Beyond Positive and Negative Affect: Achievement Goals and Discrete Emotions in the
Elementary Physical Education Classroom
University of Leuven
University of Leuven
Yves VANDEN AUWEELE
University of Leuven
Corresponding author: firstname.lastname@example.org
Beyond Positive and Negative Affect 2
This study investigated the relations between achievement goals and elementary school pupils’ discrete emotions during Physical Education (PE) class. Based on Pekrun et al’s (2006) model, three classes of
emotions were considered: positive activating, negative activating, and negative deactivating emotions. After controlling for perceived competence and gender differences, task goals, as assessed by pupils or rated by PE teachers, were positively associated with positive activating emotions and negatively related to boredom, hopelessness, and anger. Ego goals exhibited a mixed picture as they were positively associated with pride and most of the negative emotions. Furthermore, the relations between ego goals and emotions were qualified by an ego by task goal and by an ego by perceived competence interaction suggesting that ego goals were especially linked to emotional maladjustment when task goals were low and when competence perceptions were high. The discussion underscores the emotional role of achievement goals in PE classes.
Keywords: Emotions; Achievement Goals; Physical Education; Elementary School Students;
Beyond Positive and Negative Affect 3
Beyond Positive and Negative Affect: Achievement Goals and Discrete Emotions in the
Elementary Physical Education Classroom
Achievement goals are considered as the aims or purposes for students’ task engagement in achievement settings. Achievement goals create the framework within which students interpret and react to learning outcomes and academic achievements (Dweck & Leggett, 1988; Nicholls, 1984). Two types of achievement goals have received most attention with the achievement goal tradition: Task goals and ego goals. Both types of goals differ in the way competence is defined. When individuals define competence with self-referenced or absolute standards, they are presumed to pursue task goals; their focus is mainly on developing their competence, on learning, and on mastering new skills. When individuals define competence on the basis of normative standards, they are said to pursue ego goals; their aim is to demonstrate their competence relative to others and to gain favorable judgments from others (Dweck, 1986; Nicholls, 1984). Task and ego achievement goals have been found to relate to a host of different cognitive, behavioral, and affective outcomes (Duda & Ntoumanis, 2003; Matos, Lens, & Vansteenkiste, 2007; Roberts, 2001).
Given the importance and the omnipresence of emotions in educational settings (Pekrun, Goetz, Titz, & Perry, 2002), several studies have investigated the relations of achievement goals to emotions and affect-related outcomes. Most, if not all previous studies have however, used composite measures of positive and negative affect, thus obscuring possible insightful information with respect to relationships between achievement goals and specific emotions. In an attempt to move beyond positive and negative affect, Pekrun and associates (2002; 2004; Pekrun, Elliot, & Maier, 2006) have called for studies on emotions in educational settings that move beyond the dichotomy of positive and negative affect. The aim of the present study was to meet this call, thereby examining the relationship between elementary school students’ achievement goals, as perceived by themselves and as rated by their PE
teachers, and a variety of specific emotions.
Achievement Goals and Emotions in the Physical Education Context
A number of previous studies in the PE context have examined the associations between the degree of being task and ego goal oriented and positive and negative affect. These studies have shown that whereas task goals are positively associated with positive affect and inversely related to negative affect, ego goals tend to be associated with a mixed pattern of emotions, as they relate both to positive and negative affect (for reviews see Biddle, Wang, Kavussanu, & Spray, 2003; Duda & Ntoumanis,
Beyond Positive and Negative Affect 4
2003; Ntoumanis & Biddle, 1999; Roberts, 2001; see also Linnenbrink & Pintrich, 2002). It should be noted, however, that the positive association between ego goals and positive affect is relatively weak and that this relationship may be moderated by factors such as the contexts (e.g. sport settings vs. physical education or PE settings) and the type of activities (Ntoumanis & Biddle, 1999).
In addition to examining these composite outcomes of positive and negative affect, a number of studies have paid attention to specific emotions. The one discrete positive emotion that has probably been studied most extensively within the achievement goal framework is enjoyment. Previous research has shown that task goals are positively associated with enjoyment – measured either as an index of
intrinsic motivation (e.g. Biddle, Soos, & Chatzisarantis, 1999; Conroy, Kaye, & Coatsworth, 2006) or directly as an affect-based outcome (e.g. Duda & Nicholls, 1992). This is obvious because endorsement of task goals is presumed to elicit more inner sources of satisfaction as task-involved individuals are more likely to perceive their task engagement as an end in itself rather than as a means to an end (Duda, 2001).The relationship of ego goals to enjoyment is less clear, with some studies finding evidence for a positive association (e.g., Wang & Liu, 2007), but other studies failing to confirm such a relationship (e.g. Duda & Nicholls, 1992; Spray, Biddle, & Fox, 1999).
Among the negative emotions that have been studied as discrete emotions in the PE contexts are boredom and anxiety. For instance, boredom has shown a negative relationship to task goals and a no relationship to ego goals in the PE settings (Duda & Nicholls, 1992) whereas cross sectional studies conducted in the general classroom have shown a positive, although small, relationship of boredom to ego goals (Jagacinski & Duda, 2001). Likewise, anxiety has been found to relate negatively to task goals and positively to ego goals before a competition event among junior athletes (Hall & Kerr, 1997).
In the present research, we aimed to examine the relationship between achievement goals and emotions in greater detail. In doing so, we relied on Pekrun and colleagues (2002; 2004) proposed taxonomy of emotions. According to Pekrun, four different dimensions need to be differentiated to meaningfully classify emotions, that is, (1) valence, which refers to whether an emotion is desired or not (i.e., positive or negative), (2) activation, which implies whether an emotion promotes or inhibits physiological activation or arousal (i.e., activating vs. deactivating), (3) object-focus, which signifies whether an emerging emotion is linked with a test, a learning or a task-related activity (i.e., activity-related emotions that are referring to a test or to a learning process) and (4) reference of time which is referred to whether an emotion is experienced before, during, or after a given event (i.e., anticipatory, concurrent, or retrospective emotions). To illustrate, enjoyment for the forthcoming PE class could be characterized as a positive, activating, anticipatory, class-related emotion according to the dimensions of valence, activation, time reference, and object-focus, respectively. In contrast, shame because of
Beyond Positive and Negative Affect 5
poor performance in a fitness test in a PE class could be considered as a negative, activating, retrospective, test-related emotion.
The association between achievement goals and different sets of emotions according to the valence, activation, and object focus dimensions has only recently received empirical attention in a study by Pekrun and colleagues (2006) among university students. When combining the dimensions of valence and activation, these authors created four aggregated categories of emotions, that is, (1) positive activating emotions such as enjoyment, hope, and pride, (2) negative activating emotions, such as anxiety, anger, and shame, (3) positive deactivating emotions, such as relief, and (4) negative deactivating emotions, such as hopelessness and boredom. Whereas some of these emotions are more process-oriented (e.g., enjoyment), others are more outcome-oriented (e.g., shame). Pekrun et al. (2006) examined the association between achievement goals and each of the eight discrete emotions. They found task goals to be positively associated with positive activating class-related emotions and to be negatively associated with negative activating and deactivating class-related emotions such as anger and boredom, respectively. In contrast, similar to previous research on the composite scores of positive and negative affect, an inconsistent and mixed pattern of relationships was obtained for ego goals, as these goals were related to enjoyment and pride (i.e. positive activating emotions) in Study 2 but not in Study 1 and shame (i.e. a negative deactivating emotion) in Study 1 but not in Study 2.
The present study aimed to build on the initial study on discrete emotions by Pekrun in several ways. First, we aimed to test the generalizability of the observed relationships between achievement goals and specific emotions as reported by Pekrun and colleagues (2006) by examining these associations in the PE context (rather than a regular school context) and in a sample of upper elementary school pupils (rather than university students).
Second, we aimed to verify these relationships by using two different sources of information for assessing pupils’ degree of achievement goal orientation. Our purpose was to cross-validate the associations between achievement goals and PE class-related emotions by recruiting a second source of information about pupils’ achievement goals – their PE teachers. In this way, we intended to
overcome the limitations derived from mono-method studies.
Third, we sought to examine the task and ego goals relationships to emotions after taking into account perceived competence. In addition, we intended to test the moderating role of perceived competence in the relationship between achievement goals and emotions (Biddle et al., 2003). This is a critical issue, because the association between ego goals and positive activating emotions, for instance,
Beyond Positive and Negative Affect 6
might be very different under conditions of low versus high perceived competence. For instance, the experience of high perceived competence might protect ego-oriented individuals from experiencing negative emotions.
Fourth, another relatively understudied issue that we aimed to explore concerned gender differences. As Biddle and associates (2003) have noted in their review, although most of the studies that investigated the relationships of achievement goals to affect included both genders, almost none of them had investigated the differences between males and females (see also Duda & Ntoumanis, 2003; Ntoumanis & Biddle, 1999). We sought to explore this issue without formulating specific hypotheses.
In short, given the paucity of studies on the relation between achievement goals and discrete emotions in PE classes, we examined these issues in the present study. In doing so, we did not only examine the relation between achievement goals and the aggregated types of distinguished emotions (i.e., positive activating, negative activating, and negative deactivating), but we also examined their relationships with each of the distinguished specific emotions within each aggregated category. These more detailed analyses were deemed insightful, given that recent work has shown that emotions belonging in the same type or class may differently relate to achievement goals (see Pekrun et al., 2006).
We formulated the following set of hypotheses. First, in line with previous research, we expected that, after controlling for levels of perceived competence, task goals would be positively related to the aggregated category and each of the specific positive activating emotions (i.e. enjoyment, hope, and pride), and that they would be inversely related to the aggregated category and each of the specific negative activating (i.e. anxiety, anger, and shame) and negative deactivating emotions (i.e. hopelessness and boredom).
Second, consistent with previous work documenting the positive – albeit weak – relationship
between ego goals and negative affect (Biddle, et al., 2003; Duda & Ntoumanis, 2003; Ntoumanis & Biddle, 1999; Roberts, 2001), we predicted that ego goals would be positively related to each of the two aggregated categories and the five specific negative emotions. Furthermore, we had more specific hypotheses with respect to the association between ego goals and positive activating emotions. Although ego goals might show a positive relationship to the aggregated category of positive activating emotions, we expected subtle differences in the relationships between ego goals and each of the three positive activating emotions. That is, unlike enjoyment and hope, we hypothesized that pride would be positively related to ego goals, as the experience and pursuit of pride may be one of the underlying reasons for which individuals may aspire a normative goal (Biddle, et al., 2003; Urdan & Mestas, 2006)
Third, taking into account the multiple goal perspective (Barron & Harackiewicz, 2001; Pintrich, 2000), we anticipated that the relationship of ego goals to both positive and negative emotions would
Beyond Positive and Negative Affect 7
be moderated by pupils’ degree of task orientation, such that ego goals would be related to positive emotions only under conditions of high task orientation and would be related to negative emotions only under conditions of low task orientation, suggesting that the positive association of ego orientation to positive emotions are limited to those being high on task orientation. Similarly, we hypothesized that the positive association between ego orientation and negative emotions would be off set for those high in task orientation, indicating that task orientation plays a buffering role against the experience of negative emotions when ego orientation is high.
Fourth, we equally predicted that ego goals would interact with perceived competence in the prediction of negative and positive emotions. Specifically, we hypothesized that the positive association between ego goals and negative emotions would be more salient for pupils with low competence perceptions because ego-oriented pupils with doubts about their abilities would be more likely to exhibit a maladaptive motivational pattern (Roberts, 2001) and hence to experience negative emotions such as hopelessness (Dweck & Leggett, 1988) or anxiety (Hall & Kerr, 1997). Similarly, we expected that ego goals would be positively related to positive activating emotions only when a tendency to outperform others would be coupled with a strong belief in one’s competencies to do well.
Fifth, we also expected that perceived competence would be positively related to emotional adjustment and inversely related to maladjustment because perceived competence is considered a significant determinant of motivational outcomes (Bandura, 1977; Elliot & Dweck, 2005). Finally, concerning the gender differences in the amount of experience emotions, we made no a priori hypotheses because this issue has received little prior research attention (Biddle et al., 2003; Duda & Ntoumanis, 2003; Ntoumanis & Biddle, 1999). Similarly, no predictions had been articulated about the possible moderating role of gender in the relation between achievement goals and PE class-related emotions.
Participants and Procedure
ththOne hundred and fifty-seven male and 162 female upper elementary school pupils (4 – 6
grades) from four public schools in Greece participated in the study. Data were collected during regular PE class hours in the spring semester with the aid of a research assistant who previously had obtained consent from parents, the school principals, and the PE teachers. The research assistant assured the pupils that their answers would be treated in a confidential way and made clear that they could quit the session at any time. No pupil denied participation. Children responded all questionnaires on a 5-point Likert-type scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). PE teachers (N = 2)
Beyond Positive and Negative Affect 8
provided an assessment of pupils’ achievement goals one week after the children had completed the
questionnaires. Each PE teacher provided a single assessment for each of his pupils. PE teachers were well aware of pupils’ class-related behavior as data collection took place at the end of the schooling year. Consequently, PE teachers had been in touch with the pupils for about seven months, and this allowed them to get to know their pupils sufficiently well to get insight in their achievement goal orientation.
Achievement goals. The Task and Ego Orientation Sport Questionnaire (TEOSQ; Duda &
Nicholls, 1992) was adjusted to Greek language and was used to assess pupils’ task and ego goal
orientation towards the PE. After reading the stem “I usually felt successful when …” pupils responded
to seven items that assessed their task goal orientation (e.g. “… when I learn a new skill by trying hard”) and to six items that assessed their ego goal orientation (e.g. “ … when my classmates cannot do as well as me”). The internal consistency for task and ego subscales was .72 and .78, respectively.
Perceived competence. Six items were adjusted from the Intrinsic Motivation Inventory (McAuley, Duncan, & Tammen, 1989) and were used to assess children’ competence perceptions
relative to the PE class-related activities. An example item was: “I think I am pretty good at PE task-
related activities”. Internal consistency of this scale was satisfactory (α = .75).
Class-Related Emotions. The Academic Emotions Questionnaire (AEQ; Pekrun et al., 2002)
was adjusted to assess eight discrete PE class-related emotions that are assumed to emanate before, during, or after PE class hours. Pupils were prompted to think about how they regularly felt before, during, and after the PE classes in the last few months. The eight discrete emotions that were assessed belong to three different higher order classes of emotions, that is (a) positive activating emotions, which included enjoyment (9 items, e.g. “I enjoy being in PE class”; α = .73), hope (8 items,
e.g. “I am full of hope”; α = .77), and pride (9 items, e.g. “I take pride in being able to keep up with the pace of PE activities”; α = .80); (b) negative physiologically activating emotions, which included anxiety (12 items, e.g. “I feel nervous in PE class”; α = .83), anger (9 items, e.g. “I feel frustrated in PE class”; α
= .87), and shame (11 items, e.g. “When I do anything in PE class I feel like I am making a fool of myself; α = .87); and (c) negative physiologically deactivating emotions, which included hopelessness (10 items, e.g. “I have lost all hope in doing well in PE class”; α = .87) and boredom (11 items, e.g.
“During PE class hours I am getting bored”; α = .93). Following Pekrun et al.’s (2002) taxonomy of
emotions, we created aggregated categories of emotions, thereby averaging enjoyment, hope, and pride into a positive activating emotions composite score (α = .84), anger, anxiety, and shame into a
Beyond Positive and Negative Affect 9
negative activating emotions composite score (α = .90) and boredom and hopelessness into a negative
deactivating emotions composite score (α = .89).
Rated achievement goals. Apart from assessing pupils’ achievement goals through self-reports
we asked from PE teachers to provide ratings of each of their pupils’ task and ego goals. To minimize
the burden on PE teachers, each achievement goals was assessed with only two rated items, ranging on a 5-point scale from 1 (Almost never) to 5 (Very often). Task goals were assessed with the following
items, “How often does the pupil aim at” (a) “learning”; and (b) “self-improving” (α = .85); Ego-goals
were assessed with the following two items: “How often does the pupil aim at (a) “outperforming his or
her classmates”; and (b) “being among the best in his or her class group” (α = .75).
Descriptive statistics and intercorrelations are shown in Table 1. As can be seen, pupils’ self-
reports of task goals, ego goals, and competence perceptions were positively interrelated and all of them were positively correlated with the composite score of positive activating emotions as well as with each of the three discrete positive emotions. The correlational pattern for the two achievement goals with respect to the negative emotions was, however, dissimilar. Whereas task goals were unrelated to the aggregated category of negative activating emotions, ego goals were positively correlated with them. Furthermore, task goals were negatively correlated with anger, whereas ego goals were positively correlated with each of the three negative activating emotions. With respect to the category of negative deactivating emotions, task goals were negatively correlated with the composite score and each of the two discrete negative deactivating emotions, whereas ego goals were positively related to the composite score and to the emotion of boredom. With the exception of the negative deactivating emotion composite score and hopelessness, perceived competence was unrelated to any of the measured negative emotions. Concerning teachers’ ratings of pupils’ achievement goals, task goals
were positively correlated with ego goals, but there was a lack of convergence between the self-reported achievement goals and the corresponding teacher rated achievement goals. Despite this, rated task goals were positively related to the aggregate score of positive activating emotions and to two of the discrete emotions, that is, enjoyment and hope. Also, rated task goals were negatively related to the negative activating and negative deactivating emotions composite scores and to the discrete emotions of anger, shame, and hopelessness. In contrast, rated ego goals showed no relation to any of the three composite scores or any of the eight discrete emotions.
Beyond Positive and Negative Affect 10
Because the achievement goals were only rated for a subsample of the pupils (N = 152), two
sets of multivariate analysis of variance (MANOVA) were conducted to test for any gender differences, one involving the self-reported measures and one involving the rated achievement goals. The first MANOVA indicated a general gender effect for the self-reported variables, F(11, 303) = 3.57, Wilk’s λ
= .89, p <.01. After adjusting the alpha level at the .003 according to the Bonferroni procedure, follow-up univariate analyses showed that girls scored lower on ego goals and on all the negative emotions but the emotion of shame. The second MANOVA was equally found to be significant, F(2, 149) = 12.75,
Wilk’s λ = .89, p <.01. The follow-up univariate analysis showed that boys were rated higher in ego goals (M = 3.42, SD = 0.13) than girls (M = 2.47, SD = 0.14) by their PE teachers F(1, 150) = 33.62, p
<.01. Considered together, MANOVA tests indicated that boys differed from girls with respect to various emotions suggesting that gender would have to be included as a covariate in the primary analyses.
Two series of hierarchical regression analyses were performed, one involving self-reported achievement goals and one involving rated achievement goals as independent variables. In each of these series of analyses, each of the emotions was regressed on achievement goals, perceived competence, and gender in Step 1, followed in Step 2 by all the possible two-way interactions between each of the predictors. The Interaction terms were created by multiplying the centered means of the respective predictors (Cohen, Cohen, West, & Aiken, 2003). Interaction terms that failed to add a
2statistically significant increment in R according to the F statistic were trimmed from the model.
Hierarchical regression analyses were conducted for the aggregated categories of emotions and for each discrete emotion separately. The latter was done to investigate whether any of the discrete emotions that are presumed to belong to the same category would exhibit a distinct pattern of relationships to the linear combination of achievement goals, perceived competence, and gender. If any significant interaction between predictors emerged, we split the data according to the moderator into
rdrdthththree subsets, yielding a low (<33 percentile), mean (33-66 percentile), and high (66 >) percentile
group. Then, we considered the relations of the predictor to the criterion under the low, middle, and high conditions of the moderator.
Self-reported Achievement Goals
The results of the hierarchical regression analyses can be found in Table 2. With respect to the aggregated category of positive activating emotions, the overall model was significant F (4, 311) =
252.17, p <01, R = .39. Task goals, perceived competence, and ego goals were positively related to positive activating emotions, although the association of ego goals was considerably smaller. Adding