*** .18 -.07 -.05 -.02 -.09 .00 .01 .02 -.05 fpsyg.2015.01865 -.03 .07 -.01 -.05 .23 .74*** .37*** -.05 -.04 -.35** -.19** .16 .11 -.17* -.07 .03 .06 .15* -.07 -.20 -.07 .07 -.00 .05 19* French English Phys Ed SchoolNote. All coefficients are standardized correlations. * p < .05 ** p < .01 *** p < .001 doi:10.1371/journal.pone.0134660.tmotivation in writing was negatively correlated to students' achievement in fnins.2015.00094 mathematics (-.20) and intrinsic motivation in reading was positively correlated to students’ achievement in science, writing and school (.15, .19, and .14, respectively). These results are thus similar to those observed with self-concept dimensions. None of the correlations between identified, introjected or external regulations and students’ teacher-rated achievement were statistically significant at CBIC2 cost school-subject-specific level. However, the correlations between controlled motivations at the academic level were negatively related to students’ teacher-rated achievement in all school subjects. More specifically, it appears that controlled motivations are not school-subject-specific but predict achievement negatively in each specific school subject. In other words, when we measure controlled motivation in specific school subjects, we are not measuring something specific to a school subject, but rather something more global or contextual. One way to test the previous finding more stringently is to compare the obtained CTCM-1 solution to a classical CFA one where all indicators are caused solely by their respective latent constructs. More specifically, if the classical CFA solution reveals that introjected and external school subject regulations are negatively related to achievement in school subjects, this wouldPLOS ONE | DOI:10.1371/journal.pone.0134660 August 6,15 /School Subjects Specificity of LY-2523355 web Autonomous and Controlled MotivationsTable 7. Latent correlations with external criterion (i.e., students’ teacher-rated achievement) in Study 1. Mathematics Mathematics Intrinsic Identified Introjected External Science Intrinsic Identified Introjected External Writing Intrinsic Identified Introjected External Reading Intrinsic Identified Introjected External School Intrinsic Identified Introjected External -.02 .03 -.36*** -.28*** .03 .12 -.38*** -.23** .03 .06 -.35*** -.32*** -.02 .01 -.42*** -.30*** .04 .12 -.40*** -.26** .07 -.08 -.05 -.08 .15* -.02 .13 -.05 .19* -.00 -.03 -.08 .25** -.03 -.04 -.14 .14* -.05 .09 -.11 -.20** -.05 -.05 .09 -.12 .08 .05 .04 -.07 -.06 .02 .07 -.09 -.00 .01 .01 -.09 .02 .06 .02 .06 -.07 .02 .03 .15 .07 .16 -.02 -.06 -.11 -.03 -.08 .05 -.07 .06 .02 .01 -.09 .08 -.09 .17* -.11 .06 -.07 .13 -.04 .13 -.05 -.07 -.10 .02 -.13 .02 -.10 .05 -.19 .10 -.09 .14 -.16 Science Writing Reading SchoolNote. All coefficients are standardized correlations. * p < .05 ** p < .01 *** p < .001 doi:10.1371/journal.pone.0134660.tmean that the CTCM-1 solution captures a large portion of variance that is contextual and not school-subject-specific. Correlations from the classical CFA solution are in line with this idea. More specifically, in a classical CFA model, all twenty correlations between controlled motivations and achievements were found to be negative and significant for introjected (between-.14 and-.30) and external (between-.16 and-.29) motivations. Therefore, these results demonstrate that the common variance shared at the specific and contextual levels attenuates the correlations between school subject regulations and students' achievem.*** .18 -.07 -.05 -.02 -.09 .00 .01 .02 -.05 fpsyg.2015.01865 -.03 .07 -.01 -.05 .23 .74*** .37*** -.05 -.04 -.35** -.19** .16 .11 -.17* -.07 .03 .06 .15* -.07 -.20 -.07 .07 -.00 .05 19* French English Phys Ed SchoolNote. All coefficients are standardized correlations. * p < .05 ** p < .01 *** p < .001 doi:10.1371/journal.pone.0134660.tmotivation in writing was negatively correlated to students' achievement in fnins.2015.00094 mathematics (-.20) and intrinsic motivation in reading was positively correlated to students’ achievement in science, writing and school (.15, .19, and .14, respectively). These results are thus similar to those observed with self-concept dimensions. None of the correlations between identified, introjected or external regulations and students’ teacher-rated achievement were statistically significant at school-subject-specific level. However, the correlations between controlled motivations at the academic level were negatively related to students’ teacher-rated achievement in all school subjects. More specifically, it appears that controlled motivations are not school-subject-specific but predict achievement negatively in each specific school subject. In other words, when we measure controlled motivation in specific school subjects, we are not measuring something specific to a school subject, but rather something more global or contextual. One way to test the previous finding more stringently is to compare the obtained CTCM-1 solution to a classical CFA one where all indicators are caused solely by their respective latent constructs. More specifically, if the classical CFA solution reveals that introjected and external school subject regulations are negatively related to achievement in school subjects, this wouldPLOS ONE | DOI:10.1371/journal.pone.0134660 August 6,15 /School Subjects Specificity of Autonomous and Controlled MotivationsTable 7. Latent correlations with external criterion (i.e., students’ teacher-rated achievement) in Study 1. Mathematics Mathematics Intrinsic Identified Introjected External Science Intrinsic Identified Introjected External Writing Intrinsic Identified Introjected External Reading Intrinsic Identified Introjected External School Intrinsic Identified Introjected External -.02 .03 -.36*** -.28*** .03 .12 -.38*** -.23** .03 .06 -.35*** -.32*** -.02 .01 -.42*** -.30*** .04 .12 -.40*** -.26** .07 -.08 -.05 -.08 .15* -.02 .13 -.05 .19* -.00 -.03 -.08 .25** -.03 -.04 -.14 .14* -.05 .09 -.11 -.20** -.05 -.05 .09 -.12 .08 .05 .04 -.07 -.06 .02 .07 -.09 -.00 .01 .01 -.09 .02 .06 .02 .06 -.07 .02 .03 .15 .07 .16 -.02 -.06 -.11 -.03 -.08 .05 -.07 .06 .02 .01 -.09 .08 -.09 .17* -.11 .06 -.07 .13 -.04 .13 -.05 -.07 -.10 .02 -.13 .02 -.10 .05 -.19 .10 -.09 .14 -.16 Science Writing Reading SchoolNote. All coefficients are standardized correlations. * p < .05 ** p < .01 *** p < .001 doi:10.1371/journal.pone.0134660.tmean that the CTCM-1 solution captures a large portion of variance that is contextual and not school-subject-specific. Correlations from the classical CFA solution are in line with this idea. More specifically, in a classical CFA model, all twenty correlations between controlled motivations and achievements were found to be negative and significant for introjected (between-.14 and-.30) and external (between-.16 and-.29) motivations. Therefore, these results demonstrate that the common variance shared at the specific and contextual levels attenuates the correlations between school subject regulations and students' achievem.