683

The British Psychological Society

British Journal of Educational Psychology (2007), 77, 683–702 q 2007 The British Psychological Society

www.bpsjournals.co.uk

Comparing dichotomous and trichotomous approaches to achievement goal theory: An example using motivational regulations as outcome variables Vassilis Barkoukis1*, Nikos Ntoumanis2 and Nikitas Nikitaras3 1

Aristotle University of Thessaloniki, Greece The University of Birmingham, UK 3 University of Athens, Greece 2

Background. It is commonly assumed that there is conceptual equivalence between the task and ego achievement goals proposed by Nicholl’s (1989) dichotomous achievement goal theory (Nicholls, 1989), and the mastery and performance approach goals advanced by Elliot’s (1997) trichotomous hierarchical model of approach and avoidance achievement motivation. Aims. Our study examined whether this conceptual equivalence is reflected in measurement equivalence by examining the factorial structure and predictive validity of two established questionnaires that assess achievement goals based on Nicholl’s and Elliot’s approaches to achievement motivation. Sample.

Greek adolescents (N ¼ 336, M age ¼ 13:45 years, SD ¼ 1:04).

Measures. The participants completed the Task and Ego Orientation in Sport Questionnaire (Duda & Nicholls, 1992), the Approach – Avoidance Achievement Goals Questionnaire (Elliot & Church, 1997) and a Physical Education (PE) version of the SelfRegulation Questionnaire (Goudas, Biddle, & Fox, 1994). Results. Confirmatory factor analyses of a number of competing models showed that a model with five correlated independent factors had the best fit. This finding suggests that the goals measured by the two achievement goal questionnaires are related, although independent constructs. However, hierarchical regression analyses predicting regulatory styles in PE showed quite a substantial overlap between the mastery and performance approach goals proposed by Elliot (1997), and the task and ego goals, respectively, advanced by Nicholls (1989).

* Correspondence should be addressed to Vassilis Barkoukis, Department of Physical Education & Sport Science, Aristotle University of Thessaloniki, 54124, Thessaloniki, Greece (e-mail: [email protected]). DOI:10.1348/000709906X171901

684 Vassilis Barkoukis et al. Conclusions. Taken together, our results indicate that the self-referenced and comparative1 goals of the TEOSQ and AAGQ are substantially related, to the extent that they have minimal unique predictive validity; however, they are not identical constructs.

There is an increasing body of evidence to demonstrate that physical activity during childhood may contribute to short-term and lifelong health (Cale & Almond, 1992; Sallis & Patrick, 1994). For example, Sallis and Patrick (1994) and Rocchini (1999) argued that physical activity during childhood improves both psychological and physical health (e.g. positive affective states, stronger immune system, enhanced bone development). In view of these benefits, it is important to understand the regulatory mechanisms that underpin young people’s participation in physical activity. To this end, achievement goal theory (e.g. Duda & Hall, 2001; Dweck, 1986; Nicholls, 1989; Roberts, 2001), a social-cognitive approach to motivation, has been widely used to understand motivation in youth sport and physical education (PE). An important tenet of this theory is that individuals in achievement contexts hold two independent achievement goals, namely, a task and an ego goal orientation. Individuals high in task orientation engage in an achievement activity to achieve mastery and personal improvement, and use self-referenced criteria to judge their ability. Individuals high in ego orientation engage in an activity to outperform others and demonstrate superior ability (Nicholls, 1989). These individuals use normative or comparative criteria to judge their perceived ability. Task orientation has been found to relate to more adaptive motivational outcomes such as greater effort and persistence (Williams & Gill, 1995), fair play (Smith, Hall, & Wilson, 1999; White & Zellner, 1996), greater enjoyment (Duda, Chi, Newton, Walling, & Catlin, 1995) and lower anxiety (Ommundsen & Pedersen, 1999; see Duda & Hall, 2001, for a detailed review). This original (also called dichotomous) achievement goal approach has been criticized and revised by Elliot and his colleagues (e.g. Elliot, 1997; Elliot & Church, 1997; Elliot & Thrash, 2001) by incorporating key constructs from the achievement motive approach (Atkinson, 1964). Elliot argued that the dichotomous achievement goal approach does not adequately address the issue of energization of behaviour and fails to distinguish between approach and avoidance motivation. In Elliot’s hierarchical model of approach and avoidance achievement motivation, Atkinson’s achievement motives (need for achievement and fear of failure) are conceptualized as general, higher order motivational tendencies. In contrast, achievement goals are conceptualized as more concrete sources of influence to achieve specific outcomes. In the hierarchical model, achievement motives are hypothesized to have indirect effects on achievement-related outcomes through achievement goals. Elliot proposed three types of achievement goals.2 The first one is a mastery goal, which refers to a personal focus on the development of competence and task mastery. Elliot also proposed two performance goals, which in contrast to Nicholls’ (1989) ego orientation, incorporate both approach

1 We use the terms ‘self-referenced’ and ‘comparative’ to refer to the criteria employed to judge competence and not on whether the focus of attention is on the self or the task. 2 More recently, Elliot and McGregor (2001) argued that mastery goals can also have an avoidance component. With masteryavoidance goals competence, is still self-referent but the focus of the individual is to avoid a negative outcome (e.g. task failure). Elliot and McGregor indicated that mastery-avoidance goals have a more negative motivational profile than mastery-approach goals, but a more positive profile when compared with performance-avoidance goals. There is no widespread acceptance of mastery-avoidance goals (e.g. see Pintrich, 2000) and therefore, this study has not assessed them.

Dichotomous and trichotomous achievement goals

685

and avoidance tendencies. A performance-approach goal reflects involvement in an activity in order to demonstrate superiority over others, whereas a performanceavoidance goal reflects a focus on avoiding the demonstration of incompetence relative to others. Mastery and performance-approach goals construe an approach orientation while performance-avoidance goals an avoidance orientation (Elliot, 1997; Elliot & Church, 1997). The trichotomous approach to achievement motivation has been tested empirically in the education context. One of the key outcome variables in this line of research is intrinsic motivation. For example, Elliot and Harackiewicz (1996) conducted two experiments which manipulated achievement goals and examined the influence of the three achievement goal conditions on intrinsic motivation to solve puzzles. The results of two experiments indicated that the performance-avoidance condition undermined intrinsic motivation. In contrast, the effect of the performance-approach condition on intrinsic motivation was equivalent to that of the mastery condition, and significantly higher than that of the performance-avoidance condition. In a study of achievement goals in college students, Elliot and Church (1997) showed that mastery goals predicted intrinsic motivation, performance-approach goals predicted academic performance and performance-avoidance goals undermined both intrinsic motivation and performance. These findings provide further support to Elliot’s (1997) argument that mastery and performance-approach goals are more adaptive than performance-avoidance goals. This argument is also corroborated by a meta-analysis by Rawsthorne and Elliot (1999) of the experimental literature that has examined the effects of performance and mastery achievement goals on intrinsic motivation. The meta-analysis showed that the undermining effect of performance goals relative to mastery goals was evident only when the experimental procedures induced a performance-avoidance orientation. The trichotomous approach to achievement motivation proposed by Elliot’s (1997) hierarchical model has been also been tested to the PE context. For example, Cury, Elliot, Sarrazin, Da Fonse´ca, and Rufo (2002) have also demonstrated the detrimental effects of performance-avoidance goals on intrinsic motivation (i.e. free-time preparation for a basketball dribbling task). Adolescent PE students were assigned to one of the three experimental conditions (i.e. mastery, performance-approach, performance-avoidance) to perform a basketball dribbling task. The results indicated that intrinsic motivation was lower in the performance-avoidance goal condition; no differences were observed between the mastery and performance-approach goal conditions. These effects were independent of participants’ pre-experiment levels of intrinsic motivation. Competence valuation, state anxiety and task absorption were shown to have a mediational role in the achievement goals–intrinsic motivation relationship. For example, performance-avoidance, relative to performance-approach and mastery goals, reduced competence valuation and task absorption and increased state anxiety, all of which in-turn undermined intrinsic motivation. These findings were further supported by Cury, Da Fonse´ca, Rufo, Peres, and Sarrazin (2003) in a similar experiment with PE students. As noted by Grant and Dweck (2003), mastery goals and task goals have often been regarded in the literature as conceptually equivalent (e.g. Ames, 1992). For example, Elliot and Church (1997) used the term mastery goals to refer to Nicholl’s (1989) task orientation and characterized these goals as ‘similar conceptualizations with different nomenclature’ (p. 218). Similarly, it is often assumed in the literature that there is conceptual equivalence between ego and performance approach goals. For example,

686 Vassilis Barkoukis et al.

Elliot (1997) used the term ‘performance goal’ to refer to the comparative goal proposed by Nicholls. Further, both ego orientation and performance approach goals refer to approach forms of motivation focusing on the demonstration of competence relative to others. For example, Duda and Nicholls (1992) defined ego orientation as the ‘goal of establishing one’s superiority over others and the beliefs that success in school requires attempts to beat others and superior ability’ (p. 290). Hence, it could be assumed that there is conceptual equivalence between the selfreferenced (task orientation and mastery goals) and comparative approach goals (ego orientation, performance-approach goals) advanced by the dichotomous and trichotomous achievement goal approaches. However, the measurement equivalence of these goals has not been demonstrated. In order to accomplish this, we compared two established questionnaires that assess achievement goals based on the respective theoretical approaches. These questionnaires were the Task and Ego Orientation in Sport Questionnaire (TEOSQ) developed by Duda and Nicholls (1992), and the Approach and Avoidance Achievement Goals Questionnaire (AAAGQ) developed by Elliot and Church (1997). Recent research has shown that different questionnaire measures of purportedly the same achievement goal may tap different aspects of this goal. For example, Jagacinski and Duda (2001) compared three different instruments measuring the same, at face value, achievement goals. The authors found that a model assuming that the factors from the three questionnaires were independent, fit the data relatively better than alternative models which collapsed some purportedly similar factors. Such results could be explained by Grant and Dweck’s (2003) argument that different dimensions of the same achievement goal might exist. For example, Grant and Dweck suggested that comparative goals may measure outcome, ability or normative aspects of the involvement. Also, self-referenced goals might tap learning or challenge/mastery aspects of the achievement experience. Although all these goals fall within the broader definitions of comparative and self-referenced goals respectively, they can have distinct predictive ability in terms of certain motivation outcomes and achievement (Grant & Dweck, 2003). An inspection of the TEOSQ and AAAGQ items that we employed in our study also suggests that the factors of these questionnaires may tap different aspects of selfreferenced and comparative approach goals. The task orientation subscale of the TEOSQ includes items referring to effort and feelings of fun and enjoyment during sport participation, whereas the mastery subscale of the AAAGQ focuses on task mastery and challenge. Similarly, with respect to comparative approach goals, there is a discrepancy in the content of the ego orientation and performance approach goals items. The TEOSQ’s ego orientation is composed of items reflecting a mixture of normative, outcome and ability items. In contrast, the AAAGQ’s performance-approach subscale focuses exclusively on normative items (see Appendix). In view of the above arguments, the primary purpose of our study was to examine whether the self-referenced and comparative approach goals assessed by the TEOSQ and the AAAGG are equivalent. To this end, we carried out confirmatory factor analyses to compare three competing factor model structures: (a) a three-factor model with a selfreferenced goals factor (task orientation and mastery items), a comparative approach factor (ego orientation and performance approach items) and a performance avoidance factor (equivalent hypothesis model); (b) a five-factor model with five independent uncorrelated factors (alternative model 1) and (c) a five-factor model with five independent but correlated achievement goals (alternative model 2). To further examine the issue of measurement equivalence, we also looked at how the self-referenced and comparative approach goals of the TEOSQ and the AAAGQ predict

Dichotomous and trichotomous achievement goals

687

various regulatory styles in PE. As our brief review has shown so far, intrinsic motivation is a key outcome variable in the research using the trichotomous model. This is expected as both achievement goals and intrinsic motivation have mutual reliance on the construct of competence (Harackiewicz & Elliot, 1993). Competence and a sense of mastery underpin feelings of intrinsic interest in an activity (Deci & Ryan, 1985), whilst different aspects of competence conceptions give rise to different achievement goals. However, intrinsic motivation is only one self-regulatory style. According to Deci and Ryan’s selfdetermination theory, motivation should be viewed as a multidimensional concept, encompassing, besides intrinsic motivation, extrinsic motivation and amotivation. Intrinsic motivation refers to involvement in an activity for the pleasure and fun derived, while extrinsic motivation refers to activity involvement as a result of internal and external pressure or anticipation of rewards. Amotivation refers to feelings of uncontrollability and lack of intention to participate in an activity (Deci & Ryan, 1985). These dimensions represent different levels of self-determination with intrinsic motivation being the most self-determined dimension and amotivation the least selfdetermined one. Extrinsic motivation, posited in the middle of this self-determination continuum, is further divided into three dimensions representing different levels of selfdetermination. Identified regulation is the most self-determined dimension of extrinsic motivation and refers to engagement in an activity because the latter is valued as important by the individual. Next to identified regulation lies introjected regulation, a more controlling motivational regulation, which reflects participation in an activity because of internal compulsion (e.g. feelings of guilt). Finally, external regulation is the least self-determined form of extrinsic motivation and refers to involvement in an activity because of external coercion or in order to achieve rewards. In view of the multidimensional nature of motivation, the sole focus of the trichotomous-based research on intrinsic motivation is surprisingly restrictive. It is, therefore, important that research examines the relationship between mastery, performance-approach and performance-avoidance goals with the various components of extrinsic motivation and amotivation, in addition to intrinsic motivation. Research based on the dichotomous achievement goal theory (e.g. Ntoumanis, 2001a) has shown that task orientation is related to motivational variables with a high degree of selfdetermination (i.e. intrinsic motivation, identified regulation). In contrast, ego orientation is related to motivational variables with a low degree of self-determination (i.e. introjected and external regulation). Whether these findings would replicate with the three goals proposed by Elliot (1997) was the secondary purpose of our study. We hypothesized that, similar to Ntoumanis (2001a), self-referenced goals would predict high self-determined motivation, whereas comparative approach goals would predict low, and performance avoidance goals even lower, self-determined motivation. More importantly, we made a number of hypotheses to examine the issue of measurement equivalence. We hypothesized that, if mastery goals are conceptually equivalent to task orientation (Elliot, 1997), the mastery scale of the AAAGQ should not be able to predict a unique amount of variance of motivational regulations when, in a hierarchical regression analysis, it is entered after the task subscale of the TEOSQ (assuming that task orientation makes a significant prediction at Step 1).3 Alternatively,

3

We limit the interpretation of our findings to regression models with significant beta coefficients. Clearly, if two independent variables have the same null effect on a dependent variable, this cannot be used as an evidence to suggest that the two independent variables are similar.

688 Vassilis Barkoukis et al.

measurement equivalence could also be established by showing that the effect of mastery goals is significant at Step 2, whereas the effect of task orientation becomes nonsignificant).4 However, if mastery goals are able to make a unique prediction (whilst the effect of task orientation remained significant at Step 2) it would imply lack of measurement equivalence of the two self-referenced goals. Elliot and Church (1997) have claimed that measures of comparative-ability goals (i.e. ego) vary considerably in composition, with some composed entirely of positively valenced items (thereby assessing a performance-approach goal), and others composed of a combination of positively and negatively valenced items (thereby assessing both performance-approach and performance-avoidance goals). The ego subscale of the TEOSQ comprises positively valenced items only, thus potentially corresponding to a performance-approach goal. If ego orientation and performance-approach goals are conceptually identical, then similar to the hypothesis regarding self-referenced goals, the performance-approach subscale of the AAAGQ should not predict a unique amount of variance of motivational outcomes over and above any significant contribution made by the ego orientation subscale of the TEOSQ. Alternatively, measurement equivalence could also be established by showing that the effect of performance approach goals is significant at Step 2, whereas the effect of ego orientation becomes non-significant4. We also included performance-avoidance goals in the regressions. Being independent of ego orientation, we expected these goals to make a unique prediction in explaining the variance of the dependent variables.

Method Participants The sample of the study consisted of 336 Greek adolescents (166 males and 170 females; M age ¼ 13:45 years, SD ¼ 1:04) participating in a summer sports camp in southern Greece. All children were Caucasians. Informed consent was obtained from the children, their parents and the camp director prior to the data collection.

Measures Task and ego orientation in sport questionnaire (TEOSQ; Duda & Nicholls, 1992) The TEOSQ, adapted for use in PE, was employed to assess task and ego goal orientations. Task orientation was measured with 7 items (e.g. ‘I feel most successful in PE when I learn something that is fun to do’) and ego orientation was assessed with 6 items (e.g. ‘I feel most successful in PE when the others can’t do as well as me’). The responses were provided on a 5-point scale from 1 (strongly disagree) to 5 (strongly agree). The TEOSQ has been adapted to Greek PE by Papaioannou and Macdonald (1993) and is considered a valid and reliable tool for the assessment of goal orientations of Greek children (Papaioannou & Kouli, 1999; Papaioannou & Theodorakis, 1996). A confirmatory factor analysis performed on the data of the present study revealed the following fit indices: x2 ð61Þ ¼ 135:411, p , :001; CFI ¼ :91, SRMR ¼ :06, RMSEA ¼ :06. According to Hu and Bentler (1999), a good model fit is indicated by values close to or above .95 for the CFI, and close to or below .08 and .06 for the SRMR 4

We would like to thank an anonymous reviewer for pointing this out.

Dichotomous and trichotomous achievement goals

689

and the RMSEA, respectively. The CFI value in our analysis was slightly below the .95 level. However, Rigdon (1996) demonstrated that CFI is suited to more ‘exploratory’ SEM applications, while RMSEA is suited to more ‘confirmatory’ applications, such as ours. Furthermore, the SRMR value was adequate. In fact, Hu and Bentler found that, among all fit indices, SRMR is the most sensitive to misspecification in both simple and complex models, and is less sensitive to sample size and violations of distributional assumptions. In view of these findings, we consider the model fit of TEOSQ as satisfactory. Approach – avoidance achievement goals questionnaire (AAAGQ; Elliot & Church, 1997) To measure the three achievement goals proposed by Elliot (1997), we adapted to PE a scale developed by Elliot and Church (1997). This scale consists of 18 items, 6 for each achievement goal. Example items are: ‘I want to learn as much as I can in this class’ (mastery goal); ‘I am striving to demonstrate my ability relative to others in this class’ (performance-approach goal) and ‘I just want to avoid doing poorly in this class’ (performance-avoidance goal). The responses were given on a 5-point scale (1 ¼ totally disagree; 5 ¼ totally agree). Elliot and Church provided evidence for the validity and reliability of the scale. Barkoukis and Anastasiadis (2002), in a preliminary test of this scale with Greek students, reported that the exclusion of three items from the performance-avoidance subscale (‘I often think to myself, What if I do badly in this class?’, ‘I just want to avoid doing poorly in this class’ and ‘I wish this class was not graded’) improved its internal consistency. In the present study, the internal consistency of the performance-avoidance subscale was very low (a ¼ :33). After the exclusion of these items, the internal consistency of the subscale improved somewhat (a ¼ :50). A confirmatory factor analysis was performed on our data to test the factorial structure of the questionnaire without these three items. The results revealed acceptable fit indices (x2 ð86Þ ¼ 159:179, p , :001, CFI ¼ :93, SRMR ¼ :06, RMSEA ¼ :05). Self-regulation questionnaire (Ryan & Connell, 1989) A version of the Self-Regulation Questionnaire (Ryan & Connell, 1989), adapted for use in PE by Goudas et al. (1994), was used to measure the different motivational regulations. Goudas et al. also adapted the amotivation subscale of the Academic Motivation Scale (Vallerand et al., 1992). The scale we used consisted of five factors (4 items per factor). The participants responded to the stem ‘I take part in this PE class : : : ’ on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). Example items are: intrinsic motivation (e.g. ‘I take part in this PE class because it is fun’); identified regulation (e.g. ‘I take part in this PE class because I want to improve in sport’); introjected regulation (e.g. ‘I take part in this PE class because I would feel bad about myself if I didn’t’); external regulation (e.g. ‘I take part in this PE class so that the teachers won’t yell at me’) and amotivation (e.g. ‘I take part in this PE class but I can’t see what I am getting out of PE’). The instrument has been shown to have adequate psychometric properties with British (Goudas et al., 1994; Ntoumanis, Pensgaard, Martin, & Pipe, 2004) and Greek (Kiriakidis, 2005) children. Procedure The questionnaire was administered within the first days of the camp. The students were told that the purpose of the questionnaire was to obtain information about their

690 Vassilis Barkoukis et al.

experiences in school PE only. They were reassured that their responses would be confidential and were told that they could stop completing the questionnaire at any time. None of the students withdrew from the study.

Results Preliminary analyses Means, standard deviations, normality statistics and internal consistency estimates are reported in Table 1. The descriptive statistics show that the students reported high task orientation and mastery goals, and moderate ego orientation and performance (approach and avoidance) goals. Additionally, the students reported high levels of selfdetermined motivation (intrinsic motivation and identified regulation), and low levels of controlling motivation (introjected and identified regulation) and amotivation. In regard to the internal consistency of the scales, with the exception of introjected regulation (a ¼ :64) and performance-avoidance goals (a ¼ :50), the alpha values for the other variables were satisfactory. The alpha value for introjected regulation is similar to the alpha values reported by many other studies on motivation in PE classes (e.g. Ntoumanis, 2001b; Standage, Duda, & Ntoumanis, 2003). In regard to performanceavoidance goals, Smith, Duda, Allen, and Hall (2002) also reported a relatively low alpha value (a ¼ :65). Both the introjected regulation and the performance-avoidance scales were retained due to their theoretical importance, but the results pertaining to these variables should be interpreted with caution. We feel that the low alpha value for the performance-avoidance scale is not particularly worrisome, as the main purpose of the study was to examine the measurement equivalence of self-referenced and normative approach goals. Therefore, performance avoidance goals served an auxiliary role in our study. Table 1. Descriptive statistics and Cronbach’s alpha for all variables

Goal orientations Task orientation Ego orientation Mastery Performance-approach Performance-avoidance Motivational regulations Intrinsic motivation Introjected regulation Identified regulation External regulation Amotivation

M

SD

Skewness

Kurtosis

a

4.21 3.17 4.00 3.02 3.00

0.56 0.77 0.77 0.88 0.91

21.02 20.03 20.90 0.01 0.05

1.11 2 0.50 0.71 2 0.48 2 0.26

.70 .71 .80 .77 .50

5.26 3.27 5.44 2.81 2.10

1.32 1.41 1.21 1.47 1.22

20.89 0.40 20.91 0.81 1.29

0.47 2 0.34 0.77 2 0.09 1.32

.72 .64 .70 .70 .74

Note. Achievement goals and motivational regulations were measured using 5- and 7-point scales, respectively.

Correlation analysis Correlation analysis is presented in Table 2. As expected (Nicholls, 1989), task orientation was unrelated to ego orientation and both performance goals. Similarly,

Dichotomous and trichotomous achievement goals

691

the relationships between mastery goals with ego orientation and with both performance goals were very small. Moderate relationships were found between task orientation and mastery goals, ego orientation with performance-approach goals, and between performance-approach and performance-avoidance goals. However, the size of the correlations indicated that the common variance among the correlated variables was less than 25%. The correlation between ego orientation and performance-avoidance goals was small (r ¼ :16). Mastery goals and task orientation were positively correlated with intrinsic motivation and identified regulation, negatively correlated with external regulation and amotivation, and unrelated to introjected regulation. Ego orientation and the two performance goals were unrelated to intrinsic motivation and amotivation, negatively correlated with external regulation, and positively related to introjected regulation and (unexpectedly) identified regulation.

Confirmatory factor analyses (CFA) To examine the measurement equivalence of the self-referenced and comparative approach goals assessed by the TEOSQ and AAAGQ, three alternative factor model structures were tested. The first model (equivalent hypothesis model) assumed measurement equivalence of self-referenced and comparative approach goals as assessed by the two questionnaires, and proposed three factors: a self-referenced goals factor (mastery and task goals), a comparative approach factor (ego and performance approach goals) and a performance avoidance factor. The second and third (alternative) models assumed measurement divergence. Model 2 proposed five independent uncorrelated achievement goal factors, whereas model 3 proposed five independent correlated achievement goal factors. The fit indices for model 1 (x2 ð347Þ ¼ 897:431, p , :001; CFI ¼ :74, SRMR ¼ :08, RMSEA ¼ :07) and model 2 (x2 ð337Þ ¼ 762:298, p , :001; CFI ¼ :79, SRMR ¼ :12, RMSEA ¼ :06) were very poor, whereas for the third model they were satisfactory5 (x2 ð327Þ ¼ 522:707, p , :001; CFI ¼ :91, SRMR ¼ :06, RMSEA ¼ :05).6 In the last model, a moderate relationship emerged between task orientation and mastery goals (r ¼ :56), while low relationships were found between task orientation with performance-approach (r ¼ :14), and between mastery goals and performanceapproach (r ¼ :13). Performance-approach and performance-avoidance goals were relatively highly correlated (r ¼ :63). Additionally, low correlations were found between task orientation with ego orientation (r ¼ 2:12) and performance-avoidance goals (r ¼ :18). Ego orientation was moderately high related to performance-approach goals (r ¼ :68), whereas it was unrelated to mastery goals (r ¼ 2:02) and performanceavoidance goals (r ¼ :03). Finally, low relationships were found between mastery goals and performance-avoidance goals (r ¼ :28). 5

The inconsistency between the CFI value and the RMSEA and SRMR values is due to the fact that incremental fit indices (such as the CFI) are stronger when the size of the factor correlations is high. In this model, the absolute average r value was relatively low (r ¼ :277). 6 An anonymous reviewer suggested testing a four-factor model hypothesizing that the two self-referenced goals might collapse better in one factor than ego orientation and performance approach goals. The results of the CFA model with four factors (selfreferenced, ego, performance approach, performance avoidance goals) showed that the x 2 was significant (x 2 ð337Þ ¼ 651:713; p , :001) and the absolute fit indices were acceptable (SRMR ¼ :07 and RMSEA ¼ :06). However, the incremental fit indices were low and unsatisfactory (CFI ¼ :85 and NNFI ¼ :83). Hence, this model was not deemed to fit the data as well as the five-correlated factor model.

692 Vassilis Barkoukis et al. Table 2. Correlation matrix with all the variables of the study 1 1. Task orientation 2. Ego orientation 3. Mastery 4. Performance-approach 5. Performance-avoidance 6. Intrinsic motivation 7. Identified regulation 8. Introjected regulation 9. External regulation 10. Amotivation

2

3

4

5

2 .01 .50** .01 .01 .16** .46** .16** .10* .12* .41**

6

7

8

9

10

.38** .07 .58** .05 .09

.44** .12* .62** .17** .24** .60**

.02 .19** .07 .32** .29** .01 .20**

2 .14** .11* 2 .31** .12* .19** 2 .38** 2 .23** .48**

2.25** .04 2.38** .05 .01 2.38** 2.41** .23** .55**

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

Hierarchical regression analyses A series of hierarchical regression analyses was conducted to determine whether the three achievement goals postulated by the trichotomous model could predict a unique variance of the motivational regulations over and above what is predicted by the two goals of the traditional achievement goal theory. An overview of these analyses is presented in Table 3. In each of the five regressions, task and ego goal orientations were entered in the first step to ascertain their unique contribution to the amount of variance explained. In the second step, the mastery and performance-approach goals were entered to examine whether they could account for additional variance of the motivational regulations. As performance-avoidance goals were not expected to map onto either task or ego orientation, they were entered in the third step. By following this sequence of variable entry, the R2 difference from Step 1 to Step 2 would reflect only the addition of mastery and performance approach goals, thus allowing us to test our hypothesis. In regard to intrinsic motivation, the results indicated that the inclusion of task and ego achievement goals predicted a significant amount of variance (R 2 ¼ :15), with task orientation being a significant predictor (b ¼ 0:38). The addition of the two goals from the trichotomous model in Step 2 improved the prediction of intrinsic motivation (R 2 ¼ :34). Task orientation and mastery goals were both significant predictors (b ¼ 0:12 and 0.52, respectively). The addition of performance-avoidance goals in the third step did not contribute to the prediction of intrinsic motivation. Task orientation and mastery goals remained the two significant predictors of intrinsic motivation (b ¼ 0:12 and 0.51, respectively). However, the semi-partial correlation between intrinsic motivation and task orientation, that is the proportion of the intrinsic motivation variance accounted for by task orientation beyond that accounted for by the other predictors, was small (sr 2 ¼ :11), as was its effect size (f 2 ¼ :01; see Cohen, Cohen, West, & Aiken, 2003). The hierarchical regression analysis predicting identified regulation revealed that the two goal orientations predicted a significant amount of variance (R 2 ¼ :21), with both orientations being significant predictors (b ¼ 0:44 for task orientation and b ¼ 0:12 for ego orientation). The addition of mastery and performance-approach goals in Step 2 improved the prediction of identified regulation (R 2 ¼ :41). Task orientation (b ¼ 0.18), mastery goals (b ¼ 0:51) and performance-approach goals (b ¼ 0:12)

Dichotomous and trichotomous achievement goals

693

Table 3. A summary of the regression analyses predicting motivational regulations from the achievement goals proposed by the dichotomous and trichotomous achievement goal theories Variables Intrinsic motivation Step 1 Task orientation Ego orientation Step 2 Task orientation Ego orientation Mastery Performance-approach Step 3 Task orientation Ego orientation Mastery Performance-approach Performance-avoidance Identified regulation Step 1 Task orientation Ego orientation Step 2 Task orientation Ego orientation Mastery Performance-approach Step 3 Task orientation Ego orientation Mastery Performance-approach Performance-avoidance Introjected regulation Step 1 Task orientation Ego orientation Step 2 Task orientation Ego orientation Mastery Performance-approach Step 3 Task orientation Ego orientation Mastery Performance-approach Performance-avoidance External regulation Step 1 Task orientation

R2

R2change

.15

F

Fchange

b

30.81** .38 .07

.34

.34

.19

.002

.21

.41

.43

.12

.03

45.08**

36.27**

7.71** 1.47

50.24** .12 2 .01 .52 .003

2.41* 20.26 10.01** 0.06

.12 2 .01 .51 2 .01 .05

2.45* 20.20 9.85** 20.33 1.02

.44 .12

9.19** 2.59*

.18 2 .01 .51 .12

3.77** 20.36 10.48** 2.63*

.19 2 .01 .50 .05 .15

3.98** 20.19 10.24** 1.15 3.38**

.02 .19

0.47 3.53**

.004 .04 .03 .30

0.06 0.73 0.55 5.10**

.01 .05 .01 .21 .19

0.22 0.93 0.19 3.46** 3.53**

2 .14

22.74**

1.04

45.60**

.20

.02

.03

.10

t

60.50**

52.22**

59.40**

11.46**

6.35*

.07

.02

10.03**

10.80**

13.25**

12.46**

6.17**

694 Vassilis Barkoukis et al. Table 3. (Continued) Variables Ego orientation Step 2 Task orientation Ego orientation Mastery Performance-approach Step 3 Task orientation Ego orientation Mastery Performance-approach Performance-avoidance Amotivation Step 1 Task orientation Ego orientation Step 2 Task orientation Ego orientation Mastery Performance-approach Step 3 Task orientation Ego orientation Mastery Performance-approach Performance-avoidance

R2

R2change

F

Fchange

.12

.09

12.94**

19.05**

.16

.04

.06

.15

.15

13.85**

b

t

.11

2.19*

.03 .13 2 .36 .01

0.56 2.23* 25.98** 1.70

.04 .14 2 .38 .006 .21

0.75 2.48* 26.48** 0.09 3.90**

2 .25 .04

24.77** 0.88

2 .06 .08 2 .37 .06

21.11 1.39 26.32** 1.05

2 .06 .08 2 .37 .05 .01

21.09 1.41 26.31** 0.84 0.32

15.23**

11.80**

.09

.000

16.76**

13.39**

20.34**

.10

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

were the three significant predictors. Although task orientation remained significant in Step 2, its semi-partial correlation with identified regulation was small (sr 2 ¼ :16), as was its effect size (f 2 ¼ :03). The inclusion of performance-avoidance goals in Step 3 further improved the prediction of identified regulation (R 2 ¼ :43). The significant predictors were task orientation (b ¼ 0:19), mastery goals (b ¼ 0:50) and performanceavoidance goals (b ¼ 0:15). The hierarchical regression analysis predicting introjected regulation indicated that the inclusion of task and ego achievement goals predicted a significant amount of variance (R 2 ¼ :03), with ego orientation being the only significant predictor (b ¼ 0:19). The addition of the two achievement goals from the trichotomous model in Step 2 improved the prediction of introjected regulation (R 2 ¼ :10). Performanceapproach goals (b ¼ 0:30) were the only significant predictors. The inclusion of performance-avoidance goals in the third step significantly improved the prediction of introjected regulation (R 2 ¼ :12). Performance-approach (b ¼ 0:21) and performanceavoidance (b ¼ 0:19) goals were the two significant predictors of introjected regulation. In regard to external regulation, the results showed that Step 1 explained a significant amount of variance (R 2 ¼ :03), with task orientation being a negative

Dichotomous and trichotomous achievement goals

695

(b ¼ 20:14) and ego orientation a positive predictor (b ¼ 0:11). Step 2 improved the prediction of external regulation (R 2 ¼ :12). Ego orientation (b ¼ 0:14) was a positive predictor and mastery goals (b ¼ 238) were negative predictors. The addition of performance-avoidance goals in Step 3 further improved the prediction of external regulation (R 2 ¼ :16). Ego orientation (b ¼ 0:14) and performance-avoidance (b ¼ 0:21) goals were positive predictors, whereas mastery goals (b ¼ 238) were negative predictors. Lastly, the hierarchical regression analysis predicting amotivation revealed that Step 1 made a significant prediction (R 2 ¼ :06), with task orientation being a significant negative predictor (b ¼ 20:25). Step 2 improved the prediction of amotivation (R 2 ¼ :16), with mastery goals being the only significant predictors of amotivation (b ¼ 20:37). Step 3 did not improve the prediction of amotivation; mastery goals remained the only significant predictors (b ¼ 20:37).

Discussion Elliot (1997) proposed a revision of the dichotomous achievement goal theory (e.g. Nicholls, 1989) by suggesting, among other things, a distinction between two types of comparative competence-based goals, that is, performance-approach and performance-avoidance goals. Elliot’s trichotomous model (which also includes self-referenced mastery goals) has been extensively applied to education and has also attracted considerable interest in sport psychology research. Nicholls’ task orientation (as assessed by the TEOSQ) is purported to be conceptually equivalent to Elliot’s mastery goals (as assessed by the AAAGQ), whereas Nicholls’ ego orientation is assumed to be equivalent to Elliot’s performance-approach goals. However, it is not known whether this purported conceptual equivalence is reflected in measurement equivalence of the two self-referenced and the two comparative approach goals of the TEOSQ and the AAAGQ. We examined the issue of measurement equivalence in two different ways. First, we conducted CFAs to ascertain whether the items of the two self-referenced and the two comparative approach goals could load on two respectively named factors, or whether they should be represented by separate correlated or uncorrelated factors (along with a fifth performance-avoidance factor). Second, we conducted hierarchical regression analyses to look at the ability of the different achievement goals to predict controlling and self-determined motivational regulations proposed by SDT (Deci & Ryan, 1985). We assumed that if there was measurement equivalence between the self-referenced goals of the TEOSQ and the AAAGG, task orientation and mastery goals would not both have significant independent effects on motivational regulations. We made the same hypothesis for the comparative approach goals of the two questionnaires. A secondary purpose of the regression analyses was to extend past work linking task and ego achievement goals to intrinsic motivation, by including the three goals proposed by Elliot (1997) as well as measures of extrinsic motivation and amotivation. We were interested to examine whether self-referenced goals would be more likely to predict high self-determined motivation, whereas comparative approach goals would predict low, and performance-avoidance goals even lower, self-determined motivation. The CFAs revealed that task and mastery goals could not be collapsed under the same ‘self-referenced’ goal factor. Similarly, ego and performance-approach goals could not be collapsed under a ‘comparative-approach’ goal factor. Another factor structure tested,

696 Vassilis Barkoukis et al.

assuming that the five factors are independent and uncorrelated, was also not confirmed. In contrast, the alternative possibility that these four goals constitute correlated but independent factors was found to be a more plausible one. In fact, the correlations between the task and mastery factors and between the ego and performance approach factors were moderately strong, but not as strong as to imply collinearity. These results imply no measurement equivalence between the selfreferenced and comparative approach goals assessed by the TEOSQ and the AAAGQ. With regard to the predictive validity of the achievement goals, the results were more supportive of the measurement equivalence hypothesis. Specifically, in relation to intrinsic motivation, the analysis demonstrated that task orientation was a positive predictor. Furthermore, in Step 2 mastery goals also predicted intrinsic motivation in the same direction. The two comparative approach goals and performance avoidance goals did not predict intrinsic motivation. These findings are consistent with theory (Elliot, 1997; Nicholls, 1989) and previous research (Cury et al., 2003; Cury et al., 2002; Duda et al., 1995; Harackiewicz & Elliot, 1993; Ntoumanis, 2001a). Students with selfreferenced goals are intrinsically motivated because they enjoy the challenge of learning a new skill and improving on their weaknesses. Although both task orientation and mastery goals predicted independently the intrinsic motivation in Step 2, the independent effect of task orientation was quite small, as shown by the effect size of the semi-partial correlation between task orientation and identified regulation. Thus, there was some support for lack of measurement equivalence between the self-referenced goals of the TEOSQ and AAGQ, however, the evidence was weak. Identified regulation represents the most self-determined dimension of extrinsic motivation (Vallerand, 1997; Vallerand & Losier, 1999). As such, it was expected to be predicted by self-referenced goals only. The results showed that both task and mastery goals made an independent and substantial prediction. However, unexpectedly, ego orientation, performance approach goals and performance-avoidance goals also positively predicted identified regulation (in different steps). The positive beta coefficient for ego orientation and performance approach goals might reflect that students with such goals participate in PE because they value the skills developed through sport participation. These physical skills can help them to accomplish their comparative-referenced goals. In the third step, performance-avoidance goals emerged as positive independent predictors. The reason why the prediction made by performance-avoidance goals was positive is difficult to explain. It is possible that some students, although they participate in PE because they recognize its benefits, do not feel competent enough relative to their peers and strive to avoid demonstrating low competence. Further research is needed on this issue, especially in view of the problematic measurement properties of the performance avoidance subscale in this study. With regard to the measurement equivalence question, the effect of task orientation remained significant after entering mastery goals in the equation; however, the effect size associated with task orientation was small, providing only marginal evidence against the measurement equivalence hypothesis. In contrast, the effect of ego orientation was reduced to nearly zero when performance approach goals were added in the equation, implying measurement equivalence between the two normative approach goals. With regard to introjected regulation, task and mastery climate were not significant predictors. This finding makes conceptual sense as students with predominantly selfreferenced goals do not participate in PE out of internal pressure. Introjected regulation was predicted by ego orientation (in Step 1 only) and both performance-approach

Dichotomous and trichotomous achievement goals

697

(in steps 2 and 3) and performance-avoidance goals (in Step 3). These findings are in accordance with theory, emphasizing the links between ego involvement and introjected regulation (Nicholls, 1989; Ryan & Deci, 1989), and with previous research in physical activity settings (e.g. Ntoumanis, 2001a). Students with performanceapproach goals are probably motivated by conditional self-worth whereas those with performance-avoidance goals are probably motivated by feelings of guilt. An interesting finding was that when performance approach goals were entered in the regression equation, the effect of ego orientation became non-significant. Similar to the results for identified regulation, this finding indicates a substantial overlap in the measurement of the two normative approach variables. The independent significant contributions of the performance approach and avoidance goals of the AAAGQ were expected given the differences in the underlying constructs (Elliot, 1997). The findings for external regulation showed that task orientation (in Step 1 only) and mastery goals were negative predictors of external regulation. This is not surprising as students with self-referenced goals do not participate in PE because of external pressure or to demonstrate success and obtain extrinsic rewards (Duda & Ntoumanis, 2003). In contrast, ego orientation and performance-avoidance goals were positive predictors of external regulation. Consistent with previous research (e.g. Ntoumanis, 2001a), these findings indicate that students with comparative goals participate in PE for extrinsic reasons. Possibly, those with an ego orientation aim to obtain extrinsic rewards and social approval, whereas those with performance-avoidance goals are motivated by pressure and coercion. Contrary to the results for self-determined motivation (i.e. intrinsic motivation and identified regulation), the findings for external regulation showed that task orientation did not make a unique prediction after accounting for the effect of mastery goals. Performance-approach goals did not significantly predict external regulation over and above the effect of ego orientation. This indicates a substantial overlap in the measurement of the two variables.7 Similar to the results for introjected regulation, the effect of performance avoidance goal was significant and independent of the effect of the normative approach goal (in this case, of ego orientation). Lastly, with regard to amotivation, the analyses demonstrated that task orientation was a negative predictor. Furthermore, in Step 2 mastery goals also predicted amotivation in the same direction. This finding makes conceptual sense because students with self-referenced goals are not likely to experience feelings of task uncontrollability or purposelessness. Similar to the results for external regulation, task orientation did not make a unique prediction after accounting for the effect of mastery goals. In brief, the findings of the present study generally support previous research by showing that self-referenced goals (task and mastery) are positively related to high selfdetermined motivation regulations and negatively related to controlling regulatory styles. For comparative goals (ego, performance-approach, performance-avoidance) the pattern of results was less clear-cut. Specifically, as expected by both the dichotomous and the trichotomous achievement goal perspectives, the three comparative goals were

7

To examine whether performance approach goals would predict external regulation at the absence of ego orientation, we ran a regression analysis with performance approach and mastery goals as the only predictors in the equation. The beta coefficient for performance approach goals was significant (b ¼ 0:16; p , :01Þ. Thus, our interpretation regarding the overlapping effects of ego orientation and performance approach goals in predicting external regulation is valid.

698 Vassilis Barkoukis et al.

unrelated to intrinsic motivation. However, ego orientation and performance approach goals were positively linked to extrinsic regulatory styles with very different degrees of self-determination. Further, there were some conceptual inconsistencies in terms of the predictive effects of performance approach goals. Specifically, based on Elliot’s (1997) theorizing, one would expect performance avoidance goals to be positively associated with amotivation and to be unrelated or negatively related to identified regulation. This was not the case with our data. Our findings indicate the need for further examination of the role of comparative approach and avoidance goals in the self-determination continuum. In relation to the measurement equivalence of self-referenced and comparative goals of the dichotomous and trichotomous models, the pattern of results was mixed. The confirmatory factor analyses indicated that task and mastery goals could not be collapsed under the same ‘self-referenced’ goal factor. Similarly, ego and performanceapproach goals could not be collapsed under a ‘comparative-approach’ goal factor. In fact, assuming that these factors are independent but correlated provided a better fit to the data. The results of the regression analyses pointed more to the direction of measurement equivalence. Specifically, for the lowest self-determined forms of motivation (i.e. amotivation and external regulation), the predictive effects of task and mastery goals overlapped quite substantially. For the highest forms of selfdetermined motivation (i.e. intrinsic motivation and identified regulation), the predictive effects of task and mastery goals were both significant; however, the contribution of task orientation was minimal. In terms of comparative approach goals, the regression analyses showed that, for those motivational regulations (i.e. identified, introjected, external regulation) significantly predicted by ego orientation, the effects of the two comparative approach goals overlapped substantially. Some achievement goal researchers (e.g. Grant & Dweck 2003; Jagacinski & Duda, 2001) proposed the existence of different facets of self-referenced and comparative goals. Our CFA results support such claims. A closer examination of the items underlying the task and mastery goals provides some potential explanations for our findings. The items of the mastery goal subscale emphasize task mastery, task understanding and challenge (e.g. ‘I want to learn as much as possible from this PE class’; ‘I desire to completely master the tasks presented in this PE class’). In contrast, the items of the task orientation subscale reflect emphasis on individual effort and task enjoyment (e.g. ‘I feel successful in this PE class when I work really hard’; ‘I feel successful in this PE class when I learn something that is fun to do’). Similarly, there are differences in the items underlying the ego and performance approach goals. The TEOSQ’s ego orientation subscale includes normative, outcome and ability items. In contrast, the AAAGQ’s performance-approach subscale focuses exclusively on normative items. Evaluating together the results of the confirmatory factor and regression analyses, we conclude that the self-referenced and comparative goals of the TEOSQ and AAGQ are substantially related, to the extent that they have minimal unique predictive validity; however, they are not identical constructs (this conclusion stands even when interpreting factor correlations which are not attenuated by measurement error). Nevertheless, our evidence is preliminary and might be specific to the achievement goal questionnaires used or the types of outcomes assessed. Future research is needed to replicate our findings with other important motivational outcomes and with other questionnaires assessing achievement goals from the dichotomous and trichotomous achievement goal perspectives. For example, based on our earlier arguments, it would be plausible to hypothesize that task orientation and mastery goal would differ in how

Dichotomous and trichotomous achievement goals

699

they predict indices of effort and challenge-seeking. Our tentative recommendation is that achievement goal researchers should compare studies using the TEOSQ and the AAAGQ with caution, because these instruments do not tap identical facets of normative and self-referenced goals. In conclusion, we believe it is imperative that achievement goal researchers examine with greater scrutiny both the conceptualization and the measurement of self-referenced and comparative approach goals.

References Ames, C. (1992). Classrooms: Goals, structures, and student motivation. Journal of Educational Psychology, 84, 261–271. Atkinson, J. W. (1964). An introduction to motivation. Princeton, NJ: Van Nostrand. Barkoukis, V., & Anastasiadis, A. (2002). Psychometric properties of the Approach and Avoidance Achievement Goals Questionnaire. In H. Tsorbatzoudis (Ed.), Proceedings of the 7th National Congress on Sport Psychology. Thessaloniki, Greece. Cale, L., & Almond, L. (1992). Physical activity levels of young children: A review of the evidence. Health Education Journal, 51, 94–99. 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. Cury, F., Da Fonse´ca, D., Rufo, M., Peres, C., & Sarrazin, P. (2003). The trichotomous model and investment in learning to prepare for a sport test: A mediational analysis. British Journal of Educational Psychology, 73, 529–543. Cury, F., Elliot, A., Sarrazin, P., Da Fonse´ca, D., & Rufo, M. (2002). The trichotomous achievement goal model and intrinsic motivation: A sequential mediational analysis. Journal of Experimental Social Psychology, 38, 473–481. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum. Duda, J., Chi, L., Newton, M., Walling, M., & Catley, D. (1995). Task and ego orientation and intrinsic motivation in sport. International Journal of Sport Psychology, 26, 40–63. Duda, J., & Nicholls, J. (1992). Dimensions of achievement motivation in schoolwork and sport. Journal of Educational Psychology, 84, 290–299. Duda, J., & Ntoumanis, N. (2003). Correlates of achievement goal orientations in physical education. International Journal of Educational Research, 39, 415–436. Duda, J. L., & Hall, H. (2001). Achievement goal theory in sport: Recent extensions and future directions. In R. N. Singer, H. A. Hausenblas, & C. M. Janelle (Eds.), Handbook of sport psychology (2nd ed., pp. 417–443). New York: Wiley. Dweck, C. (1986). Motivational processes affecting learning. American Psychologist, 41, 1040–1048. Elliot, A. J. (1997). Integrating the classic and contemporary approaches to achievement motivation: A hierarchical model of approach and avoidance achievement motivation. In M. Maehr & P. Pintrich (Eds.), Advances in motivation and achievement (Vol. 10, pp. 143–179). London: JAI Press. Elliot, A. J., & Church, M. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72, 218–232. Elliot, A. J., & Harackiewicz, J. (1996). Approach and avoidance achievement goals and intrinsic motivation: A mediational analysis. Journal of Personality and Social Psychology, 70, 461–475. Elliot, A. J., & McGregor, H. (2001). A 2 £ 2 achievement goal framework. Journal of Personality and Social Psychology, 80, 501–519. Elliot, A. J., & Thrash, T. (2001). Achievement goals and the hierarchical model of achievement motivation. Educational Psychology Review, 13, 139–156.

700 Vassilis Barkoukis et al. Goudas, M., Biddle, S., & Fox, K. (1994). Perceived locus of causality, goal orientations, and perceived competence in school physical education classes. British Journal of Educational Psychology, 64, 453–463. Grant, H., & Dweck, C. (2003). Clarifying achievement goals and their impact. Journal of Personality and Social Psychology, 85, 541–553. Harackiewicz, J., & Elliot, A. J. (1993). Achievement goals and intrinsic motivation. Journal of Personality and Social Psychology, 65, 904–915. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. Jagacinski, C. A., & Duda, J. (2001). A comparative analysis of contemporary achievement goal orientation measures. Educational and Psychological Measurement, 61, 1013–1039. Kiriakidis, T. (2005). Hierarchical model of approach and avoidance achievement motivation: Effect of mastery and performance goals on motivational regulations. Unpublished Master Thesis. Aristotle University of Thessaloniki, Greece. Nicholls, J. (1989). The competitive ethos and democratic education. London: Harvard University Press. Ntoumanis, N. (2001a). Empirical links between achievement goal theory and self-determination theory in sport. Journal of Sport Sciences, 19, 397–409. Ntoumanis, N. (2001b). A self-determination approach to the understanding of motivation in physical education. British Journal of Educational Psychology, 71, 225–242. Ntoumanis, N., Pensgaard, A. M., Martin, C., & Pipe, K. (2004). An ideographic analysis of amotivation in compulsory school physical education. Journal of Sport and Exercise Psychology, 26, 197–214. Ommundsen, Y., & Pedersen, B. (1999). The role of achievement goal orientations and perceived ability upon somatic and cognitive indices of sport competition trait anxiety. Scandinavian Journal of Medicine and Science in Sports, 9, 333–343. Papaioannou, A., & Kouli, O. (1999). The effects of task structure, perceived motivational climate and goal orientations on students’ task involvement and anxiety. Journal of Applied Sport Psychology, 11, 51–71. Papaioannou, A., & Macdonald, A. (1993). Goal perspectives and purposes of physical education as perceived by Greek adolescents. Physical Education Review, 16, 41–48. Papaioannou, A., & Theodorakis, Y. (1996). A test of three models for the prediction of intention for participation in physical education lessons. International Journal of Sport Psychology, 27, 383–399. Pintrich, P. R. (2000). An achievement goal theory perspective on issues in motivation terminology, theory, and research. Contemporary Educational Psychology, 25, 92–104. Rawsthorne, L. J., & Elliot, A. J. (1999). Achievement goals and intrinsic motivation: A metaanalytic review. Personality and Social Psychology Review, 3, 326–344. Rigdon, E. E. (1996). CFI versus RMSEA: A comparison of two fit indices for structural equation modeling. Structural Equation Modeling, 3, 369–379. Roberts, G. (2001). Understanding the dynamics of motivation in physical activity: The influence of achievement goals on motivational processes. In G. Roberts (Ed.), Advances in motivation in sport and exercise (pp. 1–50). Champaign, IL: Human Kinetics. Rocchini, A. (1999). Raising heart healthy children. Pediatric International, 41, 597–602. Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization: Examining reasons for acting in two domains. Journal of Personality and Social Psychology, 57, 749–761. Ryan, R. M., & Deci, E. L. (1989). Bridging the research traditions of task/ego involvement and intrinsic/extrinsic motivation: Comment on Butler (1987). Journal of Educational Psychology, 81, 265–268. Sallis, J., & Patrick, K. (1994). Physical activity guidelines for adolescents: Consensus statement. Pediatric Exercise Science, 6, 302–314.

Dichotomous and trichotomous achievement goals

701

Smith, M., Duda, J. L., Allen, J., & Hall, H. (2002). Contemporary measures of approach and avoidance goal orientations: Similarities and differences. British Journal of Educational Psychology, 2, 154–189. Smith, M., Hall, H., & Wilson, P. (1999). The relationship of goal orientation and competitive climate to sportsmanship attitudes and the perceived legitimacy of intentionally injuries acts. Proceedings of the 10th European Congress of Sport Psychology. Prague, Czech Republic. Standage, M., Duda, J., & Ntoumanis, N. (2003). A model of contextual motivation in physical education: Using constructs and tenets from self-determination and goal perspective theories to predict leisure-time exercise intentions. Journal of Educational Psychology, 95, 97–110. Vallerand, R. (1997). Toward a hierarchical model of intrinsic and extrinsic motivation. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 29, pp. 271–360). San Diego: Academic Press. Vallerand, R., & Losier, G. (1999). An integrative analysis of intrinsic and extrinsic motivation in sport. Journal Applied Sport Psychology, 11, 142–169. Vallerand, R. J., Pelletier, L. G., Blais, M. R., Brie`re, N. M., Sene´cal, C., & Vallie`res, E. F. (1992). The academic motivation scale: A measure of intrinsic, extrinsic, and amotivation in education. Education and Psychological Measurement, 52, 1003–1017. White, S., & Zellner, S. (1996). The relationship between goal orientation, beliefs about causes of sport success, and trait anxiety among high school, intercollegiate, and recreational sport participants. Sport Psychologist, 10, 58–72. Williams, L., & Gill, D. (1995). The role of perceived competence in the motivation of physical activity. Journal of Sport and Exercise Psychology, 17, 363–378. Received 20 September 2005; revised version received 5 September 2006

I’m the only one who can do the play or skill (E) I can do better than my friends (E) The others can’t do as well as me (E) Others mess-up and I don’t (E) I score the most goals/points/hits etc (E) I’m the best (E) I learn a new skill and it makes me want to practice more (T) I learn something that is fun to do (T) I learn a new skill by trying hard (T) I work really hard (T) Something I learn makes me want to go and practice more (T) A skill I learn really feels right (T) I do my very best (T)

TEOSQ I feel successful in PE when: : :

Note. Pap, performance-approach goals; M, mastery goals; Pav, performance-avoidance goals; E, ego orientation; T, task orientation.

It is important to me to do better than the other students (Pap) My goal in this class is to get a better grade than most of the students (Pap) I am striving to demonstrate my ability relative to others in this class (Pap) I am motivated by the thought of outperforming my peers in this class (Pap) It is important to me to do well compared to others in this class (Pap) I want to do well in this class to show my ability to my family, friends, advisors, or others (Pap) I want to learn as much as possible from this class (M) It is important for me to understand the content of the lessons as thoroughly as possible (M) I hope to have gained a broader and deeper knowledge of physical education when I am done with this class (M) I desire to completely master the skills presented in this class (M) In a class like this, I prefer skills that are interesting, even if they are difficult to learn (M) In a class like this, I prefer skills that really challenge me so I can learn new things (M) I often think to myself ‘What if I do badly in this class?’ (Pav) I worry about the possibility of getting a bad grade in this class (Pav) My fear of performing poorly in this class is often what motivates me (Pav) I just want to avoid doing poorly in this class (Pav) I am afraid that if I ask my PE teacher a dumb question, they might not think I’m very smart (Pav) I wish this class was not graded (Pav)

AAAGQ

Appendix. The TEOSQ and AAAGQ items adapted to PE

702 Vassilis Barkoukis et al.