- Email: [email protected]
Emotionality, Emotion Regulation, and School Performance in Middle School Children
Journal of School Psychology, Vol. 40, No. 5, pp. 395 – 413, 2002 Copyright D 2002 Society for the Study of School Psychology Printed in the USA 0022-4405/02 $ – see front matter
PII S0022-4405(02)00108-5
Emotionality, Emotion Regulation, and School Performance in Middle School Children Gail Gumora and William F. Arsenio Ferkauf Graduate School of Psychology, Yeshiva University, Bronx, NY, USA This research investigated the connections of middle school students’ emotional dispositions and academic-related affect with their school performance. One hundred three 6th – 8th grade students completed three self-rated assessments regarding: (a) their academic competency; (b) affective tendencies (both mood and emotion regulation); and (c) negative emotions experienced during school-related tasks. Teachers assessed students’ positive and negative moods, and schools provided achievement test results and student grades as measures of cognitive ability/ achievement and school performance, respectively. Results indicated that although students’ emotion regulation, general affective dispositions, and academic affect were related to each other, each of these variables also made a unique significant contribution to students’ GPA, over and above the influence of other cognitive contributors. Overall, these results provide support for the role of socio-emotional factors in students’ school performance, while also clarifying some of the uniquely affective contributors (rather than relationships or goals) to that performance. D 2002 Society for the Study of School Psychology. Published by Elsevier Science Ltd Keywords: Emotional dispositions, Emotion regulation, School performance.
There is a growing awareness that social and emotional factors play an important role in students’ academic success. Research has shown that students who have more prosocial, socially responsible goals (Wentzel, 1996, 1999) or more positive relationships with adults (teachers and parents; Birch & Ladd, 1997; Pianta, 1999) than their peers subsequently perform better in academic contexts. For example, Hamre and Pianta (2001) recently found that, controlling for cognitive ability, children who had negative conflicted relationships with their kindergarten teachers had lower grades and worse work habits in 8th grade. In one model of the processes underlying these connections (Ford, 1992), academic motivation is explained ‘‘in terms of three inter-related
Received 12 December 2001; received in revised form 26 March 2002; accepted 29 May 2002. Address correspondence and reprint requests to Gail Gumora, 115 W. 86th Street, Apartment 16E, New York, NY 10024, USA; or William F. Arsenio, Ferkauf Graduate School of Psychology, Yeshiva University, 1300 Morris Park Avenue, Bronx, NY 10461, USA. E-mail addresses: [email protected], [email protected]
395
396
Journal of School Psychology
processes: personal goals, personal agency beliefs, and emotional arousal. . .. According to this model, an individual will pursue a goal to the extent emotions and personal agency beliefs are associated with the goal and will support efforts to achieve one’s goals’’ (Wentzel, 1999, p. 78). To date, however, much of the work on socio-emotional contributors to academic performance has focused on personal goals (‘‘do I want to do well in school?’’), agency beliefs (‘‘will I be able to if I try?), and/or student– teacher relationships (‘‘how do I feel about my teacher?’’). In contrast, much less is known about how students’ emotions, per se, support academic performance (‘‘how do I feel about schoolwork?).1 The research on the affective contributors to achievement that does exist has largely been confined to the role of test anxiety on test performance (e.g., Musch & Broeder, 1999; and see McDonald, 2001 for a review of research involving children). In general, it appears that worry interferes with the test performance of highly anxious students because it distracts attention from the task at hand. Vail (1981) recognized that for some students, however, the anxiety and frustrations associated with more everyday school tasks could also be problematic. Some children and adolescents may have difficulties with their emotional arousal during routine homework and classroom activities that have little to do with test anxiety, per se. Consequently, the present study was designed to examine the connections between young adolescents’ academically related emotions and their academic performance. It was expected that, overall, students who report more negative emotions during academic tasks would have lower levels of school performance (as assessed by their grade point averages). In addition to examining the connection between negative academic affect and school performance, two other issues were addressed in this research. One is whether academically related affect is just a reflection of students’ more general affective dispositions, and consequently, whether it is necessary to distinguish academic affect from more general affective contributors to school performance (see below). A second issue is whether academic affect is mostly a reflection of students’ underlying cognitive abilities and their perceived competence regarding their abilities. Although Ford’s (1992) model makes a theoretical distinction between personal
1
Although there is extensive research on how broad aspects of students’ affective and behavioral tendencies are related to their school performance (e.g., aggression, peer rejection, and school failure), most of that work does not differentiate specifically affective components of behavior from other socio-emotional components such as peer and adult relationships. By contrast, the literature on social competence has shown that the connections between children’s and adolescents’ emotion dispositions and abilities and their social behaviors are often more complex, and in need of empirical verification, than is typically assumed (e.g., Arsenio, Cooperman, & Lover, 2000; Denham, 1999).
Gumora and Arsenio
397
agency beliefs (i.e., perceived academic competence) and emotional arousal, connections between school performance and academic affect could largely be a ‘‘by product’’ of students’ cognitive abilities and academic competence. In other words, young adolescents who are cognitively able and who believe they are good at school tasks may perform well in school, and consequently, feel more positively (and less negatively) about academic tasks. In terms of the first issue (general affective dispositions vs. academic affect), although there is little work on affective dispositions and academic performance, an extensive literature has emerged on the connections between children’s emotionality and their social competence. Eisenberg et al. (1993) and Fabes and Eisenberg (1992) for example, have conducted an influential series of studies showing that children’s general negative emotionality is related to their subsequent peer behaviors, and to both peer and adult ratings of their social competence. Moreover, other studies indicate that positive emotionality is associated with lower levels of socially incompetent behaviors and higher levels of peer acceptance in children (Cole, Michel, & Teti, 1994). Collectively, this research both demonstrates that it is possible to assess broad aspects of children’s emotionality and that these emotional tendencies are related to children’s social functioning. Other studies, however, suggest that children’s emotionality is at least partly context specific. For example, Arsenio et al. (2000) distinguished between children’s emotionality during aggression and more general aspects of emotionality. Negative emotionality (especially anger) during aggression was modestly but significantly related to children’s more general angry negativity, but only aggression-related negativity and not general negativity was related to whether children were liked by peers (see also Arsenio & Killen, 1996). For the present purposes, these studies suggest that it may be important to distinguish between more general affectivity and specific affective contexts of interest (i.e., academic affect). Some students, for example, may have generally positive affective dispositions, and yet still experience during schoolwork negative emotions that do not simply stem from underlying cognitive difficulties. Finally, the connections between students’ emotionality and social competence depend, in part, on their abilities to regulate and control negative emotionality (Eisenberg & Fabes, 1992). For example, in a longitudinal study (Eisenberg et al., 1997) it was found that children who were high in negative emotionality but also high in emotion regulation were rated as functioning well socially. At the same time, however, ‘‘regulation and emotionality frequently contributed unique variance to the prediction of social functioning’’ (p. 642), suggesting that the interactive contribution of emotionality and regulation to social competence was in addition to separable contributions of emotionality and emotion regulation.
398
Journal of School Psychology
Guided by these studies on children’s emotion and social competence, we decided to assess the contributions of young adolescents’ emotion regulation, as well as their general emotionality and academic affect, to their school performance. Although emotion regulation has been assessed in several ways, one increasingly central component is effortful control which includes ‘‘individual differences in the ability to voluntarily sustain focus on a task’’ (Ahadi & Rothbart, 1994, p. 196; and see also Corno & Rohrkemper, 1985 for evidence that older students’ abilities to manage and control their effort is linked with academic performance). This ability to focus on tasks seemed especially relevant for academic contexts, and consequently emotion regulation was assessed in the present study using a measure of task orientation. Extrapolating from the emotions and social competence literature, we expected both that there would be significant connections among general emotionality, academic affect, and emotion regulation (Hypothesis 1A), and that each of these three constructs would also be related to children’s academic performance (Hypothesis 1B). In addition, however, we expected that academic affect would make a separate, unique contribution to students’ school performance, over and above the contribution of general emotionality or emotion regulation (Hypothesis 2). In other words, academic affect was not expected to be just a subset of more general emotionality, or the result of emotion regulation abilities, and consequently, it would make its own contribution to school performance. Finally, based on research indicating that emotionality and emotion regulation may have an additional interactive influence on children’s functioning (i.e., in addition to main effects for emotionality and regulation), it was also expected (Hypothesis 3) that students’ academic emotionality would interact with their emotion regulation abilities to affect school performance. Specifically, high negative academic affect would have a negative relation to academic performance unless these young adolescents could compensate for that negative emotionality with high levels of academic emotion regulation.2 As noted above, another issue is whether affective contributors add anything unique to the prediction of students’ school achievement and performance beyond common cognitive contributors. Although there are often complex bi-directional connections between affect and cognition (e.g., Lazarus, 1984), recent research also indicates that socio-emotional factors are significant predictors of children’s and adolescents’ school
2 Previous studies on emotion regulation and emotionality did not address the role of the context in which emotionality was assessed, but it seems reasonable to expect that children’s academic affect would be especially influenced by their emotion regulation abilities. For example, children who often experience negative emotions in the context of academic tasks might be more likely to become distracted and do poorly in their academic work unless this is compensated for by highs of high levels of emotion regulation level to help keep them oriented to academic tasks.
Gumora and Arsenio
399
motivation and performance even after controlling for cognitive factors (Hamre & Pianta, 2001; Wentzel, 1999). Consequently, it was expected that although there might be significant connections among cognitive and affective variables, affective contributors (emotion regulation, general emotionality, and academic affect) would make an additional, unique contribution to students’ school performance over and above the contributions made by cognitive variables (Hypothesis 4). Several cognitively oriented variables were included in the present study. Although it would have been desirable to have a measure of IQ, pragmatic limitations made it necessary to use available standardized measures of academic achievement as a rough indicator of general cognitive abilities (see Methods). In addition, a measure of students’ perceived academic competence was included. As Wentzel (1999) has described, students’ perceptions of their academic competence influence their academic motivation and, in turn, their school performance. Consequently, connections between school performance and academic affect might primarily be a result of students’ perceived academic competence: students who perceive that they are good at school tasks may perform well in school, and then feel more positively (and less negatively) about academic tasks as a result. Although no correlational study can assess causal connections, controlling for the influence of academic achievement and perceived academic competence would begin to address the potentially unique contributions of affective influences on young adolescents’ school success. In summary, this study focused on whether middle school students’ affective dispositions and abilities (i.e., emotionality and emotion regulation) are related to their school performance, even after accounting for more cognitive contributions to that performance. Although connections were expected between affective and cognitive measures, the underlying goal was to examine whether, as has sometimes been argued (Mayer & Salovey, 1997), affective dispositions and tendencies also have a unique influence on students’ school success. For the reasons discussed below, a decision was made to focus on young adolescents in middle school, and no age-related differences were expected in this narrow age range. In terms of gender, some research with younger children has shown differences in emotionality and emotion regulations in boys and girls, but these differences are inconsistent across studies. Consequently, no specific gender differences were hypothesized.
METHODS Participants The 103 middle school (6th to 8th grade) young adolescents who participated in this study (51 girls and 52 boys) attended one of two Jewish
400
Journal of School Psychology
affiliated day schools in a major Northeastern city. This particular age group (11 –14 years, M = 12.1 years) was selected because the emergence of formal operational reasoning in children of this age has been linked with an increasingly abstract ability to understand and reflect on previous affective experiences (Piaget, 1981). More specifically, the consolidation of affective and cognitive abilities at this age both makes it possible for children to reflect on their own affective tendencies, while also increasing the likelihood that such cognitive/affective schemas will routinely be applied to new situations. All of the participants were Jewish, and based on school reports, these students were mostly from middle to upper – middle class backgrounds.
MEASURES Affect-Related Measures Negative academic affect scale. The Negative Academic Affect Scale (NAAS) (Gumora, 1993) was developed to assess students’ perceptions of the frequency of negative affect that they experience while engaged in school-related tasks. The 39-item NAAS was created to measure the degree of anxiety, frustration, and anger that students feel while engaged in a variety of academic tasks ranging from homework, to class participation, to engagement in class projects. Some questions focus on students’ abilities to organize and synthesize information, others on classroom participation, and others on their test performance and teacher evaluation (with 6 of 39 items in this last category). Participants were provided with definitions of the three emotions. For example, anxiety was defined as an uncomfortable feeling similar to worry and/or nervousness. A sample item reads as follows ‘‘I (never, rarely, sometimes, often, always) experience frustration writing essays.’’ Students rated the frequency of each emotion for each of 13 specific school tasks. Items on the NAAS were scored using a 5-point Likert scale, where 1 = never, 2 = rarely, 3 = sometimes, 4 = often, and 5 = always. Higher scores indicated that participants experienced more frequent negative emotions while engaged in academic tasks. Item scores were summed to form a total score of negative academic affect. A factor analysis of participants’ responses was conducted to examine whether the NAAS was best characterized as a unitary or multidimensional scale. The Kaiser –Meyer – Olkin (KMO) measure of sampling adequacy was .80, indicating that 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 1 emerged (11.6), and most items loaded .40 or greater on this factor. Internal consistency was quite high (Cronbach’s a=.93).
Gumora and Arsenio
401
Self-reported negative mood. Participants assessed their own general emotionality using the seven-item mood scale of the Revised Dimensions of Temperament Scale (DOTS-R). Sample items include ‘‘I do not laugh and smile at a lot of things’’ and ‘‘my mood is generally cheerful’’ (reverse scored). Items are scored on a 4-point Likert scale ranging from ‘‘usually true’’ to ‘‘usually false’’. There are different versions of the DOTS-R in which various raters assess the target participant. In this study the child selfrated version was used. The DOTS-R was originally developed to measure individual differences in temperament exhibited from childhood to adulthood (Windle & Lerner, 1986), and it has been shown to be related to a number of other important measures of psychosocial functioning (Windle, 1992). Although temperament is sometimes seen as including a biological and/or genetic component (Rothbart & Bates, 1998), the authors of the DOTS-R neither provide any empirical support for biological biases of DOTS-R scales, nor do they explicitly discuss their instrument in these terms. Consequently, in this study the DOTS-R was seen as a useful assessment of broad behavioral styles without any specific claim regarding the origins of those styles. The 54 items on the child-self version are divided into nine categories: activity level-general, activity level-sleep, approach/withdrawal, flexibility/ rigidity, mood, rhythmicity-sleep, rhythmicity-eating, rhythmicity-daily habits, and task orientation. Items within each of the nine categories are summed to calculate the total score for that category. The DOTS-R does not yield a composite score. For the present purposes only, the scales for mood and task orientation were of theoretical interest. Exploratory analyses confirmed this theoretical focus: there were very few significant connections between any of the other DOTS-R scales and other study variables other than those involving the mood and task orientation scales. Positive affect and negative affect scale. The Positive Affect and Negative Affect Scale (PANAS) was selected to provide a source of information on students’ affective dispositions that was not based on self-ratings (i.e., the DOTS-R Mood). More specifically, this version of the PANAS (Eisenberg et al., 1997) measures teachers’ perceptions of students’ moods. It was designed to study the relations between common events and two independent mood factors, positive affect and negative affect (Clark & Watson, 1988), and it has been used in a number of studies on the connections between children’s emotion regulation and their social competence. Teachers rate how well a set of 16 different emotion descriptors (e.g., irritable, distressed, etc.) characterizes each child using a 5-point Likert scale (ranging from ‘‘not at all’’ to ‘‘extremely’’). All students in the study were rated by two of their teachers, which always included their English teacher and nearly always their math teacher. Overall, there was a high level of agreement between the teachers’ ratings,
402
Journal of School Psychology
both for students’ positive affect (r =.65), and, to a lesser degree, for negative affect (r =.41). These levels of inter-rater agreement were comparable to what has been observed in other research (e.g., Eisenberg et al., 1993) and, as in those studies, a decision was made to combine the two raters’ scores. Ratings were averaged within pairs by item. Although it was expected that there would be connections among the three measures of students’ general emotionality (i.e., the DOTS-R Mood and two PANAS scales), it also seemed important to include assessments from more than a single rater. Previous studies by Eisenberg et al. (1995) for example, have shown that there are only modest connections between teacher and parents ratings of children’s emotionality, and that teacher and parent ratings are linked with different aspects of children’s social competence. In the present context, it was unclear whether teachers (as outside ‘‘objective’’ raters) or students (as ‘‘privileged’’ raters) would be more able to assess students’ emotionality. Consequently, both raters were used. Emotion-regulation. As described above, a major component of recent assessments of emotion regulation involves individual differences in the voluntary ability to focus on tasks for sustained periods (Ahadi & Rothbart, 1994; Eisenberg et al., 1997). Consequently, in the present study students’ ability to regulate emotions was assessed using the eight-item Task Orientation scale of the DOT-R. Sample items include, ‘‘Once I take up something, I stay with it’’ and ‘‘I am hard to distract.’’ (See the description on the Selfreported Negative Mood above for more background). Cognition-Related Measures Academic performance. Students’ academic performance was assessed using teacher assigned grades for English and mathematics. These class grades were selected both because they are core subjects for literacy and numeracy, and because these two courses were taken by all participants (there was some variation in other course offerings depending on students’ grade level, school, and their particular elective choices). 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). Academic achievement. One of the goals of this research was to examine whether students’ affective dispositions and negative academic affect would be related to their school performance even after controlling for more cognitive influences. Unfortunately, it was not possible to include an assessment of basic cognitive abilities, such as the WISC. Instead it was necessary to rely on available assessments of school achievement at the two
Gumora and Arsenio
403
schools, the Educational Records Bureau’s (ERB) Comprehensive Testing Program III (CTP) at one site, and the Iowa Tests of Basic Skills (ITBS) at the other site. The CTP provides a standardized educational assessment instrument capable of measuring ability and achievement along the continuum of student capability and performance (Educational Testing Service, 1995). Similarly, the ITBS is a standardized achievement test battery used to measure the development of general cognitive skills from kindergarten through 9th grade. Although there are no data available on the relationship between the CTP and the ITBS, independent reviews of each support their validity as assessments of school achievement (see, e.g., Educational Testing Service, 1995; Hambleton, 1987, respectively, for reviews). Thus it seems reasonable to believe that the CTP and the ITBS measure roughly the same constructs. For the present purposes, students’ verbal and quantitative achievement scores were converted to relevant percentile scores to adjust for differences in the raw scoring systems, and the average of these two percentages was taken as an overall score of academic achievement. Academic competency scale. The Academic Competency Scale (ACS) measures students’ self-concepts of their academic competency. The ACS is one of the six subscales of the Multidimensional Self-Concept Scale (Bracken, 1992), a comprehensive assessment device designed to facilitate the clinical appraisal of individuals from 9 to 19 years old. The test can be used as an overall assessment of self-concept or as an individual measure of any of six-scaled dimensions of self-concept (i.e., Social, Affect, Academic, Family, and Physical Competence Scales). ACS items are scored on a 4point Likert scale ranging from strongly agree to strongly disagree. A sample item reads as follows, ‘‘I learn fairly easily.’’ Procedure Only those students with parental consent participated. The NAAS, ACS, and DOTS-R assessments were administered to students in small groups during school, either at lunchtime or during a library period. Teachers independently filled out the PANAS on their own time. School officials provided the information regarding students’ grades and achievement test scores.
RESULTS Preliminary analyses revealed that there was no pattern of significant differences in the study variables as a function of either gender or grade level (only 1 of 16 analyses was significant). There were also no pattern of
404
Journal of School Psychology
significant differences in the magnitude of any of the correlations among the measures (using Fisher Z tests) as a function of students’ gender, grade level, or the school they attended. Consequently, correlations among the study measures (see Table 1) are shown for the entire sample. Overall, of the 28 correlations, 17 were significant ( p < .05), including five of seven correlations involving the NAAS, and five of seven correlations involving GPA. Moreover, another six correlations were somewhat significant ( p < .10). The numerous connections among the core variables suggest that few of these links are likely to be the result of type 1 errors. In fact, the mean magnitude of all correlations was r = .24, p < .05 (treating positive and negative correlations as equivalent).
Associations Among Affect-Related Variables (Hypothesis 1A) Students who had higher levels of emotion regulation were rated by teachers as having more positive moods (PANAS-Positive) and these students reported less negative academic affect (NAAS). In addition students who assessed themselves as having more negative overall moods (DOTS-R Mood) reported higher levels of negative academic affect, were rated by their teachers as less emotionally positive (PANAS-Positive), and were somewhat more likely to be rated by teachers as having negative moods (PANAS-Negative). Moreover, students who reported more negative academic affect were somewhat more likely to be rated by teachers as Table 1 Intercorrelations Among Cognition and Affect-Related Variables (n = 103) 1 Cognition-related GPA Achievement scores Academic competence Affect-related DOTS-R-emotion regulation PANAS – positive mood PANAS – negative mood DOTS-R – negative mood NAAS * p < .05. ** p < .01. *** p < .001. + p < .10.
2
– 0.61*** –
3
4
5
6
7
8
0.39*** 0.01
0.36***
0.28**
0.16 +
0.31**
0.30**
0.21 *
0.05
0.20 *
0.26**
0.44***
0.51***
–
0.17 *
+
0.13
0.19 *
0.12
0.10
0.23**
0.12
0.34***
0.14 +
0.13 +
0.14 +
–
0.23 *
0.28** 0.15 –
–
–
+
–
Gumora and Arsenio
405
having more negative and fewer positive moods (PANAS-Positive). Somewhat surprisingly, however, teachers’ rating of students’ positive moods were not significantly related to their rating of students’ negative moods. Associations Among Cognition-Related Variables Several significant connections emerged among the cognition-related variables. The connection between GPA and achievement tests supports the traditional view that students’ academic performance is related to their basic cognitive abilities (as measured by standardized achievement tests). Furthermore, students’ perceptions of their academic competence were linked with both their GPA and achievement scores. Specifically, students who had more positive conceptions of their academic ability had higher GPAs and higher achievement scores. Associations Between Affect-Related and Cognition-Related Variables (Hypothesis 1B) Teachers’ ratings of students’ positive and negative emotionality had a number of connections with cognition-related variables. Students with more positive moods had higher GPAs, achievement scores, and somewhat higher academic competence. In contrast, more negative moods were linked with lower GPAs and somewhat with lower academic competence. In terms of students’ rating of their own moods, students with more negative moods perceived themselves as less academically competent, had lower achievement scores, and somewhat lower GPAs. Finally, students’ negative academic affect was associated with all of the cognition-related variables. Students who perceived themselves as experiencing more negative affect during academic tasks had a poorer sense of their academic competence, lower achievement scores, and, finally, lower GPAs. Regression Analyses Predicting Students’ GPA (Hypotheses 2 and 4) Although students’ affective disposition and emotion regulation were connected with their GPA and other cognition-related variables, it was also expected that the affective variables would be a significant predictor of their grade point average over and above the contributions of the cognitive variables. Moreover, it was expected that academic affect would make a unique contribution to GPA beyond the contribution of the other affective variables. To address these issues, we conducted a hierarchical regression (see Table 2). The cognition-related variables (academic achievement, academic competence) were entered on Step 1. This was followed by Step 2, in which all affect-related variables except negative academic affect were entered (i.e., PANAS-Negative, PANAS-Positive, and DOTS-R Mood, DOTS-
406
Journal of School Psychology Table 2 Cognition and Affect-Related Predictors of Students’ GPA
Predictors added at
Step 1 B
Achievement 0.55 Step 1 (R2 = 0.424, Fchange = 36.83***) scores Academic 0.24 competence 2 DOTS-RStep 2 (R = 0.572, Fchange = 8.35***) emotion reg. PANASpositive mood PANASnegative mood DOTS-Rnegative mood Negative Step 3 (R2 = 0.574, academic affect Fchange = 4.59 *)
Step 2 t
B
Step 3 t
B
t
6.80***
0.52
7.34***
0.50
7.16***
2.98**
0.31
3.93**
0.23
2.67**
0.21
3.01**
0.23
3.21**
0.27
3.76***
0.26
3.64***
0.23
3.36**
0.27
3.85***
0.18
2.35 *
0.19
2.45 *
0.18
2.14 *
R2 and Fchange are shown for the entire step; the B and t are for each predictor. The DOTS-R measures were self-ratings and the PANAS measures were teacher ratings. * p < .05. ** p < .01. *** p < .001.
R Emotion Regulation), and finally by Step 3 which included only negative academic affect (NAAS). The variables were entered in this order to determine whether affect makes an additional contribution to predicting students’ GPA beyond the more traditional focus on cognitive abilities and competencies. Moreover, the NAAS was entered last to assess whether students’ negative academic affect, in particular, had an influence on their GPA over and above the more general affective contribution made by their moods. As can be seen in Table 2, students’ cognition-related abilities were a significant predictor of their GPA (Step 1), and academic achievement, academic competence were both independent significant predictors of GPA (i.e., after accounting for the effects of the other cognition-related variable). Furthermore, each of these cognition-related variables remained a significant predictor of students’ GPA when the affect-related variables were entered on Steps 2 and 3. The affect-related variables, however, also made a significant contribution to GPA beyond the cognitive contribution (Step 2). Finally, as hypothesized, the NAAS (Step 3) was still a significant predictor of GPA even after accounting for the influence of all other cognition and affect-related variables; compared to their peers, students who reported more negative emotions in relation to routine school tasks performed worse in school and this was not the result of underlying cognitive differences (at least as measured in this study). All of the variables continued to be significant predictors, even at the final step, suggesting that the contributions of these variables were relatively independent.
Gumora and Arsenio
407
It is also noteworthy that these findings were not influenced by problems with multicollinearity. None of the correlations approached the .90 criterion that is considered problematic, and this was confirmed by the collinearity diagnostics: no variable had more than one variance proportion greater than .50 (Tabachnik & Fidell, 2001). Overall, the affect and cognition-related variables accounted for nearly 60% of the variance in students’ grade point averages. Interaction of Emotion Regulation and Negative Academic Affect on GPA (Hypothesis 3) It was originally hypothesized that students’ negative academic affect might interact with their emotion regulation abilities in the prediction of their GPA. The emotion regulation negative academic affect interaction term, however, was not significantly related to GPA (r = .12), and consequently, there was no support for this hypothesis. Interaction of Negative Academic Affect and General Emotionality on GPA Although not originally hypothesized, it is possible that negative academic affect combines with more general aspects of students’ emotionality to influence their school performance. For example, a child who experiences a combination of both negative academic affect and more general negative affectivity might be at special risk for school difficulties. Consequently, separate mood negative academic affect interaction terms were created for each three mood-related variables (the PANAS-Positive, PANAS-Negative and DOTS-R Mood scales were not highly related enough to form a composite). No interaction term was significantly related to students’ GPA, and thus, there was no evidence for a combined influence of negative academic affect and aspects of students’ more general emotionality beyond the main effects of these affective variables.
DISCUSSION This study provides additional support for the role of socio-emotional factors in students’ school performance (Birch & Ladd, 1997; Pianta, Steinberg, & Rollins, 1995; Wentzel, 1996, 1999), while also clarifying some of the uniquely affective contributors (rather than relationships or goals) to that performance. More specifically, the present results indicate that both middle school children’s emotional dispositions and their academically related affect are connected with their school success. Even after controlling for the influence of other cognitive variables, including academic achievement and academic self-efficacy, students’ emotional dispositions
408
Journal of School Psychology
(teacher and self-rated) and academic affect predicted their academic performance. Moreover, negative academic affect could be distinguished from students’ affective dispositions, and negative academic affect was a unique predictor of students’ combined grade point average in language arts and mathematics classes beyond the contributions of all other variables assessed. Although there is previous evidence that anxiety interferes with test performance (McDonald, 2001), the present findings suggest that negative academic affect is also linked with more general aspects of academic success. Compared to their peers, students in this study who reported experiencing more frequent negative affect during a variety of academic tasks, including homework, class participation, quizzes, and tests had lower grades in their two core academic classes, language arts and mathematics. Moreover, it is important to recall that factor analyses revealed the presence of only a single factor in NAAS. The lack of separate factors for tests and quizzes or other academic tasks suggests that many students have a global affective reaction to academic work. How can this connection between academic affect and school performance be explained? Given the correlational nature of this study, there are several possible routes. For example, less cognitively proficient students may do worse in school, and, as a result, have more negative emotional reactions to any school-related task. Yet, even after accounting for the influence of several cognitive contributors in the present study (i.e., academic achievement and academic self-efficacy) academic affect still added significantly to the prediction of students’ school performance. Although it is still possible that IQ (or another unassessed cognitive influence) is responsible for these connections, this is less likely given the overlap between IQ and the cognitive measures that were included (e.g., academic achievement, Sattler, 1992). Another possible explanation for the link between academic affect and school performance stems from related research on the affective contributors to children’s social (rather than academic) competence. In that work, children’s negative emotionality has been found to predict lower levels of social competence both concurrently (Eisenberg et al., 1993) and longitudinally (Eisenberg et al., 1997). In a related vein, negative emotionality in the present study was also connected with lower levels of academic performance. And, once again, regression analyses suggested that this connection was not the result of obvious cognitive influences: different aspects of emotionality added to the prediction of students’ GPA over and above the influence of the cognitive variables, and each of the three dispositional affective measures was a significant predictor of GPA at all of the steps in the regression. Although this last explanation seems consistent with the present findings, at least two important unanswered questions remain: what are the origins
Gumora and Arsenio
409
and nature of these affective contributors; and how are academic affect and the other affective contributors related? Given the inclusion of the DOTS-R temperament measures, it could be argued that children’s psychophysiology (especially biological components of temperament) affects their academic achievement. The literature on children’s affective dispositions, however, indicates both that parental socialization plays a major role in these dispositions (e.g., see Denham, 1999, for a review), and that parental influences can even alter the physiology of children’s emotional responsiveness over time (Gottman, Katz, & Hooven, 1997). To the extent that ‘‘temperament’’ was assessed in the present research, it seems more reflective of broad behavioral tendencies than largely biologically determined patterns. Instead we would propose that students’ affective dispositions and academic affect largely reflect a combination of parental and school-related influences. Some cognitively able students, for example, may feel unwanted pressures to excel in school, perhaps because the work is considered excessive or difficult, or because it interferes with non-academic personal goals. Consequently, compared to their peers, these young adolescents may experience more negative affect both during academic tasks and in general. Over time, however, as that negative affect increasingly interferes with students’ motivation and ability to focus on schoolwork, and as their grades decline, the academic context may become especially linked with negative affect. It must be acknowledged that this implicit model extends well beyond the present findings. Yet, we believe that this model has implications for future research, which could bridge the present focus on affective dispositions with work on relational contributors to school performance. Wentzel (2002), for example, recently found that the classroom socialization practices of effective teachers are quite similar to parenting styles that are linked with higher student achievement. Specifically, teachers who combine positive emotionality (i.e., encouragement and avoiding negative feedback) and high expectations (i.e., a push to excel) have students with more academically oriented goals and higher school performance. A useful combination of this ‘‘relational’’ literature and the present study would involve examining the connections among students’ affective dispositions, adult – child relationships, and school performance. It might be expected, for example, that adults who are emotionally positive will promote students’ own positive emotionality in ways that make these students more responsive to both academic expectations and to the instructional process itself. That is, students’ emotionality may mediate and/or moderate the influence of adults on students’ school performance in at least two ways: (1) by increasing what Kochanska (in a different context) called ‘‘committed compliance’’ (Kochanka & Aksan, 1995),
410
Journal of School Psychology
students’ active embracing of adults’ academic expectations and (2) by permitting adults easier access to the students’ zone of proximal development (Vygotsky, 1978), and consequently, increasing both the quantity and quality of adult – student educational exchanges. More generally, additional research is needed to examine the connections among the affective, motivational, and relational contributors to school performance. Several unexpected findings and study limitations need to be mentioned at this point. Although, as expected, students’ emotion dispositions and emotion regulation both predicted their school performance, contrary to expectations, students’ emotion regulation did not interact with their emotional dispositions to make an additional significant contribution (as it sometimes has in studies of social competence, e.g., Eisenberg et al., 1997). This discrepancy could stem from the older age group used in this study compared to research on social competence, or perhaps from the present emotion regulation instrument, which involved self ratings rather than the teacher or parent ratings typically used to assess emotion regulation. It would be helpful if future studies of emotionality and school performance included multiple raters of students’ emotion regulation. Another unexpected finding was that emotion regulation was not related to students’ GPAs in the zero-order correlations, but it was a significant predictor of GPA at all steps of the regression analyses. In other words, emotion regulation acted as a suppressor variable: ‘‘a variable can be uncorrelated with the criterion and still improve prediction by being correlated with other predictors. . .(which) implies that such variables suppress criterion-irrelevant variance in other predictors’’ (Tzelgov & Henik, 1991, p. 524). Analyses revealed that this suppressor effect was not the result of high levels of multicollinearity between emotion regulation and any other predictors, but it is still unclear how to explain this suppression. Tzelgov and Henik argue that, although suppression is more common and conceptually important than is sometimes believed, it is especially important to replicate such effects before including them in psychological models. Finally, one important limitation of the present study involves the nature of the sample. The students were highly homogeneous in terms of race (all European –American) and religious/cultural background (all Jewish), and both of the schools they attended were very academically competitive. Although it is unclear how this combination of characteristics affected the present findings, it seems likely that these students experienced a much higher than average amount of pressure to do well in school. On one hand, this could have increased negative academic affect and perhaps students’ overall negativity relative to other middle school students. Alternatively, children with higher levels of negative academic affect and overall negativity may have been more likely to avoid these academically select schools. Additional research involving more diverse populations and school settings
Gumora and Arsenio
411
is needed to examine whether the present work can be generalized to other student populations. In conclusion, this study highlights the importance of affective contributors to middle school children’s academic performance. The present data suggest that although students’ emotion regulation, general affective dispositions, and academic affect are related to each other, each of them also makes a unique, significant contribution to students’ GPAs, over and above the influence of other cognitive contributors. Future studies can build on this work both by addressing some of the present study’s limitations, and by simultaneously assessing the connections among students’ school performance, relevant adult – child relationships, and various aspects of their affective dispositions and emotion regulation abilities. ACKNOWLEDGEMENTS This article is based on a dissertation by Gail Gumora submitted in partial fulfillment of the requirements for the doctoral degree at Ferkauf Graduate School of Psychology, Yeshiva University. We would especially like to thank the children and teachers who made this research possible. A version of this paper was presented at the 2001 meeting of the Society for Research in Child Development in Minneapolis, MN. REFERENCES Ahadi, S., & Rothbart, M. (1994). Temperament, development, and the big five. In C. Halverson Jr., G. Kohnstamm, & R. Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood ( pp. 189 – 207). Hillsdale, NJ: Erlbaum. Arsenio, W., Cooperman, S., & Lover, A. (2000). Affective predictors of preschoolers’ aggression and peer acceptance: direct and indirect effects. Developmental Psychology, 36, 438 – 448. Arsenio, W., & Killen, M. (1996). Preschoolers’ conflict-related emotions during peer disputes. Early Education and Development, 7, 43 – 57. Birch, S., & Ladd, G. (1997). The teacher – child relationship and children’s early school adjustment. Journal of School Psychology 35, 61 – 79. Bracken, B. (1992). Multidimensional self concept scale. Texas: Pro-Ed. Clark, L., & Watson, D. (1988). Mood and the mundane: relations between daily events and self-reported mood. Journal of Personality and Social Psychology, 54(2), 296 – 308. Cole, P., Michel, M., & Teti, L. (1994). The development of emotion regulation and dysregulation: a clinical perspective. In N. Fox (Ed.), The development of emotion regulation: biological and behavioral considerations, Monographs of the Society for Research in Child Development, vol. 59 (pp. 73 – 100) (2 – 3, Serial No. 240). Corno, L., & Rohrkemper, M. (1985). The intrinsic motivation to learn in classrooms. In C. Ames, & R. Ames (Eds.), Research on motivation, vol. 2, (pp. 53 – 90). New York: Academic Press. Denham, S. (1999). Emotional development in young children. New York: Guilford. Educational Testing Service (1995). Comprehensive Testing Program III Technical Report. (1995). Educational Testing Service, Princeton, NJ.
412
Journal of School Psychology
Eisenberg, N., & Fabes, R. (1992). Emotion, regulation, and the development of social competence. In M. Clark (Ed.), Review of personality and social psychology: vol. 14. Emotion and social behavior ( pp. 119 – 150). Newbury Park, CA: Sage. Eisenberg, N., Fabes, R., Bernzweig, J., Karbon, M., Poulin, R., & Hanish, L. (1993). The relations of emotionality and regulation to preschoolers’ social skills and sociometric status. Child Development, 64, 1418 – 1438. Eisenberg, N., Fabes, R., Murphy, B., Maszk, P., Smith, M., & Karbon, M. (1995). The role of emotionality and regulation in children’s social functioning: a longitudinal study. Child Development, 66, 1360 – 1384. Eisenberg, N., Fabes, R., Shepard, S., Murphy, B., Guthrie, I., Jones, S., Friedman, J., Poulin, R., & Maszk, P. (1997). Contemporaneous and longitudinal prediction of children’s social functioning from regulation and emotionality. Child Development, 68, 642 – 664. Fabes, R., & Eisenberg, N. (1992). Young children’s coping with interpersonal anger. Child Development, 63, 116 – 128. Ford, M. (1992). Motivating humans: goals, emotions, and personal agency beliefs. Newbury Park, CA: Sage. Gottman, J., Katz, L., & Hooven, C. (1997). Meta-Emotion: how families communicate emotinally. Mahwah, NJ: Erlbaum. Gumora, G. (1993). The construction of the negative academic affect scale (NAAS). Unpublished manuscript. Hambleton, R. (1987). Iowa test of basic skills: forms G and H. In D. Keyser, & R. Sweetland (Eds.), Test critiques. Missouri: Test Corporation of America. Hamre, B., & Pianta, R. (2001). Early teacher – child relationships and the trajectory of children’s school outcomes through eight grade. Child Development, 72, 625 – 638. Kochanka, G., & Aksan, N. (1995). Mother – child mutually positive affect, the quality of child compliance to requests and prohibitions, and maternal control as correlates of early internalization. Child Development, 66, 236 – 254. Lazarus, R. (1984). On the primacy of cognition. American Psychologist, 39, 124 – 129. Mayer, J., & Salovey, P. (1997). What is emotional intelliegence? In P. Salovey, & D. Sluyter (Eds.), Emotional development and emotional intelligence: educational implications ( pp. 3 – 25). New York: Basic Books. McDonald, A. (2001). The prevalence and effects of test anxiety in school children. Educational Psychology, 21(2), 89 – 102. Musch, J., & Broeder, A. (1999). Test anxiety versus academic skills: a comparison of two alternative models for predicting performance in a statistics exam. British Journal of Educational Psychology, 69(1), 105 – 115. Piaget, J. (1954/1981). In: T. Brown, & C. Kaegi (Eds.), Intelligence and affectivity: their relationship during child development. Palo Alto, CA: Annual Reviews Monographs. Pianta, R. (1999). Enhancing relationships between children and teachers. Washington, DC: American Psychological Association. Pianta, R., Steinberg, M., & Rollins, L. (1995). The first two years of school: teacher – child relationships and deflections in children’s classroom adjustment. Development and Psychopathology, 7, 295 – 312. Rothbart, M., & Bates, J. (1998). Temperament. In W. Damon, & N. Eisenberg (Eds.), Handbook of child psychology, 5th ed.: vol. 3: social, emotional, and personality development ( pp. 105 – 176). New York: Wiley. Sattler, J. M. (1992). Assessment of children: revised and updated (3rd ed.). San Diego: Jerome M. Sattler, Publisher. Tabachnik, L., & Fidell, B. (2001). Using multivariate statistics. Boston: Allyn and Bacon. Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research: definition, implications, and application. Psychological Bulletin, 109(3), 524 – 536. Vail, P. (1981). Emotion: the on off switch for learning. NJ: Modern Learning Press.
Gumora and Arsenio
413
Vygotsky, L. (1978). Mind in society. Cambridge, MA: Harvard University Press. Wentzel, K. (1996). Motivation in context: social relationships and achievement in middle school. In J. Juvonen, & K. Wentzel (Eds.), Social motivation: understanding children’s school adjustment ( pp. 226 – 247). New York: Cambridge University Press. Wentzel, K. (1999). Social-motivational processes and interpersonal relationships: implications for understanding motivation at school. Journal of Educational Psychology, 91, 76 – 97. Wentzel, K. (2002). Are effective teachers like good parents? Teaching styles and student adjustment in early adolescence. Child Development, 73, 287 – 301. Windle, M. (1992). Temperament and social support in adolescence: interrelations with depressive symptoms and delinquent behaviors. Journal of Youth and Adolescence, 21, 1 – 21. Windle, M., & Lerner, R. (1986). Reassessing the dimensions of temperamental individuality across the lifespan: the revised dimension of temperament survey (DOTS-R). Journal of Adolescent Research, 1(2), 213 – 230.
PII S0022-4405(02)00108-5
Emotionality, Emotion Regulation, and School Performance in Middle School Children Gail Gumora and William F. Arsenio Ferkauf Graduate School of Psychology, Yeshiva University, Bronx, NY, USA This research investigated the connections of middle school students’ emotional dispositions and academic-related affect with their school performance. One hundred three 6th – 8th grade students completed three self-rated assessments regarding: (a) their academic competency; (b) affective tendencies (both mood and emotion regulation); and (c) negative emotions experienced during school-related tasks. Teachers assessed students’ positive and negative moods, and schools provided achievement test results and student grades as measures of cognitive ability/ achievement and school performance, respectively. Results indicated that although students’ emotion regulation, general affective dispositions, and academic affect were related to each other, each of these variables also made a unique significant contribution to students’ GPA, over and above the influence of other cognitive contributors. Overall, these results provide support for the role of socio-emotional factors in students’ school performance, while also clarifying some of the uniquely affective contributors (rather than relationships or goals) to that performance. D 2002 Society for the Study of School Psychology. Published by Elsevier Science Ltd Keywords: Emotional dispositions, Emotion regulation, School performance.
There is a growing awareness that social and emotional factors play an important role in students’ academic success. Research has shown that students who have more prosocial, socially responsible goals (Wentzel, 1996, 1999) or more positive relationships with adults (teachers and parents; Birch & Ladd, 1997; Pianta, 1999) than their peers subsequently perform better in academic contexts. For example, Hamre and Pianta (2001) recently found that, controlling for cognitive ability, children who had negative conflicted relationships with their kindergarten teachers had lower grades and worse work habits in 8th grade. In one model of the processes underlying these connections (Ford, 1992), academic motivation is explained ‘‘in terms of three inter-related
Received 12 December 2001; received in revised form 26 March 2002; accepted 29 May 2002. Address correspondence and reprint requests to Gail Gumora, 115 W. 86th Street, Apartment 16E, New York, NY 10024, USA; or William F. Arsenio, Ferkauf Graduate School of Psychology, Yeshiva University, 1300 Morris Park Avenue, Bronx, NY 10461, USA. E-mail addresses: [email protected], [email protected]
395
396
Journal of School Psychology
processes: personal goals, personal agency beliefs, and emotional arousal. . .. According to this model, an individual will pursue a goal to the extent emotions and personal agency beliefs are associated with the goal and will support efforts to achieve one’s goals’’ (Wentzel, 1999, p. 78). To date, however, much of the work on socio-emotional contributors to academic performance has focused on personal goals (‘‘do I want to do well in school?’’), agency beliefs (‘‘will I be able to if I try?), and/or student– teacher relationships (‘‘how do I feel about my teacher?’’). In contrast, much less is known about how students’ emotions, per se, support academic performance (‘‘how do I feel about schoolwork?).1 The research on the affective contributors to achievement that does exist has largely been confined to the role of test anxiety on test performance (e.g., Musch & Broeder, 1999; and see McDonald, 2001 for a review of research involving children). In general, it appears that worry interferes with the test performance of highly anxious students because it distracts attention from the task at hand. Vail (1981) recognized that for some students, however, the anxiety and frustrations associated with more everyday school tasks could also be problematic. Some children and adolescents may have difficulties with their emotional arousal during routine homework and classroom activities that have little to do with test anxiety, per se. Consequently, the present study was designed to examine the connections between young adolescents’ academically related emotions and their academic performance. It was expected that, overall, students who report more negative emotions during academic tasks would have lower levels of school performance (as assessed by their grade point averages). In addition to examining the connection between negative academic affect and school performance, two other issues were addressed in this research. One is whether academically related affect is just a reflection of students’ more general affective dispositions, and consequently, whether it is necessary to distinguish academic affect from more general affective contributors to school performance (see below). A second issue is whether academic affect is mostly a reflection of students’ underlying cognitive abilities and their perceived competence regarding their abilities. Although Ford’s (1992) model makes a theoretical distinction between personal
1
Although there is extensive research on how broad aspects of students’ affective and behavioral tendencies are related to their school performance (e.g., aggression, peer rejection, and school failure), most of that work does not differentiate specifically affective components of behavior from other socio-emotional components such as peer and adult relationships. By contrast, the literature on social competence has shown that the connections between children’s and adolescents’ emotion dispositions and abilities and their social behaviors are often more complex, and in need of empirical verification, than is typically assumed (e.g., Arsenio, Cooperman, & Lover, 2000; Denham, 1999).
Gumora and Arsenio
397
agency beliefs (i.e., perceived academic competence) and emotional arousal, connections between school performance and academic affect could largely be a ‘‘by product’’ of students’ cognitive abilities and academic competence. In other words, young adolescents who are cognitively able and who believe they are good at school tasks may perform well in school, and consequently, feel more positively (and less negatively) about academic tasks. In terms of the first issue (general affective dispositions vs. academic affect), although there is little work on affective dispositions and academic performance, an extensive literature has emerged on the connections between children’s emotionality and their social competence. Eisenberg et al. (1993) and Fabes and Eisenberg (1992) for example, have conducted an influential series of studies showing that children’s general negative emotionality is related to their subsequent peer behaviors, and to both peer and adult ratings of their social competence. Moreover, other studies indicate that positive emotionality is associated with lower levels of socially incompetent behaviors and higher levels of peer acceptance in children (Cole, Michel, & Teti, 1994). Collectively, this research both demonstrates that it is possible to assess broad aspects of children’s emotionality and that these emotional tendencies are related to children’s social functioning. Other studies, however, suggest that children’s emotionality is at least partly context specific. For example, Arsenio et al. (2000) distinguished between children’s emotionality during aggression and more general aspects of emotionality. Negative emotionality (especially anger) during aggression was modestly but significantly related to children’s more general angry negativity, but only aggression-related negativity and not general negativity was related to whether children were liked by peers (see also Arsenio & Killen, 1996). For the present purposes, these studies suggest that it may be important to distinguish between more general affectivity and specific affective contexts of interest (i.e., academic affect). Some students, for example, may have generally positive affective dispositions, and yet still experience during schoolwork negative emotions that do not simply stem from underlying cognitive difficulties. Finally, the connections between students’ emotionality and social competence depend, in part, on their abilities to regulate and control negative emotionality (Eisenberg & Fabes, 1992). For example, in a longitudinal study (Eisenberg et al., 1997) it was found that children who were high in negative emotionality but also high in emotion regulation were rated as functioning well socially. At the same time, however, ‘‘regulation and emotionality frequently contributed unique variance to the prediction of social functioning’’ (p. 642), suggesting that the interactive contribution of emotionality and regulation to social competence was in addition to separable contributions of emotionality and emotion regulation.
398
Journal of School Psychology
Guided by these studies on children’s emotion and social competence, we decided to assess the contributions of young adolescents’ emotion regulation, as well as their general emotionality and academic affect, to their school performance. Although emotion regulation has been assessed in several ways, one increasingly central component is effortful control which includes ‘‘individual differences in the ability to voluntarily sustain focus on a task’’ (Ahadi & Rothbart, 1994, p. 196; and see also Corno & Rohrkemper, 1985 for evidence that older students’ abilities to manage and control their effort is linked with academic performance). This ability to focus on tasks seemed especially relevant for academic contexts, and consequently emotion regulation was assessed in the present study using a measure of task orientation. Extrapolating from the emotions and social competence literature, we expected both that there would be significant connections among general emotionality, academic affect, and emotion regulation (Hypothesis 1A), and that each of these three constructs would also be related to children’s academic performance (Hypothesis 1B). In addition, however, we expected that academic affect would make a separate, unique contribution to students’ school performance, over and above the contribution of general emotionality or emotion regulation (Hypothesis 2). In other words, academic affect was not expected to be just a subset of more general emotionality, or the result of emotion regulation abilities, and consequently, it would make its own contribution to school performance. Finally, based on research indicating that emotionality and emotion regulation may have an additional interactive influence on children’s functioning (i.e., in addition to main effects for emotionality and regulation), it was also expected (Hypothesis 3) that students’ academic emotionality would interact with their emotion regulation abilities to affect school performance. Specifically, high negative academic affect would have a negative relation to academic performance unless these young adolescents could compensate for that negative emotionality with high levels of academic emotion regulation.2 As noted above, another issue is whether affective contributors add anything unique to the prediction of students’ school achievement and performance beyond common cognitive contributors. Although there are often complex bi-directional connections between affect and cognition (e.g., Lazarus, 1984), recent research also indicates that socio-emotional factors are significant predictors of children’s and adolescents’ school
2 Previous studies on emotion regulation and emotionality did not address the role of the context in which emotionality was assessed, but it seems reasonable to expect that children’s academic affect would be especially influenced by their emotion regulation abilities. For example, children who often experience negative emotions in the context of academic tasks might be more likely to become distracted and do poorly in their academic work unless this is compensated for by highs of high levels of emotion regulation level to help keep them oriented to academic tasks.
Gumora and Arsenio
399
motivation and performance even after controlling for cognitive factors (Hamre & Pianta, 2001; Wentzel, 1999). Consequently, it was expected that although there might be significant connections among cognitive and affective variables, affective contributors (emotion regulation, general emotionality, and academic affect) would make an additional, unique contribution to students’ school performance over and above the contributions made by cognitive variables (Hypothesis 4). Several cognitively oriented variables were included in the present study. Although it would have been desirable to have a measure of IQ, pragmatic limitations made it necessary to use available standardized measures of academic achievement as a rough indicator of general cognitive abilities (see Methods). In addition, a measure of students’ perceived academic competence was included. As Wentzel (1999) has described, students’ perceptions of their academic competence influence their academic motivation and, in turn, their school performance. Consequently, connections between school performance and academic affect might primarily be a result of students’ perceived academic competence: students who perceive that they are good at school tasks may perform well in school, and then feel more positively (and less negatively) about academic tasks as a result. Although no correlational study can assess causal connections, controlling for the influence of academic achievement and perceived academic competence would begin to address the potentially unique contributions of affective influences on young adolescents’ school success. In summary, this study focused on whether middle school students’ affective dispositions and abilities (i.e., emotionality and emotion regulation) are related to their school performance, even after accounting for more cognitive contributions to that performance. Although connections were expected between affective and cognitive measures, the underlying goal was to examine whether, as has sometimes been argued (Mayer & Salovey, 1997), affective dispositions and tendencies also have a unique influence on students’ school success. For the reasons discussed below, a decision was made to focus on young adolescents in middle school, and no age-related differences were expected in this narrow age range. In terms of gender, some research with younger children has shown differences in emotionality and emotion regulations in boys and girls, but these differences are inconsistent across studies. Consequently, no specific gender differences were hypothesized.
METHODS Participants The 103 middle school (6th to 8th grade) young adolescents who participated in this study (51 girls and 52 boys) attended one of two Jewish
400
Journal of School Psychology
affiliated day schools in a major Northeastern city. This particular age group (11 –14 years, M = 12.1 years) was selected because the emergence of formal operational reasoning in children of this age has been linked with an increasingly abstract ability to understand and reflect on previous affective experiences (Piaget, 1981). More specifically, the consolidation of affective and cognitive abilities at this age both makes it possible for children to reflect on their own affective tendencies, while also increasing the likelihood that such cognitive/affective schemas will routinely be applied to new situations. All of the participants were Jewish, and based on school reports, these students were mostly from middle to upper – middle class backgrounds.
MEASURES Affect-Related Measures Negative academic affect scale. The Negative Academic Affect Scale (NAAS) (Gumora, 1993) was developed to assess students’ perceptions of the frequency of negative affect that they experience while engaged in school-related tasks. The 39-item NAAS was created to measure the degree of anxiety, frustration, and anger that students feel while engaged in a variety of academic tasks ranging from homework, to class participation, to engagement in class projects. Some questions focus on students’ abilities to organize and synthesize information, others on classroom participation, and others on their test performance and teacher evaluation (with 6 of 39 items in this last category). Participants were provided with definitions of the three emotions. For example, anxiety was defined as an uncomfortable feeling similar to worry and/or nervousness. A sample item reads as follows ‘‘I (never, rarely, sometimes, often, always) experience frustration writing essays.’’ Students rated the frequency of each emotion for each of 13 specific school tasks. Items on the NAAS were scored using a 5-point Likert scale, where 1 = never, 2 = rarely, 3 = sometimes, 4 = often, and 5 = always. Higher scores indicated that participants experienced more frequent negative emotions while engaged in academic tasks. Item scores were summed to form a total score of negative academic affect. A factor analysis of participants’ responses was conducted to examine whether the NAAS was best characterized as a unitary or multidimensional scale. The Kaiser –Meyer – Olkin (KMO) measure of sampling adequacy was .80, indicating that 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 1 emerged (11.6), and most items loaded .40 or greater on this factor. Internal consistency was quite high (Cronbach’s a=.93).
Gumora and Arsenio
401
Self-reported negative mood. Participants assessed their own general emotionality using the seven-item mood scale of the Revised Dimensions of Temperament Scale (DOTS-R). Sample items include ‘‘I do not laugh and smile at a lot of things’’ and ‘‘my mood is generally cheerful’’ (reverse scored). Items are scored on a 4-point Likert scale ranging from ‘‘usually true’’ to ‘‘usually false’’. There are different versions of the DOTS-R in which various raters assess the target participant. In this study the child selfrated version was used. The DOTS-R was originally developed to measure individual differences in temperament exhibited from childhood to adulthood (Windle & Lerner, 1986), and it has been shown to be related to a number of other important measures of psychosocial functioning (Windle, 1992). Although temperament is sometimes seen as including a biological and/or genetic component (Rothbart & Bates, 1998), the authors of the DOTS-R neither provide any empirical support for biological biases of DOTS-R scales, nor do they explicitly discuss their instrument in these terms. Consequently, in this study the DOTS-R was seen as a useful assessment of broad behavioral styles without any specific claim regarding the origins of those styles. The 54 items on the child-self version are divided into nine categories: activity level-general, activity level-sleep, approach/withdrawal, flexibility/ rigidity, mood, rhythmicity-sleep, rhythmicity-eating, rhythmicity-daily habits, and task orientation. Items within each of the nine categories are summed to calculate the total score for that category. The DOTS-R does not yield a composite score. For the present purposes only, the scales for mood and task orientation were of theoretical interest. Exploratory analyses confirmed this theoretical focus: there were very few significant connections between any of the other DOTS-R scales and other study variables other than those involving the mood and task orientation scales. Positive affect and negative affect scale. The Positive Affect and Negative Affect Scale (PANAS) was selected to provide a source of information on students’ affective dispositions that was not based on self-ratings (i.e., the DOTS-R Mood). More specifically, this version of the PANAS (Eisenberg et al., 1997) measures teachers’ perceptions of students’ moods. It was designed to study the relations between common events and two independent mood factors, positive affect and negative affect (Clark & Watson, 1988), and it has been used in a number of studies on the connections between children’s emotion regulation and their social competence. Teachers rate how well a set of 16 different emotion descriptors (e.g., irritable, distressed, etc.) characterizes each child using a 5-point Likert scale (ranging from ‘‘not at all’’ to ‘‘extremely’’). All students in the study were rated by two of their teachers, which always included their English teacher and nearly always their math teacher. Overall, there was a high level of agreement between the teachers’ ratings,
402
Journal of School Psychology
both for students’ positive affect (r =.65), and, to a lesser degree, for negative affect (r =.41). These levels of inter-rater agreement were comparable to what has been observed in other research (e.g., Eisenberg et al., 1993) and, as in those studies, a decision was made to combine the two raters’ scores. Ratings were averaged within pairs by item. Although it was expected that there would be connections among the three measures of students’ general emotionality (i.e., the DOTS-R Mood and two PANAS scales), it also seemed important to include assessments from more than a single rater. Previous studies by Eisenberg et al. (1995) for example, have shown that there are only modest connections between teacher and parents ratings of children’s emotionality, and that teacher and parent ratings are linked with different aspects of children’s social competence. In the present context, it was unclear whether teachers (as outside ‘‘objective’’ raters) or students (as ‘‘privileged’’ raters) would be more able to assess students’ emotionality. Consequently, both raters were used. Emotion-regulation. As described above, a major component of recent assessments of emotion regulation involves individual differences in the voluntary ability to focus on tasks for sustained periods (Ahadi & Rothbart, 1994; Eisenberg et al., 1997). Consequently, in the present study students’ ability to regulate emotions was assessed using the eight-item Task Orientation scale of the DOT-R. Sample items include, ‘‘Once I take up something, I stay with it’’ and ‘‘I am hard to distract.’’ (See the description on the Selfreported Negative Mood above for more background). Cognition-Related Measures Academic performance. Students’ academic performance was assessed using teacher assigned grades for English and mathematics. These class grades were selected both because they are core subjects for literacy and numeracy, and because these two courses were taken by all participants (there was some variation in other course offerings depending on students’ grade level, school, and their particular elective choices). 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). Academic achievement. One of the goals of this research was to examine whether students’ affective dispositions and negative academic affect would be related to their school performance even after controlling for more cognitive influences. Unfortunately, it was not possible to include an assessment of basic cognitive abilities, such as the WISC. Instead it was necessary to rely on available assessments of school achievement at the two
Gumora and Arsenio
403
schools, the Educational Records Bureau’s (ERB) Comprehensive Testing Program III (CTP) at one site, and the Iowa Tests of Basic Skills (ITBS) at the other site. The CTP provides a standardized educational assessment instrument capable of measuring ability and achievement along the continuum of student capability and performance (Educational Testing Service, 1995). Similarly, the ITBS is a standardized achievement test battery used to measure the development of general cognitive skills from kindergarten through 9th grade. Although there are no data available on the relationship between the CTP and the ITBS, independent reviews of each support their validity as assessments of school achievement (see, e.g., Educational Testing Service, 1995; Hambleton, 1987, respectively, for reviews). Thus it seems reasonable to believe that the CTP and the ITBS measure roughly the same constructs. For the present purposes, students’ verbal and quantitative achievement scores were converted to relevant percentile scores to adjust for differences in the raw scoring systems, and the average of these two percentages was taken as an overall score of academic achievement. Academic competency scale. The Academic Competency Scale (ACS) measures students’ self-concepts of their academic competency. The ACS is one of the six subscales of the Multidimensional Self-Concept Scale (Bracken, 1992), a comprehensive assessment device designed to facilitate the clinical appraisal of individuals from 9 to 19 years old. The test can be used as an overall assessment of self-concept or as an individual measure of any of six-scaled dimensions of self-concept (i.e., Social, Affect, Academic, Family, and Physical Competence Scales). ACS items are scored on a 4point Likert scale ranging from strongly agree to strongly disagree. A sample item reads as follows, ‘‘I learn fairly easily.’’ Procedure Only those students with parental consent participated. The NAAS, ACS, and DOTS-R assessments were administered to students in small groups during school, either at lunchtime or during a library period. Teachers independently filled out the PANAS on their own time. School officials provided the information regarding students’ grades and achievement test scores.
RESULTS Preliminary analyses revealed that there was no pattern of significant differences in the study variables as a function of either gender or grade level (only 1 of 16 analyses was significant). There were also no pattern of
404
Journal of School Psychology
significant differences in the magnitude of any of the correlations among the measures (using Fisher Z tests) as a function of students’ gender, grade level, or the school they attended. Consequently, correlations among the study measures (see Table 1) are shown for the entire sample. Overall, of the 28 correlations, 17 were significant ( p < .05), including five of seven correlations involving the NAAS, and five of seven correlations involving GPA. Moreover, another six correlations were somewhat significant ( p < .10). The numerous connections among the core variables suggest that few of these links are likely to be the result of type 1 errors. In fact, the mean magnitude of all correlations was r = .24, p < .05 (treating positive and negative correlations as equivalent).
Associations Among Affect-Related Variables (Hypothesis 1A) Students who had higher levels of emotion regulation were rated by teachers as having more positive moods (PANAS-Positive) and these students reported less negative academic affect (NAAS). In addition students who assessed themselves as having more negative overall moods (DOTS-R Mood) reported higher levels of negative academic affect, were rated by their teachers as less emotionally positive (PANAS-Positive), and were somewhat more likely to be rated by teachers as having negative moods (PANAS-Negative). Moreover, students who reported more negative academic affect were somewhat more likely to be rated by teachers as Table 1 Intercorrelations Among Cognition and Affect-Related Variables (n = 103) 1 Cognition-related GPA Achievement scores Academic competence Affect-related DOTS-R-emotion regulation PANAS – positive mood PANAS – negative mood DOTS-R – negative mood NAAS * p < .05. ** p < .01. *** p < .001. + p < .10.
2
– 0.61*** –
3
4
5
6
7
8
0.39*** 0.01
0.36***
0.28**
0.16 +
0.31**
0.30**
0.21 *
0.05
0.20 *
0.26**
0.44***
0.51***
–
0.17 *
+
0.13
0.19 *
0.12
0.10
0.23**
0.12
0.34***
0.14 +
0.13 +
0.14 +
–
0.23 *
0.28** 0.15 –
–
–
+
–
Gumora and Arsenio
405
having more negative and fewer positive moods (PANAS-Positive). Somewhat surprisingly, however, teachers’ rating of students’ positive moods were not significantly related to their rating of students’ negative moods. Associations Among Cognition-Related Variables Several significant connections emerged among the cognition-related variables. The connection between GPA and achievement tests supports the traditional view that students’ academic performance is related to their basic cognitive abilities (as measured by standardized achievement tests). Furthermore, students’ perceptions of their academic competence were linked with both their GPA and achievement scores. Specifically, students who had more positive conceptions of their academic ability had higher GPAs and higher achievement scores. Associations Between Affect-Related and Cognition-Related Variables (Hypothesis 1B) Teachers’ ratings of students’ positive and negative emotionality had a number of connections with cognition-related variables. Students with more positive moods had higher GPAs, achievement scores, and somewhat higher academic competence. In contrast, more negative moods were linked with lower GPAs and somewhat with lower academic competence. In terms of students’ rating of their own moods, students with more negative moods perceived themselves as less academically competent, had lower achievement scores, and somewhat lower GPAs. Finally, students’ negative academic affect was associated with all of the cognition-related variables. Students who perceived themselves as experiencing more negative affect during academic tasks had a poorer sense of their academic competence, lower achievement scores, and, finally, lower GPAs. Regression Analyses Predicting Students’ GPA (Hypotheses 2 and 4) Although students’ affective disposition and emotion regulation were connected with their GPA and other cognition-related variables, it was also expected that the affective variables would be a significant predictor of their grade point average over and above the contributions of the cognitive variables. Moreover, it was expected that academic affect would make a unique contribution to GPA beyond the contribution of the other affective variables. To address these issues, we conducted a hierarchical regression (see Table 2). The cognition-related variables (academic achievement, academic competence) were entered on Step 1. This was followed by Step 2, in which all affect-related variables except negative academic affect were entered (i.e., PANAS-Negative, PANAS-Positive, and DOTS-R Mood, DOTS-
406
Journal of School Psychology Table 2 Cognition and Affect-Related Predictors of Students’ GPA
Predictors added at
Step 1 B
Achievement 0.55 Step 1 (R2 = 0.424, Fchange = 36.83***) scores Academic 0.24 competence 2 DOTS-RStep 2 (R = 0.572, Fchange = 8.35***) emotion reg. PANASpositive mood PANASnegative mood DOTS-Rnegative mood Negative Step 3 (R2 = 0.574, academic affect Fchange = 4.59 *)
Step 2 t
B
Step 3 t
B
t
6.80***
0.52
7.34***
0.50
7.16***
2.98**
0.31
3.93**
0.23
2.67**
0.21
3.01**
0.23
3.21**
0.27
3.76***
0.26
3.64***
0.23
3.36**
0.27
3.85***
0.18
2.35 *
0.19
2.45 *
0.18
2.14 *
R2 and Fchange are shown for the entire step; the B and t are for each predictor. The DOTS-R measures were self-ratings and the PANAS measures were teacher ratings. * p < .05. ** p < .01. *** p < .001.
R Emotion Regulation), and finally by Step 3 which included only negative academic affect (NAAS). The variables were entered in this order to determine whether affect makes an additional contribution to predicting students’ GPA beyond the more traditional focus on cognitive abilities and competencies. Moreover, the NAAS was entered last to assess whether students’ negative academic affect, in particular, had an influence on their GPA over and above the more general affective contribution made by their moods. As can be seen in Table 2, students’ cognition-related abilities were a significant predictor of their GPA (Step 1), and academic achievement, academic competence were both independent significant predictors of GPA (i.e., after accounting for the effects of the other cognition-related variable). Furthermore, each of these cognition-related variables remained a significant predictor of students’ GPA when the affect-related variables were entered on Steps 2 and 3. The affect-related variables, however, also made a significant contribution to GPA beyond the cognitive contribution (Step 2). Finally, as hypothesized, the NAAS (Step 3) was still a significant predictor of GPA even after accounting for the influence of all other cognition and affect-related variables; compared to their peers, students who reported more negative emotions in relation to routine school tasks performed worse in school and this was not the result of underlying cognitive differences (at least as measured in this study). All of the variables continued to be significant predictors, even at the final step, suggesting that the contributions of these variables were relatively independent.
Gumora and Arsenio
407
It is also noteworthy that these findings were not influenced by problems with multicollinearity. None of the correlations approached the .90 criterion that is considered problematic, and this was confirmed by the collinearity diagnostics: no variable had more than one variance proportion greater than .50 (Tabachnik & Fidell, 2001). Overall, the affect and cognition-related variables accounted for nearly 60% of the variance in students’ grade point averages. Interaction of Emotion Regulation and Negative Academic Affect on GPA (Hypothesis 3) It was originally hypothesized that students’ negative academic affect might interact with their emotion regulation abilities in the prediction of their GPA. The emotion regulation negative academic affect interaction term, however, was not significantly related to GPA (r = .12), and consequently, there was no support for this hypothesis. Interaction of Negative Academic Affect and General Emotionality on GPA Although not originally hypothesized, it is possible that negative academic affect combines with more general aspects of students’ emotionality to influence their school performance. For example, a child who experiences a combination of both negative academic affect and more general negative affectivity might be at special risk for school difficulties. Consequently, separate mood negative academic affect interaction terms were created for each three mood-related variables (the PANAS-Positive, PANAS-Negative and DOTS-R Mood scales were not highly related enough to form a composite). No interaction term was significantly related to students’ GPA, and thus, there was no evidence for a combined influence of negative academic affect and aspects of students’ more general emotionality beyond the main effects of these affective variables.
DISCUSSION This study provides additional support for the role of socio-emotional factors in students’ school performance (Birch & Ladd, 1997; Pianta, Steinberg, & Rollins, 1995; Wentzel, 1996, 1999), while also clarifying some of the uniquely affective contributors (rather than relationships or goals) to that performance. More specifically, the present results indicate that both middle school children’s emotional dispositions and their academically related affect are connected with their school success. Even after controlling for the influence of other cognitive variables, including academic achievement and academic self-efficacy, students’ emotional dispositions
408
Journal of School Psychology
(teacher and self-rated) and academic affect predicted their academic performance. Moreover, negative academic affect could be distinguished from students’ affective dispositions, and negative academic affect was a unique predictor of students’ combined grade point average in language arts and mathematics classes beyond the contributions of all other variables assessed. Although there is previous evidence that anxiety interferes with test performance (McDonald, 2001), the present findings suggest that negative academic affect is also linked with more general aspects of academic success. Compared to their peers, students in this study who reported experiencing more frequent negative affect during a variety of academic tasks, including homework, class participation, quizzes, and tests had lower grades in their two core academic classes, language arts and mathematics. Moreover, it is important to recall that factor analyses revealed the presence of only a single factor in NAAS. The lack of separate factors for tests and quizzes or other academic tasks suggests that many students have a global affective reaction to academic work. How can this connection between academic affect and school performance be explained? Given the correlational nature of this study, there are several possible routes. For example, less cognitively proficient students may do worse in school, and, as a result, have more negative emotional reactions to any school-related task. Yet, even after accounting for the influence of several cognitive contributors in the present study (i.e., academic achievement and academic self-efficacy) academic affect still added significantly to the prediction of students’ school performance. Although it is still possible that IQ (or another unassessed cognitive influence) is responsible for these connections, this is less likely given the overlap between IQ and the cognitive measures that were included (e.g., academic achievement, Sattler, 1992). Another possible explanation for the link between academic affect and school performance stems from related research on the affective contributors to children’s social (rather than academic) competence. In that work, children’s negative emotionality has been found to predict lower levels of social competence both concurrently (Eisenberg et al., 1993) and longitudinally (Eisenberg et al., 1997). In a related vein, negative emotionality in the present study was also connected with lower levels of academic performance. And, once again, regression analyses suggested that this connection was not the result of obvious cognitive influences: different aspects of emotionality added to the prediction of students’ GPA over and above the influence of the cognitive variables, and each of the three dispositional affective measures was a significant predictor of GPA at all of the steps in the regression. Although this last explanation seems consistent with the present findings, at least two important unanswered questions remain: what are the origins
Gumora and Arsenio
409
and nature of these affective contributors; and how are academic affect and the other affective contributors related? Given the inclusion of the DOTS-R temperament measures, it could be argued that children’s psychophysiology (especially biological components of temperament) affects their academic achievement. The literature on children’s affective dispositions, however, indicates both that parental socialization plays a major role in these dispositions (e.g., see Denham, 1999, for a review), and that parental influences can even alter the physiology of children’s emotional responsiveness over time (Gottman, Katz, & Hooven, 1997). To the extent that ‘‘temperament’’ was assessed in the present research, it seems more reflective of broad behavioral tendencies than largely biologically determined patterns. Instead we would propose that students’ affective dispositions and academic affect largely reflect a combination of parental and school-related influences. Some cognitively able students, for example, may feel unwanted pressures to excel in school, perhaps because the work is considered excessive or difficult, or because it interferes with non-academic personal goals. Consequently, compared to their peers, these young adolescents may experience more negative affect both during academic tasks and in general. Over time, however, as that negative affect increasingly interferes with students’ motivation and ability to focus on schoolwork, and as their grades decline, the academic context may become especially linked with negative affect. It must be acknowledged that this implicit model extends well beyond the present findings. Yet, we believe that this model has implications for future research, which could bridge the present focus on affective dispositions with work on relational contributors to school performance. Wentzel (2002), for example, recently found that the classroom socialization practices of effective teachers are quite similar to parenting styles that are linked with higher student achievement. Specifically, teachers who combine positive emotionality (i.e., encouragement and avoiding negative feedback) and high expectations (i.e., a push to excel) have students with more academically oriented goals and higher school performance. A useful combination of this ‘‘relational’’ literature and the present study would involve examining the connections among students’ affective dispositions, adult – child relationships, and school performance. It might be expected, for example, that adults who are emotionally positive will promote students’ own positive emotionality in ways that make these students more responsive to both academic expectations and to the instructional process itself. That is, students’ emotionality may mediate and/or moderate the influence of adults on students’ school performance in at least two ways: (1) by increasing what Kochanska (in a different context) called ‘‘committed compliance’’ (Kochanka & Aksan, 1995),
410
Journal of School Psychology
students’ active embracing of adults’ academic expectations and (2) by permitting adults easier access to the students’ zone of proximal development (Vygotsky, 1978), and consequently, increasing both the quantity and quality of adult – student educational exchanges. More generally, additional research is needed to examine the connections among the affective, motivational, and relational contributors to school performance. Several unexpected findings and study limitations need to be mentioned at this point. Although, as expected, students’ emotion dispositions and emotion regulation both predicted their school performance, contrary to expectations, students’ emotion regulation did not interact with their emotional dispositions to make an additional significant contribution (as it sometimes has in studies of social competence, e.g., Eisenberg et al., 1997). This discrepancy could stem from the older age group used in this study compared to research on social competence, or perhaps from the present emotion regulation instrument, which involved self ratings rather than the teacher or parent ratings typically used to assess emotion regulation. It would be helpful if future studies of emotionality and school performance included multiple raters of students’ emotion regulation. Another unexpected finding was that emotion regulation was not related to students’ GPAs in the zero-order correlations, but it was a significant predictor of GPA at all steps of the regression analyses. In other words, emotion regulation acted as a suppressor variable: ‘‘a variable can be uncorrelated with the criterion and still improve prediction by being correlated with other predictors. . .(which) implies that such variables suppress criterion-irrelevant variance in other predictors’’ (Tzelgov & Henik, 1991, p. 524). Analyses revealed that this suppressor effect was not the result of high levels of multicollinearity between emotion regulation and any other predictors, but it is still unclear how to explain this suppression. Tzelgov and Henik argue that, although suppression is more common and conceptually important than is sometimes believed, it is especially important to replicate such effects before including them in psychological models. Finally, one important limitation of the present study involves the nature of the sample. The students were highly homogeneous in terms of race (all European –American) and religious/cultural background (all Jewish), and both of the schools they attended were very academically competitive. Although it is unclear how this combination of characteristics affected the present findings, it seems likely that these students experienced a much higher than average amount of pressure to do well in school. On one hand, this could have increased negative academic affect and perhaps students’ overall negativity relative to other middle school students. Alternatively, children with higher levels of negative academic affect and overall negativity may have been more likely to avoid these academically select schools. Additional research involving more diverse populations and school settings
Gumora and Arsenio
411
is needed to examine whether the present work can be generalized to other student populations. In conclusion, this study highlights the importance of affective contributors to middle school children’s academic performance. The present data suggest that although students’ emotion regulation, general affective dispositions, and academic affect are related to each other, each of them also makes a unique, significant contribution to students’ GPAs, over and above the influence of other cognitive contributors. Future studies can build on this work both by addressing some of the present study’s limitations, and by simultaneously assessing the connections among students’ school performance, relevant adult – child relationships, and various aspects of their affective dispositions and emotion regulation abilities. ACKNOWLEDGEMENTS This article is based on a dissertation by Gail Gumora submitted in partial fulfillment of the requirements for the doctoral degree at Ferkauf Graduate School of Psychology, Yeshiva University. We would especially like to thank the children and teachers who made this research possible. A version of this paper was presented at the 2001 meeting of the Society for Research in Child Development in Minneapolis, MN. REFERENCES Ahadi, S., & Rothbart, M. (1994). Temperament, development, and the big five. In C. Halverson Jr., G. Kohnstamm, & R. Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood ( pp. 189 – 207). Hillsdale, NJ: Erlbaum. Arsenio, W., Cooperman, S., & Lover, A. (2000). Affective predictors of preschoolers’ aggression and peer acceptance: direct and indirect effects. Developmental Psychology, 36, 438 – 448. Arsenio, W., & Killen, M. (1996). Preschoolers’ conflict-related emotions during peer disputes. Early Education and Development, 7, 43 – 57. Birch, S., & Ladd, G. (1997). The teacher – child relationship and children’s early school adjustment. Journal of School Psychology 35, 61 – 79. Bracken, B. (1992). Multidimensional self concept scale. Texas: Pro-Ed. Clark, L., & Watson, D. (1988). Mood and the mundane: relations between daily events and self-reported mood. Journal of Personality and Social Psychology, 54(2), 296 – 308. Cole, P., Michel, M., & Teti, L. (1994). The development of emotion regulation and dysregulation: a clinical perspective. In N. Fox (Ed.), The development of emotion regulation: biological and behavioral considerations, Monographs of the Society for Research in Child Development, vol. 59 (pp. 73 – 100) (2 – 3, Serial No. 240). Corno, L., & Rohrkemper, M. (1985). The intrinsic motivation to learn in classrooms. In C. Ames, & R. Ames (Eds.), Research on motivation, vol. 2, (pp. 53 – 90). New York: Academic Press. Denham, S. (1999). Emotional development in young children. New York: Guilford. Educational Testing Service (1995). Comprehensive Testing Program III Technical Report. (1995). Educational Testing Service, Princeton, NJ.
412
Journal of School Psychology
Eisenberg, N., & Fabes, R. (1992). Emotion, regulation, and the development of social competence. In M. Clark (Ed.), Review of personality and social psychology: vol. 14. Emotion and social behavior ( pp. 119 – 150). Newbury Park, CA: Sage. Eisenberg, N., Fabes, R., Bernzweig, J., Karbon, M., Poulin, R., & Hanish, L. (1993). The relations of emotionality and regulation to preschoolers’ social skills and sociometric status. Child Development, 64, 1418 – 1438. Eisenberg, N., Fabes, R., Murphy, B., Maszk, P., Smith, M., & Karbon, M. (1995). The role of emotionality and regulation in children’s social functioning: a longitudinal study. Child Development, 66, 1360 – 1384. Eisenberg, N., Fabes, R., Shepard, S., Murphy, B., Guthrie, I., Jones, S., Friedman, J., Poulin, R., & Maszk, P. (1997). Contemporaneous and longitudinal prediction of children’s social functioning from regulation and emotionality. Child Development, 68, 642 – 664. Fabes, R., & Eisenberg, N. (1992). Young children’s coping with interpersonal anger. Child Development, 63, 116 – 128. Ford, M. (1992). Motivating humans: goals, emotions, and personal agency beliefs. Newbury Park, CA: Sage. Gottman, J., Katz, L., & Hooven, C. (1997). Meta-Emotion: how families communicate emotinally. Mahwah, NJ: Erlbaum. Gumora, G. (1993). The construction of the negative academic affect scale (NAAS). Unpublished manuscript. Hambleton, R. (1987). Iowa test of basic skills: forms G and H. In D. Keyser, & R. Sweetland (Eds.), Test critiques. Missouri: Test Corporation of America. Hamre, B., & Pianta, R. (2001). Early teacher – child relationships and the trajectory of children’s school outcomes through eight grade. Child Development, 72, 625 – 638. Kochanka, G., & Aksan, N. (1995). Mother – child mutually positive affect, the quality of child compliance to requests and prohibitions, and maternal control as correlates of early internalization. Child Development, 66, 236 – 254. Lazarus, R. (1984). On the primacy of cognition. American Psychologist, 39, 124 – 129. Mayer, J., & Salovey, P. (1997). What is emotional intelliegence? In P. Salovey, & D. Sluyter (Eds.), Emotional development and emotional intelligence: educational implications ( pp. 3 – 25). New York: Basic Books. McDonald, A. (2001). The prevalence and effects of test anxiety in school children. Educational Psychology, 21(2), 89 – 102. Musch, J., & Broeder, A. (1999). Test anxiety versus academic skills: a comparison of two alternative models for predicting performance in a statistics exam. British Journal of Educational Psychology, 69(1), 105 – 115. Piaget, J. (1954/1981). In: T. Brown, & C. Kaegi (Eds.), Intelligence and affectivity: their relationship during child development. Palo Alto, CA: Annual Reviews Monographs. Pianta, R. (1999). Enhancing relationships between children and teachers. Washington, DC: American Psychological Association. Pianta, R., Steinberg, M., & Rollins, L. (1995). The first two years of school: teacher – child relationships and deflections in children’s classroom adjustment. Development and Psychopathology, 7, 295 – 312. Rothbart, M., & Bates, J. (1998). Temperament. In W. Damon, & N. Eisenberg (Eds.), Handbook of child psychology, 5th ed.: vol. 3: social, emotional, and personality development ( pp. 105 – 176). New York: Wiley. Sattler, J. M. (1992). Assessment of children: revised and updated (3rd ed.). San Diego: Jerome M. Sattler, Publisher. Tabachnik, L., & Fidell, B. (2001). Using multivariate statistics. Boston: Allyn and Bacon. Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research: definition, implications, and application. Psychological Bulletin, 109(3), 524 – 536. Vail, P. (1981). Emotion: the on off switch for learning. NJ: Modern Learning Press.
Gumora and Arsenio
413
Vygotsky, L. (1978). Mind in society. Cambridge, MA: Harvard University Press. Wentzel, K. (1996). Motivation in context: social relationships and achievement in middle school. In J. Juvonen, & K. Wentzel (Eds.), Social motivation: understanding children’s school adjustment ( pp. 226 – 247). New York: Cambridge University Press. Wentzel, K. (1999). Social-motivational processes and interpersonal relationships: implications for understanding motivation at school. Journal of Educational Psychology, 91, 76 – 97. Wentzel, K. (2002). Are effective teachers like good parents? Teaching styles and student adjustment in early adolescence. Child Development, 73, 287 – 301. Windle, M. (1992). Temperament and social support in adolescence: interrelations with depressive symptoms and delinquent behaviors. Journal of Youth and Adolescence, 21, 1 – 21. Windle, M., & Lerner, R. (1986). Reassessing the dimensions of temperamental individuality across the lifespan: the revised dimension of temperament survey (DOTS-R). Journal of Adolescent Research, 1(2), 213 – 230.