Psychology 2013. Vol.4, No.8, 629637 Published Online August 2013 in SciRes (http://www.scirp.org/journal/psych) http://dx.doi.org/10.4236/psych.2013.48090 Copyright © 2013 SciRes. 629 A Multilevel Analysis of Classroom Goal Structures’ Effects on Intrinsic Motivation and Peer Modeling: Teachers’ Promoting Interaction as a Classroom Level Mediator Kazuhiro Ohtani1, Ryo Okada2, Takamichi Ito3, Motoyuki Nakaya4 1Department of Human Sciences, University of Osaka, Osaka, Japan 2Department of Education, University of Kagawa, Takamatsu, Japan 3Department of Education, Kyoto University of Education, Kyoto, Japan 4Departmentof Education and Human Development, Nagoya University, Nagoya, Japan Email: yamato@hus.osakau.ac.jp Received June 7th, 2013; revised July 8th, 2013; accepted July 17th, 2013 Copyright © 2013 Kazuhiro Ohtani et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This study investigated how classroom goal structures (mastery and performance goal structures) related to intrinsic motivation and peer modeling focusing on teachers’ promoting interaction as a classroom level mediator. Authors tested multilevel mediational models with a sample of 1212 Japanese elementary and junior high school students from 43 classrooms. Results provided support for classroom level media tional process in mastery goal structures. Specifically, mastery goal structures related to promoting inter action, and promoting interactions related to both intrinsic motivation and peer modeling. Limitations and future directions of the study were discussed. Keywords: Classroom Goal Structures; Promoting Interactions; Intrinsic Motivation; Peer Modeling Introduction Students’ learning is influenced by educational environment factors as well as individual factors. In particular, the classroom context, in which students spend a great deal of time every day, can be the most powerful predictor of achievement behaviors and academic outcomes. Previous motivation research has fo cused on classroom goal structures and has produced a number of findings (Kaplan, Middleton, Urdan, & Midgley, 2002). This study examined the effects of classroom goal structures in terms of teachers promoting interaction among students. Achievement Goal Theory Recent motivational research has been led by achievement goal theory (e.g., Ames, 1992; Dweck & Leggett, 1988; Maehr & Zusho, 2009; Nicholls, 1984). Research on achievement goal theory has focused on the reasons why students try to succeed and has examined two types of achievement goals: mastery goal (also called learning goal or task goal) and performance goal (also called relative ability goal or ego goal). Mastery goals are defined in terms of focusing on developing one’s competence, mastering a new skill, trying to accomplish a chal lenging task, and trying to understand learning materials. Suc cess is evaluated in terms of selfimprovement and internal norms. By contrast, performance goals are defined in terms of demonstrating competence and having high ability relative to others, striving to be better than others, and using social com parison standards to judge ability and performance. Success is derived from doing better than others and surpassing normative performance standards. In recent years, some researchers have added an approachavoidance dimension and proposed a 2 × 2 framework of achievement goals (Elliot, 2005; Elliot & McGregor, 2001). This conceptualization proposes four types of achievement goals: 1) masteryapproach goals, which focus on developing competence through task mastery; 2) mastery avoidance goals, which focus on the avoidance of taskbased and intrapersonally based incompetence; 3) performanceap proach goals, which focus on attaining competence relative to others; and 4) performanceavoidance goals, which focus on avoiding incompetence relative to others. However, few em pirical studies have focused on masteryavoidance goals. Many researchers have examined individuallevel achieve ment goals (i.e., personal goals). Personal goals refer to the spe cific goals that individuals strive to attain in achievement con texts. Personal achievement goals were found to be associated with various achievement behaviors and academic outcomes (Meece, Anderman, & Anderman, 2006; Urdan & Schoenfelder, 2006). Personal mastery goals were found to be associated with high academic selfefficacy, high levels of effort and persis tence, use of learning strategies that enhance conceptual under standing, and intrinsic motivation. In contrast, pursuing perfor manceavoidance goals is usually associated with a negative pattern of motivational behaviors such as lesser persistence, avoidance of helpseeking, and selfhandicapping. The effects of performanceapproach goals are not sufficiently clear. Al though pursuing performanceapproach goals is associated with less persistence and use of surfacelevel learning strategies, stu dents with higher levels of performanceapproach goals tend to
K. OHTANI ET AL. achieve higher academic performance (Harackiewicz, Barron, Pintrich, Elliot, & Thrash, 2002). Classroom Goal Structures In addition to personal goals, achievement goals have been treated as a contextuallevel factor of the classroom and school. Goal structures refer to messages in the learning environment that make certain goals salient (Ames, 1992). Classroom focused research has examined how teachers may create dif ferent goal structures in their classrooms by using various in structional, evaluation, and grouping strategies (Kaplan et al., 2002). Ames and Archer (1988) designed and used selfreport measures to assess the salience of mastery and performance goals in the classroom. In their study, students’ perceptions of mastery goals were related to the use of learning strategies, preference challenge, and effort attribution. In contrast, percep tions of performance goals were associated with ability attribu tion and less favorable attitudes toward the class. Ames (1992) reported that mastery goals are made salient when value is placed on the process of learning through emphasis on mean ingful learning, selfreferenced standards, and opportunities for selfdirected learning. Researchers inspired by Ames’ work have examined the ef fects of goal structures on achievement behaviors and academic outcomes. Mastery goal structures were found to be associated with academic selfefficacy (Midgley, Anderman, & Hicks, 1995; Roeser, Midgley, & Urdan, 1996), intrinsic motivation (Murayama & Elliot, 2009), and selfregulated learning strate gies (Wolters, 2004). In contrast, performance goal structures were related to academic selfconsciousness (Roeser et al., 1996), avoidance of academic helpseeking (Ryan, Gheen, & Midgley, 1998), and disruptive behavior (Kaplan, Gheen, & Midgley, 2002). However, some researchers reported no nega tive effects of performance classroom goal structures (Midgley et al., 1995; Ryan & Patrick, 2001). Classroom goal structures also have longitudinal effects on academic outcomes. Urdan and Midgley (2003) investigated the relationship between the change in goal structures from fifth to sixth grades and aca demic selfefficacy and academic achievement. Students who perceived an increase or no change in the mastery goal struc tures reported higher academic selfefficacy and attained higher academic achievement than students who perceived a decrease in the mastery goal structures. Many researchers have focused on students’ perceptions of classroom goal structures. On questionnaires, students reported their subjective perceptions of the emphasis in the classroom with respect to the purposes for engaging in academic tasks and the meanings associated with success and achievement. A ques tionnaire measuring students’ perceptions of classroom goal structures has been developed and validated (Midgley, Maehr, Hruda, Anderman, Anderman, Gheen, Kaplan, Kumar, Mid dleton, Nelson, & Urdan, 2000). Recently, classroom goal structures tend to be considered as grouplevel variables. A growing body of research distinguishes between classroom level variance and studentlevel variance and examines the effects of classroom goal structures by using multilevel analysis (Kaplan et al., 2002; Ryan et al., 1998; Turner, Midgley, Meyer, Gheen, Anderman, & Kang, 2002). Because goal structure in any classroom is shared by students in the class, it is necessary to note the difference between variance of goal structures as a characteristic of the classroom and variance of goal structure as individual students’ perceptions of the classroom. Another line of researches has focused on the relationships between classroom goal structures and teachers’ daily educa tional practices. Ames (1992) discussed how teachers’ educa tional practices influence whether classrooms have characteris tics of mastery goal structures or performance goal structures and proposed some instructional strategies. Some researchers with qualitative data revealed that teachers’ discourse and in structional strategies had original characteristics in mastery or performance goal classrooms (e.g., Patrick, 2004; Patrick, An derman, Ryan, Edelin, & Midgley, 2001; Turner, Meyer, Midg ley, & Patrick, 2003; Turner et al., 2002). For example, Turner et al. (2002) observed sixth grade mathematics classrooms and stated that teachers conveyed mastery messages to their stu dents through explicit admonitions not to feel inadequate or ashamed when they did not understand. Ames (1992) stated that teachers’ instructional strategies led to students’ mastery goals. Teachers’ Promoting Interaction as a Mediator We focused on teachers promoting interaction as a teachers’ educational practice that conveys mastery message to the stu dents. Promoting interaction is one of the social climate of classroom that affects students’ motivation and achievement, and refers to teachers’ encouraging peer interaction among students, such as sharing ideas during wholeclass lessons, working together in smallgroup activities, or informal help seeking and helpgiving during individual seatwork (Ryan & Patrick, 2001). Ryan and Patrick (2001) reported that teachers varied in the extent to which they encouraged student interac tion, and the encouragement led to students’ academic achieve ment and motivation. Patrick, Ryan, and Kaplan (2007) exam ined the relationship between fifthgrade students’ perceptions of promoting interaction (e.g., “My teacher often allows us to discuss our work with classmates”) and their motivation. In the study, students who perceived that their teachers promoted interaction reported higher levels of academic selfefficacy and more taskrelated interaction. In addition, Patrick (2004) re viewed the relevant researchers and concluded that promoting taskrelated interaction is related to students’ liking and interest of school and subject area (these are the components of intrinsic motivation). In early literature of achievement goal theory, cooperative goal structures had been considered an important factor. Ames and Ames (1984) identified three types of classroom goal structures: cooperative, competitive, and individualistic goal structures. Ames and Ames (1984) defined cooperative goal structures as situations in which the probability of one student receiving a reward is enhanced by the presence of capable others. Cooperative goal structures involve interdependence among group members, where each member shares responsibil ity for the outcome or product. However, three types of goal structures evolved into distinctions between competitive envi ronment (i.e., performance goal structures) and noncompetitive environment (i.e., mastery goal structures), and cooperative goal structures and individualistic goal structures were inte grated into mastery goal structures in the conceptualization process (Patrick, 2004). Thus, the mastery goal structures seem to include a nature of cooperative goal structures. That is, it is thought that teachers use cooperative educational strategies (i.e., promoting interaction) in mastery goal classrooms. Classroom goal structures and teachers’ educational instruc Copyright © 2013 SciRes. 630
K. OHTANI ET AL. tional practices seem to intertwine each other in the actual classrooms. In this study, we assumed the process in which classroom goal structures influence students’ motivation and achievement related behaviors through teachers’ instructional practices on the basis of Patrick, Kaplan, and Ryan’s (2011) findings and discussion. Patrick et al. (2011) stated that “when students evaluate the extent to which their teacher promotes real understanding and personal improvement (i.e., emphasizes mastery goal structure), they draw from their perceptions of the teacher’s messages about interpersonal relationships (e.g., sup port, respect, helping one another)” (p. 370). In addition, Pat rick et al. (2011) found that mastery goal structures could ac count for the variance of social climate dimensions involving teachers promoting interaction. Although the mastery goal structures and teachers promoting interaction can be related reciprocally, we assumed the process in which the mastery message in the classroom are conveyed through teachers’ edu cational practice and treated the promoting interaction as a mediator of the mastery goal structures. The Peer Modeling It is possible that teachers promoting interaction influences students’ use of selfregulated learning. It is especially likely that students could develop learning strategies mediated by their peers (i.e., classmates or friends) in mastery goal class rooms. We focused on students’ use of peer modeling as a selfregulated learning strategy. Peer modeling is the cognitive, affective, and behavioral change that derives from observing models (Schunk, 1998). In the classroom setting, children not only learn from direct interaction but also acquire knowledge and enhance their motivation by observing peers’ learning be haviors. Therefore, peer modeling plays an important role in the learning process through which children acquire skills, behav iors, and beliefs. Experimental research has revealed that peer modeling has a significant effect on children’s academic achievement and mo tivation (Schunk, 1998; Schunk & Hanson, 1985, 1989). Most of the findings about peer modeling have emerged from ex perimental studies in laboratories, and few studies have paid attention to the peer modeling that students use as learning strategies in their classrooms. However, peer modeling could play an important role in daily classroom learning activities (Brophy, 2004; Ryan, 2000). This study examined the effects of mastery goal structures mediated by teachers promoting inter action on students’ use of peer modeling in the classroom. It was possible that peer modeling was positively related to intrinsic motivation. Research of selfregulated learning has found that intrinsic motivation was associated with more use of various selfregulated learning strategies (Meece, Blumenfeld, & Hoyle, 1988; Pintrich & De Groot, 1990; Yamauchi & Ta naka, 1998; Young, 2005). This means that intrinsically moti vated students tend to use various selfregulated learning strate gies in the learning situations. No research has examined the relationship between intrinsic motivation and peer modeling. Given that peer modeling is one of the selfregulated learning strategies, however, it was expected that intrinsic motivation was positively related to peer modeling in the classroom, al though the relationship was not a main focus of this study. The Present Study We examined the effects of classroom goal structures on the students’ motivation and peer modeling in terms of teachers promoting interaction. Based on the previous findings (Mura yama & Elliot, 2009; Patrick, 2004), this study focused on in trinsic motivation as students’ motivation. We predicted that mastery goal structures would influence intrinsic motivation and peer modeling through teachers promoting interaction. As a school subject, we focus on mathematics classes. Turner and Meyer (2009) suggested that many children tend to experience difficulties in mathematical learning and consequently lose their confidence. Some researchers have examined the effects of goals structures in math classrooms (Kaplan et al., 2002; Ryan et al., 1998; Turner et al., 2002). Thus, we measured goal structures, teachers promoting interaction, intrinsic motivation, and peer modeling just in math classes. In this research, classroomlevel variances and studentlevel variances were partitioned, and we tested a multilevel media tion model. Because mastery goal structures and promoting interaction were classroomlevel variables, we expected that the above mediation process would be confirmed, especially at the classroom level. Method Participants and Procedure The sample consisted of Japanese public elementary (fifth and sixth graders) and junior high school students (seventh and eighth graders) in 43 classrooms (17 elementary, 26 junior high) in 10 schools from Kansai region. The final sample included a total of 1212 students (617 females, 591 males, 4 unspecified). Schools were recruited by contacting the board of education and then school administrators. Schools were selected by the board of education of the cities so that the diversity of students was maintained. Measures Classroom goal structures, teachers promoting interaction, peer modeling, and intrinsic motivation were all assessed with a questionnaire. All used a 6point scale, ranging from 1 (not at all true) to 6 (very true). Each variable in this study was spe cific to the domain of mathematics. Classroom goal structures. To assess classroom goal struc tures, ten items were created based on the works by Miki and Yamauchi (2005) and Midgley et al. (2000). These consisted of five items assessing mastery goal structure (sample item: “In math class, learning new things is important”) and five items assessing performance goal structure (sample item: “In math class, getting a good grade is the important thing for students”). The reliability and validity of the original scale have been documented in prior studies (Midgley et al., 2000; Miki & Ya mauchi, 2005). To assess internal consistency of the scales, coefficient omega was calculated using a withinclassroom covariance matrix. The coefficient omegas were .73 for mastery goal structure and .60 for performance goal structure. Promoting interaction. To assess teachers promoting interac tion, we developed four items based on the Ryan and Patrick’s (2001) scale. Items were: (1) “In math class, our teacher says that cooperation in the group is important,” (2) “In math class, our teacher says that talking to each other in a group is an im portant thing,” (3) “In math class, our teacher emphasized that you need to help friends when you know the answers,” and (4) “In math class, our teacher says that asking peers for help is Copyright © 2013 SciRes. 631
K. OHTANI ET AL. Copyright © 2013 SciRes. 632 important when one doesn’t know the answer.” Acceptable internal consistency was identified (ω = .77). Intrinsic motivation. To assess students’ intrinsic motivation for learning mathematics, we used Tanaka and Yamauchi’s (2000) scale. Tanaka and Yamauchi’s measure is a Japanese version of Elliot and Church’s (1997) intrinsic motivation scale. For the purpose of this study, we selected three items that have shown high factor loadings on the basis of factor analysis in Tanaka and Yamauchi (2000). Items were (1) “I think this class is interesting,” (2) “I enjoy this class very much,” and (3) “I think this class is fun.” The reliability and validity of the origi nal scale have been documented in prior studies (Elliot & Church, 1997; Tanaka & Yamauchi, 2000). The scale showed high internal consistency (ω = .93) in this study. Peer modeling. To assess peer modeling, four items were created based on Okada, Nakaya, Ito, and Ohtani’s (2010) peer modeling scale and the conceptual description of peer modeling (Schunk, 1998). Items were (1) “I try to study like my friends with high motivation to learn,” (2) “I try to follow friends’ good performance,” (3) “When I cannot solve the problem, I try to model my friends who can solve the problem,” and (4) “When studying, I try to learn like my friends do.” The scale was internally consistent (ω = .82). Results Descriptive Statistics and Correlations of Study Variables Descriptive statistics and correlations are presented in Table 1. Correlations were calculated at the classroom level and stu dent level. At the classroom level, mastery goal structure was positively related to promoting interaction (r = .77, p < .001), intrinsic motivation (r = .86, p < .001), and peer modeling (r = .55, p < .05). Performance goal structure was not strongly related to other variables except for mastery goal structure (r = −.50). The same results were obtained at the student level. Mas tery goal structure showed positive correlations with promoting interaction (r = .59, p < .001), intrinsic motivation (r = .28, p < .001), and peer modeling (r = .45, p < .001). Correlation co efficients of performance goal structure were all below .21. Analysis Plan This research tested how classroom goal structures related to intrinsic motivation and peer modeling using hierarchical linear modeling. As a first step, we conducted a preliminary analysis to check our variables’ betweenclassroom variances and intra class correlations. Next, we tested a multilevel mediation model (Zhang, Zyphur, & Preacher, 2009) in which classroom goal structures indirectly influence outcome variables through pro moting interaction. In the multilevel mediation model, vari ances of all variables were partitioned into classroom level and student level. Then, mediations were tested in both levels. Since our measurement of classroom level variables were students’ perception, our multilevel mediation model equivalent to Zhan et al.’s (2009) 111 model. In 111 model, aggregated indica tors of classroom goal structures and promoting interaction in each classroom were created for testing Level 2 mediational model. The 111 model is beneficial because it provides the information of both student and classroom level processes. The regression coefficients were reported in unstandardized coeffi cients. A multiple imputation technique was used in all subsequent analyses. The method was beneficial compared to other tech niques such as mean substitution or listwise deletion. PASW 18.0 Missing Values was used to create five imputed data sets and integrated results were reported. All subsequent analyses were performed by HLM 7 with maximum likelihood estima tion. Preliminary Analysis The analyses tested whether the variables were significantly varied among classrooms. The classroom goal structures sig nificantly varied among classrooms: mastery goal structure, χ2 (42) = 206.48, p < .001; performance goal structure, χ2 (42) = 265.91, p < .001. Intraclass correlations were 12% and 16% respectively. We also checked reliability of classroom goal structures in case that the aggregation of student rating cause biased estimation in the classroom level (Bliese, 2000). The reliabilities were relatively high, .79 and .84, respectively. We also examined whether our mediator variable (i.e., pro moting interaction) and outcome variables (intrinsic motivation and peer modeling) varied among classrooms and found the following significant betweenclass variances: for promoting interaction, χ2 (42) = 319.5, p < .001, for intrinsic motivation, χ2 (42) = 131.82, p < .001, and for peer modeling, χ2 (42) = 61.61, p < .05. Intraclass correlations were 18%, 7%, and 1%, respec tively. The reliability of promoting interaction was .86. Multilevel Mediation Model Intrinsic motivation. For intrinsic motivation, we tested a multilevel mediation model (Zhang et al., 2009). As the first step of the meditational analysis, classroom goal structures were put in the model as both classroomlevel and studentlevel predictors for the intercept of intrinsic motivation. Studentlevel variables were centered at group mean (Zhang et al., 2009). We Table 1. Descriptive statistics and Pearson productmoment correlations for the study variables. N M SD 1 2 3 4 5 N M SD 1 Mastery goal structure 1172 4.50 0.89  −.50*** .77*** .86*** .55* 43 4.48 0.35 2 Performance goal structure 1146 3.23 0.97 .05  −.11 −.25† .27 43 3.26 0.43 3 Promoting interaction 1146 4.04 1.01 .59*** .15***  .87*** .81*** 43 4.04 0.46 4 Intrinsic motivation 1157 3.32 1.58 .28*** .12*** .20***  .72** 43 3.33 0.50 5 Peer modeling 1153 3.96 1.10 .45*** .21*** .46*** .36***  43 3.96 0.25 Note: Descriptive statistics and correlation coefficients above diagonal represent classroom level, below diagonal represent student level. †p < .10, **p < .01, ***p < .001.
K. OHTANI ET AL. controlled grade (fifth grader = 0, sixth grader = 1 seventh grader = 2, eighth grader = 3) at classroom level, and gender (female = 0, male = 1) at student level because prior research has documented those variable are significant predictors of academic outcomes (Hyde & Durik, 2005; Lepper, Corpus, & Iyengar, 2005). In the classroom level, the results suggested that mastery goal structure influenced intrinsic motivation (γ = 1.32, p < .001). In contrast, performance goal structure did not reach significance, so we omitted performance goal structure from subsequent analyses at both levels (see Appendixes 1 and 2 for the final models and results). Next, we predicted that mastery goal structure was associated with promoting interaction and found significant positive relationships at classroom level (γ = 1.24, p < .001). Finally, promoting interaction was put in the model as a mediator to predict the intercept of intrinsic motiva tion. Promoting interaction showed a significant relationship to intrinsic motivation (γ = 0.38, p < .05) at the classroom level. At this time, the coefficients of mastery goal structure reduced its value from γ = 1.32, p < .001 to γ = 0.84, p < .01. The results of Sobel’s test determined that this mediation was significant (z = 2.37, p < .05). In the student level of mediational analysis, student level mastery goal structure also a positive predictor of intrinsic mo tivation (γ = 0.50, p < .001), as well as promoting interaction (γ = 0.65, p < .001). In contrast to findings at the classroom level, studentlevel promoting interaction did not show a significant relationship to intrinsic motivation (γ = 0.09, n.s.). Figure 1 and Table A1 shows the final result. Peer modeling. For peer modeling, we tested a multilevel mediation model with the same procedures as were used for intrinsic motivation. As the first step of the meditational analy sis, classroom goal structures were put in the model as both classroom and studentlevel predictors for the intercept of peer modeling. Grade and gender were controlled. In the classroom level, the results suggested that mastery goal structure influenced peer modeling (γ = 0.56, p < .001). Finally, we tested the mediation and found a significant rela tionship between promoting interaction and peer modeling (γ = 0.23, p < .05) at classroom level (see Appendix 1 for the final models)1. This mediation was significant (z = 2.40, p < .05). At classroom level, the coefficients of mastery goal structure re duced its value from γ = 0.56, p < .001 to γ = 0.27, n.s., and the value was no longer significant. This means full mediation. At student level, mastery goal structure was a positive pre dictor of peer modeling (γ = 0.50, p < .001). Then promoting interaction predicted peer modeling (γ = 0.34, p < .001). Stu dent level mediation was significant (z = 5.42, p < .001). The final results of the model are presented in Figure 2 and Table A2. Discussion The purpose of this study was to examine the effects of class room goal structures on students’ motivation and achievement related behaviors and tested the mediation effects of teachers promoting interaction. Our hypothesis was that mastery goal structures influenced intrinsic motivation and peer modeling through teachers promoting interaction. It is noteworthy that we partitioned between classroomlevel variances and studentlevel variances and examined the mediation processes. Mastery goal structures and teachers promoting interaction were classroom variables by nature. The characteristics of the classroom envi ronment and teachers’ educational practice should be shared by students in the same classroom. However, each student can perceive the characteristic differently. Thus, it is necessary to treat the grouplevel variables (i.e., goal structures and teachers promoting interaction) after dividing the variance to class roomlevel variance and studentlevel variance. Hereafter, we would discuss the mediation process considering this partition ing. At the classroom level, the hypothesized mediation process for intrinsic motivation was confirmed. That is, mastery goal structure was related to students’ intrinsic motivation, and the relationships were partly mediated by teachers promoting in teraction. This implies that teachers frequently encourage stu dents to interact with their peers in the classrooms that have the characteristics of mastery goal structures, and the interaction increases the average level of intrinsic motivation for mathe matical learning. The advantages of classrooms with mastery goal structures are mediated by teachers promoting interaction among students. In contrast, the mediation process was not fully confirmed at the student level. Although the perception of mastery goal structure was associated with the perception of teachers promoting interaction, it did not predict intrinsic moti vation significantly. Mastery goal structure was directly related to intrinsic motivation. The student level results are consistent with considerable research on the positive effects of mastery goal structures (Meece et al., 2006; Midgley et al., 1995; Mu rayama & Elliot, 2009; Roeser et al., 1996; Urdan & Midgley, 2003; Urdan & Schoenfelder, 2006; Wolters, 2004). In sum, in classrooms that emphasize mastering new skills, the average level intrinsic motivation is high. However, the effects of the characteristics of mastery goal structures shared by students in a classroom are mediated by classroomlevel variables such as teachers’ instructional strategies (i.e., promoting interaction). The hypothesized mediation process at the classroom level was also confirmed for peer modeling. Mastery goal structure related to students’ peer modeling and the relationship was me diated by teachers promoting interaction. The more classrooms are mastery oriented, the more students in the classrooms tend to model peers’ learning behaviors with their teachers’ encour agement. At the student level, the mediation process was also confirmed. Students who perceive more mastery goal structure than the average of their classmates tend to perceive the em phasis of promoting interaction and, then, report higher level of peer modeling. Peer modeling has been mainly examined in experimental studies (Schunk, 1998). In contrast, this study sug gested that students use peer modeling as a kind of learning strategy, and the use is influenced by the characteristics of classrooms. Performance goal structures did not predict teachers promot ing interaction or intrinsic motivation at either the classroom or student levels, although they did predict peer modeling. The mediation processes of performance goal structures were not totally confirmed. Some previous studies reported negative ef fects of performance goal structures (Kaplan et al., 2002; Roe ser et al., 1996; Ryan et al., 1998), and others reported no nega tive effects (Midgley et al., 1995; Ryan & Patrick, 2001). Nega tive effects of performance goal structures were not found 1Performance goal structure also reached significance (γ = 0.22, p < .01 for classroom level, γ = 0.20, p < .001 for student level). However, performance goal structure did not predict promoting interaction at both levels, so we omitted performance goal structure from the final model. Copyright © 2013 SciRes. 633
K. OHTANI ET AL. Classroom level z = 2.37 * Student level z = 1.61 0.65 *** 0.09 0.44 *** (0.50 *** ) 1.24 *** 0.38 * 0.84 ** (1.32 *** ) Promoting interactio n Masterygoal structure Intrinsic motivatio n Promoting interactio n Masterygoal structure Intrinsic motivatio n Figure 1. Multilevel mediation model predicting intrinsic motivation. Regression coefficients in parentheses represent direct effects before adding promoting interaction. *p < .05, **p < .01, ***p < .001. C lassroom level z = 2.40 * Student level z = 5.42 *** 0.37 *** (0. 50 *** ) 1.24 *** 0.23 * 0.27 (0. 56 *** ) 0.65 *** 0.34 *** Promoting interactio n Masterygoal structure Peer modeling Promoting interactio n Masterygoal structure Peer modeling Figure 2. Multilevel mediation model predicting peer modeling. Regression coefficients in parentheses rep resent direct effects before adding promoting interaction. *p < .05, ***p < .001. in this study. This study revealed that teachers promoting interaction me diated the effects of classroom mastery goal structures on in trinsic motivation and peer modeling. These results are consis tent with previous studies (Patrick, 2004; Patrick et al., 2011). It can be understood that the positive effects of promoting in teraction among students for the following two reasons. First, interaction with peers can give students resources that support their learning. The more time students spend with peers in their learning, the more easily they can ask classmates for help when confronting academic difficulties. Students can also acquire skills by observing peers. This is reflected in the relationship between promoting interaction and peer modeling in this study. Second, students can receive emotional support by interacting with peers. Some researchers suggest that positive relationships with friends and peers in the classroom play an important role in students’ academic motivation (Anderman & Freeman, 2004; Wentzel, 2005). Promoting peer interaction is an important teaching strategy that influences students’ learning and motiva tion. The main contribution of this study is to reveal classroom motivational processes in terms of teachers promoting interac Copyright © 2013 SciRes. 634
K. OHTANI ET AL. tion among students. Previous studies mainly focused on stu dents’ personal goals as a mediator of classroom effects (e.g., Anderman & Anderman, 1999; Roser et al., 1996; Midgley et al., 1995). Given that a classroom environment is shared by students, we should pay attention to classroom variables. In this study, the mediation process for intrinsic motivation was con firmed at the classroom level. Among classroom variables, tea chers’ instructional practices can have a significant impact on students’ learning. Because a classroom is a social place by na ture, instructional strategies that support peer relationships and student interaction are especially important. Therefore, it is ne cessary for educators and researchers to examine the role of teaching strategies of promoting peer interaction. This study provides some prospections for future motiva tional research. In this study, teachers’ promoting interaction mediated the effects of classroom characteristics (i.e., mastery goal structures) on the students’ intrinsic motivation and peer modeling. This findings show that interaction between teachers and students are critical for students’ achievements related be havior and motivation. Thus, future research can gain important insights regarding achievement processes by focusing on inter action between teachers and students with more details. In ad dition, this study revealed few differences between classroom level and student level. It may be profitable to pay attention to studentstudent interaction as well as teacherstudent interac tion. Limitations This research has some limitations. First, all data were col lected at one point in time. The model examined in this study assumed a causal order in which mastery goal structures influ ence teachers promoting interaction, which in turn affects in trinsic motivation and peer modeling. The simultaneous meas urement, however, makes it difficult to interpret the findings as a causal sequence. To examine the motivational processes in terms of causal sequence in more detail, a longitudinal study would add valuable information. Second, classroom variables were measured in terms of students’ ratings. We measured stu dents’ perceptions of classroom goal structures and teachers promoting interaction and partitioned the variance into two levels (classroom and individual student levels). In addition, we could measure the variables through teachers’ reports and class room observation. Some researchers have examined the effects of classroom combining students’ and teachers’ selfreported data with the observational data (Turner et al., 2002; Turner et al., 2003). Patrick (2004) found that teachers in the classroom which was rated as high mastery oriented by the students em phasized understanding and selfreferenced improvement through classroom observations. It may be that observed data and stu dents’ reported data of classroom goal structures and teachers promoting interaction have different effects on students’ moti vation. Thus, multiple measurements including students’ or tea chers’ selfreport and classroom observation would be useful to consider the effects of classroom variables. REFERENCES Ames, C. (1992). Classroom: Goals, structures, and student motivation. Journal of Educational Psycholo gy, 84, 261271. doi:10.1037/00220663.84.3.261 Ames, C., & Ames, R. (1984). Systems of student and teacher motiva tion: Toward a qualitative definition. Journal of Educational Psy chology, 76, 535556. doi:10.1037/00220663.76.4.535 Ames, C., & Archer, J. (1988). Achievement goals in the classroom: Students’ learning strategies and motivation processes. Journal of Educational Psychology, 80, 260267. doi:10.1037/00220663.80.3.260 Anderman, L. H., & Anderman, E. M. (1999). Social predictors of changes in students’ achievement goal orientation. Contemporary Educational Psychology, 25, 2137. doi:10.1006/ceps.1998.0978 Anderman, L. H., & Freeman, T. M. (2004).Students’ sense of belong ing in school. In P. R. Pintrich, & M. L. Maehr (Eds.), Advances in motivation and achievement: Vol. 13. motivating students, improving schools: The legacy of carol midgley (pp. 2763). Greenwich, CT: JAI Press. Bliese, P. D. (2000). Withingroup agreement, nonindependence, and reliability: Implications for data aggregation and analysis. In K. J. Klein, & S. W. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 349381). San Francisco: JosseyBass. Brophy, J. (2004). Motivating students to learn (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Dweck, C. S., & Leggett, E. L. (1988). A socialcognitive approach to motivation and personality. Psychological Review, 95, 256273. doi:10.1037/0033295X.95.2.256 Elliot, A. J. (2005). A conceptual history of the achievement goal con struct. In A. J. Elliot, & C. S. Dweck (Eds.), Handbook of compe tence and motivation (pp. 5272). New York: Guilford Press. Elliot, A. J., & Church, M. A. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72, 218232. doi:10.1037/00223514.72.1.218 Elliot, A. J., & McGregor, H. A. (2001). A 2 × 2 achievement goal framework. Journal of Personality and Social Psychology, 80, 501 519. doi:10.1037/00223514.80.3.501 Harackiewicz, J. M., Barron, K. E., Pintrich, P. R., Elliot, A. J., & Thrash, T. M. (2002). Revision of achievement goal theory: Neces sary and illuminating. Journal of Educational Psychology, 94, 638 645. doi:10.1037/00220663.94.3.638 Hyde, J. S., & Durik, A. (2005). Gender, competence and motivation. In A. J. Elliot, & C. S. Dweck (Eds.), Handbook of competence and motivation (pp. 375392). New York: Guilford Press. Kaplan, A., Gheen, M., & Midgley, C. (2002). Classroom goal structure and student disruptive behaviour. British Journal of Educational Psy chology, 72, 191211. doi:10.1348/000709902158847 Kaplan, A., Middleton, M. J., Urdan, T., & Midgley, C. (2002). Ach ievement goals and goal structures. In C. Midgley (Ed.), Goals, goal structures, and patterns of adaptive learning (pp. 2153). Mahwah, NJ: Erlbaum. Lepper, M. R., Corpus, J. H., & Iyengar, S. S. (2005). Intrinsic and ex trinsic motivational orientation in the classroom: Age differences and academic correlates. Journal of Educational Psychology, 97, 184 196. doi:10.1037/00220663.97.2.184 Maehr, M. L., & Zusho, A. (2009). Achievement goal theory: The past, present, and future. In K. R. Wentzel, & A. Wigfield (Eds.), Hand book of motivation at school (pp. 77104). New York: Routledge. Meece, J. L., Anderman, E. M., & Anderman, L. H. (2006). Classroom goal structure, student motivation, and academic achievement. An nual Review of Psychology, 57, 487503. doi:10.1146/annurev.psych.56.091103.070258 Meece, J. L., Blumenfeld, P. C., & Hoyle, R. H. (1988). Students’ goal orientations and cognitive engagement in classroom activities. Jour nal of Educational Psycho logy, 80, 514523. doi:10.1037/00220663.80.4.514 Midgley, C. Maehr, M. L., Hruda, L. A., Anderman, E., Anderman, L., Gheen, M., Kaplan, A., Kumar, R., Middleton, M. J., Nelson, J., & Urdan, T. (2000). Manual for the Patterns of Adaptive Learning Scale. Ann Arbor, MI: University of Michigan. Midgley, C., Anderman, E., & Hicks, L. (1995). Differences between elementary and middle school teachers and students: A goal theory approach. Journal of Early Adolescence , 15, 90113. doi:10.1177/0272431695015001006 Miki, K., & Yamauchi, H. (2005). Perceptions of classroom goal struc tures, personal achievement goal orientations, and learning strategies. Japanese Journal of Psychology, 76, 260268. Copyright © 2013 SciRes. 635
K. OHTANI ET AL. Copyright © 2013 SciRes. 636 doi:10.4992/jjpsy.76.260 Murayama, K., & Elliot, A. J. (2009).The joint influence of personal achievement goals and classroom goal structures on achievement relevant outcomes. Journal of Educational Psychology, 101, 432 447. doi:10.1037/a0014221 Nicholls, J. G. (1984). Achievement motivation: Conception of ability, subjective experience, task choice, and performance. Psychological Review, 91, 328346. doi:10.1037/0033295X.91.3.328 Okada, R., Nakaya, M., Ito, T., & Ohtani, K. (2010). Individual differ ences in peermodeling. Poster Presented at the Japanese Associa tion of Educational Psycholo gy 52n d Annual Conference, Japan. Patrick, H. (2004). Reexamining classroom mastery goal structure. In P. R. Pintrich, & M. L. Maehr (Eds.), Advances in motivation and achievement: Vol. 13. Motivating students, improving schools: The legacy of carol midgley (pp. 233263). Greenwich, CT: JAI Press. Patrick, H., Anderman, L. H., Ryan, A. M., Edelin, K., & Midgley, C. (2001). Teachers’ communication of goal orientations in four fifth grade classrooms. The Elementa ry Sch ool J ournal, 102, 3558. doi:10.1086/499692 Patrick, H., Kaplan, A., & Ryan, A. M. (2011). Positive classroom mo tivational environments: Convergence between mastery goal struc ture and classroom social climate. Journal of Educational Psychol ogy, 103, 367382. doi:10.1037/a0023311 Patrick, H., Ryan, A. M., & Kaplan, A. (2007). Early adolescents’ perceptions of the classroom social environment, motivational beliefs, and engagement. Journal of Educational Psychology, 99, 8398. doi:10.1037/00220663.99.1.83 Pintrich, P. R., & De Groot, E. V. (1990). Motivational and selfregu lated learning components of classroom academic performance. Jour nal of Educational Psycho logy, 82, 3340. doi:10.1037/00220663.82.1.33 Roeser, R. W., Midgley, C., & Urdan, T. C. (1996). Perceptions of the school psychological environment and early adolescents’ psycho logical and behavioral functioning in school: The mediating role of goals and belonging. Journal of Educational Psychology, 88, 408 432. doi:10.1037/00220663.88.3.408 Ryan, A. M. (2000). Peer groups as a context for the socialization of adolescents’ motivation, engagement, and achievement in school. Educational Psychologist, 35, 101111. doi:10.1207/S15326985EP3502_4 Ryan, A. M., & Patrick, H. (2001). The classroom social environment and changes in adolescents’ motivation and engagement during mid dle school. American Educational Research Journal, 38, 437460. doi:10.3102/00028312038002437 Ryan, A. M., Gheen, M. H., & Midgley, C. (1998). Why do some stu dents avoid asking for help? An examination of the interplay among students’ academic efficacy, teachers’ socialemotional role, and the classroom goal structure. Journal of Educational Psychology, 90, 528535. doi:10.1037/00220663.90.3.528 Schunk, D. H. (1998). Teaching elementary students to selfregulate practice of mathematical skills with modeling. In D. H. Schunk, & B. J. Zimmerman (Eds.), Selfregulated learning, from teaching to self reflective practice (pp. 137159). New York: Guilford Press. Schunk, D. H., & Hanson, A. R. (1985). Peer models: Influence on child ren’s selfefficacy and achievement. Journal of Educational Psy chology, 77, 313322. doi:10.1037/00220663.77.3.313 Schunk, D. H., & Hanson, A. R. (1989). Selfmodeling and children’s cognitive skill learning. Journal of Educational Psychology, 81, 155 163. doi:10.1037/00220663.81.2.155 Tanaka, A., & Yamauchi, H. (2000). Causal models of achievement motive, goal orientation, intrinsic interest, and academic achievement in classroom. The Japanese Journal of Psycho logy, 71, 317324. doi:10.4992/jjpsy.71.317 Turner, J. C., & Meyer, D. K. (2009). Understanding motivation in mathematics: What is happening in classrooms? In K. R. Wentzel, & A. Wigfield (Eds.), Handbook of motivation at school (pp. 527552). New York: Routledge. Turner, J. C., Meyer, D. K., Midgley, C., & Patrick, H. (2003). Teacher discourse and sixth graders’ reported affect and achievement behave iors in two highmastery/highperformance mathematics classrooms. The Elementary School Journal, 103, 357382. doi:10.1086/499731 Turner, J. C., Midgley, C., Meyer, D. K., Gheen, M., Anderman, E. M., & Kang, Y. (2002). The classroom environment and students’ reports of avoidance strategies in mathematics: A multimethod study. Jour nal of Educational Psycho logy, 94, 88106. doi:10.1037/00220663.94.1.88 Urdan, T., & Midgley, C. (2003). Changes in the perceived classroom goal structure and pattern of adaptive learning during early adoles cence. Contemporary E du cat ion al Psychology, 28, 524551. doi:10.1016/S0361476X(02)000607 Urdan, T., & Schoenfelder, E. (2006). Classroom effects on student motivation: Goal structures, social relationships, and competence be liefs. Journal of School Psychology, 44, 331349. doi:10.1016/j.jsp.2006.04.003 Wentzel, K. R. (2005). Peer relationships, motivation, and academic performance at school. In A. J. Elliot, & C. S. Dweck (Eds.), Hand book of competence and motivation (pp. 279296). New York: Guil ford Press. Wolters, C. A. (2004). Advancing achievement goal theory: Using goal structures and goal orientations to predict students’ motivation, cog nition, and achievement. Journal of Educational Psychology, 96, 236250. doi:10.1037/00220663.96.2.236 Yamauchi, H., & Tanaka, K. (1998). Relations of autonomy, selfref erenced beliefs, and selfregulated learning among Japanese children. Psychological Reports, 82, 803816. doi:10.2466/pr0.1998.82.3.803 Young, M. R. (2005). The motivational effects of the classroom envi ronment in facilitating selfregulated learning. Journal of Marketing Education, 27, 2540. doi:10.1177/0273475304273346 Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing multilevel mediation using hierarchical linear models. Organizational Research Methods, 12, 695719. doi:10.1177/1094428108327450
K. OHTANI ET AL. Appendix 1 Final models of the multilevel meditational models. Intrinsic Motivatio n Step 1 Intrinsic motivation = β0j + β1j (gender) + β2j (mastery goal structure) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + u0j, β1j = γ10, β2j = γ20 Step 2 Promoting interaction = β0j + β1j (gender) + β2j (mastery goal structure) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + u0j, β1j = γ10, β2j = γ20 Step 3 Intrinsic motivation = β0j + β1j (gender) + β2j (mastery goal structure) + β3j (promoting interaction) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + γ03 (promoting interaction) + u0j, β1j = γ10, β2j = γ20, β3j = γ30 Peer Modeling Step 1 Peer modeling = β0j + β1j (gender) + β2j (mastery goal struc ture) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + u0j, β1j = γ10, β2j = γ20 Step 2 Promoting interaction = β0j + β1j (gender) + β2j (mastery goal structure) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + u0j, β1j = γ10, β2j = γ20 Step 3 Peer modeling = β0j + β1j (gender) + β2j (mastery goal struc ture) + β3j (promoting interaction) + rij β0j = γ00 + γ01 (school grade) + γ02 (mastery goal structure) + γ03 (promoting interaction) + u0j, β1j = γ10, β2j = γ20, β3j = γ30 Appendix 2 Table A1. The result of multilevel mediation model predicting intrinsic motivation. Step 1 Step 2 Step 3 To intrinsic motivation To promoting interaction To intrinsic motivation Classroom level Coefficient SE Coefficient SE Coefficient SE Intercept (γ00) 2.85*** 0.14 3.87*** 0.14 2.92*** 0.13 Grade (γ01) 0.13* 0.06 0.10 0.06 0.09 0.06 Mastery goal structure (γ02) 1.32*** 0.20 1.24*** 0.20 0.84** 0.04 Promoting interaction (γ03)   0.38* 0.15 Random effects of intercept (u0j) Variance 0.041*** 0.067*** 0.029* The amount of variance explained in each step 67.98% 52.46% 9.17% Student level Gender (γ10) 0.47*** 0.09 −0.05 0.04 0.48*** 0.09 Mastery goal structure(γ20) 0.50*** 0.06 0.65*** 0.03 0.44*** 0.06 Promoting interaction(γ30)   0.10 0.06 The amount of variance explained in each step 7.73% 34.49% 0.01% *p < .05, **p < .01, ***p < .001. Table A2. The result of multilevel mediation model predicting peer modeling. Step 1 Step 2 Step 3 To peer modeling To promoting interaction To peer modeling Classroom level Coefficient SE Coefficient SE Coefficient SE Intercept (γ00) 3.87*** 0.09 3.87*** 0.14 3.90*** 0.08 Grade (γ01) 0.09* 0.09 0.10 0.06 0.06 0.04 Mastery goal structure (γ02) 0.56*** 0.11 1.24*** 0.20 0.27 0.17 Promoting interaction (γ03)   0.23* 0.09 Random effects of intercept (u0j) Variance 0.010† 0.067*** 0.007† The amount of variance explained in each step 46.52% 52.46% 14.21% Student level Gender (γ10) −0.14* 0.06 −0.05 0.04 −0.12* 0.05 Mastery goal structure (γ20) 0.50*** 0.03 0.65*** 0.03 0.37*** 0.05 Promoting interaction (γ30)   0.34*** 0.06 The amount of variance explained in each step 20.23% 34.49% 5.30% †p < .10, *p < .05, **p < .01, ***p < .001. Copyright © 2013 SciRes. 637
