lable at ScienceDirect

Psychology of Sport and Exercise 10 (2009) 336–343

Contents lists avai

Psychology of Sport and Exercise

journal homepage: www.elsevier .com/locate/psychsport

Beyond positive and negative affect: Achievement goals and discrete emotionsin the elementary physical education classroom

Athanasios Mouratidis a,*, Maarten Vansteenkiste b,*, Willy Lens c, Yves Vanden Auweele a

a Department of Kinesiology and Rehabilitation Studies, Tervuursevest 101, B-3001 Heverlee, Belgiumb Ghent University, Department of Psychology, H. Dunantlaan 2, Gent B-9000 Gent, Belgiumc University of Leuven, Department of Psychology, Tiensestraat 102, B 3000 Leuven, Belgium

a r t i c l e i n f o

Article history:Received 8 July 2008Received in revised form12 November 2008Accepted 13 November 2008Available online 24 November 2008

Keywords:EmotionsAchievement goalsPhysical educationElementary school students

* Corresponding authors.E-mail addresses: thanasis.mouratidis@gmail.com

vansteenkiste@ugent.be (M. Vansteenkiste), willy.lenyves.vandenauweele@faber.kuleuven.be (Y.V. Auweel

1469-0292/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.psychsport.2008.11.004

a b s t r a c t

Objectives: Based on the Achievement Goal perspective [Dweck, C. S., Leggett, E. L. (1988). A social-cognitive approach to motivation and personality. Psychological Review, 95, 256–273; Nicholls, J. G.(1984). Achievement motivation: conceptions of ability, subjective experience, task choice, and perfor-mance. Psychological Review, 91, 328–346] and on Pekrun et al.’s [(2004). Beyond test anxiety: devel-opment and validation of the test emotions questionnaire (TEQ). Anxiety, Stress, and Coping, 17, 287–316]model of discrete class-related emotions, this study investigated the relation of achievement goals todiscrete emotions in the Physical Education (PE) class.

Design: Cross-sectional.

Method: Participants were 319 Greek upper elementary school students who responded to a set ofquestionnaires assessing their achievement goal orientation, perceived competence, and the class-related emotions they experienced in the PE classes.

Results: Hierarchical regression analyses showed that, after controlling for perceived competence andgender differences, task goals were positively related to positive activating emotions and negativelyrelated to negative emotions. Ego goals exhibited a mixed picture as they were positively associated withpride and all the negative emotions. Furthermore, the relations between ego goals and emotions werequalified by an ego by task goal and by an ego by perceived competence interaction suggesting that egogoals were especially linked to emotional maladjustment when task goals were low and when compe-tence perceptions were high rather than low.

Conclusion: Unique associations between task and ego goals and specific emotions were found, renderinginsightful the disentanglement of positive and negative emotions into its components. The pursuit oftask goals might help to counteract the emotional burden associated with ego goal pursuit whereasfeeling competent to outperform when one endorses ego goals might perhaps put extra pressure on thepupils and, hence, have negative implications for their emotional adjustment.

� 2008 Elsevier Ltd. All rights reserved.

Achievement goals are considered as the aims or purposes forstudents’ task engagement in achievement settings and they createthe framework within which students interpret and react tolearning outcomes and academic achievements (Dweck & Leggett,1988). Two types of achievement goals have received most atten-tion within the achievement goal tradition: Task goals in whichcompetence is defined with self-referenced or absolute standardsand which focus on the development of competence, on learning,

(A. Mouratidis), maarten.s@psy.kuleuven.be (W. Lens),e).

All rights reserved.

and on mastering new skills and ego goals in which competence isdefined with normative standards and which focus on thedemonstration of competence mainly by outperforming others(Nicholls, 1984; Roberts, 2001). Task and ego goals have beenrelated to a host of different motivational outcomes (e.g., Duda &Ntoumanis, 2003; Matos, Lens, & Vansteenkiste, 2007).

Given the importance and the omnipresence of emotions ineducational settings (Pekrun et al., 2004), several studies haveinvestigated the relations of achievement goals to emotions andaffect-related outcomes. Most studies, however, have usedcomposite measures of positive and negative affect, thus obscuringpossible insightful information with respect to relationshipsbetween achievement goals and specific emotions (for reviews seeBiddle, Wang, Kavussanu, & Spray, 2003; Ntoumanis & Biddle,

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343 337

1999). In an attempt to move beyond positive and negative affect,Pekrun et al. (2004; see also Pekrun, Elliot, & Maier, 2006) havecalled for studies on emotions in educational settings that focus onseparate positive and negative emotions. We aimed to meet thiscall, thereby examining the relationship between elementaryschool students’ achievement goals and a set of specific positiveand negative emotions.

Achievement goals and emotions in the physical educationcontext

A number of previous studies in the PE context have examinedthe associations between the task and ego goals and positive andnegative affect. These studies have shown that whereas task goalsare positively associated with positive affect and inversely relatedto negative affect, ego goals tend to show a mixed pattern of rela-tions, as they relate both to positive and negative affect (see Biddleet al., 2003; Linnenbrink & Pintrich, 2002; cf. Ntoumanis & Biddle,1999). However, as Ntoumanis and Biddle (1999) showed, thepositive relation of ego goals to positive affect is relatively weak andmay be moderated by factors such as the context (e.g., sport settingvs. PE setting) and the type of activities.

Apart from positive and negative affect, a number of studies havepaid attention to specific emotions. The one discrete positiveemotion that has probably been studied most extensively within theachievement goal framework is enjoyment and previous researchhas shown that task goals are positively associated to enjoyment(e.g., Biddle, Soos, & Chatzisarantis, 1999; Duda & Nicholls, 1992).This is obvious because endorsement of task goals is presumed toelicit more inner sources of satisfaction as task-involved individualsare more likely to perceive their task engagement as an end in itselfrather than as a means to an end (Roberts, 2001). The relationship ofego goals to enjoyment is less clear, with some studies findingevidence for a positive association (e.g., Wang & Liu, 2007), but otherstudies failing to confirm such a relation (e.g., Duda & Nicholls,1992;Spray, Biddle, & Fox, 1999). The most studied discrete negativeemotions in the PE contexts are boredom and anxiety with boredomshowing a negative relationship to task goals and a no relationship toego goals in the PE settings (Duda & Nicholls,1992) and with anxietyshowing a negative relation to task goals and a positive one to egogoals (Hall & Kerr, 1997).

Pekrun’s taxonomy of emotions

In the present research, we aimed to examine the relationbetween achievement goals and emotions in greater detail byrelying on Pekrun et al.’s (2004) proposed taxonomy of emotions.According to Pekrun, emotions can be meaningfully classified alongfour different dimensions: (1) Valence, which refers to whether anemotion is desired or not (i.e., positive or negative); (2) activation,which implies whether an emotion promotes or inhibits physio-logical activation or arousal (i.e., activating vs. deactivating); (3)object-focus, which signifies whether an emerging emotion islinked with a test, a learning, or a task-related activity (i.e., activity-related emotions that are referring to a test or to a learningprocess); and (4) reference of time which is referred to whether anemotion 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 asa positive, activating, anticipatory, class-related emotion accordingto the dimensions of valence, activation, time reference, and object-focus, respectively. In contrast, shame because of poor performanceon 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 ofemotions according to the valence, activation, and object-focus

dimensions has only recently received empirical attention ina study by Pekrun et al. (2006) among university students. Whencombining the dimensions of valence and activation, these authorscreated three aggregated categories of emotions for class-relatedactivities, that is, (1) positive activating emotions such as enjoy-ment, hope, and pride, (2) negative activating emotions, such asanxiety, anger, and shame, and (3) negative deactivating emotions,such as hopelessness and boredom; positive deactivating emotionssuch as relief are not included in the class-related emotions as theyare particularly salient during test situations. Given that students’emotions were not assessed during a test situation in the presentresearch, we equally refrained from assessing students positivedeactivating emotions.

Pekrun et al. (2006) found task goals to be positively associatedwith positive activating class-related emotions and to be negativelyassociated with negative activating and deactivating class-relatedemotions. In contrast, and similar to previous research on thecomposite scores of positive and negative affect, an inconsistentand mixed pattern of relationships was obtained for ego goals.

Present research

Despite the 2� 2 conceptualization (Elliot & McGregor, 2001)according to which task and ego goals are further classifieddepending on how one’s competence relative to the demands ofa task is valenced (i.e., positively or negatively signifying, respec-tively, an approach or an avoidance orientation for either task orego goals), research conducted in sport and PE domain hasprovided some equivocal results within this 2� 2 achievement goalmodel (Duda, 2005). For instance, high intercorrelations betweengoals that share no common dimension have been reported (Guan,Xiang, McBride, & Bruene, 2006; Wang, Biddle, & Elliot, 2007) andsimilar relations of different achievement goals to outcomes wereobtained, suggesting a lack of functional discrimination among thefour proposed achievement goals (Carr, 2006; Guan et al., 2006;Wang et al., 2007). In this regard, Duda (2005) suggested that taskand ego goals along with the perceived competence may suffi-ciently explain the motivational phenomena in sport and PEsettings. Specifically, low competent ego-oriented individualswould be more likely to display an avoidance orientation, that is,a tendency to avoid showing incompetence, whereas the pursuit ofego goals among highly competent individuals would represent thetendency to outperform others. Thus, the present study indirectlyconsidered the pursuit of ego-approach, relative to ego-avoidancegoals, by examining whether ego-goal pursuit interacts withperceived competence in the prediction of emotions. Moreover, wedeemed it most instructive to examine possible differential rela-tions of both types of approach goals (i.e., task and ego) to one’semotional functioning, as these goals have been most heatedlydebated in the achievement goal literature (Harackiewicz, Barron,Pintrich, Elliot, & Thrash, 2002; Midgley, Kaplan, & Middleton,2001).

Following Duda’s (2005) suggestion, we aimed to build on theinitial study on discrete emotions by Pekrun in several ways. First,we tested the generalizability of the observed relationshipsbetween achievement goals and specific emotions (see Pekrunet al., 2006) by examining these associations in the PE context(rather than a regular school context) and in a sample of upperelementary school pupils (rather than university students). Second,we sought to examine the relation of task and ego goals to discreteemotions after taking into account perceived competence. Inaddition, we tested the moderating role of perceived competence inthe relationship between achievement goals and emotions (Biddleet al., 2003). The latter constitutes a critical issue, because theassociation between ego goals and positive activating emotionsmight be very different under conditions of low versus high

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343338

perceived competence (Duda, 2005, Roberts, 2001). Fourth, weexplored gender differences in the relation of task and ego goals toemotions as this issue has been relatively understudied (Biddleet al., 2003; see also Hyde & Durik, 2005). Finally, given the paucityof relevant studies on discrete emotions, we did not only examinethe relation between achievement goals and the aggregated typesof distinguished emotions (i.e., positive activating, negative acti-vating, and negative deactivating), but we also examined theirrelations with each of the distinguished specific emotions withineach aggregated category. These more detailed analyses weredeemed insightful, given that recent work has shown that emotionsbelonging in the same type or class may differently relate toachievement goals (see Pekrun et al., 2006).

We formulated the following hypotheses. First, in line withprevious research (Biddle et al., 2003; Duda & Ntoumanis, 2003),we expected that, even after controlling for perceived competence,task goals would be positively related to the aggregated categoryand each of the specific positive activating emotions (i.e., enjoy-ment, hope, and pride), and that they would be inversely related tothe 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; cf. Ntoumanis & Biddle, 1999), we predicted thatego goals would be positively related to each of the two aggregatedcategories and the five specific negative emotions. Furthermore, weexpected subtle differences in the relationships between ego goals andeach of the three positive activating emotions with pride exhibiting thestronger positive relation to ego goals, as the experience and pursuit ofpride may be one of the reasons underlying the pursuit of a normativegoal (Biddle et al., 2003; Urdan & Mestas, 2006).

Third, taking into account the multiple goal perspective (Barron& Harackiewicz, 2001), we anticipated that the relationship of egogoals to both positive and negative emotions would be moderatedby pupils’ degree of task orientation, such that ego goals would bedifferently related to positive and negative emotions under condi-tions of high and low task orientation, respectively. Specifically,taking into account the orthogonality of task and ego goals (Duda,2001; Roberts, 2001) and the studies which confirmed througha person-centered approach that individuals with a strong task andego goal orientation show adaptive motivational patterns relativeto those who score high only on the ego goal orientation (e.g., Wang& Biddle, 2001; Wang, Chatzisarantis, Spray, & Biddle, 2002), wehypothesized that ego goals would be positively related to positiveaffect and that they would show no relation to negative affect whentask goals are concurrently high.

Fourth, in line with Duda (2005), we equally predicted that egogoals would interact with perceived competence. We hypothesizedthat ego-oriented pupils with doubts about their abilities would bemore likely to experience negative emotions such as hopelessness(Dweck & Leggett, 1988) or anxiety (Hall & Kerr, 1997). Similarly, weexpected that ego goals would be positively related to positiveactivating emotions only when a tendency to outperform otherswould be coupled with a strong belief in one’s competencies to dowell (Roberts, 2001).

Fifth, we also expected a unique positive relation of perceivedcompetence to emotional adjustment and a negative one to mal-adjustment because perceived competence is considered a signifi-cant determinant of motivational outcomes (Bandura, 1977).Finally, concerning the gender differences in the amount of expe-rienced emotions and the moderating role of gender in the rela-tions of achievement goals to PE class-related emotions, we madeno a priori hypotheses because this issue has received little priorresearch attention (Biddle et al., 2003; Hyde & Durik, 2005; Ntou-manis & Biddle, 1999).

Method

Participants and procedure

Three hundred and nineteen (49.2% males) elementary schoolpupils (4th–6th grades) from four Greek public schools participatedin the study. Data were collected during regular PE class hours inthe spring semester with the aid of a research assistant whopreviously had obtained consent from parents, the school princi-pals, and the PE teachers. Pupils were informed that their answerswould remain confidential and that they could quit the session atany time. No pupil denied participation. Participants responded allquestionnaires on a 5-point Likert-type scale ranging from 1(Strongly Disagree) to 5 (Strongly Agree).

Measures

Achievement goalsThe Task and Ego Orientation Sport Questionnaire (TEOSQ;

Duda & Nicholls, 1992) was used to assess pupils’ task and ego goalorientation towards the PE. After reading the stem ‘‘I usually feltsuccessful when .’’ pupils responded to six items that assessedtheir task goal orientation (e.g., ‘‘. when I learn a new skill bytrying hard’’; a¼ .73) and to six items that assessed their ego goalorientation (e.g., ‘‘. when my classmates cannot do as well as me’’;a¼ .78). Confirmatory factor analyses (CFA) showed that a two-factor model comprising task and ego goals yielded a reasonable fitto the data S–Bc2 (64)¼ 79.72, p¼ .01, CFI¼ .944, SRMR¼ .068,RMSEA¼ .044, (90% Confidence Interval: .022–.063), with thestandardized loadings ranging from .44 to .75.

Perceived competenceSix items (a¼ .75) were adjusted from the Intrinsic Motivation

Inventory (McAuley, Duncan, & Tammen, 1989) to assess children’scompetence perceptions relative to the PE class-related activities.Students were asked to indicate their degree of agreement with theitems purporting to assess their competence perceptions (e.g., ‘‘Ithink I am pretty good at PE task-related activities’’).

Class-related emotionsFollowing a translation – back translation process by two

independent experts, as suggested by Hambleton (1994) weadapted the Academic Emotions Questionnaire (AEQ; Pekrun et al.,2004). The AEQ was used to assess eight discrete PE class-relatedemotions that are assumed to emanate before, during, or after PEclass hours. Pupils were prompted to think about how they regu-larly felt before, during, and after the PE classes in the last fewmonths. The eight discrete emotions that were assessed belong tothree different higher order classes of emotions, that is (a) positiveactivating emotions, which included enjoyment (7 items, e.g., ‘‘Ienjoy being in PE class’’; a¼ .74), hope (8 items, e.g., ‘‘Concerningthe PE class, I am full of hope’’; a¼ .77), and pride (9 items, e.g., ‘‘Itake pride in being able to keep up with the pace of PE activities’’;a¼ .80); (b) negative activating emotions, which included anxiety(12 items, e.g., ‘‘I feel nervous in PE class’’; a¼ .83), anger (9 items,e.g., ‘‘I feel frustrated in PE class’’; a¼ .87), and shame (11 items,e.g., ‘‘When I do anything in PE class I feel like I am making a fool ofmyself; a¼ .87); and (c) negative physiologically deactivatingemotions, which included hopelessness (10 items, e.g., ‘‘I have lostall hope in doing well in PE class’’; a¼ .87) and boredom (11 items,e.g., ‘‘During PE class hours I am getting bored’’; a¼ .93). FollowingPekrun et al.’s (2004) taxonomy of emotions, we created aggre-gated categories of emotions, thereby averaging enjoyment, hope,and pride into a positive activating emotions composite score(a¼ .84), anger, anxiety, and shame into a negative activatingemotions composite score (a¼ .90), and hopelessness and boredom

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343 339

into a negative deactivating emotions composite score (a¼ .89).Previous studies indicated adequate validity of the scale (Pekrunet al., 2004, 2006). Similarly, Confirmatory Factor Analysis of theAEQ with parceled items and as it has been adapted for thepurposes of the current study provided evidence for its factorialstructure: S–Bc2 (242, N¼ 315)¼ 364.98, p< .01, CFI¼ .967,SRMR¼ .081, RMSEA¼ .040, (90% Confidence Interval: .031–.048;range of standardized loading for first-order latent factors: .60–.94).

Results

Preliminary analyses

To examine the potential overlap between achievement goalsand specific emotions, we conducted a set of Principal ComponentAnalyses (PCA) with varimax rotation on the task goal items, egogoal items, and each of the separate emotion items; this allowed usto test the empirical distinctiveness of the two achievement goalfrom each of the studied discrete emotions. Results from PCAprovided evidence for such distinctiveness for all emotions anddescriptive statistics and intercorrelations of the variables of thestudy are shown in Table 1. Next, a multivariate analysis of variance(MANOVA) was conducted to test for any gender differences.MANOVA test revealed gender differences, F(11, 303)¼ 3.57, Wilk’sl¼ .89, p< .01, suggesting that gender can better be included asa covariate in the primary analyses.

Primary analyses

Having deleted the potential outliers and having checked theassumptions of normality, homoscedasticity, and multicollinearity(Cohen, Cohen, West, & Aiken, 2003), we performed a series ofhierarchical regression analyses in which, each emotion wasregressed on achievement goals, perceived competence, gender,and the task by ego goals interaction term in Step 1, followed in Step2 by all the remaining two-way interactions. The interaction termswere created by multiplying the centered means of the respectivepredictors and unless they would add a statistically significantincrement in R2 according to the F statistic they were trimmed fromthe model. The task by ego interaction term was retained in allanalyses to test the multiple goal perspective (Barron & Harack-iewicz, 2001; Duda, 2001; Wang & Biddle, 2001). Hierarchicalregression analyses were conducted for the aggregated categoriesof emotions and for each discrete emotion separately. The latterwas done to investigate whether any of the discrete emotions thatare presumed to belong to the same category would exhibita distinct pattern of relations to the achievement goals, perceivedcompetence, and gender. If any significant interaction between

Table 1Descriptive statistics and bivariate correlations between measured variables.

Variables M SD 1 2 3 4

1. Task goals 4.19 0.70 –2. Ego goals 2.92 0.98 .20** –3 Perceived competence 3.73 0.78 .36** .31** –4. Positive activating emotions 3.94 0.60 .50** .30** .50** –

5. Enjoyment 4.19 0.65 .48** .16** .42** .86**6. Hope 3.91 0.68 .47** .26** .45** .88**7. Pride 3.76 0.75 .36** .35** .43** .88**

8. Negative activating emotions 2.17 0.75 �.12* .18** �.08 �.16**9. Anxiety 2.33 0.74 �.03 .21** .00 .0110. Anger 1.94 0.87 �.19** .13** �.11 �.28**11. Shame 2.25 0.84 �.09 .17** �.09 �.15**

12. Negative deactivating emotions 1.88 0.86 �.21** .11* �.12* �.30**13. Hopelessness 1.95 0.88 �.16** .10 �.13* �.27**14. Boredom 1.81 0.93 �.24** .12* �.09 �.30**

Note: *p< .05; **p< .01.

predictors emerged, the relation between the predictor of primaryinterest and the emotion under consideration was plotted at high,medium, and low values of the moderator to provide a visualindication of the nature of the interaction. To further test whetherthe relation between the predictor and the dependent variable wasstatistically significant at high, mean, or low levels of the moder-ator, we performed a test of simple slopes (Cohen et al., 2003).

The results of the hierarchical regression analyses can be foundin Table 2. With respect to the aggregated category of positiveactivating emotions, the overall model was significant F (5,283)¼ 72.42, p< 01. Task goals and perceived competence werepositively related whereas ego goals showed no relation to positiveactivating emotions; contrary to our hypothesis, the ego by taskgoals interaction was nonsignificant. A significant two-way inter-action between ego goals and perceived competence was found inStep 2, F (6, 282)¼ 63.53, p< .01. This interaction is illustrated inFig. 1 and indicates, contrary to our expectations, that the relationbetween ego goals and positive activating emotions was significant(and positive) when perceived competence was low (b¼ .18; simpleslope test: t (282)¼ 3.17, p< .01) but not when it was medium(b¼ .08; simple slope test: t[282]¼ 1.71, p> .05) or high (b¼�.02;simple slope test: t[282]¼�0.83, p> .05).

With respect to the three discrete positive emotions, the patternof relations for task goals and perceived competence mirrored thepositive association that was obtained for the aggregated category.In contrast, but as predicted, ego goals were only positively relatedto pride, suggesting that ego goals may not be universally linkedwith all positive emotions. Furthermore, the ego goal by perceivedcompetence interaction that was obtained for the aggregatedcategory of positive activating emotions in Step 2 was found forenjoyment and pride but not for hope. The post-hoc probing test ofslopes revealed a significant relation of ego goals to pride whenperceived competence was low (b¼ .33; t[282]¼ 5.04, p< .01) ormedium (b¼ .20; t[282]¼ 3.67, p< .01) but not when it was high(b¼ .06; t[282]¼ 0.31, p> .05). Unlike pride however, no significantrelation of ego goals to enjoyment was found either whenperceived competence was low b¼ .06, t[282]¼ 1.21, p> 05,medium b¼�.02, t[282]¼�0.44, p> .05, or high b¼�.10,t[282]¼�1.78, p¼ .08 suggesting that the interaction between egogoals and perceived competence for enjoyment was carried, by thedifferences between the slopes of low and high levels of perceivedcompetence. Table 2 also shows gender differences for positiveactivating emotions and for hope but not for enjoyment or pride –although the pattern of relation was similar for all the threeemotions. This finding indicates that relative to boys, girls tended toreport less hope but similar amounts of enjoyment and pride.

With respect to the aggregated category of negative activatingemotions, the overall model was significant F (5, 286)¼ 13.22,

5 6 7 8 9 10 11 12 13 14

–.66** –.62** .64** –�.34** �.09 �.01 –�.18** .07 .12* .91** –�.42** �.20** �.14* .92** .74** –�.32** �.11* .01 .93** .78** .76** –�.45** �.19** �.16** .91** .79** .90** .80** –�.41** �.18** �.15** .88** .78** .84** .79** .95** –�.45** �.20** �.15** .85** .71** .87** .74** .95** .80** –

Table 2Beta-coefficients of hierarchical multiple regression analyses for emotions based on pupils’ achievement goals.

Predictors PAE Enjoyment Hope Pride NAE Anxiety Anger Shame NDAE Hopelessness Boredom

Step 1Task goals .43** .49** .40** .24** �.34** �.20** �.41** �.28** �.41** �.33** �.44**Ego goals .08 �.02 .01 .20** .25** .23** .18** .25** .16** .13* .18**Perceived competence .41** .34** .38** .35** �.08 �.03 �.07 �.10 �.15* �.15* �.13*Gender �.09* �.07 �.10* �.05 �.18** �.18** �.22** �.09 �.16** �.17** �.13*Task goals� ego goals .02 .04 .05 �.04 �.13* �.10 �.12* �.14* �.16** �.12* �.19**Adjusted R2 .55** .48** .45** .36** .17** .09** .23** .12** .24** .18** .25**

Step 2Task goals .44** .50** .41** .25** �.34** �.21** �.42** �.28** �.41** �.34** �.44**Ego goals .08 �.02 .01 .20** .25** .24** .19** .25** .17** .13* .18**Perceived competence .41** .34** .38** .36** �.08 �.03 �.07 �.10 �.16* �.15* �.13*Gender �.09* �.08 �.10* �.06 �.17** �.17** �.21** �.09 �.16** �.16** �.13*Ego goals� Task goals .09 .10 .08 .05 �.21** �.19** �.18** �.18** �.24** �.18** �.27**Ego goals� Competence �.14** �.11* �.06 �.18** .14* .17* .13* .08 .14* .11 .15*

F change in R2 8.93** 4.68 1.41 10.99** 5.11** 6.73* 4.84* 1.58 5.37** 3.27 6.10*Adjusted R2 .57** .49** .45** .38** .19** .11** .24** .12** .25** .18** .27**

Note. PAE¼ Positive activating emotions composite score; NAE¼Negative activating emotions composite score; NDAE¼Negative deactivating emotions composite score.Gender was dummy coded (0¼males, 1¼ females). *p< .05. **p< .01.

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343340

p< .01. As we hypothesized, task goals were negatively related andego goals were positively related to negative activating emotions. Incontrast, perceived competence was unrelated to negative acti-vating emotions. Also, the task by ego interaction was significant.This interaction is displayed in Fig. 2. As can be noticed ego goalswere positively related to negative activating emotions when taskgoals were low (b¼ .33; t[286]¼ 3.76, p< .01) or medium (b¼ .25;t[286]¼ 3.95, p< .01), but not when they were high (b¼ .16;t[286]¼ 0.34, p> .05). Thus, it was the combination of high egogoals with the absence of a strong task goal orientation that wasassociated with the experience of more negative activatingemotions. In addition, an ego by perceived competence interactionwas found in Step 2, F (6, 285)¼ 12.03, p< .01. Fig. 3 shows,contrary to our expectations, that the relation between ego goalsand negative activating emotions was significant when perceivedcompetence was high (b¼ .34; t[285]¼ 4.42, p< .01) or medium(b¼ .25; t[285]¼ 4.09, p< .01) and not when it was low (b¼ .16;t[285]¼ 1.43, p> .05).

Separate regression analyses for each of the three negativeactivating emotions showed that task goals were negatively relatedand ego goals were positively related to anger, anxiety, and shame;perceived competence showed no relation to any of the threenegative activating emotions. In addition, and similar to the nega-tive activating emotions composite score, a task by ego interactionwas found for anxiety (but only in Step 2), anger and shame. Egogoals were positively related to anxiety, anger, and shame when

Fig. 1. The ego by perceived competence interaction for positive activating emotions.

task goals were low (b¼ .36; t[285]¼ 3.87, p< .01, b¼ .26;t[286]¼ 3.06, p< .01, and b¼ .34; t[285]¼ 3.74, p< .01, respec-tively) or medium (b¼ .24; t[285]¼ 3.66, p< .01, b¼ .18;t[286]¼ 3.02, p< .01, and b¼ .25; t[285]¼ 3.88, p< .01, respec-tively) but not when they were high (b¼ .11; t[285]¼ 0.35, p> .05,b¼ .11; t[286]¼ 0.81, p> .05, and b¼ .16; t[285]¼ 1.32, p> .05,respectively).

Moreover, the ego by perceived competence interaction thatwas found for the negative activating emotions composite scorewas found also for anxiety and anger but not for shame. Tests ofmain slopes showed that ego goals were positively related toanxiety and anger when perceived competence was high (b¼ .36;t[285]¼ 4.34, p< .01 and b¼ .28; t[285]¼ 3.71, p< .01, respec-tively) or around the mean (b¼ .24; t[285]¼ 3.66, p< .05 andb¼ .19; t[285]¼ 3.16, p< .01, respectively) but not when it was low(b¼ .11; t[285]¼ 0.88, p> .05 and b¼ .09; t[285]¼ 0.79, p> .05,respectively). Also, gender was negatively related to negative acti-vating emotions (as well as to anxiety and anger) suggesting thatfemales were less likely to experience those negative emotionsthan males for PE classes.

With respect to the aggregated category of negative deactivatingemotions, the overall model was significant F (5, 279)¼ 19.15,p< .01. As expected, task goals were negatively related and egogoals were positively related to negative deactivating emotions (i.e.,hopelessness and boredom). In addition, the significant task by egointeraction suggested that ego goals were positively related tonegative deactivating emotions when task goals were low (b¼ .27;t[279]¼ 3.42, p< .01) or around the mean (b¼ .16, t¼ 2.72, p< .01),but not when they were high (b¼ .06; t[279]¼�0.23, p> .05). Thesame pattern of interactions was found for both hopelessness andboredom when these emotions were separately considered (seeTable 2) as ego goals were positively related to hopelessness andboredom when task goals were low (b¼ .20; t[279]¼ 2.48, p< .05and b¼ .30; t[278]¼ 3.83, p< .01, respectively) or medium (b¼ .13;t[279]¼ 2.02, p< .05 and b¼ .18; t [278]¼ 2.96, p< .01, respec-tively) and not when they were high (b¼ .05; t[279]¼�0.06,p> .05 and b¼ .06; t[278]¼�0.39, p> .05, respectively). Further-more, a significant ego by perceived competence interaction wasfound for negative deactivating emotions in Step 2 F(6, 278)¼ 17.10,p< .01 suggesting that ego goals were unrelated to negativedeactivating emotions when perceived competence was low(b¼ .07; t[278]¼ 0.46, p> .05) but not when it was high (b¼ .27;t[278]¼ 3.58, p< .01) or around the mean (b¼ .17; t[278]¼ 2.84,p< .01). When hopelessness and boredom were separatelyregressed, the ego by perceived competence interaction was found

Fig. 2. The task by ego interaction for negative activating emotions.

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343 341

only for boredom F(6, 277)¼ 17.98, p< .01with ego goals beingpositively related to boredom when perceived competence washigh (b¼ .29; t[277]¼ 3.87, p< .01) or around the mean (b¼ .18;t[287]¼ 3.11, p< .01) and not when it was low (b¼ .07;t[287]¼ 0.55, p> .05).

Significant gender differences were also found for both negativedeactivating emotions with females reporting less hopelessnessand boredom than males. As Table 2 demonstrates, perceivedcompetence was negatively associated with negative deactivatingemotions composite score as well as with hopelessness andboredom.

Discussion

This paper investigated the pattern of associations betweenachievement goals, perceived competence, and discrete PE class-related emotions among upper elementary school students. In linewith previous research (Ntoumanis & Biddle, 1999), task goals werefound to relate positively to the aggregated category of positiveactivating emotions and to each of the three separate positiveemotions, that is, enjoyment, hope, and pride. Notably, and similarto previous findings in the PE (Vlachopoulos & Biddle, 1997) andeducational contexts (Pekrun et al., 2006), this positive associationwas found to hold over and above the positive relation of perceivedcompetence to the three positive activating emotions.

Therefore, the current results provided further evidence for thegeneralizability of the claim that task goals and positive emotionsare closely linked regardless of the level of perceived (or actual)competence. Apparently, task-oriented pupils are more likely to

Fig. 3. The ego by perceived competence interaction for negative activating emotions.

feel hopeful, perhaps because the PE classes offers them newopportunities for skill-development, and proud, perhaps becausethey have learned some new exercise skills; they also seem to enjoythe learning process more. Future research might want to examinein more detail the precise sources of the positive emotional expe-riences that are associated with being task-oriented. Furthermoreand in line with our expectations, task goals were found to relatenegatively to negative activating emotions. These findings partlyreplicate previous studies which showed a negative relationbetween task goals and anger (Pekrun et al., 2006), although otherstudies (including Pekrun et al., 2006), failed to find consistentrelations between ego goals and anxiety (e.g., Biddle et al., 2003;Hall & Kerr, 1997) or shame.

The negative relationship between task goals and negativedeactivating emotions emerged for the composite score and thetwo specific negative deactivating emotions, that is, boredom andhopelessness. Similar negative relationships between task goalsand boredom were found among high school students (Duda &Nicholls, 1992) and college students (Spray et al., 1999) partici-pating in PE activities. We should note, however, that the negativerelationship of task goals to hopelessness in the current study wasnot shown in Pekrun et al.’s (2006) study but was consistent to theNtoumanis and Biddle’s (1999) meta-analysis in which a negativerelation between task goals and low-arousal negative emotions wasfound. In short, the pattern of emotions associated with task goalsprovides further evidence of the adaptive role of task goals onpupils’ emotional adjustment, which is considered important forsustaining intrinsic motivation (Roberts, 2001).

Consistent with our hypothesis and previous findings in thesport and the PE domain (Biddle et al., 2003; Ntoumanis & Biddle,1999), ego goals were positively associated with all the negativeactivating (i.e., anger, shame, and anxiety) and negative deacti-vating emotions (i.e., hopelessness and boredom). The pursuit ofego goals thus seems to be associated with both a high-arousingand low-arousing negative emotional involvement. It should behighlighted, however, that the association between ego goals andnegative emotions was qualified by an ego by task goal interaction.This interaction suggests that pupils characterized by thecombined pursuit of high ego and low task goals were those whoexhibited the least favorable emotional pattern. In contrast, thepositive association between ego goals and negative emotions wasoffset when pupils endorsed task goals, suggesting that task goalsplay a buffering role against the experience of negative emotionsassociated with ego goals in the PE class. This finding is in line withthe multiple goal perspective (Barron & Harackiewicz, 2001) andshows that it is the absence of task goals in combination withstrong ego goal endorsement which may make pupils vulnerableto emotional adjustment during schooling or sport activities(Roberts, 2001).

The disentanglement of the composite score of positiveemotions into its components was insightful, as it was found thatthe positive link of ego goals was true only for the emotion of pride(see also Pekrun et al., 2006). Perhaps, one could conjecture thatsimilar relations would have been revealed in past research ifdifferent emotions had been separately analyzed (Vlachopoulos &Biddle, 1997). It remains to be investigated, however, why ego-oriented persons might experience more pride. Perhaps, attainmentof pride may be one of the underlying reasons for the adoption ofego goals, such that one anticipates feeling proud of oneself if onesucceeds in outperforming others (Atkinson, 1964; Urdan & Mestas,2006; Vansteenkiste et al., 2008). Alternatively, given that thepresent findings are correlational, the reciprocal relationshipbetween ego goals and pride may denote that pupils who feel proudabout previous accomplishments are more likely to adopt ego goals,such that the positive association between ego goals and pride israther a reflection of previous success (Brophy, 2005). Future

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343342

longitudinal research needs to sort out how pride and ego goals arerelated to each other over time.

The argument that pride over successful performance may leadone to adopt ego goals might be less likely, however, as ego goalswere found to relate only to pride when perceived competence waslow rather than high. This interaction was contrary to our predic-tion that the adaptive emotional role of ego goals would only beevidenced when competence perceptions are high (Dweck & Leg-gett, 1988). Similarly, perceived competence failed to buffer againstthe experience of negative emotions when ego goals were high asego goals were positively related to most negative emotions whenperceived competence was high.

Although puzzling, the interaction between ego goals andperceived competence emerged quite consistently across fewpositive and negative emotions. Although previous studies showedthat the relation of ego goals to affect (Vlachopoulos & Biddle, 1997)was not moderated by perceived competence, no study to dateshowed that ego goals were more predictive for emotionaladjustment when competence was low rather than high. Thesefindings are thus in need of replication. Perhaps, the endorsementof high ego goals makes one more prone for emotional malad-justment when one holds high rather than low levels of compe-tence because the self-worth of highly competent individuals ismore at stake during their activity engagement; most likely,because they are aware of their competencies to outperform others,they feel a stronger pressure to do better than their peers. Such self-worth concerns might be experienced as emotionally draining asthey put more pressure on high competence individuals to do wellduring the PE classes. Future research might do well in tearing apartpupils’ ego goals from the reasons underlying the pursuit of thesegoals, (Elliot, 2005). Such empirical disentanglement might appearfruitful in better understanding under which circumstances egogoals are related to particular emotional experiences.

Finally, two other main effects are noteworthy. First, consistentwith the a priori hypothesis, perceived competence was positivelyrelated to emotional adjustment indicating the motivationalimportance of the subjective feelings of competence (Bandura,1977). It should be underscored however, that the relation ofperceived competence to emotions is much more complicated. Forinstance, highly competent pupils may attribute poor performanceor failure to external causes such as others or bad luck and thisattribution may lead to experience an emotion of anger (i.e., another-directed negative activating emotion). Others may attributelow performance to internal, controllable causes such as lack ofeffort and this attribution may lead them to experience an emotionof shame, (i.e., an inner-directed negative activating emotion)(Weiner, 1986). Second, unlike past studies which reported nogender differences (Biddle et al., 2003), the present study demon-strated that females reported less negative emotions but also lesshope than males. Perhaps, girls’ tendency to report less negativeemotions than boys may reflect females’ socialization experiencesand their ability to deal more effectively with negative emotions inunfavorable situations (Davis, 1995).

Limitations

Several limitations should be kept in mind when interpretingthe current findings. First, as the data are correlational, noconclusion can be made regarding the causality of the observedrelation as emotional adjustment can operate as an antecedent ofoptimal motivation rather than vice versa (see Linnenbrink & Pin-trich, 2002). Second, despite our endeavor to provide discriminantvalidity between achievement goals and emotions, some concep-tual overlap might still drive the strength of the observed associ-ations in the current study. Third, although most of the detectedassociations were theoretically sound or could be meaningfully

explained post-hoc, it is still possible that some of these relation-ships capitalized on chance. Finally, compared to negativeemotions, much more variance was explained for positiveemotions, perhaps because the achievement goals are positivelyformulated just as the positive activating emotions are.

Conclusion

The present research was set up in part to investigate whetherthe examination of the different discrete emotions rather thancollapsing across negative and positive emotions to formcomposite scores, as has usually been the case, would provide onegreater insight in the emotional experiences that go along withpursuing task and ego goals. We believe it does, as ego goals wereuniquely positively related to pride (but not to enjoyment andhope), as perceived competence was negatively related to negativedeactivating but not to negative activating emotions and as egogoals and perceived competence interacted in the prediction ofenjoyment and pride (but not hope) and of anxiety, anger, andboredom (but not shame or hopelessness). Hence, it would bemore instructive if researchers and education practitioners wouldavoid collectively considering all positive and negative emotions asequivalents.

References

Atkinson, J. (1964). An introduction to motivation. New York: Van Nostrand Reinhold.Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change.

Psychological Review, 84, 191–215.Barron, K., & Harackiewicz, J. (2001). Achievement goals and optimal motivation:

testing multiple goal models. Journal of Personality and Social Psychology, 80,706–722.

Biddle, S. J. H., Wang, C. K. J., Kavussanu, M., & Spray, C. M. (2003). Correlates ofachievement goal orientations in physical activity: a systematic review ofresearch. European Journal of Sport Sciences, 3, 1–20.

Biddle, S., Soos, I., & Chatzisarantis, N. (1999). Predicting physical activity intentionsusing goal perspectives and self-determination theory approaches. EuropeanPsychologist, 4, 83–89.

Brophy, J. (2005). Goal theorists should move on from performance goals. Educa-tional Psychologist, 40, 167–176.

Carr, S. (2006). An examination of multiple goals in children’s physical education:motivational effects of goal profiles and the role of perceived climate inmultiple goal development. Journal of Sports Sciences, 24, 281–297.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd Ed.). Mahwah, NJ: Erlbaum.

Davis, T. L. (1995). Gender differences in masking negative emotions: ability ormotivation? Developmental Psychology, 31, 660–667.

Duda, J. L. (2001). Achievement goal research in sport: pushing the boundaries andclarifying some misunderstandings. In G. C. Roberts (Ed.), Advances in motiva-tion in sport and exercise (pp. 129–182). Champaign, IL: Human Kinetics.

Duda, J. L. (2005). Motivation in sport: the relevance of competence and achieve-ment goals. In A. Elliot, & C. Dweck (Eds.), Handbook of competence and moti-vation (pp. 318–335). New York: Guilford.

Duda, J. L., & Nicholls, J. G. (1992). Dimensions of achievement motivation inschoolwork and sport. Journal of Educational Psychology, 84, 290–299.

Duda, J. L., & Ntoumanis, N. (2003). Correlates of achievement goal orientations inphysical education. International Journal of Educational Research, 39, 415–436.

Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation andpersonality. Psychological Review, 95, 256–273.

Elliot, A. J. (2005). A conceptual history of the achievement goal construct. InA. J. Elliot, & C. S. Dweck (Eds.), Handbook of competence and motivation (pp. 52–72). New York: Guilford.

Elliot, A. J., & McGregor, J. A. (2001). A 2 � 2 achievement goal framework. Journal ofPersonality and Social Psychology, 80, 501–519.

Guan, J., Xiang, P., McBride, R., & Bruene, A. (2006). Achievement goals, social goals,and students’ reported persistence and effort in high school physical education.Journal of Teaching in Physical Education, 25, 58–74.

Hall, H. K., & Kerr, A. W. (1997). Motivational antecedents of precompetitive anxietyin youth sport. The Sport Psychologist, 11(1), 24–42.

Hambleton, R. K. (1994). Guidelines for adapting educational and psychologicaltests: a progress report. European Journal of Psychological Assessment, 10,229–244.

Harackiewicz, J. M., Barron, K. E., Pintrich, P. R., Elliot, A. J., & Thrash, T. M. (2002).Revision of achievement goal theory: necessary and illuminating. Journal ofEducational Psychology, 94, 638–645.

Hyde, J., & Durik, A. (2005). Gender, competence, and motivation. In A. Elliot, &C. Dweck (Eds.), Handbook of competence and motivation (pp. 375–391). NewYork: Guilford.

A. Mouratidis et al. / Psychology of Sport and Exercise 10 (2009) 336–343 343

Linnenbrink, E. A., & Pintrich, P. R. (2002). Achievement goal theory and affect: anasymmetrical bidirectional model. Educational Psychologist, 37, 69–78.

Matos, L., Lens, W., & Vansteenkiste, M. (2007). Achievement goals, learning strat-egies and language achievement among Peruvian high school students. Psy-chologica Belgica, 47, 51–70.

McAuley, E., Duncan, T., & Tammen, V. V. (1989). Psychometric properties of theintrinsic motivation inventory in a competitive sport setting: a confirmatoryfactor analysis. Research Quarterly for Exercise and Sport, 60, 48–58.

Midgley, C., Kaplan, A., & Middleton, M. (2001). Performance-approach goals: goodfor what, for whom, under what circumstances, and at what cost? Journal ofEducational Psychology, 93, 77–86.

Nicholls, J. G. (1984). Achievement motivation: conceptions of ability, subjectiveexperience, task choice, and performance. Psychological Review, 91, 328–346.

Ntoumanis, N., & Biddle, S. J. (1999). Affect and achievement goals in physicalactivity: a meta-analysis. Scandinavian Journal of Medicine and Science in Sports,9, 315–332.

Pekrun, R., Elliot, A. J., & Maier, M. A. (2006). Achievement goals and discreteachievement emotions: a theoretical model and prospective test. Journal ofEducational Psychology, 98, 583–597.

Pekrun, R., Goetz, T., Perry, R. P., Kramer, K., Hochstadt, M., & Molfenter, S. (2004).Beyond test anxiety: development and validation of the test emotions ques-tionnaire (TEQ). Anxiety, Stress, and Coping, 17, 287–316.

Roberts, G. C. (2001). Understanding the dynamics of motivation in physicalactivity: the influence of achievement goals on motivational processes. InG. Roberts (Ed.), Advances in motivation in sport and exercise (pp. 1–50).Champaign, IL: Human Kinetics.

Spray, C. M., Biddle, S. J. H., & Fox, K. R. (1999). Achievement goals, beliefs about thecauses of success and reported emotion in post-16 physical education. Journal ofSports Sciences, 17, 213–219.

Urdan, T., & Mestas, M. (2006). The goals behind performance goals. Journal ofEducational Psychology, 98, 354–365.

Vansteenkiste, M., Smeets, S., Soenens, B., Lens, W., Matos, L., & Deci, E.L. Newinsights into the performance-approach goal controversy: the differentialrelations of autonomous and controlled regulation of performance-approachgoals to educational outcomes. Paper presented in the AERA annual meeting,2008, March, New York.

Vlachopoulos, S., & Biddle, S. J. H. (1997). Modeling the relation of goal orientationto achievement-related affect in physical education: does perceived abilitymatter? Journal of Sport & Exercise Psychology, 19, 169–178.

Wang, C. K., Biddle, S. J. H., & Elliot, A. J. (2007). The 2� 2 achievement goal frameworkin a physical education context. Psychology of Sport and Exercise, 8, 147–168.

Wang, C. K., Chatzisarantis, L. D., Spray, C. M., & Biddle, S. J. (2002). Achievementgoal profiles in school physical education: differences in self-determination,sport ability beliefs, and physical activity. British Journal of EducationalPsychology, 72, 433–445.

Wang, C. K. J., & Biddle, S. J. H. (2001). Young people’s motivational profiles in physicalactivity: a cluster analysis. Journal of Sport & Exercise Psychology, 23, 1–22.

Wang, C. K. J., & Liu, W. C. (2007). Promoting enjoyment in girls’ physical education:the impact of goals, beliefs, and self-determination. European Physical EducationReview, 13, 145–164.

Weiner, B. (1986). An attributional theory of motivation and emotion. New York:Springer-Verlag.