Applied Research method

THE JOURNAL OF GENETIC PSYCHOLOGY, 175(1), 76–90, 2014 Copyright C© Taylor & Francis Group, LLC ISSN: 0022-1325 print / 1940-0896 online DOI: 10.1080/00221325.2013.806293

Coping with Negative Emotions: Connections with Adolescents’ Academic Performance and Stress

William F. Arsenio and Samantha Loria Yeshiva University

ABSTRACT. The authors assessed connections among adolescents’ emotional dispositions, negative academic affect, coping strategies, academic stress, and overall grade point average (GPA). A total of 119 ninth through 12th-grade students completed assessments for (a) overall positive and negative moods, (b) GPA, and (c) academically related variables involving stress, negative emotions, and engaged and disengaged coping strategies. Greater negative academic affect and disengaged coping were related to lower GPAs, and disengaged coping mediated the connection between negative aca- demic affect and GPA. By contrast, higher academic stress was related to students’ overall moods, negative academic affect, and disengaged coping; disengaged coping mediated the connection be- tween academic stress and negative overall moods. Discussion focused on the especially problematic nature of disengaged academic coping.

Keywords academic performance, academic stress, coping styles, negative affect

Research on the connections between children’s social and emotional competence has undergone enormous growth over the past couple of decades (for reviews, see Denham, 1998; Saarni, Campos, Camras, & Witherington, 2006). In general, studies have shown that children’s negative emotional tendencies (including mood and temperament) and lower emotion-related abilities (including emotion knowledge and regulation) are connected with greater peer rejection and less competent interactions with peers (e.g., Arsenio, Cooperman, & Lover, 2000) and adults (e.g., Valiente & Eisenberg, 2006).

An early review by Parker and Asher (1987) highlighted the educational implications of this social-emotional emphasis: peer rejected children were at least two and a half times more likely to quit high school than their nonrejected peers. Since then much has been learned about how both peer–peer (e.g., Ladd, Birch, & Buhs, 1999; Wentzel, 2005) and student–teacher relationships (Pianta, 1999) influence children’s and adolescents’ academic trajectories. To date, however, much of the emphasis has been on how social rather than emotional competence, per se, relates to school outcomes (although see Trentacosta & Izard, 2007). As Schutz and Pekrun (2007) observed, “in spite of the emotional nature of classrooms, inquiry on emotions in educational contexts, outside of a few notable exceptions. .. has been slow to emerge” (p. 3).

Received December 17, 2012; accepted May 6, 2013. Address correspondence to William F. Arsenio, Ferkauf Graduate School of Psychology, Yeshiva University, 1300

Morris Park Avenue, Bronx, NY 10461, USA; [email protected] (e-mail).

COPING WITH NEGATIVE EMOTIONS 77

The present study is part of a larger effort to examine how children’s moods (both general and in academic contexts) and their coping–emotional regulation are related to their academic performance and perceived academic stress. Following Oatley, Keltner, and Jenkins (2006), affect includes “phenomena that have anything to do with emotions, moods, dispositions, and preferences” (p. 29). An initial study on these topics (Gumora & Arsenio, 2002) addressed the relations among middle school children’s affective tendencies, cognitive abilities (based on standardized school achievement tests), and school performance (the mean of mathematics and English grades). The primary focus was on examining how children’s emotional states are related to their cognitive and school performance, and whether emotions have any unique influence on school performance. In addition to assessing children’s overall moods, a measure (the Negative Academic Affect Scale [NAAS]; Gumora & Arsenio, 2002) was developed to assess children’s self-reported negative emotions in the context of various academic tasks, ranging from oral and written classroom projects to tests and quizzes.

Results revealed that middle school students who experienced more negative affect during academic tasks and more negative overall moods had lower school grades than their peers. Perhaps most important, even after accounting for cognitive abilities (academic achievement test scores), students’ moods were still related to their grades, and negative academic affect, in particular, was uniquely associated with students’ grades. More specifically, cognitive variables accounted for 42% of the variance in students’ grade point average (GPA), and affective variables accounted for an additional 15% of the variance in GPA, beyond any shared variance that affective variables may have had with cognitive variables. Overall, students with more negative moods and more negative feelings about school-related tasks had lower grades than their peers, and this did not appear to be the result of lower cognitive abilities (i.e., not simply that being cognitively less able = doing poorly in school = feeling bad).

In their discussion of the processes that might underlie these findings, Gumora and Arsenio (2002) argued that children’s abilities to regulate their emotions might play a central role in how children’s affective tendencies influence school performance. For example, in one influential longitudinal study, Eisenberg et al. (1997) found that children who were high in negative emo- tionality but who were also able to regulate those emotions effectively had relatively high levels of adaptive social functioning. In a related vein, Valiente, Lemery-Chalfant, Swanson, and Reiser (2008) recently conducted a series of studies in which they found that effortful control (EC), which they viewed as an index of children’s regulatory abilities, plays a key role in children’s social and academic performance. Specifically, they found that school-aged children who were high in EC (involving voluntary control of attention and behavior) had higher GPAs and showed greater academic improvement over the course of a year than their peers.

Stress, Coping, and Academic Performance

Recently, some researchers (see subsequent citations) have sought to understand these observed connections between children’s regulatory abilities and academic performance by turning to related research on adolescents’ coping styles. One seminal review of stress during adolescence defined coping as “conscious volitional efforts to regulate emotion, cognition . . . and the envi- ronment in response to stressful events or circumstances” (Compas, Connor-Smith, Saltzman, Thomsen, & Wadworth, 2001, p. 89). Moreover, these authors argued that certain characteristic ways of coping become firmly established by adolescence, and that these coping styles have

78 ARSENIO & LORIA

major influence on subsequent adult functioning. Three broad categories of volitional coping were outlined: primary and secondary engagement control coping, and disengagement coping (two other categories deal with involuntary ways of dealing with stress). In brief, primary engage- ment (or active) coping includes problem solving, and emotional regulation–expression (e.g., “I try to think of different ways to change the problem or fix the situation.”), whereas secondary positive coping includes cognitive restructuring, positive thinking, acceptance, and distraction (e.g., “I think about ways to laugh about so it won’t seem so bad.”). Finally, disengagement coping includes attempts to avoid, deny, or use wishful thinking (e.g., “I try to believe it never happened.”).

Initial research on these coping styles (e.g., Compas et al., 2001; Connor-Smith, Compas, Wadsworth, Thomsen, & Saltzman, 2000) found that adolescents with higher levels of both types of engagement coping had generally lower levels of internalizing and externalizing disorders (as assessed by both parent and teachers), whereas adolescents’ with higher levels of disengagement coping scores often had higher internalizing and externalizing disorders. Subsequent studies revealed that adolescents’ coping styles are related to their abilities to manage various forms of pediatric pain (e.g., Compas et al., 2006), as well as more general environmental stressors (Wadsworth, Rieckmann, Benson, & Compas, 2004).

More recently, Valiente, Lemery-Chalfant, and Swanson (2009) found that the coping styles assessed by Compas et al.’s Responses to Stress Questionnaire (RSQ, 2001) are related to chil- dren’s academic competence, as well as other broad aspects of their psychosocial functioning. Specifically, engagement coping (combining primary and secondary control coping) was “linked to children’s adjustment and mediates the relation between parent’s affective responses and chil- dren’s adjustment” (Valiente et al., 2009, p. 174) with disengagement coping also playing an additional role in these connections. Engagement coping was also related to academic compe- tence, although somewhat surprisingly there was no evidence that coping mediated any of the connections between academic competence and the other major study variables.

One possible explanation for the limited connections between coping and academic perfor- mance in the Valiente et al. (2009) study, is that their RSQ measure focused exclusively on peer-related social stressors (e.g., having problems with a friend). The RSQ often focuses on peer-related stress, and peer relations certainly play a significant role in children’s academic performance (e.g., Parker & Asher, 1987). At the same time, however, it may be important to know how adolescents’ cope with specifically academic stressors (e.g., papers, oral reports, lots of school work). For example, does it matter whether an adolescent copes with academic stress by denying or avoiding it (e.g., “I try not think about it”) instead of by sharing feelings (“I let someone else know how I feel”) or trying to address the problem directly (“I’d do something to fix the problem”)? This focus on academic stress is especially relevant for the relatively affluent youth who participated in this study (see Method section). Recent research, for example, has shown that adolescents from affluent families are at especially high risk for substance abuse, depression, and anxiety (e.g., Luthar & Barkin, 2012), and that excessive pressure to achieve aca- demically (Luthar & Latendresse, 2005) plays a major role in these difficulties. To date, however, there is surprisingly little research on academic stress, per se, and its relations with academic performance and other psychosocial variables.

In summary, previous research has shown that students who have higher levels of nega- tive general moods and negative academically related affect have lower levels of academic performance (Gumora & Arsenio, 2002; see also Arsenio, Abdo, & Gumora, 2011). In their

COPING WITH NEGATIVE EMOTIONS 79

discussion of these findings, Gumora and Arsenio proposed that children’s regulatory–coping abilities may have an important influence on how children’s affective tendencies influence school performance. A separate line of research has shown that adolescents’ coping styles, in fact, do affect their ability to manage various forms of stress, including peer-related stres- sors that affect academic performance. To our knowledge, however, little is known about how adolescents’ cope with academic stress, in particular, and whether academic coping styles are related to academic performance and previously observed affective correlates of academic performance.

Consequently, this study was designed to assess several interrelated hypotheses. A first set of hypotheses addressed the associations among adolescents’ general moods, and their aca- demic stress, coping, negative affect, and school performance. It was hypothesized (Hypothe- sis 1) that students’ lower GPAs would be associated with higher negative and lower positive overall moods, higher negative academic affect, higher academic stress, and less use of adap- tive forms of academic coping (i.e., more disengaged coping and less of either form of en- gagement coping). It was also expected (Hypothesis 2) that negative academic affect would make a separate, unique contribution to students’ GPA beyond the contribution of overall moods.

In addition to these correlational connections, it was hypothesized (Hypothesis 3) that ado- lescents’ academic coping styles would influence the connections between negative emotionality (both academic and general) and their GPAs and levels of academic stress. In general, it was expected that higher levels of effective coping (i.e., more primary and secondary engagement and less disengagement coping) would mediate the connections between negative emotionality and students’ GPA and between negative emotionality and academic stress. Although no specific age- related differences were hypothesized, analyses were conducted to examine potential differences in younger (ninth- and 10th-grade students) and older (11th- and 12th-grade students) adoles- cents. Finally, based on previous findings (e.g., Gumora & Arsenio, 2002), no gender differences were expected in the relations among the study measures.

METHOD

Participants

The participants were 119 adolescents (72 girls and 47 boys; M age = 15.40 years, SD = 1.16 years) who attended a high school in a suburban community near a major metropolitan area in the Northeastern United States. The community ranked in the top 2% nationally (2006) in terms of average family income ($217,529) and 80% of the students were European American, 15% Asian American, and 5% either Latino or African American. This particular age and socioeconomic status group was selected because of evidence that adolescents in affluent communities are under particular stress to perform well academically (e.g., Luther & Latendresse, 2005).

Procedure

Only those students with parental consent as well as adolescent assent participated. The assess- ments were administered to students during first half of health or community-building classes.

80 ARSENIO & LORIA

Students were given a brief description of the purpose of the study and were informed that their answers would be completely anonymous because they would not write their names on any of the forms (all of the forms for a single student were connected by a common identification number). Students then completed the package of questionnaires and returned them to the examiner.

Measures

NAAS

The 39-item NAAS (Gumora & Arsenio, 2002) was developed to assess students’ perceptions of their negative affect (anxiety, frustration, and anger) while engaged in a variety of academic tasks, among them homework, class participation, and engagement in class projects. Students rated the frequency of these three emotions for each of 13 specific school tasks using a 5-point Likert-type scale ranging from 1 (never) to 5 (always). Questions focused on students’ abilities to organize and synthesize information, classroom participation, and their test performance and teacher evaluation (with 6 of 39 items in this last category). For example, one item read “I (never, rarely, sometimes, often, always) experience frustration writing essays.” Higher scores indicated that participants experienced more frequent negative emotions while engaged in academic tasks. Items scores were summed to form a total score of negative academic affect.

An original factor analysis examined whether the NAAS was best characterized as a unitary or multidimensional scale. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was .80; the data were appropriate for factor analysis. An inspection of the Scree plot of the eigenvalues revealed that the NAAS was unidimensional; a single eigenvalue with a value greater than one emerged (11.6), and most items loaded .40 or greater on this factor. Internal consistency was quite high (Cronbach’s α = .93). Cronbach’s alpha for the present study was .90.

Positive and Negative Affect Schedule–Child Version

This instrument, an adaptation of the widely used Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988), was developed as an age-appropriate measure of children’s and adolescents’ perceptions of their moods (Laurent et al., 1999). The Positive and Negative Affect Schedule–Child Version (PANAS-C), similar to the adult version of the PANAS, assesses orthogonal dimensions of positive and negative affect: “a two-factor solution best described the structure of the PANAS-C, consistent with its parent measure” (Laurent et al., 1999, p. 334). Across replication studies, Cronbach’s alpha coefficients were .94 and .92 for positive affect and .90 and .89 for negative affect, respectively. Good convergent and discriminant validity were also reported in relation to other measures of children’s and adolescents’ emotion- related functioning (e.g., depressive and anxious symptoms).

Adolescents in this study were presented with a list of 27 emotion terms: 12 positive in valence (e.g., interested, happy, cheerful) and 15 negative (sad, nervous, afraid). Adolescents assessed “the extent [to which] you have felt this way during the past few weeks” using a 5-point Likert- type scale ranging from 1 (very slightly or not at all) to 5 (extremely). The negative scale was the sum of all ratings for negative emotions, whereas positive scale was the sum of all positive emotions.

COPING WITH NEGATIVE EMOTIONS 81

Academic performance

Students’ academic performance was assessed using a self-report measure of overall GPA (mean GPA) across all academic subjects. Letter grades were transformed to a common 5 point- scale (A = 4, B = 3, C = 2, D = 1, and F = 0), and plus and minus grades were assigned relevant fractional values (e.g., B+ = 3.3). This mock report card methodology was developed by Pierce, Hamm, and Vandell (1999) and refined by Valiente, Lemery-Chalfant, and Castro (2007). Students’ self-reported grades have been found to correlate .80 and higher with actual report card grades (Graham, Updegraff, Tomascik, & McHale, 1997).

Academic coping instrument

This instrument was adapted from the original RSQ, which was developed to assess a broad range of adolescents’ possible responses to stress. Development of the RSQ was guided by a seminal literature review on coping in adolescents (e.g., Compas et al., 2001) and several confirmatory studies have demonstrated the connections of the RSQ with a variety of measures of adolescents’ psychosocial functioning (Connor-Smith et al., 2000). The full instrument includes 5 scales (57 items) that assess both adolescents’ voluntary coping and involuntary engagement, with the latter including items such as “When I have problems with other kids, I feel it in my body.”

The present study only included the three scales that assess more voluntary, controlled forms of coping as it was felt that these scales assessed types of coping that were subject to active volitional efforts on the part of the adolescent. This focus on voluntary control, as well changes in wording to focus on academic coping, per se, rather than just social coping, were acknowledged as acceptable by the instrument creator (B. Compas, personal communication, February, 2006).

Adolescents were presented with the following initial statement:

Even when things are going well for teenagers, almost everyone has tough times with some parts of their schoolwork. We want to find out how things have been going for you lately. Please put a check mark next to all the things on this list that have been a problem or have created stress for you in school in the last year or so.

Students could then check off as many items as they wanted from a list that included 10 items, such as talking to teachers, doing oral reports in class, or assignments not under- stood. The total number of boxes checked (from 0 to 10) was used as a measure of adolescents’ perceived academic stress.

Following this task, adolescents were presented with several pages listing what people some- times do, feel, or think when something stressful with school work happens. Adolescents then judged how likely they were to engage in 30 possible coping strategies (e.g., “I try to think of different ways to change the problem or fix the situation.” “I try not to think about it, I just try to forget it.”) using a 4-point Likert-type scale ranging from 1 (not at all) to 4 (a lot). Participants’ responses were categorized into three larger categories: primary control engagement coping (e.g., problem solving and emotional regulation), secondary control engagement coping (e.g., cognitive restructuring and positive thinking) and disengagement coping (e.g., denial and distraction). Each category score was the sum of responses for the items in that category. Previous studies (Com- pas et al., 2000) indicated that primary and secondary control engagement coping are related to

82 ARSENIO & LORIA

TABLE 1 Descriptive Data for All Measures

Measure M SD Possible range

Cumulative GPA 3.44 0.38 0–4.3 Negative academic affect 89.78 24.45 39–195 PANAS positive 41.73 8.23 12–60 PANAS negative 33.93 9.16 15–75 Primary positive coping 23.58 4.78 9–36 Secondary positive coping 30.04 6.40 12–48 Disengaged coping 18.48 4.31 9–36 Total academic stressors 5.52 1.94 0–10

Note. GPA = grade point average; PANAS = Positive and Negative Affect Schedule.

psychologically adaptive functioning, whereas disengagement coping is related to maladaptive functioning.

RESULTS

Preliminary Analyses Involving Age & Gender

Preliminary analyses revealed that there was no pattern of significant differences in the relations among the main study variables (see Table 1) as a function of either gender nor grade level (i.e., no patterns of moderation). Consequently, neither gender nor grade level are mentioned further.

Connections Among Study Variables (Table 2)

Mood-related variables

Adolescents with more positive general moods (PANAS positive) reported lower levels of negative general moods (PANAS negative, r = –.28, p < .05). Higher levels of negative academic

TABLE 2 Correlations Among Variables (n = 119)

1 2 3 4 5 6 7 8

1. GPA — −.21∗ −.07 −.03 .08 −.07 −.33∗∗∗ .05 2. Negative academic affect — −.24∗∗ .43∗∗∗ .20∗ .09 .61∗∗∗ .59∗∗∗ 3. PANAS positive — −.28∗∗ .18∗ .38∗∗∗ −.21∗ −.16∗ 4. PANAS negative — −.07 −.11 .40∗ .29∗∗∗ 5. Primary positive coping — .37∗∗∗ .03 .14 6. Secondary positive coping — −.09 −.02 7. Disengaged coping — .42∗∗ 8. Academic stress —

Note. GPA = grade point average; PANAS = Positive and Negative Affect Schedule. ∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

COPING WITH NEGATIVE EMOTIONS 83

affect were associated with more negative general moods (r = .43, p < .001) and with less positive general moods (r = –.24, p < .01).

Coping variables

Adolescents with higher levels of primary positive coping also had higher levels of secondary positive coping (r = .37, p < .001). Disengaged coping, however, was not related to either primary positive (r = .14, p = ns) or secondary positive coping (r = –.02, p = ns).

GPA and other variables

Adolescents’ GPA was associated with two of the seven other study variables. As expected (Hypothesis 1) participants with higher GPAs experienced less negative academic affect (r = –.21, p < .05) and were less likely to engage in disengaged coping (r = –.33, p < .001). Contrary to expectations, however, adolescents’ GPAs were not related to their positive general (r = –.07, p = ns) or negative general moods (r = –.03, p = ns), their primary positive (r = .08, p = ns) or primary secondary coping (r = –.07, p = ns) or to their academic stress (r = .05, p = ns).

Academic stress and other variables

Although adolescents’ academic stress was not related to their GPAs, academic stress was strongly related to more negative academic affect (r = .59, p < .001). In addition, adolescents with higher levels of academic stress had more negative general moods (r = .29, p < .001), less positive general moods (r = –.16, p < .05), and were more likely to use disengaged forms of academic coping (r = .42, p < .001). Supplemental regression analyses revealed that all of the significant associations involving academic stress and other variables were linear only (i.e., there were no curvilinear relations, which might indicate that moderate levels of stress were optimal or less problematic).

In summary, 16 of 28 correlations among the study variables were significant, but these patterns varied depending on the specific types of variables being assessed. It should also be noted that adolescents with higher levels of negative academic affect were substantially more likely to use disengaged forms of coping affect (r = .61, p < .001). Finally, additional partial correlations revealed that even after controlling for general moods (PANAS positive and negative), greater negative academic affect was, as expected (Hypothesis 2) related to students’ lower GPA (partial r = .23, p < .01), as well as to higher academic stress (partial r = .54, p < .001) and greater use of disengaged coping (partial r = .52, p < .001; see subsequent regressions).

Regression analysis predicting GPA

A regression analysis was conducted to provide a more detailed picture of the relative and combined influences of students’ general moods, and academic stress, affect, and coping strate- gies, on their GPAs. Variables were entered in four steps, with the first block including general moods (positive and negative), followed by a second block that included academic stress, a

84 ARSENIO & LORIA

TABLE 3 Hierarchical Regression Predicting Adolescents’ Grade Point Average (GPA)

Step 1 Step 2 Step 3 Step 4

b t b t b t b t

Step 1 (R2 = .007, Fchange = 0.39) PANAS positive −.08 −0.82 −.08 −0.77 −.11 −1.12 −.16 −1.61 PANAS negative −.05 −0.54 −.07 −0.66 .04 0.38 .12 1.21

Step 2 (R2 = .009, Fchange = 0.32) Academic stress .06 .57 .16 1.59 .18 1.65

Step 3 (R2 = .092, Fchange = 11.74∗∗∗) Negative academic affect −.40 −3.43∗∗∗ −.26 −2.06∗

Step 4 (R2 = .101, Fchange = 4.71∗∗) Primary positive coping .13 1.32 Secondary positive coping .02 .23 Disengaged coping −.37 −3.32∗∗∗

Note. For total model, F(6, 112) = 3.37, p < .05. R2 and Fchange are shown for each step, β and t are for each predictor at that step. PANAS = Positive and Negative Affect Schedule. ∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

third block that included negative academic affect, and a fourth block that included the three types of academic coping strategies. General moods were entered first because these were seen as reflecting broad aspects of adolescents’ affective dispositions, and academic stress followed because it was expected to exert a pervasive influence on academic performance (e.g., Luthar & Latendresse, 2005). Negative academic affect and coping strategies were entered third and fourth because they were seen as reflecting more specific responses to the broader challenges resulting from academic stress and general moods.

As can be seen in Table 3, adolescents’ general moods were not a significant predictor of GPA when entered on the first step, and neither was academic stress when it was entered on the second step. However, when negative academic affect was added (step 3) it predicted significantly lower GPAs, (accounting for about 9% of GPA variance). Finally, when adoles- cents’ coping strategies were entered on the fourth step, they made an additional significant contribution to predicting GPA (about 10% of the variance), over and above the influence of negative academic affect. Overall, the total model was significant, with the four blocks predicting about 20% of the variance in adolescents’ GPA. However, looking at the individ- ual contributions of each variable on the final step (i.e., considering all variables simultane- ously), only negative academic affect and disengaged coping were unique significant predictors of GPA after accounting for the influence of other variables. In summary, adolescents’ neg- ative academic affect and use of disengaged coping, in particular, were related to having a lower GPA.

Regression analysis predicting academic stress

Contrary to expectations (part of Hypothesis 1), academic stress was not related to adolescents’ GPA. Yet, academic stress, per se, is known to be associated with problematic internalizing and externalizing difficulties in affluent youth, such as those who participated in this study (Luthar

COPING WITH NEGATIVE EMOTIONS 85

TABLE 4 Hierarchical Regression Predicting Adolescents’ Academic Stress

Step 1 Step 2 Step 3

β t β t β t

Step 1 (R2 = .08, Fchange = 4.73∗∗) PANAS positive −.07 −0.70 −.01 −0.12 −.01 −0.04 PANAS negative .26 2.70∗∗ .05 0.47 .05 0.54

Step 2 (R2 = .26, Fchange = 44.49∗∗∗) Negative academic affect .56 6.67∗∗∗ .54 5.99∗∗∗

Step 3 (R2 = .03, Fchange = 1.75) Primary positive coping .19 2.16∗ Secondary positive coping −.10 −1.10 Disengaged coping −.03 −0.34

Note. For total model, F(6, 112) = 10.66, p < .001. R2 and Fchange are shown for each step, β and t are for each predictor at that step. PANAS = Positive and Negative Affect Schedule. ∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

& Latendresse, 2005; see also Luthar & Barkin, 2012). Consequently, a decision was made to conduct a second regression analysis to provide more information on the potential affective and coping contributors to adolescents’ academic stress. Variables were entered in three steps: block 1 included general moods (positive and negative), block 2 included negative academic affect, and block 3 included the three types of academic coping strategies.

As can be seen in Table 4, for academic stress, (step 1) adolescents’ general moods (espe- cially negative mood), and, in addition (step 2), their negative academic affect added signifi- cantly to predicting academic stress. Finally, academic coping (step 3) did not add to the total model. Overall, the total model was significant, with the three blocks predicting 37% of the variance in adolescents’ academic stress. An examination of the unique contributions of the individual variables at the final step, however, provided a somewhat different picture. Specifi- cally, higher negative academic affect was uniquely related to higher levels of academic stress, whereas higher levels of primary academic coping were uniquely related to lower academic stress.

Negative emotionality, coping, and GPA: mediation analysis (hypothesis 3)

This analysis addressed whether the connections between students’ negative emotionality (negative academic affect or negative general moods, respectively) and their GPA was mediated by students’ use of certain coping strategies. Only disengaged coping, negative academic affect, and GPA met the minimal criteria for conducting this analysis (Baron & Kenny, 1986), that is, the presence of significant connections among all 3 relevant variables.

The resulting regression analysis confirmed that negative academic affect was significantly related to adolescents’ GPA (Table 5, step 1). However, when both negative academic affect and disengaged coping were simultaneously assessed (step 2), the contribution of negative academic affect was nearly reduced to zero. In fact, 98% of the influence of negative academic affect on students’ GPA was mediated by their disengaged coping, indicating near complete mediation.

86 ARSENIO & LORIA

TABLE 5 Summary of Analyses of the Mediational Role of Children’s Disengaged Coping in the Prediction of

Adolescents’ GPA by Negative Academically Related Affect

Step Effects in model sr2 t

1 Negative academic affect .044 −2.31∗ 2 Negative academic affect .001 −0.14 2 Disengaged doping .07 −2.85∗∗

Note. sr2 (squared semipartialed correlation coefficient) is the percentage of unique variance that is shared by two variables (controlling for any other variable entered in that regression step). The first step in the model indicates the relation of negative academic affect to grade point average (GPA), and the second step indicates the independent contributions of the negative academic affect and disengaged coping when both are simultaneously regressed on adolescents’ GPA. ∗p < .05. ∗∗p < .01.

Negative emotionality, coping, and academic stress: mediation analysis (hypothesis 3)

The final analyses assessed whether the connections between students’ negative moods (neg- ative academic affect and negative general moods, respectively) and their academic stress was mediated by students’ use of certain coping strategies. Two sets of variables met the Baron and Kenny criteria for conducting mediation analysis: (a) negative academic affect, academic stress, and disengaged coping and (b) general negative moods, academic stress, and disengaged coping. No tables are included for these two analyses; they follow the same basic organization as depicted in Table 5.

Regression analyses indicated that disengaged coping did not mediate the significant con- nections between negative academic affect and academic stress: negative academic affect was a significant predictor of academic stress at step 1, t(1, 116) = 7.97, p < .001, and when negative academic affect and disengaged coping were simultaneously entered at step 2, negative academic affect remained a significant predictor of academic stress, t(1, 115) = 5.63, p < .001, and disen- gaged coping did not significantly predict academic stress, t(1, 115) = 0.99, p = ns. By contrast, a separate analysis revealed that disengaged coping did mediate the connection between general negative moods and academic stress. At step 1 of this regression, negative general mood on its own was a significant predictor of academic stress, t(1, 116) = 3.22, p < .01, but when both negative general mood and disengaged coping were entered together at step 2, negative general mood was no longer a significant predictor of academic stress, t(1, 115) = 1.53, p = ns, whereas disengaged coping was a significant predictor of academic stress, t(1, 115) = 3.96, p < .001.

DISCUSSION

The present findings underscore the importance of examining the socioemotional dimensions of students’ school-related experiences (Pianta, 1999; Valiente et al., 2008) and how adolescents cope with or regulate these affectively charged experiences (Valiente et al., 2009). As expected: (Hypothesis 1), there were significant connections among adolescents’ emotional dispositions, academic stress, GPA, and their academic coping styles; and (Hypothesis 2) negative academic

COPING WITH NEGATIVE EMOTIONS 87

affect made a unique contribution to adolescents’ GPA beyond any contribution of overall moods. Moreover, these results are among the first to demonstrate that the techniques adolescents use to cope with uniquely academic (rather than peer) stress are related to their affective dispositions both in and out of the classroom, as well as to their academic performance and level of academic stress. Adolescents’ use of disengaged academic coping (e.g., avoidance and denial), in particular, emerged as an especially problematic way of regulating negative emotions (Hypothesis 3). These broad findings, however, differed depending on the context in which adolescents’ emotional dispositions were assessed (i.e., academic vs. nonacademic) and whether the focus was on academic performance (GPA) or academic stress.

Using the same affective measures included in this study, Gumora and Arsenio (2002) found that middle school children with more general negative moods, somewhat more positive general moods, and more negative academic affect had lower GPAs than their classmates. By contrast, we found that only adolescents’ higher negative academic affect was related to lower GPAs: general moods had no connections with academic performance. Yet higher levels of adolescents’ academic stress (not assessed in Gumora & Arsenio, 2002), were related to both more negative academic affect, as well as higher negative and lower positive general moods. Surprisingly, however, adolescents’ GPA was not related to their level of academic stress (even in a curvilinear relationship where moderate levels of stress might be related to higher GPAs).

The differences in GPA-related findings in Gumora and Arsenio (2002) and the present study may stem from the different age groups in the two studies, that is, middle school versus high school participants. In comparison with middle school students, older adolescents (as in this study) may be better able to regulate their general moods (Thompson & Meyer, 2007), so that general moods are no longer related to academic performance. By contrast, negative feelings about school-related tasks may require complex regulatory abilities that continue to develop into the college years. Consequently, the connection between negative academic affect and school performance could extend from middle school through high school.

Unfortunately, explanations for the present findings involving academic stress are complicated by the lack of directly relevant research (e.g., Gumora and Arsenio [2002] did not assess stress). Yet, as noted previously, findings for high school students’ academic stress resembled those for middle school students’ GPA in that both were strongly related to negative academic affect, and to more negative and less positive general moods. Again, age-related improvements in older adolescents’ emotion regulation abilities might result in a connection between general moods and academic stress without those influences then also affecting GPA. With age adolescents may get better at controlling and containing the effects of general moods and academic stress so that they do not spill over and influence actual academic performance (which would also explain why GPA and academic stress were unrelated). Ideally, it would have been desirable to use path analyses or structural equation modeling to examine the overall connections among the study variables more fully. The relatively limited number of study participants, however, precluded these analyses (Klem, 1997), leaving some important questions about the broad connections among adolescents’ academic stress, school performance, and affective tendencies for future research.

Perhaps the most striking findings from this study were those involving adolescents’ academic coping styles. Adolescents who used more disengaged academic coping (i.e., denial and avoidance of academic problems) not only had lower GPAs but also reported higher levels of academic stress (although academic stress and GPA were unrelated). Furthermore, adolescents’ use of

88 ARSENIO & LORIA

disengaged academic coping fully mediated the connection between higher negative academic affect and lower GPAs; that is, higher negative academic affect was related to more disengaged coping, and, in turn, it was greater disengaged academic coping which lead to lower academic performance (GPA). Additionally, adolescents’ disengaged coping significantly mediated the connection between negative general moods and higher levels of negative academic affect. In both cases then, negative moods increased disengaged coping and disengaged coping influenced school-related measures, although only negative academic affect was involved in students’ GPAs and only general negative moods were involved in academic stress.

These findings are consistent with an extensive literature linking adolescents’ voluntary at- tempts to use more engaged and less disengaged forms of coping with a range of positive psychosocial outcomes, from lower externalizing and internalizing problems (Compas et al., 2001) to reduced levels of pediatric pain (Compas et al., 2006). More specifically, our results may help to clarify Valiente et al.’s (2009) recent efforts to examine children’s RSQ-assessed coping styles in relation to parental affect and children’s academic performance competence. Although Valiente et al. found that coping was related to academic competence, coping did not mediate any of the connections between children’s academic competence and any other study variable (coping did, however, mediate other connections not involving academic competence). By con- trast, adolescents’ coping style in the present study mediated connections between adolescents’ affective states and both their academic performance.

The most likely explanation for these study differences is that Valiente et al. (2009) focused on how children cope with peer stressors, whereas our explicit focus on academic coping pro- vided a somewhat clearer picture of how adolescents cope with academic stressors. Other study differences, however, such as the exclusive importance of disengaged coping in this research, are much harder to understand. Although it seems reasonable that negative emotional dispositions might increase a somewhat hopeless, disengaged way of coping with academic stress, it is un- clear why we found no evidence linking active, engaged forms of coping with positive academic competence. This pattern was especially unexpected given the key role that engaged forms of RSQ-related coping have played in most studies in this literature.

It is also important to acknowledge several study limitations, beginning with whether the present findings would generalize to high school students in less affluent communities, where the stress to excel academically is likely to be less salient (Luthar & Latendresse, 2005). Without additional research it is difficult to know whether the same patterns of findings (especially involving disengaged coping) are typical of most high school students’ experiences.

Other concerns include the exclusive use of self-report measures and difficulties arising from the correlational design. Future researchers would benefit from the inclusion of additional affec- tive and coping information from the perspective of adolescents’ teachers, peers, and parents. Additionally, standardized achievement tests and IQ data would be useful for beginning to ad- dress questions about the direction of effects in the present findings. For example, most of this discussion is guided by the notion that adolescents’ moods and coping styles influence their academic performance, when it also possible that relative academic performance (or underlying cognitive abilities) affect adolescents’ moods and choice of coping styles. Gumora and Arsenio (2002) attempted to address this concern by controlling for middle school students’ academic achievement (a proxy for cognitive ability) prior to examining the links between their GPA and moods. Ideally, however, initial cognitive ability can be controlled in a longitudinal design, as when Hamre and Pianta (2001) found that even after controlling for cognitive abilities, children

COPING WITH NEGATIVE EMOTIONS 89

who had negative relationships with their kindergarten teachers had lower grades and work habits in Grade 8. A related longitudinal approach could clarify whether adolescents’ moods and coping abilities (rather than relationships with teachers) have a continuing and cumulative influence on academic performance over and above initial cognitive abilities.

Overall, however, the present findings underscore the importance of continuing to examine the various emotional contexts of children’s and adolescents’ academic experiences and how they relate to school functioning (Schutz & Pekrun, 2007). By the time adolescents reach high school they have accumulated almost a decade of affective experiences and reactions regarding their academic abilities and performance. The ways in which they understand, interpret, and learn to cope with those affective experiences are likely to influence both their expectations about how they will feel about additional schooling and, potentially, about learning in general.

AUTHOR NOTES

William F. Arsenio is a professor of psychology in the clinical program at the Ferkauf Graduate School of Psychology, Yeshiva University. His research interests focus on the role of affective influences on children’s and adolescents’ academic performance and peer relations. Samantha Loria is a recent graduate of the school-clinical child psychology program at the Ferkauf Graduate School of Psychology, Yeshiva University.

REFERENCES

Arsenio, W., Abdo, N., & Gumora, G. (2011, March). Adolescents’ academic performance and stress: Relations with Eqi, mindfulness, and academic coping styles. Poster presented at the Biennial meeting of the Society for Research in Child Development, Montreal, Quebec, Canada.

Arsenio, W., Cooperman, S., & Lover, A. (2000). Affective predictors of preschoolers’ aggression and peer acceptance: Direct and indirect effects. Developmental Psychology, 36, 438–448.

Baron, R., & Kenny, D. (1986). The moderator-mediator distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182,

Compas, B., Boyer, M., Stanger, C., Colletti, R., Thomsen, A., Dufton, L., & Cole, D. (2006). Latent variable analysis of coping, anxiety/depression, and somatic symptoms in adolescents with chronic pain. Journal of Consulting and Clinical Psychology, 74, 1132–1142.

Compas, B., Connor-Smith, J., Saltzman, H., Thomsen, A., & Wadsworth, M. (2001). Coping with stress during childhood and adolescence; Problems, progress, and potential in theory and research. Psychological Bulletin, 127, 87–127.

Conner-Smith, J., Compas, B., Wadsworth, M., Thomsen, A., & Saltzman, H. (2000). Responses to stress in adoles- cence: Measurement of coping and involuntary stress responses. Journal of Consulting and Clinical Psychology, 6, 976–992.

Denham, S. (1998). Emotional development in young children. New York, NY: Guilford. Eisenberg, N., Fabes, R., Shepard, S., Murphy, B., Guthrie, I., Jones, S., . . . Maszk, P. (1997). Contemporaneous and

longitudinal prediction of children’s social functioning from regulation and emotionality. Child Development, 68, 642–664.

Gumora, G., & Arsenio, W. F. (2002). Emotionality, emotion regulation, and school performance in middle school children. Journal of School Psychology, 40, 395–413.

Hamre, B., & Pianta, R. (2001). Early teacher-child relationships and the trajectory of children’s school outcomes through 8th grade. Child Development, 72, 625–638.

Klem, L. (1997). Path analysis. In L. Grimm & P. Yarnold (Eds.), Reading and understanding multivariate statistics (pp. 65–94), Washington DC: American Psychological Association.

90 ARSENIO & LORIA

Ladd, G., Birch, S., & Buhs, E. (1999). Children’s social and scholastic lives in kindergarten: Related spheres of influence? Child Development, 70, 1373–1400.

Laurent, J., Catanzaro, S., Rudolph, K., Joiner, K., Potter, K., Lambert, S., . . . Gathright, T. (1999). A measure of positive and negative affect for children: Scale development and preliminary validation. Psychological Assessment, 11, 326–338.

Luthar, S., & Barkin, S. (2012). Are affluent youth truly “at risk”? Vulnerability and resilience across three diverse samples. Development and Psychopathology, 24, 429–449.

Luthar, S., & Latendresse, S. (2005). Children of the affluent: Challenges to well-being. Current Directions in Psycho- logical Science, 14, 49–53.

Oatley, K., Keltner, D., & Jenkins, J. (2006) Understanding emotions (2nd ed.). Oxford, England: Basil Blackwell. Parker, J., & Asher, S. (1987). Peer relations and later personal adjustment: Are low-accepted children at risk? Psycho-

logical Bulletin, 102, 357–389. Pianta, R. (1999). Enhancing relationships between children and teachers. Washington, DC: American Psychological

Association. Pierce, K., Hamm, J., & Vandell, D. (1999). Experiences in after-school programs and children’s adjustment in first-grade

classrooms. Child Development, 70, 756–767. Saarni, C., Campos, J., Camras, L., & Witherington, D. (2006) Emotional development: action, communication, and

understanding. In W. Damon & R. Lerner (Ser. Eds.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology (6th ed.), vol. 3: Social, emotional, and personality development (pp. 226–299). New York, NY: Wiley.

Schutz, P. A., & Pekrun, R. (2007). Emotions in education. New York, NY: Wiley. Thompson, R., & Meyer, S. (2007). Socialization of emotion regulation in the family. In J. Gross (Ed.), Handbook of

emotion regulation (pp. 249–268). New York, NY: Guilford. Trentacosta, C., & Izard, C. (2007). Kindergarten children’s emotion competence as a predictor of their academic

competence in first grade. Emotion, 7, 77–88. Valiente, C., & Eisenberg, N. (2006). Parenting and children’s adjustment: The role of children’s emotion regulation.

In D. Snyder, J. Simpson, & J. Hughes (Eds.), Emotion regulation in families: Pathways to dysfunction and health (pp. 123–142). Washington DC: American Psychological Association.

Valiente, C., Lemery-Chalfant, K., & Castro, K (2007). Children’s effortful control and academic competence. Merrill- Palmer Quarterly, 53, 1–25.

Valiente, C., Lemery-Chalfant, K., & Swanson, J. (2009). Children’s response to daily stressors: Relations with parenting, effortful control, and adjustment. Journal of Child Psychology and Psychiatry, 50, 707–717.

Valiente, C., Lemery-Chalfant, K., Swanson, J., & Reiser, M. (2008). Prediction of children’s academic competence from their effortful control, relationships, and classroom participation. Journal of Educational Psychology, 100, 67–77.

Wadsworth, M., Rieckmann, T., Benson, M., & Compas, B. (2004). Coping and response to stress in Navajo adolescents: Psychometric properties of the response to stress questionnaire. Journal of Community Psychology, 32, 392–411.

Watson, D., Clark, L., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070.

Wentzel, K. (2005). Peer relationships, motivation, and academic performance at school. In A. Elliot & C. Dweck (Eds.), Handbook of competence and motivation (pp. 279–296). New York, NY: Guilford.

Copyright of Journal of Genetic Psychology is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.