Motivation at school: Differentiation between and within school subjects matters in the prediction of academic achievement

Highlights

• This study contrasts the effects of different types of motivation on achievement.

• Intrinsic and identified regulation are differentiated across school subjects.

• Introjected and external regulations are undifferentiated across school subjects.

Abstract

School motivation is a multidimensional concept. It can be qualitatively defined by various sources of regulation as well as by the school subject to which it pertains. Based on self-determination theory, we postulate that motivation types vary in terms of quality (from lower to higher quality these types are: external, introjected, identified, and intrinsic) and that higher motivational quality predicts positive outcomes. In this study, we examined school subject differentiation in motivational quality and prediction patterns of academic achievement. Results from bi-factor ESEM examining differences in motivational quality within a subject (French, math, and English as a second language) showed that high general levels of motivation in math and English predicted achievement, and more so in the corresponding school subject. Intrinsic motivation for a school subject was generally positively associated with achievement, but only in the corresponding school subject, whereas introjected and external regulations for most school subjects negatively predicted achievement in the corresponding school subject, but also in the other ones. Results from bi-factor ESEM examining differences in motivation levels for distinct school subjects for a given motivation type showed that general levels of intrinsic and external regulations across school subjects predicted achievement positively and negatively, respectively, in all school subjects, while intrinsic motivation, but also identified regulation, had positive subject-specific associations with achievement. The specificity of intrinsic and identified motivations and non-specificity of introjected and external motivations point toward various recommendations in school motivation research and practice. While assessment of autonomous motivations should be subject-specific, assessment of controlled motivations could be general with no loss of predictive power.

Introduction

According to some motivational theories, such outcomes as achievement and performance are best predicted by motivational variables measured at the same level of generality (Bandura, 1997, Vallerand, 1997). For example, a measure of motivation specific to math, and not a global measure of motivation at school, should be the best predictor of math achievement (Huang, 2012). Indeed, global measures do not consider the complexity and variation of self-perceptions, and this could impair the ability to understand and predict behavior (Ajzen, 2005, Marsh and Yeung, 1998). Therefore, scales measuring important areas of life would be more useful than global scales for understanding the consequences pertaining to area-specific self-related constructs. In addition, when motivation toward a specific school subject, for example, is being measured, it is expected to be less correlated with outcomes (e.g., achievement) in non-corresponding school subjects (Bong, 2002, Guay et al., 2010).

Such specificity principles (i.e., level of specificity–global vs. specific–and area of specificity–maths vs. French–, for example) imply that knowledge of the determinants of students’ achievement relies on an understanding of subject-bound dynamics. In various studies stemming from different motivational theories (i.e., self-efficacy theory, achievement goal theory, expectancy value theory, and self-concept theory), researchers have applied these principles (see Wentzel & Wigfield, 2009, for a literature review of each theory) and have shown that a large portion of motivational variance is specific to school subjects (Bong, 2001, Shen et al., 2008). However, our survey of the field reveals a dearth in self-determination theory (SDT; Ryan & Deci, 2007) research testing the specificity principles, limiting the understanding of motivational dynamics. More specifically, motivations derived from SDT have been assessed in various school subjects, but only a few researchers have done so simultaneously across school subjects (Chanal and Guay, 2015, Guay et al., 2010), possibly because of the degree of complexity of such a research endeavor. Indeed, within SDT, qualitatively distinct motivation types exist that may differ among school subjects as well as in their relations to outcomes within a given school subject. In this study, we analyze, among those proposed by SDT, four types of motivation (intrinsic, identified, introjected, and external) toward three school subjects (French, math, and English as a second language). Our main research endeavor, which aligns with a collective effort to better understand how motivation predicts achievement (Guay et al., 2010, Linnenbrink and Pintrich, 2002, Mega et al., 2014), is to uncover fundamental processes in the way each type of motivation toward various school subjects predicts academic achievement in these subjects and others. Our analysis could lead to important discoveries regarding area (types of school subjects) and level specificity (global vs. specific) of motivation and broaden our understanding of student motivation and the associated outcomes. More precisely, the predictive power of various types of motivation could be increased when simultaneously taking into account (1) their specificity to school subjects as well as (2) their communalities in terms of global motivation (see Fig. 1).

In SDT, motivation is defined as the reasons underlying a behavior. Applied to education, it refers to the reasons students engage in various school activities (Ryan & Deci, 2000). It is possible to distinguish among various types of motivation that differ in terms of self-determination (i.e., the extent to which a behavior originates from the self). Intrinsic motivation refers to engaging in an activity for its own sake, for the pleasure and satisfaction it provides (Ryan & Deci, 2000). Extrinsic motivation refers to engaging in an activity for instrumental reasons rather than for its intrinsic qualities. According to SDT, there are various types of extrinsic motivation that differ in terms of self-determination. From low to high self-determination, these are external regulation, introjected regulation, identified regulation, and integrated regulation (Ryan & Deci, 2000).

External regulation occurs when a behavior is motivated by the desire to obtain a reward or avoid punishment. Introjected regulation refers to behaviors performed in response to internal pressures, such as obligation or guilt: the individual somewhat endorses the reasons for doing something, but in a controlled manner. Identified regulation is observed when individuals identify with the reasons for performing a behavior, or when they personally find it important. This is a self-determined form of extrinsic motivation, because the behavior originates from the self in a non-contingent manner. Integrated regulation occurs when the identified regulation is congruent with other values and needs. The behavior is therefore performed because it is part of who the person is. However, this form of regulation requires individuals to have formed a coherent identity (Deci, Ryan, & Williams, 1996), such that they can identify with the importance of a behavior and reciprocally assimilate that identification with other aspects of their coherent sense of self. Consequently, this type of extrinsic motivation is not assessed in studies on children and adolescents such as this one.

Thus, in SDT, motivation types are located along a self-determination continuum reflecting motivational quality, rather than motivational intensity. Motivation types are therefore expected to relate to each other in a quasi-simplex-like pattern, with stronger positive correlations between adjacent motivations than between distant ones. For example, identified regulation and intrinsic motivation should be positively and moderately correlated, and this correlation should be stronger than the one between intrinsic motivation and introjected regulation. In previous research, the self-determination continuum was supported for types of motivation toward school in general (Otis et al., 2005, Ryan and Connell, 1989, Vallerand et al., 1989, Vallerand et al., 1992, Vallerand et al., 1993). Furthermore, this continuum also reflects how each motivation type affects various school outcomes. For example, students who endorse autonomous types of motivation (intrinsic and identified regulation) are more persistent and cognitively involved in their tasks, experience more positive emotions, and have better grades, whereas students who are motivated in a controlled fashion (introjected and external) are less persistent, are more distracted, experience more negative emotions (anxiety), and obtain lower grades (Guay et al., 2016, Guay et al., 2008). With these findings, researchers have underscored the importance of developing intrinsic and identified motivations rather than introjected or external regulations during the school years.

Based on theory and results, many researchers have claimed that motivational quality matters, more so than motivational intensity (Ratelle et al., 2007, Vansteenkiste et al., 2009). However, statistical issues arise from research testing this proposition, which was done within the confines of confirmatory factor analyses (CFA; Guay, Morin, Litalien, Valois, & Vallerand, 2015) or profile analyses (Ratelle et al., 2007) that do not offer a stringent test of this hypothesis. In these two methods, type-specific factors are estimated without prior removal of the variance shared among all items of the scale. In other words, although SDT motivation items assess various types of motivation, the scores on the same items may also reflect a general factor of motivation. This factor would represent motivation intensity, which could differentially predict achievement in comparison with specific components. In previous research testing employees’ work motivation, researchers have estimated a global factor in order to assess quantity of self-determined motivation and has supported the hypothesis that qualitatively distinct motivation types predict work outcomes over and above the G-factor (Howard, Gagné, Morin, & Forest, 2016). Similar findings have been observed in physical activity as well (Gunnell & Gaudreau, 2015), although motivational intensity predicted goal-consistent behavior more strongly than motivational quality did.

There are two approaches to the differential examination of school motivation. The first is to examine motivation across various school subjects. As previously stated, this has been done from several theoretical standpoints (e.g., goal theory, self-efficacy theory, self-concept, and the expectancy-value model) focusing primarily on such disciplines as writing, reading and math (Bong, 2001, Green et al., 2007). We refer to this motivational differentiation as between-subject differentiation. Another approach is to examine school motivation as a multidimensional concept that illustrates varying reasons underlying engagement in a given school subject. SDT substantiates this approach by distinguishing among various types of motivation. We refer to this motivational differentiation as within-subject differentiation.

In this study, we wanted to combine both approaches by measuring each type of motivation for each school subject, while taking into account the common variance (i.e., a global factor of motivation) associated with these items in each school subject. We also wanted to test how these components fared in the prediction of academic achievement. Conducting such an analysis to untangle specific and common variance is relevant because, if motivational quality matters more than motivational intensity, each type of motivation should be a stronger predictor of achievement than the G_w-factor (see Fig. 1; a global factor related to within-subject differentiation, representing a global indicator of various types of motivation within a specific subject). Thus, to support previous research stressing the importance of motivation quality, our analyses should first show that, compared with introjected and external regulations, intrinsic and identified motivations within a given school subject will predict achievement more positively in this subject. Second, the G_w-factor of motivation should not be a better predictor of achievement in a corresponding subject than intrinsic or identified motivation because it is expected mostly to mirror motivational intensity.

As previously mentioned, few SDT researchers have examined types of school motivation toward various school subjects simultaneously. Some support has been obtained for an effect of between-subject differentiation (i.e., area specificity of motivation) with respect to intrinsic motivation toward reading, math, social studies and science (Gottfried, 1985, Gottfried, 1990). More specifically, students engage in various activities that schools offer as an opportunity to discover at a relatively early age which activities they do or do not enjoy. Students’ intrinsic motivation differentiation across school subjects is evidenced by correlational patterns, whereby intrinsic motivation for a given school subject is more strongly associated with other motivational constructs within that school subject than with motivational constructs for other school subjects (Gottfried, 1985, Gottfried, 1990). Other researchers (Eccles et al., 1993, Green et al., 2007) have also shown that “valuing of school subject” had lower between-subject correlations than more “trait-like” academic constructs, such as school anxiety.

In this study, we wanted to replicate the school subject differentiation effect obtained for intrinsic motivation and extend our focus to how identified, introjected, and external regulations are differentiated across subjects and in their relations to educational outcomes. Because these regulations are phenomenologically distinct from intrinsic motivation, we postulated that intensity in differentiation effects (or area specificity) would differ across motivation types. We expected school subject differentiation to be stronger at the higher end of the self-determination continuum (intrinsic motivation) and lower as self-determination declines. Intrinsic motivation should be more area-specific in its predictions than the facets of extrinsic motivation because intrinsic motivation originates autotelically, arising from the inherent satisfaction in the action and presumably energizing behavior circumscribed to the interest. In contrast, extrinsic motivation relies on contingent outcomes that are separable from the action. Identified regulation should be less area-specific in its predictions than intrinsic motivation. This is because, although this regulatory process is somewhat tied to the inherent characteristics of the activity, it is nevertheless governed mostly by the endorsement of cultural values (Deci & Ryan, 1985). In fact, students may understand relatively early that reading, writing, and math are all important for their development as individuals, and that identified goals enabled by success in one subject can be pursued in other ones as well. Weaker area specificity in identified motivation should translate into higher cross-subject outcome prediction (e.g., identified motivation in French predicting math grades). Finally, we posited that introjected and external regulations would predict outcomes across school subjects equally because they involve management of internal and external impetuses that should operate independently of the school subject, affecting outcomes in all subjects at once. This hypothesis is in line with previous findings obtained in children and adolescent samples showing that correlations among intrinsic motivations toward various school subjects are weaker than correlations among identified regulations, which are in turn weaker than correlations among introjected and external regulations (Chanal and Guay, 2015, Guay et al., 2010).

Because autonomous motivations (intrinsic and identified) are hypothesized to be more specific, we postulated that their relations with achievement would be stronger in corresponding school subjects than in non-corresponding ones. We also expected to find more positive relations with achievement in a given school subject for these motivations than for controlled ones (introjected and external), reflecting the higher quality of autonomous motivations. In addition to hypotheses testing area-specific predictions of motivation, we also tested level-specific hypotheses by extracting a G_b-factor (see Fig. 1), which is a global factor in between-subject differentiation models estimated among items of a single type of motivation for three school subjects (e.g., general introjected regulation across French, math, and English), and examining how its general prediction compared to that of specific factors. The G_b-factor extracted from intrinsic or identified regulation should be less correlated with achievement in the three school subjects than subject-specific factors because area-specific predictions should entail better predictions from the according-level factors. However, because we expect introjected and external regulations to be less differentiated across subjects, the G_b-factor related to these constructs—a higher-level, global indicator of these motivation types—should be more correlated with achievement than subject-specific factors.

In this study, we aimed to examine high school students’ motivation differentially by estimating the effects of both within- and between-subject motivational differentiation while taking into account the common variance across motivation types. We believe that the within- and between-subject differentiation effects of motivation, estimated while taking into account general factors, are so conceptually central that our findings may be used to refine our understanding of the relations between important constructs involved in students’ achievement and of how determinants affect students’ levels of autonomous and controlled motivations in learning situations. More specifically, at the within-subject level, if a G_w-factor is more predictive of achievement than the specific factors, such findings might call into question the focus on motivation quality advocated by SDT. Furthermore, until now, most researchers have considered types of motivation (intrinsic, identified, introjected, and external) in a given school subject as equally specific. However, if some types of motivation are differentiated across school subjects, and thus more area-specific, researchers should be specific in their assessment and should design interventions that are unique to each school subject. If some are undifferentiated, we can dispense with measuring them specifically, which can help reduce the length of some questionnaires. In addition, future research and interventions on undifferentiated types of motivation should be directed toward antecedents that are not subject-specific.

In this study, two general perspectives are proposed under which many different hypotheses between types of regulation for the three school subjects and achievement in those school subjects are tested. These hypotheses are all presented in Table 1. In the within-subject differentiation perspective, we expected the following relations for each school subject: (a) the G_w-factor (see Fig. 1) will be related to achievement in corresponding subjects in a positive manner (+), but not associated with achievement in other school subjects (0); (b) intrinsic motivation in a given school subject will be associated positively and moderately to achievement in the corresponding school subject (++), but not associated to achievement in non-corresponding school subjects (0); (c) a similar pattern is expected for identified regulation, but the magnitude of specific relations should be smaller than for intrinsic motivation (+); (d) introjected regulation should be associated negatively to achievement in all subjects (−; undifferentiated), but to a lesser degree than external regulation (−−). For the between-subject differentiation perspective, we expected the following relations for each regulation: (a) the G_b-factor (see Fig. 1) should be associated (positively (+) or negatively (−) depending on the regulation type) to achievement in school subjects, but in an undifferentiated way; (b) intrinsic and identified regulations should have specific factor relations that are subject-specific and positive, though stronger for intrinsic (++) than for identified regulation (+); (c) introjected and external regulations should have specific factor relations that are negative and undifferentiated, though stronger for external (−−) than for introjected (−). Although the two perspectives (i.e., within and between-subject differentiation) lead to similar predictions between motivation and achievement, it is important to highlight that the ways there are tested are sharply different. In within-subject analyses, each regulation type competes with the others as well as with motivation intensity (G_w-factor) to predict achievement, whereas in the between-subject, only one motivation type is assessed, but has to compete with its measurement in other subjects as well as with its global level at school (G_b-factor). Thus, both hypothesis offer the possibility to test more stringently the differentiation of types of regulation.