variables.

Regression analyses were computed to determine the predictive ability of homeschool dissonance and the cognitive variables on academic cheating. Given that student demographic characteristics were not meaningfully associated with home-school dissonance, the cognitive variables, or academic cheating, the regression models only examined home-school dissonance and each cognitive variable (i.e., academic efficacy, mastery goal orientation, performance approach goal orientation, and performance avoidance goal orientation) as predictors of academic cheating. Stepwise regression analyses were conducted to determine how much variance in academic cheating each predictor variable could account for. The first regression model contained home-school dissonance as the predictor of academic cheating and emerged statistically significant, F(1, 658) = 116.04, p < .01. Home-school dissonance accounted for 15% of the variance in academic cheating. The initial standardized β coefficient for homeschool dissonance was .39 (t = 10.77, p < .01). With the stepwise regression, academic efficacy was entered in the second step as it was believed to be the strongest cognitive factor. The second regression model containing both home-school dissonance and academic efficacy also emerged statistically significant in predicting academic cheating, F(2, 657) = 81.00, p < .01. Here, home-school dissonance and academic efficacy accounted for 20% of the variance in academic cheating, with academic efficacy accounting uniquely for 5%. The standardized 0 coefficients for homeschool dissonance and academic efficacy in the second model were .35 (t = 2.19, p < .01) and -.22 (t = -6.23, p < .01), respectively. Taken together, the goal orientation variables, though each statistically significant in its prediction of academic cheating, only accounted for a combined 5% of the variance in academic cheating reports, with roughly 4% of variance accounted for by performance approach goal orientation variable (β = .13, t = 2.84, p < .01). In the last stepwise regression analyses, the regression model remained statistically significant F(5, 654) = 43.70, p < .01, with the combination of home-school dissonance and all four cognitive variables explaining 25% of the variance in academic cheating reports.

In addition to these regression analyses, several linear regression analyses were performed to determine whether home-school dissonance significantly predicted the proposed mediators for the association between home-school dissonance and academic cheating. The linear regression analyses showed that home-school dissonance was a significant predictor of each proposed mediator including academic efficacy (β = -.16, t = -4.20, p < .01), mastery goal orientation (β = -.13, t = -3.30, p < .01), performance approach goal orientation (β = .26, t = 6.97, p < .01), and performance avoidance goal orientation (β = .24, t= 6.32, p < .01).

With each predictive association being significant for each mediator variable, along with each mediator variable significantly predictive of academic cheating, four independent Sobel analyses were computed to examine the mediation of the home-school dissonance—academic cheating association. Though several quantitative researchers have argued for the use of bootstrapping to examine mediation over Sobel analyses (Hayes, 2009; MacKinnon et al., 2007), limited training in bootstrapping techniques and limited access to the statistical software containing such an analytic technique (i.e., R, M-Plus, PROCESS) precluded the use of the preferred bootstrapping mediation analysis. Nonetheless, the sample size in the current study allows for the use of the Sobel statistical method, which can minimize Type I error and maintain adequate power so long as the sample size is large (Preacher & Hayes, 2008). Specifically the Sobel analysis detects significant mediation after an exploration of the associations between independent variable and the mediating variable and the mediating variable and the dependent variable has occurred (Baron & Kenny, 1986; Preacher & Hayes, 2004, 2008). This causal steps approach requires the computation of unstandardized regression coefficients between the independent variable and mediator, the mediator and dependent variable and the independent variable and dependent variable. Sobel detects mediation once these criteria are met.

Given that there were four cognitive factors examined as mediating variables in the current study (i.e., academic efficacy, mastery goal orientation, performance approach orientation, and performance avoidance orientation), four sets of regression coefficients—based on the causal steps approach—were imputed into an online Sobel analysis calculator. This imputation allowed statistical evidence to emerge to either support or refute the presence of mediation for the home-school dissonance and academic cheating association. Figures 1-4 illustrate each cognitive variable serving as a significant mediator of the association between home-school dissonance and academic cheating, with the mediation of mastery goal orientation achieving only marginal significance.

The purpose of this study was to examine home-school dissonance as a significant predictor of academic cheating among urban middle school students. To better understand this association between a specific context-related variable such as home-school dissonance and academic cheating, several cognitive variables were included in the study as potential mediators. The means for home-school dissonance and academic cheating were below the scale midpoints of 3, indicating that both factors were not highly reported or endorsed by the middle school students. Means for the cognitive variables (i.e., academic efficacy, mastery goal orientation, performance approach goal orientation, and performance avoidance goal orientation) were much higher (i.e., at or above the scale midpoint). Here, academic efficacy and mastery goal orientation were endorsed highly among middle school students while performance approach and performance avoidance goal orientations were moderately endorsed.

Both home-school dissonance and each of the cognitive predictors were statistically associated with academic cheating. In testing the mediation for home-school dissonance and academic cheating association, three of the four mediation models emerged statistically significant. Specifically, academic efficacy, performance approach, and performance avoidance goal orientations proved to be statistically significant mediators of the homeschool dissonance—academic cheating association. Specifically, scores on home-school dissonance were associated with higher academic efficacy reports, which were significantly associated with lower reports of academic cheating among the middle school sample. Also, scores on home-school dissonance were associated with lower scores on both performance approach and performance avoidance, which, in turn, were associated with academic cheating reports.

Given the literature and its articulation of performance approach and performance avoidance goal orientations fostering maladaptive schooling behaviors such as academic cheating (Bong, 2008; Koul, 2012; Kumar, 2006; Murdock & Anderman, 2006; Tas & Tekkaya, 2010), the positive associations between performance approach and performance avoidance goal orientations and academic cheating are consistent with previous findings. It is likely that the need to impress others by either appearing smart (i.e., performance approach) or not looking incompetent (i.e., performance avoidance) may fuel the decision to cheat in school. The data suggest that such a decision could be further exacerbated by the difficulties some middle school students may face as a result of perceived home-school dissonance. In a similar vein, academic efficacy and mastery orientation were negative associated with academic cheating, two findings which also fit the literature (Barzegar & Khezri, 2012; Koul, 2012). Here, high academic efficacy implied that middle grade students have strong beliefs in their ability to achieve at school, while a high mastery goal orientation implied a strong focus on comprehension and doing well in school. Study data suggest that the low homeschool dissonance reports among these middle grade students may have facilitated the development of higher academic efficacy and mastery goal orientation and thus, lower academic cheating reports.

The current study has limitations that should be addressed in future research. To begin, the nature of this study was correlational and thus, causality cannot be implied throughout the findings. Regarding statistical analyses, future research should look to employ more refined mediation analyses such as bootstrapping and multiple mediator analyses via the use of structural equation modeling to simultaneously assess the presence of the multiple mediators while having better control over Type I error and higher statistical power to detect significant mediation effects (Preacher & Hayes, 2008). The lack of program software availability or training in this quantitative method, coupled with the relative size of the current study sample also influenced the decision to execute Sobel analyses. Moreover, the Sobel analysis continues to be viewed as a viable mediation analysis technique (Boone, 2012). However, future research examining the effects of homeschool dissonance should aligned itself with the quantitative research literature that has moved toward alternative mediation techniques such as bootstrapping (Hayes, 2009; MacKinnon et al., 2007; Shrout & Bolger, 2002).

In addition, academic cheating was selfreported and it is likely that students may have attempted to underestimate the degree to which they engage in these behaviors. Future studies should include a social desirability scale to determine the degree to which academic cheating reports are reflective of socially desirable responding by students. Similarly, future research should access school-level grade reports rather than rely on self-reported data from middle grade students who, at the time of research participation, may not be fully aware or accurate about their GPA or current grades. Moreover, given the relatively low percentage of variance accounted for by the predictors in the current study, additional data should be collected to discern whether academic cheating can be predicted by other cognitive (i.e., racial/ethnic identity) and contextual factors (i.e., classroom goal orientation). Finally, the home-school dissonance measure—in its current form—does not indicate what the perceived differences are between home and school. Future studies should look to develop and/or utilize more specific instrumentation to assess possible distinctions in the cultural, structural, and procedural of middle grade students’ home experiences and those within their schools.

In terms of middle grade schooling, the findings that students have low reports of home-school dissonance and academic cheating are encouraging. Though generalizability is limited as this was not a national sample of middle grade students, these findings, coupled with the self-reports of students’ academic performance, suggest that many middle grade students are doing moderately well in middle school. Educators are encouraged to continue those instructional and administrative practices that bridge the home and schooling experiences of middle grade students. Such practices, as the literature reports (e.g., frequent and positive student-teacher interactions, incorporating aspects of students’ cultural value-based behaviors and learning preferences into the classroom) are associated with their academic efficacy and their orientation toward developing mastery within their classes. Additionally, teachers should promote greater home-school continuity by assessing students’ learning preferences from their parents, thereby facilitating the transition to middle school. Determining the use of technology usage at home for middle grade students and incorporating lessons that contain such information would also demonstrate some commitment to bridging the perceived gap between what teachers have their students do in class versus the learning and knowledge retrieval and building processes that students use outside of school. Such practices, combined with a stronger teacher-student rapport, would reduce home-school dissonance even further and thus, maximize the cognitive factors that precede optimal performance.