Summary

Some Stress Is Good Stress: The Challenge-Hindrance Framework, Academic Self-Efficacy, and Academic Outcomes

Justin Travis University of South Carolina Upstate

Alyssa Kaszycki and Michael Geden North Carolina State University

James Bunde University of South Carolina Upstate

Historically, most investigations involving stress have assumed its undesirability, and deleterious effects have been identified across a variety of domains. Recently, however, researchers in management and health have differentiated between types of stress, and revealed a more complicated picture as a result. Specifically, stressors perceived as goal-relevant and manageable (i.e., challenging) are thought to increase motivation, performance, and well-being, while stressors viewed as goal-relevant but unman- ageable (i.e., hindering) are believed to hamper performance and occasion maladaptive behaviors. Empirical support for this theoretical framework has accumulated in employment settings, but the model has yet to be adequately extended to higher education. The current study used a longitudinal design and multiple academic outcomes to explore the challenge-hindrance distinction in a large, diverse student sample. Students from 2 Southeastern institutions (N � 853) were assessed for challenge stress (e.g., class difficulty, high expectations), hindrance stress (e.g., ambiguous expectations, favoritism), academic self-efficacy (ASE), grade point average (GPA), hours withdrawn, and transfer intentions. Results were generally theory-consistent, as ratings of challenge and hindrance stress were associated with positive and negative academic outcomes, respectively. ASE did not moderate the challenge–GPA relationship, but emerged as an independent predictor of academic functioning. Implications for stress researchers, educators, and academic decision-makers are discussed.

Educational Impact and Implications Statement This study provides empirical support for the conceptual distinction between challenging and hindering stress in the academic milieu, and demonstrates differential relations between these two constructs and a range of student outcomes (e.g., GPA, transfer intentions, hours withdrawn). The implications of the challenge-hindrance framework for academic environments, research, and inter- ventions are discussed.

Keywords: stressors, self-efficacy, academic performance

The study of stress and performance has a rich, varied, and multidisciplinary history. In higher education specifically, the re- sults of relevant investigations document the deleterious effects of perceived stress on students’ academic performance (Akgun &

Ciarrochi, 2003; Richardson, Abraham, & Bond, 2012), persis- tence/retention (Cox, Reason, Nix, & Gillman, 2016; Johnson, Wasserman, Yildirim, & Yonai, 2014), and experiences of strain (Hudd et al., 2000). While the term “stress” generally refers to a psychological arousal process resulting from demands on an indi- vidual’s coping resources, “stressors” are often conceptualized as the environmental conditions that occasion such arousal (Lazarus, 2006). There is reason for concern regarding the experience and impact of such circumstances among modern college students (Sax, 1997). Indeed, almost 30% of the 84,760 undergraduates surveyed during the American College Health Association’s (ACHA) National College Health Assessment II (NCHA-II) reported that stress had impacted their performance (i.e., grades or course withdrawals) over the last 12 months. These findings highlight the importance of (a) systematic investigations into the stressors of present-day college students, and (b) the explo- ration of stressor-performance relations in the academic milieu.

This article was published Online First April 2, 2020. X Justin Travis, Department of Psychology, University of South Caro-

lina Upstate; Alyssa Kaszycki and Michael Geden, Department of Psy- chology, North Carolina State University; James Bunde, Department of Psychology, University of South Carolina Upstate.

Michael Geden is now at Modern Hire, Cleveland, Ohio. All data collection received IRB approval and informed consent. This

project did not receive funding from any entity. Correspondence concerning this article should be addressed to

Justin Travis, Department of Psychology, University of South Carolina Upstate, 800 University Way, Spartanburg, SC 29303. E-mail: jtravis@ uscupstate.edu

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Journal of Educational Psychology © 2020 American Psychological Association 2020, Vol. 112, No. 8, 1632–1643 ISSN: 0022-0663 http://dx.doi.org/10.1037/edu0000478

1632

Although models of the stressor-performance relationship have generally prioritized the “dark side” of stress, recent theoretical and empirical contributions challenge the assumption that stress is invariably problematic. In particular, a body of management re- search informed by the work of Richard Lazarus (2006) distin- guishes between stressors that negatively impact performance, called “hindrance” stressors, and those that positively impact per- formance, called “challenge” stressors. From this perspective, whether a stressor is appraised as either (a) challenging and doable, or (b) hindering and unmanageable is an important consideration (LePine, Podsakoff, & LePine, 2005; Podsakoff, LePine, & LeP- ine, 2007). Although the former may drive a “rising to the occa- sion” and consequently increase performance, the latter is more likely to induce feelings of helplessness and a motivation to avoid or withdraw, ultimately lowering performance. Assuming that the circumstances in question do in fact occasion arousal (i.e., are indeed considered stressful), research suggests that these alternate appraisals may produce meaningfully disparate psychological re- sponses. For instance, challenge stressors may be accompanied by heightened attentiveness and performance, whereas hindrance stressors may produce anger, leading to deviant or counterproduc- tive behavior (Rodell & Judge, 2009).

In light of this challenge-hindrance distinction and its relative neglect in the higher education literature, the aims of the present study are (a) to synthesize findings relevant to the topic across disciplines, (b) to construct an empirical model of the challenge- hindrance stressor-performance framework based on this integra- tion, and (c) to test the model in a representative sample of college students.

By way of introduction, we will first discuss the challenge- hindrance stressor framework in light of previous findings. Then, we briefly review pertinent work in higher education, identifying gaps in stressor-performance research. Finally, we propose a po- tential moderator of the challenge-academic performance relation- ship (academic self-efficacy [ASE]), and introduce two additional academic outcome variables of interest (in addition to grade point average [GPA])— hours withdrawn and intent to transfer.

The Challenge-Hindrance Framework

Organizational psychologists have historically conceptualized stressors as both a negative antecedent leading to poor perfor- mance, as well as a negative outcome resulting from poor perfor- mance. Early theorizing in the area consequently centered on the notion that stress is inherently negative; however, empirical tests of this assumption yielded mixed and contradictory results (Pod- sakoff et al., 2007). To reconcile these inconsistent findings, researchers developed a new framework for understanding the stressor-performance relationship, drawing on Lazarus and Folk- man’s (1984) model of the stress process.

For contemporary management and educational scholars, the most influential aspect of Lazarus’ theory is likely the “two-stage appraisal process.” This component suggests that an individual first evaluates a stimulus as stressful or not, based on personal relevance (“primary appraisal”). If the stimulus is deemed to be irrelevant, no stress ensues and thus no further appraisals are required. Alternatively, primary appraisals consisting of loss, threat, or challenge naturally result in perceived stress.

If a stimulus is determined to be stressful during primary ap- praisal, the individual evaluates relevant resources and potential coping strategies (i.e., “secondary appraisal”). If the stressor is judged to be surmountable, then the circumstance in question may be viewed as a challenge or growth opportunity, which could boost performance. If the stressor is considered uncontrollable or insur- mountable, it will be perceived as a hindrance or threat, which will likely damage performance. Obviously, this two-dimensional “challenge-hindrance” model theoretically undermines the charac- terization of stressors as universally problematic. Although some kinds of stressful stimuli are inherently damaging to performance, others may serve as motivating and ultimately beneficial opportu- nities for growth. In sum, various stressors may have systemati- cally different and diametrically opposed consequences for perfor- mance, depending on the appraisal of those stressors.

One might question, however, whether there is compelling reason to believe that hindrance and challenge stress represent fundamentally different dimensions—as opposed to varying levels of a unitary construct. Perhaps the distinction merely divides stressors into those that are more (hindrance) and less (challenge) intense and/or unmanageable, with differential associations reflect- ing only the curvilinear relationships routinely found between stress/arousal and various aspects of performance. There are sev- eral reasons to believe that this is not the case. First, it would seem that the two categories are logically separable when goal-relevance and level of control are considered alongside intensity and man- ageability. For instance, demand ambiguity and unfair treatment are likely to be considered uncontrollable—and therefore hinder- ing— even when relatively inconsequential. “Busy work” is a hindrance stressor even when perceived as doable, precisely be- cause it provides little in the way of goal achievement or need satisfaction. These kinds of demands could become less stressful (or not at all stressful) if levels of manageability and intensity were varied, but could likely never— due to their very nature—“trans- form” into challenges. Similarly, a goal-relevant demand under one’s control does not cease to be a challenge simply because the difficulty and stakes—and the accompanying stress level—is per- ceived to be high. One might “rise to the occasion” specifically in response to the reinforcing stimulation associated with a genuine, formidable challenge (i.e., one with an uncertain outcome, met only with considerable difficulty) in an area of importance. There are limits, of course; an apparently impossible demand would constitute a hindrance stressor independent of all other character- istics. It is only at this extreme of utter unmanageability, however, that we would expect otherwise meaningful distinctions to lose relevance.

Second, the I/O psychology literature provides empirical evi- dence supporting the differentiation of stressors. For instance, Cavanaugh, Boswell, Roehling, and Boudreau (2000) reported a two-factor challenge-hindrance structure for work stress (manag- ers), as did Boswell, Olson-Buchanan, and LePine (2004) (univer- sity staff employees). Van den Broeck, De Cuyper, De Witte, and Vansteenkiste (2010) found multifactor models superior to the one-factor solution for job demands. Although both job challenges and job hindrances are taxing to the employee, they are in general only moderately correlated and exhibit not only differential, but opposite relations with a host of work-related variables (e.g., job satisfaction, performance, turnover; Cavanaugh et al., 2000; Lep- ine et al., 2005; Podsakoff et al., 2007). Notably, part-time MBA

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1633STRESSORS ON CAMPUS

students not only rated challenge stressors as more challenging than hindering, and hindrance stressors as more hindering than challenging—they did so despite judging the two kinds of de- mands as similarly stressful (LePine et al., 2005). Based on the above, we would argue that the challenge-hindrance framework is not only theoretically compelling, but also relatively well estab- lished in the vocational realm. One task of the current study is, then, to examine whether the distinction also obtains in the aca- demic milieu.

In fact, surprisingly few studies have explored this perspective in higher education, despite its fruitful application to other do- mains. At least one recent, comprehensive review (Richardson et al., 2012) fails to distinguish between types of stressors, and conceptualizes academic stress simply as negative emotionality with academic origins. It is relatively unsurprising, therefore, that the authors report small negative correlations between general/ academic stress and GPA. The actual effects of stress perception on performance may have been obscured by a failure to distinguish between types of stress, such that the theoretically positive impact of challenge stressors and the theoretically negative impact of hindrance stressors partially counteracted each other, resulting in a general estimate of effect that actually corresponds to neither relationship.

In one of the few studies applying the challenge-hindrance stressor framework to the collegiate environment, Zhu, He, & Wang, 2017 examined the effect of stressor perceptions on aca- demic performance in a sample of Chinese university students. Although their results are generally supportive of the two- dimensional stressor model, the primary outcome variable—aca- demic achievement—was assessed via self-report. It is conse- quently difficult to evaluate the strength of this evidence for a model intended to predict, among other things, objective perfor- mance.

In the only study we could find using the challenge-hindrance framework to predict GPA, LePine, LePine, and Jackson (2004) examined motivation to learn and exhaustion as mediators of the stressor-performance relationship. In this study, the outcome of interest was “learning performance,” operationalized as the stu- dent’s semester GPA. Importantly, these researchers also con- trolled for “past learning performance” (the average of all previous semester GPAs), a historically strong predictor of future GPA. Overall, LePine et al. (2004) describe a pattern of results consistent with the challenge-hindrance framework. Specifically, learning performance and motivation to learn were positively associated with challenge stress (e.g., “The difficulty of the work required in your classes”) and negatively related to hindrance stress (e.g., “The degree to which favoritism rather than performance affects final grades in your classes”). Interestingly, both forms of stress were positively associated with exhaustion, supporting the notion that challenge stress, despite contributing to strain, may also be moti- vational. Although the authors found that exhaustion and motiva- tion to learn partially mediated the stressor-performance relation- ship, these effects were smaller than the direct effects of stress perceptions. Notably, both hindrance and challenge stress pre- dicted semester GPA independent of previous academic perfor- mance. In sum, the results of this study support the applicability of the challenge-hindrance framework to the academic milieu, with potential relevance for both objective and subjective outcomes.

In general, the extant education literature on stress and perfor- mance is characterized by excessive reliance on homogenous, specific, and/or small samples. For instance, relatively recent stud- ies have included 107 atypical college freshmen in an introductory seminar course (Zajacova, Lynch, & Espenshade, 2005), 696 in- troductory management students (LePine et al., 2004), 246 man- agement students (Zhu et al., 2017), 579 first-year students (Krumrei-Mancuso, Newton, Kim, & Wilcox, 2013), 469 nontra- ditional students enrolled in adult-tailored degree programs (San- dler, 2000), 530 female first-year students (Rayle, Arredondo, & Kurpius, 2005), and 203 students evaluating stress from an intro- ductory course (Struthers, Perry, & Menec, 2000).

These studies have been vital in the exploration of the domain, often capitalizing on the control afforded by certain levels of homogeneity (in terms of gender, class type, class standing, etc.). Still, the evaluation of hindrance and challenge stressors in rela- tionship to academic outcomes would be optimally informative if executed with representative samples of college students, enrolled across disparate departments, programs, and majors. A truly rep- resentative U. S. sample, heterogeneous across demographic and academic variables, would allow for a more generalizable under- standing of the experience and response to stress among American college students, and could reliably inform interventions designed to apply to this diverse population (or indicate that a varied, tailored approach is optimal).

The Role of Academic Self-Efficacy

Despite considerable empirical support (LePine et al., 2005; Podsakoff et al., 2007), the challenge-hindrance framework has also been subject to theoretical scrutiny. As González-Morales and Neves (2015) point out, the model put forth by Lepine et al. (2005) fails to appreciate the subjectivity involved in stressor perceptions. The widely cited conceptualization of Lazarus and Folkman (1984) hinges on these subjective judgments, implying that hin- drance and challenge are not features of the stressors themselves, but are the product of individual appraisals of those stressors.

Interestingly, the extent to which stressor perceptions differ across individuals appears related to the type of stressor under examination. Specifically, certain stressors are almost always viewed as hindering, such as situational constraints and unfair treatment, but no such consensus exists in the domain of challenge stressors. For instance, one student may see a heavy workload as a challenge and therefore rise to the occasion, while another may doubt her ability to cope, appraise the stressor as a hindrance, and perform poorly as a result. In this case, both students experienced the same objective demand, but evaluated their abilities differ- ently, resulting in vastly disparate consequences for performance. Alternatively, all (or most) students are likely to perceive insuffi- cient guidance or resources as hindering, given that task ambiguity and situational constraints are largely divorced from personal control. This notion—that the perception of theory-consistent chal- lenges is variable but that of theory-consistent hindrances is uni- versal— can be found in the conceptual writings of some stress scholars (Lazarus, 2006; Lazarus & Folkman, 1984) and is sup- ported by empirical research (Gilboa, Shirom, Fried, & Cooper, 2008).

A potential proxy for students’ appraisals of their own scholastic abilities is ASE. Self-efficacy is a self-perception construct cap-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1634 TRAVIS, KASZYCKI, GEDEN, AND BUNDE

turing an individual’s belief in her capabilities, and is generalizable across contexts (e.g., work or school; Bandura, 1977). Bandura’s theory of self-efficacy has generated robust empirical support, with numerous studies supporting the role of efficacy beliefs in medi- ating ability-performance relationships (see Bandura, 1982). Im- portantly, self-efficacy’s role in self-regulatory processes is high- lighted in models of goal-setting (Locke & Latham, 2002) and stress appraisal (Lazarus, 2006), and self-efficacy is a critical variable in the literature on student performance (Pajares, 1996; Zimmerman, 2000). Lazarus (2006) theorized:

The more confident we are of our capacity to overcome obstacles and dangers, the more likely we are to be challenged rather than threatened and vice versa, a sense of inadequacy promotes threat. Because confidence in ourselves varies greatly among different people, indi- viduals differ in whether they are more prone to experience threat or challenge. We can think of this tendency as a personality trait and a concept like self-efficacy (Bandura, 1982) applies (p. 77).

In addition to Lazarus (2006) and Bandura (1982) explicitly connecting self-efficacy to the stressor appraisal process, Chemers, Hu, and Garcia (2001) found that self-efficacy was positively associated with challenging, as opposed to hindering, stressor appraisals. This suggests that students’ perceptions of their own abilities could be crucial in predicting how they will respond to certain demands. For instance, students low in ASE may perceive stressors that “should” be considered challenging as in fact hin- dering; that is, a low opinion of their own abilities may cause them to view virtually any academic requirement as insurmountable.

Once a student perceives a challenge stressor as a hindrance, she may be at risk for a variety of negative academic outcomes—in addition to a compromised GPA. In the following section, two supplementary outcome measures of interest are introduced.

Transfer Intentions and Withdrawal

In addition to GPA, we chose two indicators of student academic involvement— hours withdrawn and transfer inten- tions—as outcome variables. These were selected for three reasons: First, these indicators represent an expanded sampling (in addition to GPA) from the theoretical domain of student performance, and are thus important outcomes in their own right. Second, the sparse research on retention and withdrawal in higher education has largely focused on using demographic characteristics to statistically predict “at risk” groups (Miller, 1997; Okun, Karoly, Martin, & Benshoff, 2009). Although predicting withdrawal is important, this strategy does not illu- minate the underlying mechanisms that explain observed dif- ferences in retention (and is consequently of limited utility from the perspective of theory and practical intervention). Finally, these outcomes provide additional opportunities to evaluate theoretically relevant relationships between various types of stressors and higher education phenomena. Universities are interested in a variety of student outcomes, from academic performance, to extracurricular involvement, to drop-out rates; therefore, the present study includes transfer intentions and hours withdrawn as outcome variables, in addition to GPA, to evaluate stressor effects on a broader range of student function- ing.

The Present Study

The empirical and theoretical foundations of the proposed model can be summarized as follows: Stress has traditionally been conceptualized as problematic for academic functioning, and does in fact predict negative outcomes when broadly construed (Akgun & Ciarrochi, 2003; Cox et al., 2016; Hudd et al., 2000; Johnson et al., 2014; Richardson et al., 2012). Previous investigations in the educational realm have generally failed, however, to consider research supporting the distinction between two categories of stress— hindrance and challenge—and their potentially disparate (and opposing) implications for various types of performance (LePine et al., 2005; Podsakoff et al., 2007); this conceptualization has been successfully applied to workplace psychology (Lazarus, 2006, but the idea of “positive” stress has yet to gain traction in the education literature). The experience and impact of stressors is thought to depend significantly on secondary appraisals— evalua- tions of one’s ability to cope—and the modern college student may face an array of demands with varying implications for such judgments (e.g., challenging, interesting assignments vs. percep- tions of unfair treatment). Secondary appraisals are expected to impact academic outcomes by driving levels of effort, attention, and persistence, and by motivating differentially effective coping behaviors. It is also plausible that the consideration of stress may help to clarify the role of ASE, an established but complicated predictor of college performance, and that ASE may be relevant to the educational application of the challenge-hindrance framework. Given the nature of secondary appraisal, a moderating role (i.e., students who believe they can meet academic challenges will perform especially well when relevant demands are construed as such) appears to us most logically promising. Our model, there- fore, predicts that (a) challenge stressors will be associated with positive academic outcomes, (b) hindrance stressors will be asso- ciated with negative academic outcomes, (c) ASE will be associ- ated with positive academic outcomes, and (d) ASE will moderate the challenge stressor–performance relationship.

The present study first contributes to the higher education liter- ature by using the challenge-hindrance framework to predict an objective measure of academic performance (GPA) in a diverse and generalizable student sample.

Hypothesis 1a: Challenge stressors will have a positive rela- tionship with GPA.

Hypothesis 1b: Hindrance stressors will have a negative rela- tionship with GPA.

Although we were unable to identify any empirical work apply- ing the challenge-hindrance framework to student retention, sev- eral reviews in the management literature have discussed its ap- plicability to employee withdrawal and attrition. For example, meta-analytic regression analyses of more than 180 samples found that hindrance stressors positively predicted turnover intentions and actual turnover, whereas challenge stressors were negatively predictive of both outcomes (Podsakoff et al., 2007). Given these findings, along with the lack of relevant student-centered litera- ture, we expect challenge and hindrance stressors to demonstrate the same general pattern of relationships in our college sample.

Hypothesis 2a: Challenge stressors will be negatively predic- tive of transfer intentions.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1635STRESSORS ON CAMPUS

Hypothesis 2b: Challenge stressors will be negatively predic- tive of hours withdrawn.

Hypothesis 3a: Hindrance stressors will be positively predic- tive of transfer intentions.

Hypothesis 3b: Hindrance stressors will be positively predic- tive of hours withdrawn.

Academic self-efficacy has emerged as a significant predictor of GPA and retention in previous investigations (Multon, Brown, & Lent, 1991; Richardson et al., 2012; Robbins, Lauver, Le, Davis, Langley, & Carlstrom, 2004). Thus, we expect to replicate these robust relationships in the current study.

Hypothesis 4a: Academic self-efficacy will be positively pre- dictive of GPA.

Hypothesis 4b: Academic self-efficacy will be negatively pre- dictive of hours withdrawn.

Hypothesis 4c: Academic self-efficacy will be negatively pre- dictive of transfer intentions.

Finally, ASE was used to explore the potential effects of sec- ondary appraisals. We predict that academic self-efficacy will moderate the effects of challenge stressors on GPA, enhancing those effects for students with high ASE and reducing them for students with low ASE.

Hypothesis 5: ASE will moderate the relationship between challenge stress and grade point average, such that higher ASE will result in a stronger positive relationship between chal- lenge stress and GPA.

Method

Sample and Procedure

Participants (N � 853) included students enrolled at two south- eastern institutions, one private college and one public university, recruited from a representative array of classes 4 to 10 weeks after the start of the study semester. Stressor and ASE items referred to current functioning, experiences, and beliefs, so it was important to assess in the midst of an academic term. The majority of data were collected in the fourth and fifth weeks of the semester (n � 669); however, some data (n � 184) were collected around Week 10. The sample characteristics of the early wave (Mage � 21.92, Propmale � .27, MGPA � 2.99) displayed some moderate differ- ences from the later wave (Mage � 20.86, Propmale � .39, MGPA � 2.70). As this could introduce systematic error, comparisons of subsequent models were conducted, revealing no meaningful dif- ferences between analyses including and excluding the second batch of data.

After providing informed consent, participants completed a bat- tery of study questionnaires (described below). At the end of the term, semester GPAs and class withdrawals were acquired from the records departments of the relevant institutions. The majority of participants identified as female (62%), with an average age of 21 (SD � 4.63). The sample was generally representative in terms of university standing (27% freshman, 22% sophomores, 28%

juniors, 23% seniors), and participants were enrolled for an aver- age of 15 credit hours at the beginning of the study semester. Every academic department was included as part of the recruitment effort, resulting in widespread institutional representation (Top 5: psychology n � 153, business n � 148, computer science/systems n � 76, criminal justice/prelaw n � 62, and interdisciplinary n � 61). Thirty-nine percent of participants reported at least one change of major. Data on race were collected from a subset of participants (n � 362), and revealed the following (62% Cauca- sian, 29% African American, 5% Hispanic, and 4% Asian). Al- though this composition deviates from national statistics in several respects (a higher percentage of African Americans and fewer Caucasians and Hispanics), it matches regional data relatively well (U.S. Census Bureau, 2019).

Measures

Hindrance and challenge stress. Perceptions of hindrance (five items) and challenge (five items) stress were evaluated using LePine, LePine, and Jackson’s (2004) stressor scales. Previous research with students provides evidence of the internal consis- tency (�s � .70) and validity of these scales (Flinchbaugh, Luth, & Li, 2015; LePine et al., 2004; Zhu et al., 2017). Participants rated the level of stress associated with each circumstance on a scale from 1 (no stress) to 5 (a great deal of stress), referring to the current semester. Sample items from the challenge stressor scale include “The difficulty of the work required in your classes” and “The number of projects/assignments in your classes.” Sample items from the hindrance stressor scale include “The inability to clearly understand what is expected of you in your classes” and “The amount of time spent on ‘busy work’ for your classes.” Internal consistencies for the challenge (� � .93) and hindrance (� � .83) stressor scales were acceptable.

Academic self-efficacy. Following Bandura’s (2006a) recom- mendations for scale construction in self-efficacy research, a mea- sure of general self-efficacy was adapted to reflect beliefs specif- ically relevant to collegiate performance. Items were constructed to be applicable to the academic context in general—not tied to particular tasks that might vary across classes and departments. Participants were asked to rate, on a 5-point scale (1 � not at all confident to 5 � extremely confident), how confident they were in their abilities to: “Complete coursework when it is due,” “Manage my schoolwork effectively,” and “Learn the material if I try.” The internal consistency of this three-item scale was acceptable (� � .79), and standardized factor loadings for the items were moderate to strong (.77, .87, and .52, respectively).

Transfer intention. Participants were asked to report their level of agreement (1 � strongly disagree to 5 � strongly agree) with two statements: “I am looking to transfer to another institu- tion” and “It is likely that I will leave or transfer from (name of institution).” The internal consistency of this two-item scale was acceptable (� � .96) and standardized factor loadings for the items were strong (.97 and .88, respectively).

GPA. Semester GPAs (on a 4-point scale) were acquired from the relevant institutional departments at the conclusion of the term (i.e., after grades were finalized).

Withdrawal. At the conclusion of the term, institutional re- cords reported hours associated with any uncompleted course (i.e., no credit was assigned) in which a participant initially enrolled.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1636 TRAVIS, KASZYCKI, GEDEN, AND BUNDE

Results

Analyses were conducted using the lavaan package (Rosseel, 2012) in R (R Core Team, 2016). A total of 853 participants completed the survey; however, there were missing data for GPA (11%) and withdrawal (16%) due to technical challenges at one site (specifically, this institution [n � 89] was unable to access withdrawal data, and GPA records were only reported for a subset of participants). Additionally, across all item responses there was a low proportion of missing data (total � 118, .81%) with a maximum of missing data for the fourth challenge item (1.8%). The missing data was not missing completely at random (MCAR)—the Hawkins test for normality and homoscedasticity (p � .001) and the nonparametric test for homoscedasticity (p � .007) were significant (computed using the MissMech package; Jamshidian, Jalal, & Jansen, 2014). Instead, the data were assumed to be missing at random (MAR), and missing data were handled using full-information maximum likelihood. Participants were screened for careless responding (Meade & Craig, 2012) using Mahalanobis distance at 95% confidence, resulting in the removal of 75 participants (11%), a proportion consistent with previous research (Kam & Meyer, 2015; Meade & Craig, 2012) and high enough to have a detrimental impact on factor structure (Schmitt & Stuits, 1985). The measurement model included students with missing GPA and withdrawal (n � 763) while the structural model used listwise removal on these missing outcome variables (n � 639).

Continuous variables were assessed for multivariate normality using Mardia’s (1970, 1974) and Henze-Zirkler’s (1990) tests in the MVN package (Korkmaz, Goksuluk, & Zararsiz, 2014), with the resulting large and significant deviations from normality (p � .001) supporting the use of Satorra-Bentler scaled test statistics and robust standard error estimates. The use of robust estimators helps control for Type II but not Type I error, which may remain a problem for strongly positively skewed variables such as with- drawal (which has a limited number of observations greater than its minimum).

The criteria for goodness-of-fit metrics were root mean square approximation of error (RMSEA) with a 90% confidence inter- val �.08, comparative fit index (CFI) � .90, and non-normed fit index (NNFI) � .95 (Hooper, Coughlan, & Mullen, 2008; Hu & Bentler, 1999; Steiger, 2007). These somewhat liberal cutoffs were chosen due to the heterogeneity of the study sample. The means, standard deviations, and correlations between latent and manifest

variables are presented in Table 1. Overall, the internal consisten- cies for the challenge stressor, hindrance stressor, and ASE scales were satisfactory.

Measurement Model

Following a two-step confirmatory approach (Anderson & Gerbing, 1988), a measurement model with challenge stressors, hindrance stressors, ASE, and turnover intentions was assessed prior to fitting structural models. Indicators were specified to their respective latent variables, and all variables were freely allowed to covary. Given the large sample size, the chi-square test was nearly guaranteed to reject the null (Bentler & Bonett, 1980), encouraging primary emphasis on alternative goodness-of-fit metrics. Overall fit for the model was moderate, �2(84) � 442.71, p � .001, CFI � .95, RMSEA � 0.075 (.068, .082), NNFI � .94, with all items displaying sufficient factor loadings (�.30; see Table 2). The measurement model fit excluding individuals with missing re- sponses for GPA and withdrawal was essentially the same. Given that challenge and hindrance stressors were highly correlated (r � .86), one plausible alternative model involved combining chal- lenge and hindrance stressors into a single factor. The one-factor model �2(87) � 642.18, p � .001, CFI � .92, RMSEA � 0.091 (.085, .098), NNFI � .90, demonstrated significantly inferior fit, ��2(3) � 176.78, p � .001, compared WITH the two-factor model, supporting the separation of the two constructs.

Structural Models

The proposed structural model (see Figure 1), excluding mod- eration, demonstrated acceptable model fit, �2(106) � 433.90, p � .001, CFI � 0.95, RMSEA � 0.070 (.063, .076), NNFI � .93, with the GPA R2 � .15, withdrawal R2 � .04, and transfer intentions R2 � .03 (see Appendix for residual and factor variances). Chal- lenge stress was positively associated with GPA, (� � .20, p � .038) and negatively associated with withdrawal (� � .22, p � .065), providing support for Hypothesis 1a and partial support for Hypothesis 2b, but was not significantly related to transfer inten- tions (� � .18, p � .105; see Table 3). Hindrance stress was negatively associated with GPA (� � .26, p � .013), supporting Hypothesis 1b, but did not show a statistically significant relation- ship to transfer intentions (� � .17, p � .155) and hours with- drawn (� � .18, p � .121).

Table 1 Descriptive Statistics, Correlations, and Reliabilities

Measure M 95% CI SD 1 2 3 4 5 6

1. Challenge 3.22 [3.14, 3.29] 1.06 (.93) 2. Hindrance 2.66 [2.60, 2.73] .94 .72��� (.83) 3. Academic SE 3.77 [3.71, 3.83] .81 .19��� .20��� (.79) 4. Transfer intentions 2.14 [2.05, 2.22] 1.19 .06 .02 .09� (.96) 5. Withdrawal .39 [.32, .46] .87 .04 .06 .15��� .11�� — 6. GPA 2.84 [2.78, 2.90] .79 .08�� .16��� .30��� .02 .28��� —

Note. GPA � grade point average. Calculated on data before imputation and after careless responder removal. Cronbach’s alpha displayed on diagonals for challenge, hindrance, and academic self-efficacy (SE). Reliability for transfer intentions was assessed using the Spearman-Brown estimate as only two items were included (Eisinga, Grotenhuis, & Pelzer, 2013). � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1637STRESSORS ON CAMPUS

Academic self-efficacy was positively associated with GPA (� � .33, p � .001), and negatively associated with transfer intentions (� � .13, p � .005) and withdrawal (� � .17, p � .001; supporting Hypothesis 4a, Hypothesis 4b, and Hypothesis 4c). We assessed Hypothesis 5 by creating a latent variable com- posed of indicator variables for all item combinations between challenge stress and ASE (n � 15), double-mean centering their values, and then regressing the interaction construct on GPA in an unconstrained model. The interaction construct had no additional impact on GPA (� � .05, p � .172), thus failing to support Hypothesis 5 (see Figure 2).

The structural equation model was assessed for measurement in- variance across class standing, as it is possible that students’ percep-

tions change systematically over the course of their college careers. There was insufficient data for comparing all four class standings (i.e., freshman, sophomore, junior, senior), so a binary indicator was used for comparing early (freshman and sophomores) and advanced (junior and senior) career students. There was no significant effect on factor loadings, ��2(11) � 11.19, p � .428, so although the groups were significantly different in their intercepts, ��2(13) � 38.82, p � .001, and means, ��2(4) � 89.77, p � .001, class standing did not influ- ence our interpretation of results.

Discussion

In the two-stage model proposed by Lazarus, challenge and hindrance stress are differentially associated with performance. The purpose of the current study was to evaluate this framework in higher education, using a large, diverse sample and multiple, objective academic outcomes. Specifically, we predicted that per- ceptions of challenge stressors would be positively related to GPA and negatively related to transfer intentions and withdrawal, whereas hindrance stressors would be negatively associated with GPA and positively associated with transfer intentions and with- drawal. Further, we predicted a moderating role for ASE in the challenge-performance relationship.

Table 2 Path Coefficients for the Measurement Model

Latent variable b SE

Challenge Item 1 1.00 .86 — Item 2 1.01 .89 .02 Item 3 .84 .78 .03 Item 4 .99 .85 .03 Item 5 1.02 .86 .03

Hindrance Item 1 1.00 .63 — Item 2 .88 .54 .08 Item 3 1.14 .70 .08 Item 4 1.38 .84 .08 Item 5 1.14 .71 .08

Academic self-efficacy Item 1 1.00 .82 — Item 2 1.15 .86 .07 Item 3 .69 .52 .05

Transfer intentions Item 1 1.00 .91 — Item 2 1.12 .97 .19

Note. Standardized estimates are standardized across the latent variables. b � unstandardized estimate; � standardized estimate; SE � standard error.

GPA

Challenge

Stress

Hindrance

Stress

Academic

Self-

Efficacy

Transfer

Intentions

Hours

Withdrawn

+

-

-

+ -

-

- +

+

Figure 1. Hypothesized conceptual model. “�” refers to predicted pos- itive relationships. “ ” refers to predicted negative relationships. GPA � grade point average.

Table 3 Path Coefficients for the Structural Model

Outcome Predictor b � SE p

GPA Challenge .17 .20 .08 .038 Hindrance .31 .26 .12 .013 Academic self-efficacy .36 .33 .05 .001

Hours withdrawn Challenge .26 .22 .14 .065 Hindrance .30 .18 .20 .121 Academic self-efficacy .25 .17 .07 .001

Transfer intentions Challenge .20 .18 .12 .105 Hindrance .26 .17 .18 .155 Academic self-efficacy .17 .13 .06 .005

Note. Standardized estimates are standardized across the latent variables. GPA � grade point average; b � unstandardized estimate; � � standard- ized estimate; SE � standard error; p � significance level.

GPA

Challenge

Stress

Hindrance

Stress

.87

Academic

Self-

Efficacy

Transfer

Intentions

Hours

Withdrawn -.30

-.23

.33

-.17

-.13

.20 -.22

-.18

-.26 .18

.17

Figure 2. Structural equation model with the standardized regression weights (N � 638), �2(106) � 433.90, p � .001, CFI � .95, RMSEA � .070 (.063, .076), NNFI � .93. GPA � grade point average.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1638 TRAVIS, KASZYCKI, GEDEN, AND BUNDE

Challenge stress, hindrance stress, and ASE were all significant predictors of GPA. These results are consistent with the theoretical models of Lazarus (Lazarus, 2006) and Bandura (Bandura, 1982), as well as the meta-analytic results linking challenge and hin- drance stressors to work performance (LePine et al., 2005; Pod- sakoff et al., 2007). Given the strong positive relationship between hindrance and challenge stress in our sample (r � .72), it is especially notable that each of the three predictors accounted for unique variance in GPA.

Consistent with hypotheses, ASE was negatively associated with transfer intentions, whereas challenge stress and ASE were negatively predictive of withdrawal. As goal-setting theory would suggest (Locke & Latham, 1990), self and environmental appraisals were associated with task persistence (measured subjectively or objec- tively). Assuming that degree completion is a primary goal of most students, perceptions of challenge stress and self-efficacy appeared to facilitate the motivational process via increased persistence, which may help to explain their associations with performance as well.

Despite generally small to medium direct effects, in expected directions, the remaining hypothesized relationships failed to reach statistical significance. One hypothesized explanation for this re- sult, which was found, notably, only among “persistence” out- comes (i.e., transfer intentions and withdrawal), involved class standing. It seems plausible that for seniors, who were included in all analyses, relationships between stress perceptions and persis- tence would be attenuated in light of a powerful, proximal met- agoal (i.e., graduation). Indeed, one might expect virtually all relationships with persistence to disappear in the semester prior to degree completion, given the potential of variance constriction in both hours withdrawn and transfer intentions. Notably, this would not necessarily constitute a sign of persistence in the positive sense, but could reflect a problematic honoring of “sunk costs.” There were slight differences among study variables across class years, however, our test of structural invariance did not support this explanation in the current study. Nevertheless, we encourage future researchers to examine the possibility with sufficient sample sizes in each class year.

The expected moderating effect for ASE was not observed, and an inspection of the bivariate associations in Table 1 suggests a potential, partial explanation. First, as previously noted, the relationship be- tween hindrance and challenge stressors was strong and positive. While this makes their unique (and opposite) associations with GPA all the more notable, and suggests that their nonoverlapping variance is of critical importance, it also indicates that in general, participants reporting high levels of the “good” kind of stress also reported high levels of the “bad” kind. Further, the significant negative relationships between ASE and both kinds of stress suggest that the construct may have failed to moderate the stress-performance relationship partially because the hypothesized advantages for high ASE students were less relevant than expected; that is, coping ability is less important when one has somewhat less stress to cope with. Although the effect in question was small, it does point toward one avenue by which ASE might contribute to academic outcomes—through suppressing general levels of perceived stress independent of the actual nature or extent of the relevant demands. It is notable that the items on both the hindrance and challenge stressor scales do not ask participants to what degree they have experienced certain circumstances, but to what degree they were stressed by such circumstances. This is consistent with the underlying theoretical model, but it seems plausible that a high ASE

student might respond to a challenge stressor question not by saying “yes, I have been stressed by the difficulty of work in my classes, but I have what it takes to rise to the challenge,” but by instead indicating that “the work in my classes is difficult, but it doesn’t stress me out because I know I can handle it.”

This is related to one additional point regarding the hindrance and challenge stressor scales. It was noted above that there is little debate regarding the status of certain hindrance stressors, but that perceptions of challenge stressors vary markedly due to differ- ences in secondary appraisals (see González-Morales & Neves, 2015 for discussion). That individual differences could play an important role in primary appraisals of the circumstances de- scribed in the stressor scales, for both hindrance and challenge items, should also be considered. For instance, one item asks participants how much stress they experience from “the amount of time spent working on projects/assignments for your classes.” Whether or not this represents a potentially stressful circumstance requiring an evaluation of coping resources, and how potentially stressful it might in fact be, depends on whether and how much you care about the class, whether and how much you care about your grade in the class, and a variety of other factors reflecting levels of motivation, personal values, and goals. A student could truthfully report low levels of hindrance and challenge stress because he doesn’t care at all about college and is thus unperturbed by its injustices and demands, while a grade-conscious perfection- ist might report universally high challenge stress, simply because of her need and expectation for flawlessness in the academic arena.

The comparison of a proposed model to viable alternatives is important for establishing the validity of that model (MacCallum, Wegener, Uchino, & Fabrigar, 1993); however, the temporal se- quence of data collection in this study renders most alternative spec- ifications implausible (e.g., class withdrawals could not influence reported stressors, as the latter were assessed while all participants were still enrolled). The modeling of structural relationships with stressors as a single construct (traditional conceptualization) would seem a reasonable alternative; however, the one-factor measurement model proved inferior to the two-factor model, leaving few plausible alternatives to test. One construct that should be included in future models (whenever possible) is previous academic performance, given its expected relationship with current performance and potential rel- evance for stress perceptions. Unfortunately, we were unable to ac- quire this information for the current study.

Conclusions

In sum, these results support the possibility that some stressful stimuli can enhance academic performance—provided the stress is perceived as both goal-relevant and controllable. For example, a high workload should, according to the challenge-hindrance framework, predict added effort—assuming that (a) success in the class is important to the student, and (b) the student believes that success hinges, at least in part, on her efforts. This latter point highlights the issue of control, a notion consistently emphasized in motivation theory (e.g., Bandura, 2006b; Rotter, 1966; Ryan & Deci, 2006). If a stressor is perceived as wholly uncontrollable, helplessness and inactivity are more likely (and perhaps more logical) responses than redoubled effort. For instance, a deeply held belief that an instructor “has it in for you” might lead one to conclude, quite understandably, that enhanced academic efforts, at

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1639STRESSORS ON CAMPUS

least, are futile (e.g., “What’s the point?”). As Vroom’s (1964) expectancy theory suggests, the perception that outcome will re- flect effort appears vital for maintaining and strengthening moti- vation. Indeed, many “hindrance” stressors are so categorized precisely because they are assumed to compromise this link.

Once a stressor is deemed goal-relevant and controllable, subse- quent performance may best be understood through goal-setting and the high-performance cycle. In response to challenge stress, improved performance results from increased persistence, effort, directed atten- tional focus, and planning (Locke & Latham, 1990). For instance, in the current study, “difficulty of work” was classified as a challenge stressor— one that would likely demand each of the behaviors listed above, but one for which these behaviors would also, presumably, be rewarded. Conversely, if a student feels “hindered” by implacable favoritism (e.g., the professor has decided to “fail me no matter what”), the mechanisms by which stress could potentially improve performance are likely to be dismissed as irrelevant and ineffectual, if they are entertained at all.

Distinguishing between hindrance and challenge stressors is an important step forward in clarifying the mixed findings for stressor-performance relationships in the domain of organizational behavior. Although this distinction has recently been applied to the study of students’ perceived stress and academic performance, these investigations were generally limited by restricted outcomes, small sample sizes, and/or homogenous sampling. This study sought to build on this previous work using the expanded criteria of GPA, hours withdrawn, and transfer intentions, and to do so using a more representative sample than has been employed to date. By including objective measures of student performance and withdrawal alongside a subjective measure of transfer intentions, the proposed model predicts criteria important to higher education decision makers. Furthermore, by avoiding sample restrictions (e.g., by major [e.g., all psychology students] or class standing [e.g., all freshmen]), the framework was examined in a sample that better generalizes to American university students.

Several limitations of the current study warrant mention. First, our theoretical model prompts causal speculation, but the correlational nature of all data must be kept in mind. Second, an inevitable tradeoff was made between generalizability and control. Although a diverse sample was a priority, it is possible that this diversity introduced additional noise. For instance, there could be differences in the effects of stressor types by year in school, gender, ethnicity, and college major, and limiting a sample to students enrolled in the same pro- gram—perhaps even taking the same classes with the same profes- sors—may allow for more comparable environmental demands (e.g., similar assignments, grading schemes, or classroom environments). Another limitation is the relatively small sample size in relation to its heterogeneity, which was likely a factor in the generally moderate- to-weak goodness-of-fit for the proposed models. The sample size also prevented further investigation into how demographic variables may influence these relationships, and likely resulted in low power for evaluating rare events (e.g., turnover intentions and withdrawal).

Another concern involves participant perceptions of LePine et al.’s (2005) stressor items. Although the goal of our study was to demon- strate the utility of differentiating between stressor types and extend- ing this framework into the educational domain, we agree with González-Morales and Neves (2015) that greater attention should be paid to whether or not “challenge” stressors (often workload and scope of responsibility) are actually appraised as challenges by par-

ticipants (an assumption that is rarely tested). If participants did not evaluate the challenge stressor items as manageable, this would likely have attenuated relationships in our model. Challenge and hindrance stressors had medium direct effects on all three outcomes, but three of the four relationships with hours withdrawn and transfer intentions failed to reach statistical significance. The large error present in estimating stressor relationships with our persistence outcomes may have resulted from failure to distinguish explicitly between primary and secondary appraisals. Future research that assesses stress experi- enced from certain scenarios and appraisals of perceived ability to cope—the role we hypothesized for ASE—will allow for finer- grained evaluation of how these different perceptions influence aca- demic phenomena.

Practical Implications

Below, we have highlighted several implications of the current study. Our investigation targeted higher education, but assuming our theoretical framework generalizes across developmental stages and instructional settings, educators and administrators responsible for the design of learning environments across the entire educa- tional spectrum might benefit from consideration of the challenge- hindrance distinction.

First, educators should experiment with academic environments that intentionally facilitate challenge and limit hindrance. As re- viewed above, much of the academic stress literature focuses on the negative effects of stressors, and therefore (understandably) recommends their removal whenever possible. Although we obvi- ously agree that elimination of true obstacles to success (“bad” stress) is a worthy and important goal, educators might also at- tempt to capitalize on the impact of “good” stress. In short, targeting hindrances will likely alleviate some distress, and remove certain concrete barriers to success, but this may not be sufficient to optimize student motivation. Instead, we think that goal-setting theory and the transactional theory of stress can be applied to the design of learning environments, potentially resulting in approach- oriented behaviors associated with superior student experiences and performance. This may have the double-edged effect of re- ducing the psychological distress and strain associated with hin- drances, while at the same time enhancing the motivation and well-being linked to appropriate challenge.

One promising area involves the application of work design principles to the learning context. For example, Cotton, Dollard, and De Jonge (2002) and Chambel and Curral (2005) have adopted work design theories (e.g., demands-control model; Karasek, 1979) to investigate the influence of school demands and resources on psychological states, and ultimately school outcomes. This work supports the notion that school characteristics (framed as job/work characteristics) influence psychological states, and that these states impact well-being and performance—much like the influential job characteristics model (Hackman & Oldham, 1976). These studies (and the current one) extend work motivation and job design theories into the academic context, and could inform structural changes to various learning environments.

Second, students may be presented with strategies for reducing or eliminating the aversiveness of stress. Academic researchers com- monly measure stress in a way that captures both the experience of stress and its deleterious effects. For example, one typical measure, the perceived stress scale (PSS; Cohen, Kamarck, & Mermelstein, 1983), asks respondents how often “. . . have you been upset because

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1640 TRAVIS, KASZYCKI, GEDEN, AND BUNDE

of something that happened unexpectedly?” and “. . . have you felt confident about your ability to handle your personal problems?” Participants also indicate, for instance, how often they have “. . . dealt successfully with irritating life hassles.” When stress assessment is approached in this fashion, the aggregate scale score is composed of (at minimum) various negative emotional reactions, efficacy beliefs, and perceptions of positive resolution.

Given this heterogeneous conceptualization of stress, academic interventions frequently seek to impact both the nature and frequency of stressful experiences as well as the student response to stress. The student stress-performance literature has increasingly acknowledged this distinction, yet research has only recently investigated the possi- bility that challenging demands might promote adaptive coping strat- egies in the academic context (Robotham, 2008). Based on the above, we recommend that designers of learning environments attempt to (a) incorporate challenging and difficult work when congruent (and only when congruent) with learning objectives; (b) make the links between challenges and goals clear and self-relevant, tailoring demands when possible and appropriate; and (c) ensure that institutional resources are available and known to students. Obviously, these recommendations do not ensure that “objective” challenges will improve performance; however, they are low-cost interventions that should increase the likelihood that students will be profitably challenged in the classroom and experience academic demands positively (notably, the benefits of difficult but attainable challenges at work are clear; Locke & Latham, 1990). Further, the facilitation of challenge appraisals is intended to promote a generally superior problem-focused (vs. emotion-focused) coping strategy. Of course, problem-focused coping is only possible within an individual’s capabilities, and educators should be mindful of student aptitude and achievement when conceptualizing and imple- menting requirements (Lazarus, 2006).

Recommendations for Future Research

One obvious arena for future research involves the investigation of individual differences in appraisals. Although this study con- ceptualized challenge and hindrance stressors as objectively clas- sifiable, people likely vary in the extent to which they deem a particular stressor to be goal-relevant and within their control. Individual difference variation in stress appraisals should be con- sidered when stress items and assessments are constructed. For instance, our measure captured the experience of stress in response to a demand; future research might include subjective measures of this type in concert with more objective indicators (e.g., the actual number of hours per week spent on classwork). Further, the analysis of discrepancies between objective demands and subjec- tive perceptions— of workload, for instance—may provide insight into the actual appraisal process. It is also critical for future work to identify individual differences that predict such discrepancies.

Last, experimental changes to educational procedures should be undertaken, consistent with the theoretical framework presented above. We have made suggestions in light of the extant literature, but identifying the ideal response to the challenge-hindrance distinction will require careful variation of demands across diverse students, institutions, and contexts, and the systematic measurement of objec- tive and subjective student outcomes resulting from these manipula- tions. It seems that some academic stress might indeed be “good” and some “bad” (or at least “better” and “worse”), and determining

precisely how this realization can be used to increase student success and well-being is an urgent, empirical question.

References

Akgun, S., & Ciarrochi, J. (2003). Learned resourcefulness moderates the relationship between academic stress and academic performance. Edu- cational Psychology, 23, 287–294. http://dx.doi.org/10.1080/ 0144341032000060129

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411– 423. http://dx.doi.org/10.1037/0033-2909.103.3.411

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. http://dx.doi.org/10.1037/ 0033-295X.84.2.191

Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122–147. http://dx.doi.org/10.1037/0003-066X.37.2.122

Bandura, A. (2006a). Guide for constructing self-efficacy scales. In F. Pajares & T. Urdan (Eds.), Adolescence and education: Vol. 5. Self-efficacy and adolescence (pp. 307–337). Greenwich, CT: Information Age.

Bandura, A. (2006b). Toward a psychology of human agency. Perspectives on Psychological Science, 1, 164 –180. http://dx.doi.org/10.1111/j.1745- 6916.2006.00011.x

Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588 – 606. http://dx.doi.org/10.1037/0033-2909.88.3.588

Boswell, W. R., Olson-Buchanan, J. B., & LePine, M. A. (2004). The relationship between work-related stress and work outcomes: The role of felt-challenge and psychological strain. Journal of Vocational Behavior, 64, 165–181. http://dx.doi.org/10.1016/S0001-8791(03)00049-6

Cavanaugh, M. A., Boswell, W. R., Roehling, M. V., & Boudreau, J. W. (2000). An empirical examination of self-reported work stress among U.S. managers. Journal of Applied Psychology, 85, 65–74. http://dx.doi .org/10.1037/0021-9010.85.1.65

Chambel, M. J., & Curral, L. (2005). Stress in academic life: Work characteristics as predictors of student well-being and performance. Applied Psychology, 54, 135–147. http://dx.doi.org/10.1111/j.1464- 0597.2005.00200.x

Chemers, M. M., Hu, L. T., & Garcia, B. F. (2001). Academic self-efficacy and first year college student performance and adjustment. Journal of Educational Psychology, 93, 55– 64. http://dx.doi.org/10.1037/0022- 0663.93.1.55

Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. http://dx.doi.org/10.2307/2136404

Cotton, S. J., Dollard, M. F., & De Jonge, J. (2002). Stress and student job design: Satisfaction, well-being, and performance in university students. International Journal of Stress Management, 9, 147–162.

Cox, B. E., Reason, R. D., Nix, S., & Gillman, M. (2016). Life happens (outside of college): Non-college life-events and students’ likelihood of graduation. Research in Higher Education, 57, 823– 844. http://dx.doi .org/10.1007/s11162-016-9409-z

Eisinga, R., Grotenhuis, M., & Pelzer, B. (2013). The reliability of a two-item scale: Pearson, Cronbach, or Spearman-Brown? International Journal of Public Health, 58, 637– 642. http://dx.doi.org/10.1007/ s00038-012-0416-3

Flinchbaugh, C., Luth, M. T., & Li, P. (2015). A challenge or a hindrance? understanding the effects of stressors and thriving on life satisfaction. International Journal of Stress Management, 22, 323–345. http://dx.doi .org/10.1037/a0039136

Gilboa, S., Shirom, A., Fried, Y., & Cooper, C. (2008). a meta-analysis of work demand stressors and job performance: Examining main and moderating effects. Personnel Psychology, 61, 227–271. http://dx.doi .org/10.1111/j.1744-6570.2008.00113.x

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1641STRESSORS ON CAMPUS

González-Morales, M. G., & Neves, P. (2015). When stressors make you work: Mechanisms linking challenge stressors to performance. Work and Stress, 29, 213–229. http://dx.doi.org/10.1080/02678373.2015.1074628

Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Perfor- mance, 16, 250 –279. http://dx.doi.org/10.1016/0030-5073(76)90016-7

Henze, N., & Zirkler, B. (1990). A class of invariant consistent tests for multivariate normality. Communications in Statistics Theory and Meth- ods, 19, 3595–3617. http://dx.doi.org/10.1080/03610929008830400

Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation modelling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6, 53– 60.

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alterna- tives. Structural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/ 10705519909540118

Hudd, S. S., Dumlao, J., Erdmann-Sager, D., Murray, D., Phan, E., Soukas, N., & Yokozuka, N. (2000). Stress at college: Effects on health habits, health status and self-esteem. College Student Journal, 34, 217–228.

Jamshidian, M., Jalal, S. J., & Jansen, C. (2014). Missmech: An R package for testing homoscedasticity, multivariate normality, and missing com- pletely at random (MCAR). Journal of Statistical Software, 56, 1–31. http://dx.doi.org/10.18637/jss.v056.i06

Johnson, D., Wasserman, T., Yildirim, N., & Yonai, B. (2014). Examining the effects of stress and campus climate on the persistence of students of color and White students: An application of Bean and Eaton’s psycho- logical model of retention. Research in Higher Education, 55, 75–100. http://dx.doi.org/10.1007/s11162-013-9304-9

Kam, C. C. S., & Meyer, J. P. (2015). How careless responding and acquiescence response bias can influence construct dimensionality: The case of job satisfaction. Organizational Research Methods, 18, 512–541. http://dx.doi.org/10.1177/1094428115571894

Karasek, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285–308. http://dx.doi.org/10.2307/2392498

Korkmaz, S., Goksuluk, D., & Zararsiz, G. (2014). An R package for assessing multivariate normality. The R Journal, 6, 151–162. http://dx .doi.org/10.32614/RJ-2014-031

Krumrei-Mancuso, E. J., Newton, F. B., Kim, E., & Wilcox, D. (2013). Psychosocial factors predicting first-year college student success. Jour- nal of College Student Development, 54, 247–266. http://dx.doi.org/10 .1353/csd.2013.0034

Lazarus, R. S. (2006). Stress and emotion: A new synthesis. New York, NY: Springer. Retrieved from http://www.ebrary.com.prox.lib.ncsu.edu

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York, NY: Springer Publishing Co.

LePine, J. A., LePine, M. A., & Jackson, C. L. (2004). Challenge and hindrance stress: Relationships with exhaustion, motivation to learn, and learning performance. Journal of Applied Psychology, 89, 883– 891. http://dx.doi.org/10.1037/0021-9010.89.5.883

LePine, J. A., Podsakoff, N. P., & LePine, M. A. (2005). A meta-analytic test of the challenge stressor– hindrance stressor framework: An expla- nation for inconsistent relationships among stressors and performance. Academy of Management Journal, 48, 764 –775. http://dx.doi.org/10 .5465/amj.2005.18803921

Locke, E. A., & Latham, G. P. (1990). Work motivation and satisfaction: Light at the end of the tunnel. Psychological Science, 1, 240 –246. http://dx.doi.org/10.1111/j.1467-9280.1990.tb00207.x

Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation. A 35-year odyssey. American Psychologist, 57, 705–717. http://dx.doi.org/10.1037/0003-066X.57.9.705

MacCallum, R. C., Wegener, D. T., Uchino, B. N., & Fabrigar, L. R. (1993). The problem of equivalent models in applications of covariance

structure analysis. Psychological Bulletin, 114, 185–199. http://dx.doi .org/10.1037/0033-2909.114.1.185

Mardia, K. V. (1970). Measures of multivariate skewness and kurtosis with applications. Biometrika, 57, 519 –530. http://dx.doi.org/10.1093/ biomet/57.3.519

Mardia, K. V. (1974). Applications of some measures of multivariate skewness and kurtosis in testing normality and robustness studies. Sankhya Series B, 36, 115–128.

Meade, A. W., & Craig, S. B. (2012). Identifying careless responses in survey data. Psychological Methods, 17, 437– 455. http://dx.doi.org/10 .1037/a0028085

Miller, J. C. (1997). Variables affecting the decision to withdraw from liberal arts and science courses. Community College Review, 25, 39 –54. http://dx.doi.org/10.1177/009155219702500304

Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of self- efficacy beliefs to academic outcomes: A meta-analytic investigation. Journal of Counseling Psychology, 38, 30 –38. http://dx.doi.org/10 .1037/0022-0167.38.1.30

Okun, M. A., Karoly, P., Martin, J. L., & Benshoff, A. (2009). Distin- guishing between exogenous and endogenous intent-to-transfer students. Journal of College Student Retention, 10, 507–524. http://dx.doi.org/10 .2190/CS.10.4.f

Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational Research, 66, 543–578. http://dx.doi.org/10.3102/ 00346543066004543

Podsakoff, N. P., LePine, J. A., & LePine, M. A. (2007). Differential challenge stressor-hindrance stressor relationships with job attitudes, turnover intentions, turnover, and withdrawal behavior: A meta-analysis. Journal of Applied Psychology, 92, 438 – 454. http://dx.doi.org/10.1037/ 0021-9010.92.2.438

Rayle, A. D., Arredondo, P., & Kurpius, S. E. R. (2005). Educational Self-Efficacy of college women: Implications for theory, research, and practice. Journal of Counseling and Development, 83, 361–366. http:// dx.doi.org/10.1002/j.1556-6678.2005.tb00356.x

R Core Team. (2016). R: A language and environment for statistical computing (Version 3.3.2) [Computer software]. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R- project.org/

Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students’ academic performance: A systematic review and meta-analysis. Psychological Bulletin, 138, 353–387. http://dx.doi.org/ 10.1037/a0026838

Robbins, S. B., Lauver, K., Le, H., Davis, D., Langley, R., & Carlstrom, A. (2004). Do psychosocial and study skill factors predict college out- comes? A meta-analysis. Psychological Bulletin, 130, 261–288. http:// dx.doi.org/10.1037/0033-2909.130.2.261

Robotham, D. (2008). Stress among higher education students: Towards a research agenda. Higher Education, 56, 735–746. http://dx.doi.org/10 .1007/s10734-008-9137-1

Rodell, J. B., & Judge, T. A. (2009). Can “good” stressors spark “bad” behaviors? The mediating role of emotions in links of challenge and hindrance stressors with citizenship and counterproductive behaviors. Journal of Applied Psychology, 94, 1438 –1451. http://dx.doi.org/10 .1037/a0016752

Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48, 1–36. http://dx.doi.org/10.18637/jss .v048.i02

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80, 1–28. http:// dx.doi.org/10.1037/h0092976

Ryan, R. M., & Deci, E. L. (2006). Self-regulation and the problem of human autonomy: Does psychology need choice, self-determination, and will? Journal of Personality, 74, 1557–1585. http://dx.doi.org/10.1111/ j.1467-6494.2006.00420.x

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1642 TRAVIS, KASZYCKI, GEDEN, AND BUNDE

Sandler, M. E. (2000). Career decision-making self-efficacy, perceived stress, and an integrated model of student persistence: A structural model of finances, attitudes, behavior, and career development. Research in Higher Education, 41, 537–580. http://dx.doi.org/10.1023/A:1007032525530

Sax, L. J. (1997). Health trends among college freshmen. Journal of American College Health, 45, 252–262. http://dx.doi.org/10.1080/ 07448481.1997.9936895

Schmitt, N., & Stuits, D. M. (1985). Factors defined by negatively keyed items: The result of careless respondents? Applied Psychological Mea- surement, 9, 367–373. http://dx.doi.org/10.1177/014662168500900405

Steiger, J. H. (2007). Understanding the limitations of global fit assessment in structural equation modeling. Personality and Individual Differences, 42, 893– 898. http://dx.doi.org/10.1016/j.paid.2006.09.017

Struthers, C. W., Perry, R. P., & Menec, V. H. (2000). An examination of the relationship among academic stress, coping, motivation, and perfor- mance in college. Research in Higher Education, 41, 581–592. http:// dx.doi.org/10.1023/A:1007094931292

U.S. Census Bureau. (2019). Quick facts—Population estimates. Retrieved from https://www.census.gov/quickfacts/fact/table/U.S./PST045216

Van den Broeck, A., De Cuyper, N., De Witte, H., & Vansteenkiste, M. (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. http:// dx.doi.org/10.1080/13594320903223839

Vroom, V. H. (1964). Work and motivation. New York, NY: Wiley. Zajacova, A., Lynch, S. M., & Espenshade, T. J. (2005). Self-efficacy,

stress, and academic success in college. Research in Higher Education, 46, 677–706. http://dx.doi.org/10.1007/s11162-004-4139-z

Zhu, Y., He, W., & Wang, Y. (2017). Challenge– hindrance stress and academic achievement: Proactive personality as moderator. Social Behavior and Per- sonality, 45, 441– 452. http://dx.doi.org/10.2224/sbp.5855

Zimmerman, B. J. (2000). Self-efficacy: An essential motive to learn. Contemporary Educational Psychology, 25, 82–91. http://dx.doi.org/10 .1006/ceps.1999.1016

Appendix

Table of Unstandardized Observed Residual Variances and Factor Variances From the Final Structural Model

Received March 17, 2019 Revision received January 24, 2020

Accepted February 16, 2020 �

Table A1

Latent variable Observed residual variance Factor variance

Challenge .991 Item 1 .396 Item 2 .326 Item 3 .511 Item 4 .409 Item 5 .419

Hindrance .503 Item 1 .925 Item 2 1.07 Item 3 .813 Item 4 .455 Item 5 .736

Academic self-efficacy .605 Item 1 .301 Item 2 .283 Item 3 .568

Transfer intentions 1.10 Item 1 .180 Item 2 .026

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

1643STRESSORS ON CAMPUS