Examining the Effects of Core Self-Evaluations and Perceived Organizational Support on Academic Burnout Among Undergraduate Students

Abstract

Academic burnout is a prevalent issue that has debilitating effects on students and refers to the phenomena of long-term fatigue and loss of interest in schoolwork, and is characterized by a student’s lack of engagement, dulled emotions, and feelings of helplessness. This survey-based study examined the predictive ability of two popular constructs in organizational psychology research, core self-evaluations, and perceived organizational support, to explain students’ academic burnout. Extending the Job Demands-Resources model and Conservation of Resources theory to the university context, the study investigated whether core self-evaluations and perceived organizational support would similarly predict burnout for 199 undergraduate students in a university setting as they do employees in work settings. Confirmatory factor analysis was employed to assess the factor structure of the variables, and moderated multiple regression was employed to test the hypotheses. Results indicated that that core self-evaluations and perceived organizational support were individually strong predictors of burnout, and that perceived organizational support had a small moderating effect on the core self-evaluations-burnout relationship. Implications and potential applications of these results are discussed as a means to mitigate the negative effects of academic burnout experienced by so many college students.

Keywords

Academic burnout perceived organizational support core self-evaluations college student wellness

Introduction and literature review

A highly prevalent issue among college students (Jacobs & Dodd, 2003; Pisarik, 2009), academic burnout refers to the phenomenon of long-term fatigue and loss of interest in schoolwork and is generally characterized by lack of engagement, dulled emotions, and feelings of helplessness (Gold, 1988). Given the negative consequences it can have on student’s learning and overall collegiate experience, particularly in light of the substantial financial investments involved in higher education; identifying predictors and means of reducing the threat of academic burnout would be helpful to students and university personnel alike. Borrowing existing theories and empirical findings in academic literature on employees in work settings, this study employs the Job Demands-Resources model (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001; Schaufeli & Bakker, 2004) as a framework to examine the relationships among core self-evaluations (CSE), perceived organizational support (POS), and academic burnout in university students. CSE is a latent construct that underlies the traits of self-esteem, self-efficacy, emotional stability, and locus of control (Chang, Ferris, Johnson, Rosen, & Tan, 2012; Judge, Locke, & Durham, 1997), whereas POS is the degree to which an employee (or in this case, student) feels that they and their individual contributions are valued by their organization (or in this case, school; Rhoades & Eisenberger, 2002).

A relationship between low CSE and academic burnout among university students has been previously determined (Lian, Sun, Ji, Li, & Peng, 2014). While the POS-academic burnout relationship has to date not been specifically examined in a university setting, logic would dictate that an inverse one exists similar to what has been found in work settings (Jawahar, Stone, & Kisamore, 2007). One purpose of this study is to empirically examine this relationship in a university setting. An additional purpose is to examine the interplay among these three variables; more specifically, the study is an investigation of whether POS has a moderating effect on the CSE–academic burnout relationship. To the extent that academic burnout and its antecedents are better understood on the basis of scientific study, more effective strategies can be developed to combat its debilitating effects. Justification for such research is perhaps best summarized by Pisarik (2009), who stated,

Considering the array of potential negative effects associated with burnout among college students, it would behoove administrators, faculty members, and student affairs professionals who attempt to develop academic climates that facilitate greater academic performance, retention, and psychological well-being, to obtain a better understanding of the prevalence and correlates of burnout on their campuses. (p. 1240)

Academic burnout

The construct of burnout was first conceptualized as the experience of exhaustion, cynicism, and inefficacy at work (Maslach, Jackson, & Leiter, 1996). However, Bianchi, Truchot, Laurent, Brisson, and Schonfeld (2014) pointed out that burnout may not be exclusive to these settings. Rather, the authors suggest that burnout can be experienced in many different environments, including the academic context. Schoolwork can be considered “work” from a psychological standpoint because it includes obligatory activities (e.g., attendance, homework, exams, etc.) and performance reviews (e.g., evaluations and grades). As Noh, Shin, and Lee (2013) found, burnout in students similarly manifests itself as it does in employees and can be measured by emotional exhaustion, depersonalization, and academic inefficacy. More specifically, emotional exhaustion encompasses mental and emotional fatigue a student experiences as a result of extensive academic demands, depersonalization refers to a student’s lack of personal attachment to or motivation to complete work, and inefficacy refers to a student’s general feelings of academic incompetence (Noh et al., 2013).

According to the Job Demands-Resources model, burnout is influenced by both demands (i.e., physically or psychological draining components of the work) and resources (i.e., physically or psychologically helpful aspects of the work; Demerouti et al., 2001). Heavy workloads and emotional demands (Schaufeli & Bakker, 2004; Van den Broeck, De Cuyper, De Witte, & Vansteenkiste, 2010) and work–home conflict (Huynh, Xanthopoulou, & Winefield, 2014; Van den Broeck et al., 2010) are such demands that have been found to predict higher levels of burnout in workers. Within the academic context, demands including academic workload and work pressure have demonstrated similar relationships with burnout among students (Boyd, Bakker, Pignata, Winefield, Gillespie, & Stough, 2011; Salanova, Schaufeli, Martínez, & Bresó, 2010).

Recent research has applied Conservation of Resources (COR) theory (Hobfoll, 1989) to extend the concept of resources to include personal resources, which refer to individual characteristics or qualities related to perceived or real control over a given situation or context. Personal resources such as self-efficacy, self-esteem, and optimism have been shown to also buffer individuals from the experience of burnout (Ouweneel, Le Blanc, & Schaufeli, 2011; Xanthopoulou, Bakker, Demerouti, & Schaufeli, 2007). In an academic context, research has found that the academic burnout that plagues students may worsen yet as these students engage in more self-handicapping behaviors and are less likely to seek academic help than are students experiencing burnout (Shih, 2013).

As academic burnout is a multi-dimensional phenomenon that is a function of demands on job resources and personal resources, understanding and identifying predictors may allow for early detection and prevention. In this study, CSE and perceptions of organizational support were examined as potential personal resources that may predict academic burnout.

Core self-evaluations

Judge, Bono, and Thoresen (2003) defined CSE as the, “fundamental assessments that people make about their worthiness, competence, and capabilities” (p. 305). A latent, higher order personality construct, CSE is manifested through four of the most commonly researched traits in psychology: self-esteem, generalized self-efficacy, emotional stability, and internal locus of control. In their search for the most appropriate indicators of CSE, Judge et al. (1997) ultimately identified these four first-order traits because they met the following criteria that they established: (a) evaluation focus (i.e., they were evaluative rather than descriptive), (b) fundamentality (i.e., they were source traits that underlie surface-level traits), and (c) breadth or scope (i.e., they were broader than secondary traits; Judge & Bono, 2001; Judge et al., 1997). To note, CSE is not a simple additive function of these traits; rather, each of these traits is a unique manifestation of underlying CSE.

Self-esteem can be defined as an individual’s broad evaluation of their self-worth (Judge et al., 1997). Generalized self-efficacy, a modified version of Bandura’s (1982) definition of self-efficacy, can be defined as an individual’s broad appraisal of their ability to complete tasks successfully under varying circumstances (Judge, Locke, Durham & Kluger, 1998). Emotional stability, a term often used interchangeably with its inverse neuroticism (Barrick & Mount, 1991), can be defined as having a “tendency to be confident, secure, and, steady” (Judge & Bono, 2001, p. 80) and a lower susceptibility to negative emotions such as anger, anxiety, and depression. Locus of control can be defined as an individual’s overall attribution of life events and can be divided into two classifications, internal or external. Individuals who possess an external locus of control believe that their circumstances are determined by outside forces (e.g., luck, a higher power, fate, etc.), whereas individuals who possess an internal locus of control believe that they exert control over a broad array of their life’s circumstances (Judge et al., 1997).

One theoretical approach used to explain how CSE influences outcomes is the approach–avoidance framework, which focuses on one’s sensitivity to stimuli (positive and negative) and how they in turn react to them (approach or avoid; Ferris, Rosen, Johnson, Brown, Risavy, & Heller, 2011). Certain personality traits are characterized by tendency for a heightened sensitivity to positive stimuli, negative stimuli, or both. In an approach–avoidance framework, high CSE individuals would be more sensitive to positive stimuli and less sensitive to negative stimuli, and strong approach tendencies and weak avoidance tendencies. For example, a high CSE student would be less likely to internalize harsh critiques of their academic coursework and would utilize criticism as an opportunity for growth, whereas a low CSE student may view the same feedback as a demotivating personal affront. As demonstrated by Judge et al. (1998), individuals having the former reaction (i.e., high CSE) not only demonstrate more resiliency and resolve to improve their situations but also greater overall life satisfaction.

Other research supports this general finding. Brunborg (2008) found that high-CSE individuals perceived lower levels of job stress than did low CSE individuals because they were more likely to actively attempt to change a stressful situation rather than allow it to negatively affect them. While a low CSE individual might be more prone to believe that they are unable to meet the challenges of a stressful environment and heavy workload and thus more likely to experience job burnout; high-CSE individuals are more immune to such burnout because of their fortitude and proactive approach to manage the situation at hand (Maslach, Schaufeli, & Leiter, 2001). A behavioral pattern of persevering and effectively managing work challenges is directly linked to job and career success, as Judge and Hurst’s (2007) longitudinal study found. Moreover, as Judge et al. (1997) found in their original study of CSE, high CSE is predictive of job satisfaction as well as performance. This study is an extrapolation of the predictive ability of CSE from a business to an academic environment, namely, whether it can predict a student’s perceptual reaction to the academic demands and stressors they face.

Perceived organizational support

The construct of POS was first introduced by Eisenberger, Huntington, Hutchison, and Sowa (1986) who defined it as “the degree to which employees believe that their organization values their contributions and cares about their well-being and fulfills socio-emotional needs” (Eisenberger et al., 1986, p. 502). According to organizational support theory, such perceptions of this employer–employee relationship are naturally occurring (Rhoades & Eisenberger, 2002) and effect the latter’s reciprocal response in a variety of forms including increased levels of effort, commitment, and performance and decreased levels of stress, absenteeism, and turnover (Eisenberger, Armeli, Rexwinkel, Lynch, & Rhoades, 2001; Levy, 2013). Attitudinally, low levels of POS are related to employees having cynical views of their organization (Dean, Brandes, & Dharwadkar, 1996) and overall decreased commitment to it (Eisenberger et al., 2001; Rhoades & Eisenberger, 2002). Positive or negative, these responses are a function of an employee’s perception of their organization’s engagement in activities that outwardly demonstrate its support and concern for the employee. In essence, POS can be thought of as the extent to which an employee believes that their organization truly wants them to be part of it based on the organization’s acting on in their best interest, particularly in ways that are beyond the employee’s own control.

Rhoades and Eisenberger’s (2002) meta-analysis identified three categories of organization activities that serve as positive antecedents to an employee’s POS. Fairness of treatment is akin to a Greenberg’s (1990) notion of organizational justice, or the extent to which rewards are distributed fairly, and employees are involved in decision-making that affects them. Due to the frequent direct interaction that is typical between supervisor and employee, the latter’s perception of the former’s favorable or unfavorable orientation toward them generalizes to the entire organization. As such, the degree of supervisor support, manifested by the supervisor’s generally demonstrating concern for the employee’s well-being and fostering quality, supportive, and trusting relationships, serves as a proxy for the support that the employee perceives is manifested by the entire organization. The final category identified by Rhoades and Eisenberger (2002) is rewards and job conditions, that include a host of factors that make an employee feel valued and fulfilled, including rewards, recognition, and pay; degree of autonomy and role stressors; and training opportunities and job security. Taken together, when an employee perceives a high degree of fair treatment, sees their supervisor taking voluntary action to demonstrate their genuine concern for their well-being, and appreciates organizational rewards and working conditions as satisfying and motivating; they are likely to believe that their organization truly wants them as members and thus will report a high level of POS. Although the reciprocal relationships between an employee’s perception of their organization’s favorable treatment toward them and a host of mutually beneficial socioemotional, commitment, and performance outcomes appears intuitive, meta-analytic research (Rhoades & Eisenberger, 2002) has provided clear empirical support for them and thus the importance of POS to organizations and employees alike.

Similar to how an employee formulates POS based on their organization’s treatment of them, logic would dictate that a college student would similarly develop perceptions of the extent to which their university is invested in their personal welfare and promoting an environment in which they can thrive academically and personally. Applying Rhoades and Eisenberger’s (2002) categories of an organization’s POS-inducing activities to a university, fairness of treatment could be exhibited by a professor assigning a reasonable workload, establishing clear assessment criteria and grading fairly against them, and notifying and/or involving students in decisions that impact them. Likewise, supervisor support could be determined by a professor’s manifesting a sincere interest in a student’s personal life, making an effort to help students who are struggling, and providing motiving feedback and inspirational encouragement. Finally, rewards and job conditions might come in the form of a professor fostering a comfortable and respectful classroom environment that breeds engagement, offering a student latitude in shaping assignments, and reiterating the value that a course and eventual degree from the institution will bring the student. Receiving these activities favorably, the student would report a higher level of POS, which in turn should result in the same positive individual and organizational outcomes that have been observed in work settings (Eisenberger et al., 1986, 2001; Rhoades & Eisenberger, 2002).

Study rationale and hypotheses

Academic burnout is a serious issue for many college students, as it can lead to a variety of negative outcomes that preclude students’ having an optimal and satisfying learning and overall life experience during their college years. To the extent that college administrators, counselors, and faculty can better identify the factors that make some students more susceptible to academic burnout, the more readily they can intervene and make adjustments. This study considers the role that two variables can play in predicting academic burnout. The first, CSE, is a dispositional variable that reflects an individual’s overall assessment of herself. If CSE is high, meaning the student possesses high self-worth, confidence, and a general belief that they are capable of overcoming obstacles and succeeding; it would suggest that they might be less prone to experiencing academic burnout. A high level of CSE might buffer the stress and challenges of the student’s heavy academic demands. From an approach–avoidance perspective, such students may be more prone to approach positive stimuli and avoid (or at least resist or effectively manage) negative stimuli otherwise typically associated with burnout. On the other hand, a “weaker” student who is low on CSE may lack the resiliency and inner fortitude to resist or avoid the negative stimuli and thus be more prone to experience academic burnout. Thus, Hypothesis 1 is stated as follows: There will be a significant inverse relationship between CSE and academic burnout; in other words, as CSE increases, academic burnout will decrease.

The second variable, POS, is a function of the environment and reflects the extent to which a student feels supported and valued by their professor(s) and in turn, university. If the student believes that the university is genuinely invested in actively promoting their well-being and success, these positive feelings and encouragement will buffer the pressure of academic demands that can lead to academic burnout. Thus, Hypothesis 2 is stated as follows: There will be a significant inverse relationship between POS and academic burnout; in other words, as POS increases, academic burnout will decrease.

While the first two hypotheses essentially serve as confirmations of previous research findings, the primary goal of this study is to examine the possible moderating effect of POS on the CSE–academic burnout relationship. Even if the overall CSE–academic burnout relationship is strong across all students, a question arises whether it will be the same for students experiencing high levels of POS as it would be for other students who experience low levels.

If POS is low, a fairly strong inverse CSE–academic burnout relationship would be expected to exist. A student with a low level of CSE, in the absence of a buffering effect of POS, may not be sufficiently equipped to overcome adversity and thus be more prone to suffer from academic burnout. However, if the same low-CSE student reported a high level of POS, their more supportive environment might enable them to better contend with academic demands and stressors thus making them less susceptible to burnout. For high CSE students, POS may have less effect since these individuals already possess the fortitude and resilience to overcome negative stimuli. At the same time, perhaps high CSE students will be naturally better equipped than low CSE students to leverage the benefits of an environment that cultivates POS and thus even further inure themselves from the threat of burnout. As a means to explore these possibilities and directionalities, Hypothesis 3 is formally stated as follows: POS will have a significant moderating effect on the relationship between CSE and academic burnout, i.e., the strength of the CSE–academic burnout relationship will vary as a function of the level of POS.

Method

Participants

A total of 199 (157 females and 42 males) full-time, undergraduate students at a mid-sized university in the southeastern United States voluntarily signed up to participate in the study through an online research participation management system; the gender breakdown of participants generally reflected that of the fairly female dominated campus population. The mean age of participants was 19.96 years (S = 2.12), and all completed the study. As an incentive, participants’ names were entered into a drawing to win one of a number of $10.00 gift cards from a popular local retailer. Most students also received extra credit or credit for fulfilling a research participation requirement for a psychology course in which they were concurrently enrolled.

Procedure

All participant data were collected during the last month of a fall semester, when it was assumed that students might be most susceptible to academic burnout. Participants completed a 38-item online survey that was comprised of separate scales to assess their individual levels of CSE, POS, and academic burnout. The order of scales was counterbalanced to combat order effects. Completion of the survey took between 15 to 25 minutes, and participants were reminded multiple times that their responses were anonymous and would only be presented as part of an aggregate.

Materials

Core self-evaluations

CSE was measured with the Core Self-Evaluations Scale (CSES) (Judge et al., 2003), a 12-item Likert-type survey with responses that ranged from 1 (strongly disagree) to 5 (strongly agree). Examples of items are as follows: I am confident I get the success I deserve, and I complete tasks successfully. The mean and standard deviation were 43.27 and 7.36, respectively, and a Cronbach’s α of .85 suggested good reliability.

Perceived organizational support

POS was measured with a modified eight-item Likert-type version of the Survey of Perceived Organizational Support (SPOS; Eisenberger et al., 1986). The items were selected following the procedure used by Kinnunen, Feldt, and Makikangas (2008) who used the eight highest loading items from Eisenberger et al.’s (1986) original scale. In their meta-analysis in which they identify multiple short versions used, Rhoades and Eisenberger (2002) note that such practice is permissible, as the original SPOS was measuring a single factor. In this study, we used five of the same items used by Kinnunen et al. (2008). Examples of these items are as follows: The university values my contribution to its well-being and The university would ignore any complaint from me. Three different items were selected because they were thought to be more relevant to students than Kinnunen et al.’s items; however, these items were still among the top 12 highest loading in Eisenberger et al.’s (1986) scale development.

As a means to adapt the SPOS for student respondents, participants were told in the instructions to consider their university as the “organization” when that term appeared in the items. In addition, three items were slightly modified to make them more relevant to a university setting (e.g., Even if I did the best job possible, the organization would fail to notice was modified to Even if I did the best job possible academically, the organization would fail to notice). Responses ranged from 0 (strongly disagree) to 6 (strongly agree) and the mean and standard deviation were 31.93 and 8.28, respectively, and a Cronbach’s α of .88 suggested good reliability.

Academic burnout

Academic burnout was measured with items from the School Burnout Scale (SBS; Salmela-Aro, Kiuru, Leskinen, & Nurmi, 2009), a nine-item Likert-type survey with responses that ranged from 0 (strongly disagree) to 6 (strongly agree). The SBS contains three factors: exhaustion at schoolwork, measured by four items (e.g., I feel overwhelmed by my schoolwork); cynicism toward the meaning of school, measured by three items (e.g., I feel that I am losing interest in my schoolwork); and sense of inadequacy at school, measured by two items. (e.g., I often have feelings of inadequacy in my schoolwork). The respective Cronbach’s α for the scales were .77, .80, and .59, respectively. However, when combined, the Cronbach’s α for the overall burnout scale was a more impressive .87, suggesting good reliability. The mean and standard deviation for the overall scale were 28.42 and 11.08, respectively.

Data analyses

The scale data were first analyzed using the Statistical Package for Social Sciences 25 (SPSS) to examine the descriptive statistics and correlations among the variables. As analyses found no significant gender differences on these variables or order effects among them, the data were combined for each scale.

Next, as all variables were normally distributed, a confirmatory factor analysis using MPlus 5 was performed to assess the factor structure of the variables. After establishing evidence for a three-factor model, SPSS was again used to test hypotheses. Bivariate correlations were examined to examine CSE’s and POS’s individual effects on academic burnout, and moderated multiple regression was used to test whether POS moderated the relationship between CSE and burnout. In addition, the moderated multiple regression analysis was followed by supplemental simple slope analyses in order to detect the specific nature of the interaction.

Results

Confirmatory factor analysis

A confirmatory factor analysis was conducted to compare the proposed three-factor structure to a one-factor model. For model fit, Hu and Bentler (1999) recommend the following cutoffs: comparative fit index (CFI)/Tucker–Lewis Index (TLI)≥.95, root mean square error of approximation (RMSEA)≤.06, and standardized root mean square residual (SRMR)≤.08. The three-factor model met the SRMR cutoff and nearly met the RMSEA cutoff; however, it was below the recommended threshold for CFI (χ² = 731.76 (df = 371); CFI = .85; RMSEA = .07, and SRMR = .07). The one-factor model did not meet any of the fit indices cutoffs (χ² = 1296.99 (df = 377); CFI = .63; RMSEA = .11, and SRMR = .11). In a direct comparison of the models, the three-factor model was a significantly better fit to the data than the one-factor model (Δχ² = 565.23, p<.001). Descriptive statistics for the CSE, POS, and burnout scales, as well as the subscales of burnout, are presented in Table 1.

Table 1.

Descriptive statistics for CSE, POS, and burnout.

Variable	Mean (SD)	Skew	Kurtosis	1	2	3	4	5	6
1. CSE	3.61 (.61)	−.01	1.68	.85
2. POS	3.99 (1.04)	−.25	2.18	.38*	.88
3. Burnout	3.12 (1.24)	−.41	1.48	−.70*	−.43*	.87
Burnout Subscales^a
4. Inadequacy	6.10 (3.12)			−.69*	−.39*	.85*	.59
5. Cynicism	9.13 (4.50)			−.61*	−.39*	.90*	.73*	.80
6. Exhaustion	13.20 (5.14)			−.57*	−.33*	.84*	.59*	.58*	.77

CSE: core self-evaluations; POS: perceived organizational support.

^aBurnout subscales were combined and the overall measure of burnout was used in the tests of the hypotheses.

*Significant at the p <.05 level; N = 199.

Bivariate correlational analysis

As predicted, a strong negative and statistically significant relationship was found between CSE and academic burnout (r = −.70, p<.05). Thus, Hypothesis 1, which stated that individuals with high levels of CSE would be less susceptible to academic burnout, was supported. Hypothesis 2, which stated that individuals who reported higher levels of POS would report lower levels of academic burnout, was also supported (r = −.42, p<.05).

Moderated multiple regression

To test Hypothesis 3, that POS would have a moderating effect on academic burnout, a moderated regression analysis was performed as outlined by Aiken and West (1991). The logic of moderated regression here is that, while CSE academic significantly predicts academic burnout, the relationship would presumably vary as a function of POS. In other words, although CSE should predict academic burnout regardless of a student’s POS, the CSE–academic burnout relationship will be differentially stronger or weaker depending on the exact level of POS.

The key to Aiken and West’s (1991) moderated regression procedure is the creation of a predictor term that represents the interaction between the predictor (CSE) and the moderator (POS) variables. This is done by creating the cross products of these variables, here represented by CSE × POS. Using a hierarchical regression procedure, the outcome variable (academic burnout) is then regressed on the original predictor (CSE) and the moderator (POS) on the first step, and then the interaction term (CSE × POS) on the second. If a significant ΔR² is found after the addition of the interaction term, one may conclude the presence of a moderator (here, POS).

As recommended by Aiken and West (1991), individual CSE and POS scores were centered to mitigate the threat of multicollinearity. Although the correlation between CSE and POS was only moderate (r=.38, p<.05), the threat of multicollinearity stems from the interaction term being a direct function of them. After centering each CSE and POS score (denoted as CSE_c and POS_c), the cross product of these centered scores became the centered interaction term, CSE_c × POS_c. As prescribed by Aiken and West (1991), the outcome variable was not centered.

On the first step, academic burnout was regressed on CSE_c and POS_c. As expected, this produced a significant adjusted R² (.51, p<.05) and significant β weights for both the CSE (β=−.63) and POS (β = −.18) predictors. To test Hypothesis 3, the CSE_c × POS_c interaction term was then added on the second step and produced a statistically significant ΔR² (.02, p<.05) thus supporting Hypothesis 3. The standardized β weight for CSES_c × POS_c, was −.13 and statistically significant at the p<.05 level. These results confirmed that, while CSE (β = −.62) and POS (β = −.21) both independently explained a significant amount of academic burnout variance, POS still did indeed have a significant moderating effect on the CSE–academic burnout relationship. The results of this moderated regression analyses can be found in Table 2.

Table 2.

Moderated regression analyses.

Variable	R	Adjusted R²	ΔR²	B	SE	β	t
Step 1	.72*	.51*
CSE				−.95	.08	−.63	−11.82*
POS				−.24	.07	−.18	−3.33*
Constant				28.42	.55
Step 2	.73*	.53*	.02*
CSE				−.93	.08	−.62	−11.58*
POS				−.28	.07	−.21	−3.86*
CSE × POS				−.02	.01	−.13	−2.68*
Constant				28.48	.56

CSE, POS, and CSE × POS are centered. CSE: core self-evaluations; POS: perceived organizational support.

*Significant at the p <.05 level; N = 199.

Simple slope analysis

To determine the specific nature of a significant moderated relationship, Aiken and West (1991) recommend a follow-up procedure referred to as simple slope analysis in which the predictor–outcome (here, CSE–academic burnout) relationship is examined at three different levels of the moderator (POS). Again, using CSE as a predictor, three separate new cross-products were formed by the product of CSE and POS at specified low, medium, and high levels of POS.

This led to the creation of three new separate regression equations: (1) one standard deviation below the POS mean (−8.28; denoted as POS_lo), (2) at the POS mean (0; POS_med), and (3) one standard deviation above the POS mean (8.28; denoted POS_hi).

Regression analyses were conducted at each of the three levels of POS. As seen in Table 3, CSE, POS, and the CSE × POS interaction had statistically significant β weights at each of the low, medium, and high levels of POS.

Table 3.

Simple slope analyses: Regressions.

Variable	B	SE	β	t
Regression of school burnout on CSE at POS_lo (one standard deviation below the mean).
CSE	−.775	.104	−.515	−7.479*
POS_lo	−.278	.072	−.208	−3.856*
CSE × POS_lo	−.018	.007	−.171	−2.680*
Constant	31.152	.843
Regression of school burnout on CSE at POS_med
CSE	−.927	.080	−.616	11.575*
POS_med	−.278	.072	−.208	−3.858*
CSE × POS_med	−.018	.007	−.134	−2.680*
Constant	28.848	.56
Regression of school burnout on CSE at POS_hi (one standard deviation above the mean)
CSE	−1.078	.092	−.716	−11.705*
POS_hi	−.278	.072	−.208	−3.858*
CSE × POS_hi	−.018	.007	−.163	−2.680*
Constant	26.544	.796

CSE: core self-evaluations; POS: perceived organizational support.

*Significant at the p <.05 level, N = 199.

The pattern across all levels was the same, such that CSE was the strongest predictor (β = −.52, −.62, and −.72 respectively for POS_lo, POS_med, and POS_hi), followed by POS (β = −.21, −.21, and −.21), and the CSE × POS interaction term (−.17, −.13, −.16). Thus, it appears that while CSE and POS individually predict academic burnout at the three levels of POS, POS similarly moderates the CSE–academic burnout relationship across them as well.

To visually examine the relative differences among the three equations, the simple slopes were plotted. Following Aiken and West’s (1991) guidelines, academic burnout was regressed on CSE using the CSE unstandardized coefficient and constant from the previous regressions at POS_lo, POS_med, and POS_hi. As seen in Figure 1, all three equations have negative slopes which reflect the general hypothesized inverse relationship between CSE and academic burnout. However, whereas the level of academic burnout is similar (high) for POS_lo, POS_med, and POS_hi at low levels of CSE (i.e., two standard deviations below the mean), the CSE–academic burnout relationship changes across levels of POS at higher levels of CSE (i.e., two standard deviations above the mean).

Figure 1.

Visual depiction of simple slope analysis Y = −.93X −.28Z −.02XZ + 28.85 (Y=Academic Burnout, X= CSE, Z=POS, XZ=CSE × POS).

The relationship was strongest for POS_hi, followed by POS_med and POS_lo. In other words, while the bivariate relationship between CSE and academic burnout is a strong and statistically significant one, the relationship is even stronger when POS is high. Because the CSE × POS interaction term was statistically significant in the initial analysis, by logic, the β weights of CSE are significantly different, thus requiring no further testing (Aiken & West, 1991).

Discussion

Summary and interpretation of findings

An issue which plagues many college campuses today, academic burnout typically leads to students having a diminished interest in their schoolwork. In addition, academic burnout is generally accompanied by a lack of engagement, dulled emotions, and feelings of helplessness (Gold, 1988). As a result of these outcomes, the “college experience” for afflicted students can be severely compromised as they are inhibited from maximizing their academic and personal development during these critical formative years. The negative emotional outcomes aside, the considerable financial costs associated with tuition and college living expenses force one to consider the exact extent to which academic burnout comprises the return on the huge investments made in students’ academic and professional futures. Determining predictors of and means of reducing risk for burnout could potentially provide extremely beneficial information for universities, professors, and students to manage its threat. To help address this issue, this study examined academic burnout’s relationship with two variables that have been helpful in predicting similar constructs in organizational settings. One of these variables, CSE, is a dispositional construct, while the other, POS, is a function of the environment.

As predicted, a strong negative and statistically significant relationship was found between CSE and academic burnout. High CSE indicates an individual’s possession of high self-worth, confidence, and a general belief that she is capable of overcoming obstacles and succeeding. Based on the integration of COR theory (Hobfoll, 1989) and the JD-R model (Demerouti et al., 2001), CSE functions as a personal resource which protects the individual from academic burnout. Such a student may be sufficiently “strong” and firmly grounded and thus may not fall prey to the same burnout threats as easily as a “weaker” low CSE student might. Or, if the legitimate heavy academic demands are indeed perceived by high CSE students, their CSE may serve as somewhat of an emotional buffer and better equip them to assertively and productively manage the stresses and challenges they are facing. Also, from an approach–avoidance perspective, it can be thought that high CSE students may simply be more prone to approach positive stimuli they encounter on campus (e.g., social events, intrinsic motivation to learn, friendships, etc.) and avoid focusing on the negative factors that might be associated with burnout. On the other hand, a “weaker” student who is low on CSE may lack the resiliency and inner fortitude to contend with negative environmental stimuli and thus be more susceptible to academic burnout.

The second hypothesis, that POS would be negatively related to academic burnout, was also supported. According to the JD-R model (Demerouti et al., 2001), POS may reflect an environmental resource because it reflects the extent to which a student feels supported and valued by their school. To the extent that the student has positive perceptions in this regard, they are less apt to feel the pressure and other negatives that can lead to academic burnout. From an approach–avoidance perspective, when POS is high, there is likely more to approach and less to avoid regarding coursework. If a student perceives that an instructor is supportive and values their efforts and contribution, the student would be more likely to “approach” academic work than in a situation where such support was not present, or one they would be more inclined to “avoid.”

The support of the first two hypotheses are consistent with COR theory (Hobfoll, 1989), the JD-R model (Demerouti et al., 2001), and previous empirical findings. To further understand the complex interplay of personal and environmental resources on burnout, this study also explored whether POS had a moderating effect on the CSE–academic burnout relationship. The logic here was that, perhaps a high-CSE student may respond to high or low level of POS differently than a low-CSE student would. Using the above rationale that low CSE students might require or benefit relatively more from a supportive environment than would high CSE students (or that high CSE might be more immune to unsupportive environments than low CSE students), moderated multiple regression analysis was used to test POS as a potential moderator in the CSE–academic burnout.

From a traditional statistical significance perspective, there was evidence that POS did moderate the CSE–academic burnout relationship at the p <.05 level. Although this effect was not nearly as strong a predictor as CSE or POS on their own, simple slope analyses showed that the CSE significantly predicted academic burnout across all levels of POS. However, the strongest relationship was found when POS was high suggesting that high CSE students perhaps react even more (favorably) strongly to highly supportive environments than low CSE. Rather than thinking of high CSE as primarily an asset because of its potential to serve as buffer to unsupportive environments, these results suggest that it might make more sense to see the high CSE students benefit more than low CSE students from a supportive environment because they are drawn more to it. Perhaps the supportive environment reinforces their already high levels of CSE and makes them even more resistant to academic burnout. From an approach–avoidance paradigm, this would make sense.

At the same time, a growing body of authors such as Barrett (2016) question whether such small increases in adjusted R² as we observed (.02) are truly indicative of significant findings even if they meet the conventional p <.05 criterion. In his analysis of published data with such small but statistically significant magnitudes of change, Barrett provides empirical evidence that these findings are trivial to the point of meaningless. Others such as Wasserstein, Schirm, and Lazar (2019) make even stronger cases against equating statistical inference with scientific inference merely on the basis of a result’s observed probability. These authors are part of a broader trend that challenges researchers to be more diligent and thorough in when inferring and reporting scientific findings. As new optimal methods evolve, researchers should remain modest and honest (Wasserstein et al., 2019) in interpreting their results. Heeding this advice, it would be most prudent to say that our findings of a statistically significant interaction effect between CSE and POS should be taken as it is – a result that requires further research and replication.

Implications and applications

The findings of this study could ostensibly be applied in an attempt to decrease the incidence and or severity of academic burnout in college students and thus foster a more positive and satisfying learning experience for them. As CSE and POS are both strong predictors of academic burnout and can be easily measured, early assessments could identify students most at risk. Earlier detection of students’ individual susceptibilities to burnout would make it easier for educators to intervene and hopefully better equip students to ward off burnout.

As it is a variable that is directly influenced by the external environment, a student’s level of POS is something that university educators and administrators could possibly affect by creating a more supportive academic setting. More specifically, they could engage in behaviors that are positively related antecedents to POS. Three categories of such behaviors, fairness of treatment, supervisor support, and rewards and job conditions, are frequently studied inducers of POS in organizational settings but could easily be applied to an academic one. For example, professors could exhibit fairness of treatment in the classroom by treating all students with the same high level of respect, giving valid and fair exams, grading objectively and not showing favoritism, and generally taking student opinions and concerns into account when making decisions. Supervisor support can easily be applied in this setting by professors taking actions such as providing constructive feedback on assignments, providing helpful resources and direction, holding and inviting students to office hours, offering encouragement and support, and generally manifesting a genuine interest in students’ progress. Examples of rewards and job conditions in a university classroom could include the professor maintaining a professional environment, delivering clear and meaningful lectures, providing structured syllabi, assigning reasonable workloads, rewarding quality work with high grades, and providing suggestions for how students can improve. These are just a few examples, and perhaps future research can more completely identify such behaviors that increase students’ POS, which in turn decreases academic burnout.

CSE is a more difficult variable to influence, as it is a latent individual trait that is less influenced by external circumstantial factors. However, because the inverse relationship between CSE and academic burnout is so strong, students who score low on the CSES can be identified as “at greater risk” for academic burnout. Theoretically, if administrators, professors, and other university staff could more quickly identify low CSE students, they could perhaps place even greater attention on promoting an environment that creates strong POS for this population that immediately starts off at greater risk for burnout. Although it is a latent trait, perhaps behavioral and cognitive interventions could strengthen students’ levels of CSE. For example, coaching students how to transition from an external locus of control to an internal one and how to deal more constructively with criticism might strengthen CSE. Likewise, building students’ self-esteem and confidence in their abilities could increase their CSE and thus make them less susceptible to burnout. Although perhaps not formally stated as such, many universities’ existing offerings to students were likely created precisely to build CSE or concepts related to it (i.e., self-esteem, confidence, resilience etc.). Continuing to promote counseling services and collegiate life experiences that foster the development and growth of students is a built-in way to grow CSE in students and equip them with the mindset and style to better resist threats of academic burnout.

Limitations and directions for future research

While this study found support for its hypotheses, there are several limitations. Because data collection consisted of a self-report survey, there is the potential for decreased validity due to the social desirability effect. This is particularly true in this study where two of the variables’ scales, CSE and academic burnout, have fairly positive and negative connotations respectively. Social desirability can be formally defined as the tendency of research subjects “to underreport socially undesirable behavior and overreport socially desirable behavior. They distort their answers towards the social norm in order to maintain a socially favorable self-presentation” (Krumpal, 2013, p. 2027). While some participants may have intentionally or unintentionally distorted their answers to make themselves look more favorable, others may have simply reported inaccurate perceptions. How participants interpret items and their personal interpretation of terms and scales may not be consistent across the sample, to the point where one person whose “true” level of CSE is lower than another person’s yet the latter reported a higher level of CSE than the former. Future research could possibly employ other raters (e.g., participants’ friends, professors, etc.) to gain a more complete assessment of CSE. Likewise, incorporating others’ measures of POS, such as having a consultant assess the classroom climate, might add another dimension to understanding the true level of POS.

As this was a nonexperimental study, causal relationships cannot be ascertained. For example, the high correlation between CSE and academic burnout indicates only a strong relationship between the two variables, not that CSE is a direct cause of academic burnout. One might plausibly ask whether academic burnout actually influences CSE in the other direction, that is, a student who is feeling burned out then in turn begins to believe that she is less supported and experience less self-esteem. Perhaps future research can be conducted in a laboratory environment where variables can be manipulated to better determine these relationships.

Because the relationships between CSE and academic burnout, and POS and academic burnout had such strong initial correlations, a ceiling effect may have occurred when determining a moderating variable. As the initial bivariate relationship between CSE and academic burnout was already quite high (r= −.69), there was little room for a moderating effect to account for additional variance beyond that.

The current research collected data at the end of the semester when the target population was likely to be experiencing academic burnout. Presumably, the end of the semester is a time of increased work demands (i.e., major assignments are due, final exams are approaching, etc.). Since CSE is a stable individual dispositional construct (Judge et al., 1997) and thus less likely to vary across situations, such end-of-semester demands are unlikely to influence it. Empirical and meta-analytic research suggests that stress and other work demands are less important for shaping POS than work resources (Eisenberger, Cummings, Armeli, & Lynch, 1997; Kurtessis, Eisenberger, Ford, Buffardi, & Stewart, 2017). Nonetheless, future research should examine the role of time on the relationships between CSE, POS, and burnout.

Future researchers may consider measuring CSE and POS in relation to other variables in an academic setting. As most POS and CSE research is conducted in professional organizational settings, other variables related to them (e.g., productivity, organizational justice, organizational commitment, etc.) are specific to those settings.

Footnotes

Article Notes

References

Aiken

L. S.

West

S. G.

(1991). Multiple regression: Testing and interpreting interactions. Thousand Oaks, CA: SAGE.

Bandura

(1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122–147. doi:10.1037/0003-066x.37.2.122

Barrett

(2016). Hierarchical multiple linear regression and the correct interpretation of the magnitude of a deviation R-square (ΔR²). Cognadev Technical Report Series (Report No. 6). Middlesex, England: Cognadev UK Ltd. Retrieved from https://www.pbarrett.net/cognadev/techseries/Cognadev_Technical_Report_6_Deviation_R-Square_17_June_16.pdf

Barrick

M. R.

Mount

M. K.

(1991). The big five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1–26. doi:10.1111/j.1744-6570.1991.tb00688.x

Bianchi

Truchot

Laurent

Brisson

Schonfeld

(2014). Is burnout solely job-related? A critical comment. Scandinavian Journal of Psychology, 55, 357–361. doi:10.1111/sjop.12119

Boyd

C. M.

Bakker

A. B.

Pignata

Winefield

A. H.

Gillespie Ni ., & Stough

(2011). A longitudinal test of the job demands-resources model among Australian university academics. Applied Psychology: An International Review, 60, 112–140. doi:10.1111/j.1464-0597.2010.00429.x

Brunborg

(2008). Core self-evaluations: A predictor variable for job stress. European Psychologist, 13, 96–102. doi:10.1027/1016-9040.13.2.96

Chang

Ferris

Johnson

Rosen

C. C.

Tan

J. A.

(2012). Core self- evaluations: A review and evaluation of the literature. Journal of Management, 38, 81–128. doi:10.1177/0149206311419661

Dean

Brandes

Dharwadkar

(1996). Organizational cynicism. The Academy of Management Review, 23, 341. doi:10.2307/259378

10.

Demerouti

Bakker

A. B.

Nachreiner

Schaufeli

W. B.

(2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86, 499–512. doi:10.1037//0021-9010.86.3.499

11.

Eisenberger

Armeli

Rexwinkel

Lynch

P. D.

Rhoades

(2001). Reciprocation of perceived organizational support. Journal of Applied Psychology, 86, 42–51. doi:10.1037//0021-9010.86.1.42

12.

Eisenberger, R., Cummings, J., Armeli, S., & Lynch, P. (1997). Perceived organizational support, discretionary treatment, and job satisfaction. Journal of Applied Psychology, 82, 812–820. doi:10.1037/0021-9010.82.5.812

13.

Eisenberger

Huntington

Hutchison

Sowa

(1986). Perceived organizational support: A review of the literature. Journal of Applied Psychology, 71, 500–507. doi:10.1037//0021-9010.71.3.500

14.

Ferris

Rosen

Johnson

Brown

Risavy

Heller

(2011). Approach or avoidance (or both?): Integrating core self-evaluations within an approach/avoidance framework. Personnel Psychology, 64, 137–161. doi:10.1111/j.1744-6570.2010.01204.x

15.

Gold

(1988). Recognizing and coping with academic burnout. Contemporary Education, 59, 142–145. doi:10.1177/0013164485454022

16.

Greenberg

(1990). Organizational justice: Yesterday, today and tomorrow. Journal of Management, 16, 399–432. doi:10.1177/014920639001600208

17.

Hobfoll

S. E.

(1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513–524. doi:10.1037//0003-066x.44.3.513

18.

L-T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. doi:10.1080/10705519909540118

19.

Huynh

J. Y.

Xanthopoulou

Winefield

A. H.

(2014). The Job Demands-Resources Model in emergency service volunteers: Examining the mediating roles of exhaustion, work engagement and organizational connectedness. Work & Stress, 28, 305–322. doi:10.1080/02678373.2014.936922

20.

Jacobs

S. R.

Dodd

D. K.

(2003). Student burnout as a function of personality, social support, and workload. Journal of College Student Development, 44, 291–303. doi:10.1353/csd.2003.0028

21.

Jawahar

Stone

Kisamore

(2007). Role conflict and burnout: The direct and moderating effects of political skill and perceived organizational support on burnout dimensions. International Journal of Stress Management, 14, 142–159. doi:10.1037/1072-5245.14.2.142

22.

Judge

T. A.

Bono

J. E.

(2001). Relationship of core self-evaluations traits—self- esteem, generalized self-efficacy, locus of control, and emotional stability—with Job Satisfaction and job performance: A meta-analysis. Journal of Applied Psychology, 86, 80–92. doi:10.1037//0021-9010.86.1.80

23.

Judge

T. A.

Bono

J. E.

Thoresen

C. J.

(2003). The core self-evaluations scale: Development of a measure. Personnel Psychology, 56, 303–331.doi:10.1111/j.1744-6570.2003.tb00152.x

24.

Judge

Hurst

(2007). Capitalizing on one's advantages: Role of core self-evaluations. Journal of Applied Psychology, 92, 1212–1227. doi:10.1037/0021-9010.92.5.1212

25.

Judge

T. A.

Locke

E. A.

Durham

C. C.

(1997). The dispositional causes of job satisfaction: A core evaluations approach. Organizational Behavior, 19, 151–188. https://doi.org/10.1037/0021-9010.83.1.17

26.

Judge

Locke

Durham

Kluger

(1998). Dispositional effects on job and life satisfaction: The role of core evaluations. Journal of Applied Psychology, 83, 17–34. doi:10.1037//0021-9010.83.1.17

27.

Kinnunen

Feldt

Makikangas

(2008). Testing the effort-reward imbalance model among Finnish managers: The role of perceived organizational support. Journal of Occupational Health Psychology, 13, 114–127. doi:10.1037/1076-8998.13.2.114

28.

Krumpal

(2012). Determinants of social desirability bias in sensitive surveys: A literature review. Quality & Quantity, 47, 2025–2047. doi:10.1007/s11135-011-9640-9

29.

Kurtessis, J. N., Eisenberger, R., Ford, M. T., Buffardi, L. C., & Stewart, K. A. (2017). Perceived organizational support: A meta-analytic evaluation of organizational support theory. Journal of Management, 43(6), 1854–1884. doi:10.1177/0149206315575554

30.

Levy

(2013). Industrial/organizational psychology (4th ed., pp. 316–317). New York, NY: Worth.

31.

Lian

Sun

Peng

(2014). Moving away from exhaustion: How core self-evaluations influence academic burnout. PLoS ONE, 9, 1–5. doi:10.1371/journal.pone.0087152

32.

Maslach

Jackson

Leiter

(1996). MBI: The Maslach Burnout Inventory: Manual. Palo Alto, CA: Consulting Psychologists Press.

33.

Maslach

Schaufeli

Leiter

(2001). Job burnout. Annual Review of Psychology, 52, 397–422. doi:10.1146/annurev.psych.52.1.397

34.

Noh

Shin

Lee

(2013). Developmental process of academic burnout among Korean middle school students. Learning and Individual Differences, 28, 82–89. doi:10.1016/j.lindif.2013.09.014

35.

Ouweneel

Le Blanc

P. M.

Schaufeli

W. B.

(2011). Flourishing students: A longitudinal study on positive emotions, personal resources, and study engagement. The Journal of Positive Psychology, 6, 142–153. doi:10.1080/17439760.2011.558847

36.

Pisarik

C. T.

(2009). Motivational orientation and burnout among undergraduate college students. College Student Journal, 43, 1238–1252.

37.

Rhoades

Eisenberger

(2002). Perceived organizational support: A review of the literature. Journal of Applied Psychology, 87, 698–714. doi:10.1037//0021-9010.87.4.698

38.

Salanova

Schaufeli

Martínez

Bresó

(2010). How obstacles and facilitators predict academic performance: The mediating role of study burnout and engagement. Anxiety, Stress & Coping: An International Journal, 23, 53–70. doi:10.1080/10615800802609965

39.

Salmela-Aro

Kiuru

Leskinen

Nurmi

(2009). School burnout inventory (sbi): Reliability and validity. European Journal of Psychological Assessment, 25, 48–57. doi:10.1027/1015-5759.25.1.48

40.

Schaufeli

W. B.

Bakker

A. B.

(2004). Job demands, job resources, and their relationship with burnout and engagement: A multi-sample study. Journal of Organizational Behavior, 25, 293–315. doi:10.1002/job.248

41.

Shih

(2013). The effects of autonomy support versus psychological control and work engagement versus academic burnout on adolescents' use of avoidance strategies. School Psychology International, 34, 330–347. doi:10.1177/0143034312466423

42.

Van den Broeck

De Cuyper

De Witte

Vansteenkiste

(2010). Not all job demands are equal: Differentiating job hindrances and job challenges in the Job Demands-Resources model. European Journal of Work and Organizational Psychology, 19, 735–759. doi:10.1080/13594320903223839

43.

Wasserstein

R. L.

Schirm

A. L.

Lazar

N. A.

(2019). Moving to a world beyond “p<0.05. The American Statistician, 73, 1–19. doi:10.1080/00031305.2019.1583913

44.

Xanthopoulou

Bakker

A. B.

Demerouti

Schaufeli

W. B.

(2007). The role of personal resources in the job demands-resources model. International Journal of Stress Management, 14, 121–141. doi:10.1037/1072-5245.14.2.121