2

Achievement Goals and Rasch Analysis

Running Head: Achievement Goals and Rasch Analysis

A Different Look at Achievement Goals in Physical Education:

A Rasch Analysis Approach

Athanasios Mouratidis

Willy Lens

University of Leuven

Maarten Vansteenkiste

Ghent University

Georgios D. Sideridis

University of Crete

Abstract

In two cross-sectional studies, we investigated the extent to which elementary (Study 1) and middle school (Study 2) students pursue similar, yet distinct mastery- and performance-related goals in physical education. We found that students were more likely to endorse outcome goals next to mastery-related goals and ability goals next to normative goals. Rash modeling suggested that students were more likely to endorse mastery-approach goals compared to learning and outcome goals and they were more likely to favor ability goals compared to performance-approach goals. Differential item functioning analyses further showed that autonomously motivated students were more likely to endorse learning goals and mastery-approach goals and less likely to endorse outcome goals compared to less autonomously motivated students. Concerning performance-related goals, they were also more likely to endorse ability goals and less likely to endorse normative goals compared to controlled motivated students. Results are discussed within the achievement goal framework (Dweck, 1986; Elliot, 1999) and the self-determination theory (Ryan & Deci, 2002).

Keywords: physical education; achievement goals; self-determination theory; Rasch model


A Different Look at Achievement Goals in Physical Education:

A Rasch Analysis Approach

Within the achievement goal perspective, different types of achievement goals are said to create the framework within which students construe and react to achievement tasks and their outcomes (Dweck & Leggett, 1988). Initially, two class of goals were proposed: Mastery goals which are said to direct behavior in achievement tasks towards the development of competence mainly by means of improvement, learning, and challenge-seeking and performance goals which are said to direct such behavior towards demonstration of competence, by outperforming others or by avoiding performing worse than others (Dweck & Leggett, 1988; Elliot, 1999; Nicholls, 1984; Elliot, 1999). Over the past decade, several achievement goal researchers have argued and shown that the dichotomous framework needs refinement. For instance, the earlier approach-avoidance distinction (Atkinson, 1968) was incorporated in the mastery-performance dichotomy such that a 2x2 (i.e., mastery vs. performance X approach vs. avoidance) achievement goal model was developed (Elliot & McGregor, 2001). Furthermore, Grant and Dweck (2003) proposed two additional classes of achievement goals, that is, outcome goals and ability goals and, most recently, Sideridis (2008) suggested refining performance-approach goals by introducing ought-approach and ought-avoidance goals as a special subclass of performance goals that are characterized by a sense of pressure.

The latter classes of achievement goals are likely to have been introduced because the 2x2 model and its four types of achievement goals did not cover the full range of possible achievement goals that students might pursue in achievement settings. Furthermore, it is likely that some of these newly proposed goals (e.g., ability goals or outcome goals) are more strongly endorsed by students. Indeed, Brophy (2005) argued that performance goals are not frequently endorsed by students and, hence, may lack ecological validity. In the present study, we used a Rasch-model analysis to examine the odds of endorsing these different types of mastery-like and performance-like goals as a function of people’s latent trait score. Because the probability of endorsing particular types of mastery-like and performance-like goals over others is directly compared, such analyses may help to answer the question which achievement goals are most likely driving students’ achievement strivings.

A second aim of the present research involved testing whether the likelihood of endorsing each of these achievement goals would be different for students reporting different levels of autonomous versus controlled reasons for their participation in Physical Education (PE) class-related activities. To answer these questions we conducted two cross-sectional studies, one with elementary and one with middle school students.

Achievement Goal Perspective

Under the initial, dichotomous definition of achievement goals (Ames, 1992; Dweck, 1986), the aims and the reasons were interwoven such that mastery-oriented students were said to be focused on mastering the requirements of the task at hand (i.e., aim) in order to develop their skills or to seek challenge (i.e., reason), whereas performance-oriented students would be focused on outperforming others or avoiding worse performance than other (i.e., aim) as to show their ability or to prove their self-worth (i.e., reason).

Starting from this dichotomous framework, achievement goal researchers have undertaken several attempts to refine the definition of mastery and performance goals (e.g., Elliot, 1999; Elliot & Harackiewicz, 1996). In particular, Elliot and associates (1999, Elliot & McGregor, 2001) suggested that achievement goals should be exclusively defined according to the type of reference criteria (i.e., mastery-, intra-personal or interpersonal) with which individuals construe competence. Elliot and colleagues (1999; 2005; Elliot & Thrash, 2001; Elliot & Fryer, 2007) thus argued that aims form the conceptual centerpiece of achievement goals and that overarching further reasons (e.g., to show one’s ability) behind aims should be left out from the definition. This clear-cut definition of achievement goals enabled researchers to suggest some additional achievement-related goals that would be also defined either as pure aims or as aims which are explicitly intertwined with some underlying reasons.

Within this definition of achievement goals as pure aims, Grant and Dweck (2003) have recently proposed two additional classes of achievement goals, outcome goals and ability goals. Outcome goals focus on the attainment of certain levels of performance (e.g., to jump higher than 2 meter; score at least 75/100) and ability goals on the validation of one’s ability using only task-referential criteria without using any social comparisons. Grant and Dweck (2003) theorized and adequately showed that outcome and ability goals are both related to mastery goals (as their standard of evaluation is task-referenced) and performance goals (as their focus is on the outcome or the validation of one’s competence). Grant and Dweck (2003) thus proposed a distinction among (a) mastery related goals, consisted of learning goals (i.e., goals focusing on developing new skills and acquiring new knowledge) and challenge-seeking goals (i.e., goals pertaining to seeking out challenging tasks), (b) ability goals (i.e., goals aiming at validating one’s ability), (c) outcome goals (i.e., goals aspiring to the attainment of certain outcomes), and (d) normative goals, consisted of normative ability goals (i.e., goals focusing on validating one’s ability relative to others) and normative outcome goals (i.e., goals aiming at outperforming others). In a series of studies with university students, Grant and Dweck (2003) showed unique pattern of relations of mastery-related goals (i.e., learning and challenge-seeking goals), outcome goals, ability goals, and normative goals (i.e., normative ability and normative outcome goals) to motivational outcomes.

Similarly and in line with the differentiation between aims and underlying reasons, Sideridis (2008) attempted to refine performance-approach goals by introducing ought-approach goals as a special subclass of performance-approach goals. Ought-approach goals refer to the pursuit of normative competence out of a sense of obligation and pressure; one has the feeling that has no other choice than to outperform others. Although Sideridis (2008) distinguished at the theoretical level between the achievement goal of outperforming others and the underlying obligation that might characterize such goal-pursuit, the ought-approach goal items yielded a simultaneous reference to the pursuit of normative competence and the experience of pressure during this pursuit (e.g., ‘I feel I ought to better than my classmates’). Thus, aims and underlying reasons were not empirically distinguished.

In short, the number of achievement goals that have been proposed and studied over the past few years has steadily increased. Although this increase was in part due to conceptual refinements, some of the studied achievement goals (e.g., outcome goals) were likely included in the list of achievement goals because they yielded strong ecological validity. So, they were worth being examined on their own. Some other achievement goals, such as performance-approach goals, have on the contrary been criticized for being lowly subscribed by students (e.g., Brophy, 2005).

In this paper, we aimed to add to the achievement goal literature by simultaneously studying all of these recently proposed achievement goals as few, if any, studies have undertaken such an attempt. Specifically, we aimed to investigate to what extent students pursue in the PE class the following seven goals: (a) mastery-approach goals (Wang et al., 2007), (b) learning goals, (c) challenge-seeking goals, (d) outcome goals, (e) ability goals (Grant & Dweck, 2003), (f) performance-approach goals (Wang et al., 2007), and (g) ought-approach goals (Sideridis, 2008). Although mastery-approach goals are conceptually similar with learning goals, we included them because we wanted to use mastery-approach and performance-approach goals as these have been defined by Elliot and McGregor (2001) as the anchors for assessing the other types of goals. We were particularly interested in examining the extent to which these newly introduced goals (e.g., ability goals or outcome goals) are endorsed by students. To address this question we conducted two cross-sectional studies in which we examined with a more sophisticated analysis – a Rasch model analysis – participants’ relative likelihood in endorsing achievement goals that are assumed to have a common conceptual basis.

Rasch analysis allows one to test the odds of endorsing each of these goals as a function of students’ preference for these goals (i.e., latent goal trait). Specifically, rather than asking students to indicate their agreement or disagreement with each set of achievement goals, we provided the students a free-choice list of achievement goals that they might pursue in the PE class and we asked them to select, if any, the goals they usually endorse in PE. In this respect, we addressed the possibility of students’ indifference to pursue any particular achievement goal in their PE classes (see Brophy, 2005; Dowson & McInerney, 2003). Apart from examining the probability of endorsing various mastery- and performance-related goals along the respective continuum of students’ latent trait, we investigated whether the endorsability of the different achievement goals would vary as a function of students’ autonomous versus controlled reasons to engage in PE activities. This issue is discussed next.

Self-regulation of Motivated Behavior in Physical Education

According to SDT (Deci & Ryan, 2000; Ryan & Deci, 2002), students exhibit volitional or self-determined forms of motivation when they initiate or persist in activities which (a) are perceived as interesting and pleasant, thereby providing inherent satisfaction (intrinsic motivation), (b) are aligned with their personally endorsed values, goals, and needs (integrated regulation), or (c) are realized as personally important (identified regulation). In contrast, students exhibit controlled motivation when they participate in activities for reasons that are (d) either internally pressuring such as enhancing feelings of self-worth or avoiding feelings of guilt or shame (introjected regulation), or are (e) externally pressuring such as obtaining some rewards or avoiding punitive outcomes (external regulation). SDT assumes that students who undertake activities for intrinsic, integrated, or identified reasons are autonomously motivated because they feel that they are agents of their own acts. Conversely, students who are engaged for introjected or external reasons are motivated by controlling reasons because they feel some kind of internal or external pressure for their task engagement and goal striving (Ryan & Deci, 2002).

The literature on SDT in PE and sport settings has clearly demonstrated that compared to controlled motivation, autonomous forms of motivation are associated with intrinsic motivation, integrated functioning, and well-being (Wilson & Rodgers, 2008). For instance, experimental studies which manipulated students’ autonomous motivation indirectly by satisfying or thwarting individuals’ need for autonomy and cross-sectional studies have shown that autonomous (versus controlled) motivation is positively related to adjustment (see Pelletier & Sarrazin, 2008), well-being (McDonough & Crocker, 2007; Standage, Duda, & Ntoumanis, 2005), and intrinsic motivation (Chatzisarantis, Hagger, Biddle, Smith, & Wang, 2003; Hagger, Chatzisarantis, Culverhouse, & Biddle, 2003; Ntoumanis, 2005; Pelletier, Fortier, Vallerand, & Brière, 2001; Standage & Treasure, 2002; Wilson Rodgers, Blanchard, & Gessell, 2003).

Relevant to the present study is the consistent finding that mastery goals are positively associated with autonomous forms of motivation. Theoretically speaking, autonomous motivation is conceptually related to mastery goals (Deci & Ryan, 2000; Dweck & Leggett, 1988). Indeed, several studies conducted in the PE and the sport domain have shown that mastery goals are related to autonomous motivation, namely intrinsic motivation and identified regulation, and that performance goals are related to controlled forms of motivation, namely introjected and external regulation (Nien & Duda, 2008; Ntoumanis, 2001; Petherick & Markland, 2008). The present research aimed to extend these findings by examining the link of autonomous and controlled motivation to mastery-related goals, outcome goals, ability goals, and normative goals by testing whether the odds of endorsing each of these goals would be different among students reporting high versus low autonomous reasons for their participation in PE.

The Present Research

Starting from the more recently proposed different classes of achievement goals (Elliot, 2005; Grant & Dweck, 2003; Sideridis, 2008), we aimed to use Rasch-analyses to provide new insight in these evolutions. We had three aims in our study. First, with the help of Rasch model analysis we tested the odds of endorsing mastery- and performance-related goals as a function of students’ preference for these goals (i.e., their latent goal trait). Second, given that one needs to model achievement goals as a function of an underlying trait before one can conduct Rasch-analyses, we intended to examine how the different achievement goals would fall apart in factor analyses. We were especially interested in where ability goals and outcome goals would fall as both of them use a task- or a self-referenced standard of competence evaluation but still focus on the outcome or the validation of one’s competence. Third, we investigated whether the likelihood of endorsing any goal would differ between students reporting autonomous rather than controlling reasons for their engagement in PE class activities.

We hypothesized that despite their subtle conceptual differences mastery-approach aim-only goals (Wang et al., 2007), learning goals, and challenge-seeking goals (Grant & Dweck), would be positively interrelated. We also anticipated that outcome goals would be loaded on the same factor with the mastery-related goals because although by definition outcome goals are assumed to focus on the outcome rather than to the process, the evaluative criteria they use are still task-referential. Accordingly, we predicted that performance-approach aim-only goals (Elliot, 2005; Wang et al., 2007), and ought-approach goals (i.e., the performance-approach goal which are undertaken for some internally pressuring reasons) would be positively correlated. We hypothesized that ability goals would be positively related to performance-related goals because, although they use a self- or task-referenced standard of evaluation, they do instill a sense of self-validation relative to one’s competencies (Nicholls, 1984; Roberts, 2001; Ryan, 1982).