Content uploaded by Virginie Hospel
Author content
All content in this area was uploaded by Virginie Hospel on Sep 12, 2018
Content may be subject to copyright.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 1
Abstract
The current study was carried out within the framework of self-determination theory and
aimed to investigate specific, additive and combined effects of teachers’ autonomy support
and structure on students’ engagement. Using multilevel analyses, main effects and interaction
of autonomy support and structure provided at the classroom level were tested on behavioral,
cognitive and emotional engagement. 744 ninth grade students from 51 classes completed a
questionnaire about their engagement during language classes and their perceptions of the
teacher’s provision of autonomy support and structure. The results highlight the links between
classroom context, especially structure, and the three components of engagement. Autonomy
support has a complementary role as it was associated with emotional engagement. These
results improve our understanding of the relationships between learning environment and
engagement and provide more accurate indications to teachers and educators regarding the
most effective ways to enhance students’ engagement.
Keywords: classroom learning environment, structure, autonomy support, engagement,
multilevel analyses.
Hospel, V., & Galand, B. (2016). Are both classroom autonomy support and structure
equally important for students' engagement? A multilevel analysis. Learning and
Instruction, 41, 1-10. Doi : http://dx.doi.org/10.1016/j.learninstruc.2015.09.001
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 2
Are both classroom autonomy support and structure equally important for students’
engagement? A multilevel analysis
1. Introduction
Student engagement has attracted the attention of many researchers and education
professionals in recent years (Fredricks & McColskey, 2012). According to self-determination
theory (SDT; Deci & Ryan, 2008), engagement is the reflection of the positive development
of an individual. In the context of schooling, engagement describes the level of energy or
effort students invest in learning activities which has positive consequences, notably on
achievement and well-being (Reeve, 2002; Skinner, Furrer, Marchand, & Kindermann, 2008).
Engagement is considered to be a malleable state influenced by contextual factors (Fredricks,
Blumenfeld, & Paris, 2004). Improving our understanding of the effects of these factors is
important in the design of learning environments that foster student engagement and, in turn,
achievement.
SDT emphasizes the role of different dimensions of the social context in enhancing or
diminishing student engagement (Skinner et al., 2008). Recently, there has been much
discussion of the relationships between the dimensions of autonomy support and structure,
and their respective contributions to engagement (Jang, Reeve, & Deci, 2010; Vansteenkiste et
al., 2012). Autonomy support refers to the amount of psychological freedom teachers allow
students in determining their own behaviors (Assor, Kaplan, & Roth, 2002). It consists in
supporting students in the pursuit of their own goals and in creating congruence between
students’ motives and classroom activities (Reeve, Jang, Carrell, Jeon, & Barch, 2004).
Teachers support autonomy by offering choices and rationales for mandatory activities, by
highlighting meaningful learning goals, by presenting interesting activities, by adopting
students’ perspectives and by avoiding the use of control (Jang et al., 2010; Reeve et al., 2004;
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 3
Skinner & Belmont, 1993). Structure refers to the amount and the clarity of information given
to students about how to satisfy teachers’ expectations and achieve the desired educational
outcomes (Jang et al., 2010; Skinner & Belmont, 1993). Teachers provide structure by
communicating expectations, by providing guidance, optimal challenges, and feedbacks
(Reeve, 2006; Vansteenkiste et al., 2012). SDT posits that both dimensions are important for
engagement, but there is little evidence in support of this claim (Stroet, Opdenakker &
Minnaert, 2013). Moreover, the results of the few studies that include both dimensions show
substantial inconsistencies. For instance, Jang and colleagues (2010) found a positive link of
autonomy support, but not of structure, with engagement. Skinner and Belmont (1993) found
the opposite: Engagement was significantly enhanced only by structure. More work needs to
be done to determine the relative importance of each dimension on engagement, and the value
of combining them.
SDT postulates that autonomy support and structure are contextual characteristics affecting
individual functioning. Many studies carried out in the SDT framework performed data
analyses at the (student) individual level, and did not allow testing of learning environment
effects (Marsh et al., 2012). Using a multilevel analytical framework, the present study aimed
to investigate the main effects and interaction of autonomy support and structure at the
classroom level on student engagement. Given the multidimensional nature of engagement,
we investigated this question by distinguishing behavioral, cognitive and emotional
engagement (Archambault, Janosz, Fallu, & Pagani, 2009; Shernoff, 2013).
1.1. Autonomy support and structure
SDT holds that teachers’ autonomy support and structure contribute to the enhancement of
academic engagement by fulfilling basic psychological needs. Autonomy support is
hypothesized as fulfilling the need for autonomy, meaning the experience of a sense of
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 4
volition. Structure is hypothesized as fulfilling the need for competence, meaning feeling
effective (Dupont, Galand, Nils & Hospel, 2014; Vansteenkiste et al., 2012).
Two main conceptions of the relations between autonomy support and structure have been
proposed in the literature. On the one hand, they have sometimes been conceptualized as two
opposed dimensions: Autonomy support is provided by removing structure and vice-versa
(see Vansteenkiste et al., 2012). This conceptualization has been challenged for its
interpretation of autonomy support as laissez-faire, or a lack of guidance (Reeve, 2002;
Vansteenkiste et al., 2012). On the other hand, some authors have stressed that, according to
SDT, autonomy support and structure should be conceptualized as distinct orthogonal
dimensions, complementary and mutually supportive. Recent empirical studies support this
latter conception (Jang et al., 2010; Sierens, Vansteenkiste, Goossens, Soenens, & Dochy,
2009; Vansteenkiste et al., 2012). This allows the examination of the most efficient
combination of autonomy support and structure to promote students’ engagement. Teachers
can provide high or low levels of both dimensions to students, or a high level of one
dimension and a low level of the other (Jang et al., 2010). However, it is unclear how
autonomy support and structure enhance engagement in the most effective way. The effects of
each dimension could be cumulative (additive effect): Each dimension makes its own positive
contribution to engagement, and providing both would be particularly effective. One specific
dimension could be more crucial for engagement than the other (specific effect). Providing
students with the second dimension would have no significant effects beyond the effects of
the first dimension. The positive effect of one dimension could be related to the presence or
absence of the other (combined or interactive effect). The effect of one dimension on
engagement could be accentuated when the level of the other dimension is high. Conversely,
the provision of one dimension could compensate for the absence of the other.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 5
Existing studies have left three important questions unanswered regarding the relationships
between autonomy support/structure and engagement:
1°) SDT states that providing both dimensions is important in enhancing engagement, as
they tend to fulfill specific needs (see Dupont et al., 2014; Skinner & Belmont, 1993). But
what is the relative weight of autonomy support and structure? Do they have additive, specific
or combined effects on engagement?
2°) Are autonomy support and structure related the same way to each component of
engagement? SDT seems to postulate that the relationships are similar.
3°) SDT claims that social context affects individual functioning. Is teacher provision of
autonomy support and structure at the classroom level associated with student engagement at
the individual level?
A review of the available evidence regarding those questions is presented below.
1.2. Do autonomy support and structure have additive, specific or combined effects on
engagement?
Most studies have focused on autonomy support and highlighted its positive role for
engagement (Assor et al., 2002; Reeve et al., 2004; Vansteenkiste, Simons, Lens, Sheldon, &
Deci, 2004), while Nie & Lau (2009) focused on structure and found a positive link with
engagement. A few studies have investigated the effects of both dimensions simultaneously.
Using students’ ratings of teachers’ autonomy support and structure, some authors found only
a main effect of structure (Skinner & Belmont, 1993); a main effect of structure plus an
interaction between autonomy support and structure (Sierens et al., 2009); or independent
main effects of both dimensions (Tucker et al., 2002) on engagement. Using observers’ ratings
of teachers’ autonomy support and structure, Jang and colleagues (2010) found only a main
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 6
effect of autonomy support on students’ self-reported engagement. These contradictory
findings may be due to the component of engagement investigated in these studies.
1.3. Do autonomy support and structure have identical effects on each component of
engagement?
Most scholars view engagement as a multidimensional construct composed of behavioral,
cognitive and emotional components (Archambault et al., 2009; Fredricks & McColskey,
2012; Shernoff, 2013). Behavioral engagement refers to students’ actions towards learning
and school activities such as participation, attendance, etc. Emotional engagement refers to
positive and negative affective reactions toward school, teachers, etc. Cognitive engagement
consists in psychological involvement in learning, including students’ use of learning and self-
regulated strategies (Fredricks et al., 2004).
Regarding behavioral engagement, studies including both autonomy support and structure
found only a positive main effect of structure (Skinner & Belmont, 1993; Wang & Eccles,
2013). Regarding cognitive engagement, results differ. Sierens and colleagues (2009) found a
positive main effect of structure on self-regulated learning and a significant interaction: Self-
regulated learning was higher when structure was combined with a moderate or high level of
autonomy support. Wang and Eccles (2013) found a positive effect of autonomy support on
the use of self-regulated strategies. These authors did not investigate interactions between
autonomy support and structure. Regarding emotional engagement, only a main effect of
autonomy support was found on positive emotions (Wang & Eccles, 2013). No effects of
structure were found on interest (Kunter et al., 2007). Vansteenkiste and colleagues (2012)
found that test anxiety was reduced when teachers used both autonomy support and structure,
in comparison with teachers who used a low level of both or a high level of only one of them,
suggesting a combined effect of both dimensions. Specific relationships with autonomy
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 7
support and structure could therefore exist depending on the investigated components of
engagement. Since these studies analyzed these effects at the individual level, do their
findings truly reflect learning environment effects?
1.4. Is teacher provision of autonomy support and structure at the classroom level
linked with student engagement at the individual level?
Assessment of the effects of the learning environment on individual students must be based
on analyses carried out at the environment level and not at the individual level (Marsh et al.,
2012). Individual perceptions could reflect individual differences or idiosyncratic bias and
interpretations rather than contextual influences (Galand, Philippot, & Frenay, 2006).
Aggregation of the individual perceptions of the students in the same class gives a more
accurate measure of the learning environment shared by students (Lüdtke, Robitzsch,
Trautwein & Kunter, 2009; Marsh et al., 2012). Multilevel analyses aim to differentiate such
levels of analysis.
Among the studies cited in the section 1.2., only a few authors (Jang at al., 2010; Nie &
Lau, 2009; Kunter et al., 2007) have used multilevel analyses to test the effects of autonomy
support and structure at the classroom level on student engagement. However, Jang and
colleagues (2010) focused on a global measure of engagement and did not control for student
characteristics and class composition (e.g., gender ratio). According to findings from teacher
effectiveness research, observed effects at the classroom level could mix effects related to the
teacher and to class composition (De Fraine, Van Damme, Van Landeghem, Opdenakker, &
Onghena, 2003). Class composition variables should be controlled. Nie and Lau (2009)
controlled for class composition, but did not investigate the effects of autonomy support and
did not distinguish the components of engagement. Kunter and colleagues (2007) focused on
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 8
structure. Further studies are thus needed to get a more accurate picture of the effects of
autonomy support and structure at the class level on the components of engagement.
1.5. Aims and hypotheses
Using a multilevel analytical framework, the aims of the current study were to: (a) assess
the magnitude of the classroom effect on behavioral, cognitive and emotional engagement; (b)
examine the associations between autonomy support and structure at the classroom level and
each component of engagement at the student level.
According to SDT, autonomy support and structure are both important in fostering
engagement. As previously stressed, these dimensions may have specific, additive or
combined effects on engagement. These effects may differ following the component of
engagement considered. Even if SDT does not provide precise predictions on these topics,
from the results of previous studies, we expected:
- Hypothesis 1: a positive association between structure, rather than autonomy support,
and behavioral engagement, consistent with Skinner and Belmont (1993) and Wang
and Eccles (2013).
- Hypothesis 2: a positive link between structure and cognitive engagement as well as an
interaction (combined effect) between autonomy support and structure on this
component of engagement, consistent with Sierens and colleagues’ findings (2009).
The combination of a high level of structure and a high level of autonomy support is
expected to be positively associated with a higher level of cognitive engagement.
- Hypothesis 3 (a): a positive association between autonomy support and positive
emotions, consistent with Wang & Eccles (2013). As these authors did not test
interactions between both dimensions, an interaction may also exist.
- Hypothesis 3 (b): an interaction between autonomy support and structure on negative
emotions, consistent with Vansteenkiste and colleagues’ study (2004).
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 9
Individual characteristics and class composition variables were controlled in the analyses.
As SDT provides no details regarding the effects of these variables on engagement, no
specific hypotheses were formulated.
2. Method
2.1. Sample
Participants were 744 French-speaking students from 51 classes. They were in grade 9 (M
age = 15.14; SD = 0.94). Forty-five percent were girls; 51 % repeated a year at least once
(consistent with the PISA 2012 report; Fédération Wallonie-Bruxelles, 2014); 40% were in
general and 60 % in vocational education
1
. They came from 10 schools located in several
cities in Belgium and had varied socio-cultural backgrounds.
2.2. Procedure and measures
Students were invited to fill in a questionnaire assessing their perceptions of teacher
autonomy support and structure, and their engagement during their French lessons. The
questionnaire was administered by a researcher during regular class time. Students were
informed that they were free to take part in the study and that the information would be kept
confidential. Passive parental consent procedure was used to avoid bias in sample
characteristics and to obtain a wide variety of social backgrounds (Pokorny, Jason, Schoeny,
Townsend, & Curie, 2001).
Behavioral engagement was measured by means of 21 items which together formed one
factor (α = .90). It assessed different students’ behaviors: participation, effort, following
1
In the French-speaking part of Belgium, the choice of general or vocational education is related to
students’ specific paths (e.g., general education makes it possible to enter university). Students in
general and vocational education often differ in their academic levels (Fédération Wallonie-Bruxelles,
2014). Hence, we controlled for it in analyses.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 10
instructions, withdrawal (reverse), and attendance during French class (e.g., “During French
lessons, when the teacher asks the class a question, I try to answer”) (Hospel & Galand,
2010). Emotional engagement consisted in a measure of the frequency of positive emotions
(hope, curiosity, happiness; seven items, α = .78) and negative emotions (anger, sadness,
stress, shame, boredom; 11 items, α = .84) felt during the French lessons. The items came
from a validated French version (Galand & Philippot, 2005) of the Differential Emotion Scale
(Izard, Dougherty, Bloxom, & Kotsh, 1974). Cognitive engagement consisted in students’
self-regulated learning (i.e. management of effort strategies: “In French lessons I take care not
to let myself get distracted”; five items, α = .78), and the use of deep processing strategies
(e.g., “When we have a new topic in French I try to make connections with what I already
know”; five items, α = .73; Galand, Raucent, & Frenay, 2010). Students responded to the
items of these scales using a 5-point Likert scale (0 = never to 4 = very often).
Individual socio-demographic characteristics. Students reported their gender (0 = girls, 1=
boys), grade retention (0 = never retained, 1 = retained once or more) and socio-economic
status (SES) measured through cultural (e.g., Do you get scientific books at home?) and
material resources (e.g., Do you have your own bedroom?) that students have at home (0 =
they do not have this resource; 1 = they have it). They were based on measures used in PISA
studies (OECD, 2006).
Class composition was measured through several indicators: the gender ratio, the ratio of
retained students, and the mean class SES (i.e. cultural and material resources). They were
calculated on the basis of students’ reports aggregated by class (using the mean of students’
responses, Marsh et al., 2012). Educational track was also included (0 = general track; 1 =
vocational track).
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 11
Autonomy support and structure provided by the French teacher in the classroom were
measured through the aggregated students’ perceptions (i.e. the mean of the responses given
by the students of the same class, Marsh et al., 2012). The items were inspired by existing
scales (Belmont, Skinner, Wellborn, & Connell, 1988; Reeve & Halusic, 2009). Duplicates
were removed. Items focusing on the description of teacher behaviors (rather than on
students’ interpretation of teacher intentions/behaviors, e.g., “I feel that my teacher…”) were
selected. Items were translated and adapted to fit with the French lessons. The results of
factorial analysis supported the distribution of the items between the two scales as expected.
Autonomy support included six items (α = .76), which refer to the choice given to the
students, the opportunities given for students’ initiative-taking, teachers’ acknowledgement of
students’ perspective and feelings (e.g., “the teacher gives us the opportunity to work at our
own pace”, “the teacher encourages us to think up original things”). Structure included six
items (α = .71), which refer to the communication by the teacher of his/her expectations, the
guidance and the constructive feedbacks given (e.g., “before a test the teacher explains in
detail the criteria he will use to assess it”, “after a test the teacher checks whether each student
has understood the mistakes he or she made”). Students responded to the items of these scales
using a 5-point Likert scale (0= totally false to 4 = totally true).
The reliability of the aggregated measures was evaluated through an indicator based on the
intra-class correlation and the mean class size (ICC (2), see Lüdtke, Trautwein, Kunter &
Baumert, 2006). The value of ICC (2) for autonomy support and structure was .88. It was
above the critical value of .70 and indicated a satisfactory reliability for both measures.
2.3. Data analytic strategy
Multilevel analyses (HLM7 software) were performed. A step by step procedure was used.
Significance testing was undertaken at the 5% level. First, models without predictors (null
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 12
models) were run to estimate the partition of variance between and within classes. Second,
models including individual characteristics (at individual level) and class composition (at
class level) were run. The variables introduced at this step were grand-centered (see
Kyriakides & Creemers, 2008). Gender and grade retention were not centered as the value
zero was meaningful for these variables (see Opdenakker & Van Damme, 2001). Third,
teacher autonomy support and structure were added simultaneously in the models, controlling
for the variables introduced at step 2. As this study aimed at testing the relationships between
the learning environment and student engagement, autonomy support and structure were
introduced at the classroom level (level 2) but not at the individual level (level 1). According
to Marsh and colleagues (2012), the most appropriate measure of the learning climate consists
in the aggregated students’ perceptions introduced at level 2. Students’ individual perceptions
introduced at level 1 do not reflect the classroom learning environment. Classroom autonomy
support and structure were standardized (i.e. transformed into Z scores) and their interaction
term was computed (autonomy support*structure). Reduction in the residual between-
classroom variance for each step is presented in the tables. Given that this calculation can be
problematic (i.e. adding some variables can negatively contribute to the explanation of the
variance), the proportional reduction in mean squared prediction error for the model variance
at level 2 at each step is reported in the text (see Snijders & Bosker, 1994, p. 353). Finally,
effects size was calculated using the formula 𝛿 = 𝛾
√(𝜏00+ 𝜎2), “where γ is the association
between the predictor and outcomes variables, and the denominator is the standard deviation
of the outcome variable, where τ00 and σ² are the between- and within-groups variances,
respectively (…)” (Reyes, Brackett, White, & Salovey, 2012, p.706). δ is interpreted the same
way as Cohen’s (1988) d (0.2 = small; 0.5 = moderate; 0.8 = large).
3. Results
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 13
The means, standard deviations and correlations for the variables under research are
reported in Table 1. Moderate correlations (Cohen, 1988) were found between dependent
variables, and a high correlation was found between behavioral engagement and self-
regulation. Autonomy support and structure were highly positively correlated.
Insert Table 1 about here
3.1. Share of the variance in engagement related to the classroom environment
Null models were run to determine the share of between-class variance (i.e. given by the
intra-class coefficient, ICC) in students’ engagement. Regarding behavioral engagement, 13%
of the variance (ICC = 0.13) lay between classes. Regarding cognitive engagement, 10% of
the variance in self-regulation (ICC = 0.10) and 12% of the variance in the use of deep
strategies (ICC = 0.12) lay between classes. Regarding emotional engagement, 7% of the
variance in positive (ICC = 0.07) and 5% of the variance in negative (ICC = 0.05) emotions
lay between classes. This significant between-class variance shows that student engagement
differs depending on the class attended by the student.
3.2. Controlling for individual characteristics and class composition
Regarding individual characteristics (see Models 1 in Tables 2, 3, 4, 5 and 6), girls reported
significantly more behavioral (δ = .43) and cognitive (use of deep strategies: δ = .23; self-
regulation: δ = .18) engagement than did boys. The cultural and/or material resources the
students reported having at home were positively related to their self-reported behavioral (δ
cultural resources = .41; δ material resources = .49) and cognitive engagement (self-regulation: δ cultural
resources = .24; δ material resources = .37; use of deep strategies: δ cultural resources =.45; δ material resources =
.36) during French lessons. The student cultural resources were positively related to their
positive emotions (δ = .32). The introduction of individual characteristics into the models
reduced the error for the prediction of the mean behavioral engagement of a randomly drawn
class by 30.1%; of the mean use of deep processing strategies of a randomly drawn class by
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 14
21%; of the mean self-regulated learning of a randomly drawn class by 8% compared with the
empty model. It did not affect the reduction of error for the prediction of emotional
engagement.
Few significant effects of class composition (see Models 2 in Table 2, 3, 4, 5 and 6) were
found. In classes with a higher ratio of girls, students reported more behavioral (δ = .41) and
cognitive engagement (i.e., use of deep processing strategies; δ = .44). In vocational tracks,
students reported more positive emotions (δ = .27). Including class composition variables in
the models reduced the error for the prediction of the mean behavioral engagement of a
randomly drawn class by 34.2% (which was a reduction of 4.1%, compared with the model
including the individual characteristics); of the mean use of deep processing strategies of a
randomly drawn class by 25% (which was a reduction of 4%, compared with the model
including the individual characteristics); of the mean positive emotions of a randomly drawn
class by 12.5% compared with the empty model. It did not affect the reduction of error for the
prediction of self-regulation and negative emotions.
These results highlight the importance of including control variables, especially individual
characteristics, when testing the relationships between teacher autonomy support/structure
and student engagement.
3.3. Effects of classroom autonomy support and structure
In a preliminary analysis, we checked if students’ perceptions of structure and autonomy
support differed between classes by investigating the partition of between-class variance for
each dimension. Thirty-one percent of the variance in perception of structure and 31% of the
variance in perception of autonomy support were attributed to the class level (ICC = 0.31).
Students’ perception of autonomy support and structure was not only related to their personal
experiences, but also to the class they belonged to. Consequently, aggregated measures of
autonomy support and structure were introduced at the class level in the models.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 15
- Hypothesis 1: When considered together, only structure – and not autonomy support –
is positively associated with student behavioral engagement.
Only structure was significantly associated with behavioral engagement (δ = .25; see Table
2). Structure provided by the teacher at the class level was positively related to behavioral
engagement at the student level.
Insert Table 2 about here
- Hypothesis 2: Structure and the interaction between autonomy support and structure
are positively associated with cognitive engagement.
Regarding the use of deep processing strategies, the results were non-significant. Only
structure was significantly associated with self-regulation (δ = .19; see Table 4). A higher
level of teacher structure provided at the class level was related to more frequent use of self-
regulation strategies among students. No significant interactions were found.
Insert Table 3 and 4 about here
- Hypothesis 3 (a): When considered together, autonomy support – but not structure -
and is positively associated with positive emotions.
Autonomy support (δ = .14) and structure (δ = .22) were both positively associated with
positive emotions. Teacher provision of autonomy support and structure was associated with
more positive emotions among students. An interaction between both dimensions was found
(δ = .08; see Table 5 and Figure 1). Examination of this interaction showed that teacher
provision of high levels of both structure and autonomy support was related with more
positive emotions among students.
- Hypothesis 3 (b): The interaction between autonomy support and structure is
negatively associated with negative emotions.
Autonomy support was significantly associated with negative emotions (δ = .14; see Table
6). Teacher provision of autonomy support at the classroom level was negatively linked with
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 16
student report of negative emotions. The interaction between autonomy support and structure
was significant for negative emotions (δ = .06). The negative association between autonomy
support and negative emotions was stronger when structure was high (see Figure 2).
Insert Figure 1 and 2 and Table 5 and 6 about here
The introduction of autonomy support and structure in the models reduced the error of the
prediction for the mean behavioral engagement of a randomly drawn class by 63% (which
was a reduction of 28.8%, compared with the model including classroom composition); for
the mean self-regulation of a randomly drawn class by 47% (which was a reduction of 39%,
compared with the model including the individual characteristics); for the mean positive
emotions of a randomly drawn class by 53 % (which was a reduction of 40.5%, compared
with the model including classroom composition); for the mean negative emotions of a
randomly drawn class by 23% compared with the empty model.
4. Discussion
Given the pivotal role of engagement for achievement and academic success, the
identification of the classroom factors related to engagement is a major issue. SDT (Deci &
Ryan, 2008) highlights the role of such factors, notably the provision of autonomy support and
structure, on engagement (Jang et al., 2010; Skinner et al., 2008). By using multilevel analyses,
the current study shows that students’ engagement is not only related to individual
characteristics but also to class variables. Most of between-class differences in students’
behavioral, cognitive and emotional engagement were linked to teacher provision of autonomy
support and structure in the class. Class composition appears less crucial. These results are
consistent with teacher effectiveness research showing the importance of class level and
especially of teachers’ practices for students’ learning (e.g., Kyriakides & Creemers, 2008).
They are also consistent with Jang and colleagues (2010) and Nie and Lau’s findings (2009)
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 17
who found an association between autonomy support / structure provided at the class level and
student engagement.
4.1. Associations with components of engagement
As little is known about the respective importance of autonomy support and structure for
the components of engagement, the aim of the current study was to test the associations
between autonomy support and structure at the classroom level and students’ engagement. Do
both dimensions be related to engagement (“additive effect”), only one of them (“specific
effect”) or do they interact (“combined effects”)? Consistently with previous studies based on
student perceptions of teacher practices (Sierens et al., 2009), our results show that autonomy
support and structure are highly and positively correlated. As expected (Hypothesis 1), a
specific relationship between structure, but not autonomy support, and behavioral engagement
is found. The fact that teacher provide more structure at the classroom level is positively
associated with higher student behavioral engagement. It is consistent with previous findings
(Skinner & Belmont, 1993; Wang & Eccles, 2013)
A specific link between structure and cognitive engagement and an interaction of
autonomy support and structure were expected (Hypothesis 2). The results are non-significant
regarding the use of deep processing strategies. A significant association between structure
and self-regulation is found, consistent with Sierens and colleagues (2009). Teacher provision
of structure at the classroom level is positively related to the report of the use of self-
regulation strategies by the students. We did not find the expected interaction with autonomy
support. As explained in the following discussion, this may be due to methodological
differences between studies. Different results appear according to the indicator of cognitive
engagement used: Structure has a significant relationship only with self-regulation but not
with the use of deep processing strategies.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 18
Regarding emotional engagement, we expected a specific association between autonomy
support and positive emotions (Hypothesis 3a). However, both autonomy support and
structure are associated with positive emotions, suggesting an additive effect. An interaction
is also found. Wang and Eccles (2013) found only a main effect of autonomy support, but they
did not test the interaction with structure. Our results show that the level of structure is also
important: The fact that teacher combined high levels of both autonomy support and structure
is related to higher levels of positive emotions. However, our results are contrary to Kunter’ s
and colleagues (2007) who found no effect of structure on interest. This could be due to
differences in measures used: These authors focused on interest, combining the dimension of
positive emotion and value commitment, while our measure of positive emotions includes
several discrete emotions.
We found the expected interaction between autonomy support and structure on negative
emotions (Hypothesis 3b). This is consistent with Vansteenkiste and colleagues’ findings
(2012). We also found a specific link between autonomy support and negative emotions. The
teacher provision of autonomy support is negatively related to student report of negative
emotions. The decomposition of the interaction show that the fact that the teacher provide
high levels of both autonomy support and structure is related to lower levels of student
negative emotions. The specific link found between autonomy support and negative emotions
is contrary to Vansteenkiste and colleagues’ study. However, they did not take the class level
into account. Moreover, they focused on test anxiety while we used a global measure of
negative emotions. The contextual variables which have a significant effect may differ
depending on the emotion investigated (Frenzel, Pekrun, & Goetz, 2007).
4.2. Relative importance of structure and autonomy support
These results suggest the presence of specific but also additive and combined effects of
autonomy support and structure, depending on the indicator of engagement used. Globally, a
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 19
specific relationships of structure is found with several indicators of engagement. The fact
that teacher provide structure in the classroom is positively associated with the level of
students behavioral, cognitive (i.e. self-regulated learning) and emotional (i.e. positive
emotions) engagement. Autonomy support has only a specific association with emotional
engagement. Student report of positive emotions is positively associated with teacher
provision of autonomy support in the classroom. Associations between both structure and
autonomy support appear with emotional engagement, suggesting an additive effects of both
dimensions. An interaction between autonomy support and structure also appears in relation to
emotional engagement, suggesting a combined effect. The fact that teacher provide a
combination of high levels of autonomy support and structure is positively related to positive
emotions and negatively related to negative emotions. This highlights the interest of
considering both dimensions in intervention to foster emotional engagement.
The effect sizes found show that what teachers do explains a large part of the differences in
student engagement between classes. This is the case for the various components of
engagement (except the use of deep processing strategies). This suggests that teachers may
play a role for his/her students, and that it may be key to draw the attention of teachers to the
benefits of providing their students with structure and, complementarily autonomy support.
Our study shows a number of specific relationships between autonomy support or structure
depending and the components of engagement. SDT does not make specific predictions about
the associations between each dimension and the components of engagement and no
explanations were proposed in previous studies. According to cognitive load theory (CLT),
given our limited cognitive resources, providing clear guidance reduces cognitive load related
to the learning tasks and allows students to focus their attention on relevant information,
which facilitates learning (Kirschner, Sweller, & Clark, 2006). By helping students to focus
their cognitive resources on the lesson, structure could facilitate the use of cognitive strategies
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 20
to deal with the work at hand. It could enhance behavioral engagement by focusing students’
attention on the task and preventing students doing others things (e.g., chatting).
These results suggest that autonomy support may be good for engagement but structure
may have a more pivotal role. These findings are consistent with CLT and educational
effectiveness research, which stress the value of providing structured activities, information
about learning content and progression from simple to complex activities for achievement
(Kirschner et al., 2006; Klahr & Nigam, 2006). Our findings suggest that the positive effect of
structure on achievement could be partially explained through its effects on engagement.
4.3. Limitations and suggestions for future research.
It is difficult to compare directly the result of the current study with previous studies in
SDT conducted at the individual level. The theoretical interpretation of the results of those
studies is difficult to establish. To what extent do individual perceptions reflect teacher
behaviors, a common shared environment in the classroom, differential experience of this
environment, idiosyncratic or biased interpretation, or more general individual differences?
Studies relying solely on individual perceptions do not allow assessment of the effects of the
learning environment (Lüdtke et al., 2009). Consequently, it seems hazardous to derive
recommendations for educational practices from these results, especially as these findings
have not been replicated in multilevel or experimental studies. Our study is one of the few to
investigate both autonomy support and structure at the class level, and to test their interaction
on student engagement. More studies using a similar design are needed to replicate our
results.
Even if the use of measures at the class level is an improvement compared to measures at
the individual level to study the effects of learning environment on students, the present
multilevel study still relies on cross-sectional data. Consequently, the directionality of the
effects found remains elusive. Teacher practices may have an effect on student engagement as
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 21
well as student engagement may influence teacher willingness to provide autonomy support
or structure (Skinner, Kindermann, Connell, & Wellborn, 2009). A third variable, not
measured in the study, may also explain the associations found between variables in
correlational studies. For instance, the academic composition of the class may influence
teacher practices (e.g. teachers may adapt their expectancies and practices according to the
class mean level of achievement; see Dumay & Dupriez, 2007). Some elements of classroom
composition were taken into account in the present study (including the ratio of retained
students, a proxy for academic composition), but this cannot rule out completely this kind of
issue. To overcome these limits, longitudinal and experimental studies are needed. Using
longitudinal design, the effects of autonomy support and structure at a specific time could be
more accurately evaluated by controlling for previous engagement level. Such design makes it
possible to test reciprocal effects between autonomy support/structure and engagement
(Skinner & Belmont, 1993). Experimental design makes it possible to determine the causality
of some relationships (West, Cham & Liu, 2014).
Given the limited sample of the current study, school level was not investigated. This could
be considered in analyses although school effects were found to be less influential than class
effects on student engagement (Vezeau et al., 2010). As we focused on students’ perceptions,
using teachers’ reports could provide a complementary approach in assessing the effects of the
learning environment.
Contextual variables, which have a significant effect, may differ depending on the emotion
investigated (Frenzel et al., 2007). As emotional engagement was measured through global
measures of positive and negative emotions, further studies may investigate the effects of
autonomy support and structure on discrete emotions.
To better understand the role of structure on engagement, the mediation effects of students’
needs-satisfaction could be investigated. Structure could be important since it may contribute
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 22
to fulfilling not only the need for competence but other needs as well, or since the need for
competence is more critical for engagement (Dupont et al., 2014). The fulfillment of some
needs may be more crucial depending on students’ characteristics (e.g., prior highly engaged
students, see Opdenakker & Minnaert, 2014). These considerations could supply more
precision in the formulation of SDT regarding the importance or the balance of different needs
in different conditions (Sheldon & Niemiec, 2006).
Further studies carried out in SDT should include contextual aspects other than autonomy
support, such as structure. A third aspect of teaching – involvement – should be included as,
according to SDT, it also plays a role in students’ engagement (Vansteenkiste et al., 2012).
The specific and combined effects of different aspects of teaching should also be given more
attention in SDT to better determine the most efficient way to enhance students’ engagement,
and academic success. In terms of getting a clearer picture of the way classroom context
impacts engagement, the results of the present study suggest that the components of
engagement should be distinguished, as results could differ from one component to another.
Regarding practical implications, these results suggest that the learning environment
shaped by the teacher at the class level really matters for students’ engagement, beyond the
effects of students’ characteristics and class composition. They support the idea that providing
students with structure is important to sustain their engagement. Autonomy support could be
beneficial, especially regarding emotional engagement during lessons.
4.4. Conclusion.
As postulated by SDT, the current study stresses the associations between classroom
environment, especially teacher structure, and student behavioral, cognitive and emotional
engagement. Autonomy support has a complementary role as it is associated with emotional
and cognitive engagement. These results highlight the importance of not restricting further
studies carried out within the SDT framework to autonomy support, and of including other
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 23
dimensions of the social context. They underline the importance of using multilevel analyses
to better assess the effects of the learning environment on students and to provide more
precise guidance to teachers and educators regarding the most effective ways to enhance
student engagement.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 24
References
Archambault, I., Janosz, M., Fallu, J.-S., & Pagani, L. S. (2009). Student engagement and its
relationship with early high school dropout. Journal of Adolescence, 32(3), 651-670.
doi: 10.1016/j.adolescence.2008.06.007
Assor, A., Kaplan, H., & Roth, G. (2002). Choice is good, but relevance is excellent:
Autonomy-enhancing and suppressing teacher behaviours predicting students'
engagement in schoolwork. British Journal of Educational Psychology, 72(2), 261-
278. doi: 10.1348/000709902158883
Belmont, M., Skinner, E., Wellborn, J., & Connell, J. (1988). Teacher as social context: A
measure of student perceptions of teacher provision of involvement, structure, and
autonomy-support. Rochester, NY: University of Rochester.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.): Hillsdale,
NJ: Lawrence Erlbaum.
De Fraine, B., Van Damme, J., Van Landeghem, G., Opdenakker, M.-C., & Onghena, P.
(2003). The effect of schools and classes on language achievement. British
Educational Research Journal, 29(6), 841-859. doi: 10.1080/0141192032000137330
Deci, E. L., & Ryan, R. M. (2008). Self-determination theory: A macrotheory of human
motivation, development, and health. Canadian Psychology/Psychologie canadienne,
49(3), 182-185. doi: 10.1037/a0012801
Dumay, X., & Dupriez, V. (2007). Does the school composition effect matter? Some
methodological and conceptual considerations. Les cahiers de recherche en éducation et
formation, 60.
Dupont, S., Galand, B., Nils, F., & Hospel, V. (2014). Social context, self-perceptions and
student engagement: A SEM investigation of the self-system model of motivational
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 25
development (SSMMD). Revista Electronica de Investigacion Educativa y
Psicopedagogica, 12(1), 5-32. doi: 10.14204/ejrep.32.13081
Fédération Wallonie-Bruxelles (2014). Indicateurs de l’enseignement 2013 [Educational
system indicators 2013]. Bruxelles : Ministère de l’enseignement obligatoire.
Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the
concept, state of the evidence. Review of Educational Research, 74(1), 59-109. doi:
10.3102/00346543074001059
Fredricks, J. A., & McColskey, W. (2012). The measurement of student engagement: A
comparative analysis of various methods and student self-report instruments. In S. L.
Christenson, A. L. Reschly & C. Wylie (Eds.), Handbook of Research on Student
Engagement (pp. 763-782): Springer US.
Frenzel, A. C., Pekrun, R., & Goetz, T. (2007). Perceived learning environment and students'
emotional experiences: A multilevel analysis of mathematics classrooms. Learning
and Instruction, 17(5), 478-493. doi: 10.1016/j.learninstruc.2007.09.001
Galand, B., & Philippot, P. (2005). L'école telle qu'ils la voient: Validation d'une mesure des
perceptions du contexte scolaire par les élèves du secondaire [School as they see it:
Validation of a measure assessing student perceptions of school context]. Canadian
Journal of Behavioural Science, 37(2), 138-154. doi: 10.1037/h0087251
Galand, B., Philippot, P., & Frenay, M. (2006). Structure de buts, relations enseignants-élèves
et adaptation scolaire des élèves : une analyse multi-niveaux [Goal structures, teacher-
student relationships, and academic adjustment : A multilevel analysis]. Revue
française de pédagogie, 155, 57-72.
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 26
Galand, B., Raucent, B., & Frenay, M. (2010). Engineering students' self-regulation, study
strategies, and motivational believes in traditional and problem-based curricula.
International Journal of Engineering Education, 523-534.
Hospel, V., & Galand, B. (2010). Toward a more complex conceptualization of behavioural
engagement: A pattern analysis on participation, compliance, withdrawal, disruptive
behaviour, and absenteeism. Paper presented at the Summer school preceding the
"International Conference on Motivation", Porto.
Izard, C. E., Dougherty, F. E., Bloxom, B. M., & Kotsh, N. E. (1974). The differential emotion
scale: A method of measuring the meaning of subjective experience of discrete
emotions. Nashville, TE : Vanderbilt University.
Jang, H., Reeve, J., & Deci, E. L. (2010). Engaging students in learning activities: It is not
autonomy support or structure but autonomy support and structure. Journal of
Educational Psychology, 102(3), 588-600. doi: 10.1037/a0019682
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction
does not work: An analysis of the failure of constructivist, discovery, problem-based,
experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75-86. doi:
10.1207/s15326985ep4102_1
Klahr, D., & Nigam, M. (2004). The equivalence of learning paths in early science
instruction. Effects of direct instruction and discovery learning. Psychological
Science, 15, 661-667. doi: 10.1111/j.0956-7976.2004.00737.x
Kunter, M., Baumert, J., & Köller, O. (2007). Effective classroom management and the
development of subject-related interest. Learning and Instruction, 17(5), 494-509. doi:
10.1016/j.learninstruc.2007.09.002
Kyriakides, L., & Creemers, B. P. M. (2008). Using a multidimensional approach to measure
the impact of classroom-level factors upon student achievement: A study testing the
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 27
validity of the dynamic model. School Effectiveness and School Improvement, 19(2),
183-205. doi: 10.1080/09243450802047873
Lüdtke, O., Trautwein, U., Kunter, M., & Baumert, J. (2006). Reliability and agreement of
student ratings of the classroom environment: A reanalysis of TIMSS data. Learning
Environments Research, 9(3), 215-230. doi: 10.1007/s10984-006-9014-8
Lüdtke, O., Robitzsch, A., Trautwein, U., & Kunter, M. (2009). Assessing the impact of
learning environments: How to use student ratings of classroom or school
characteristics in multilevel modeling. Contemporary Educational Psychology, 34(2),
120-131. doi: 10.1016/j.cedpsych.2008.12.001
Marsh, H. W., Lüdtke, O., Nagengast, B., Trautwein, U., Morin, A. J. S., Abduljabbar, A. S.,
& Köller, O. (2012). Classroom climate and contextual effects: Conceptual and
methodological issues in the evaluation of group-level effects. Educational
Psychologist, 47(2), 106-124. doi: 10.1080/00461520.2012.670488
Nie, Y., & Lau, S. (2009). Complementary roles of care and behavioral control in classroom
management: The self-determination theory perspective. Contemporary Educational
Psychology, 34(3), 185-194. doi: 10.1016/j.cedpsych.2009.03.001
OECD (2006). Programme for International Student Assessment (PISA).
Opdenakker, M.-C. & Minnaert, A. (2014). Learning environment experiences in primary
education: Their importance to academic engagement. In D. Zandvliet, P. den Brok, T.
Mainhard & J. van Tartwijk (Eds.), Interpersonal relationships in education: Fron
theory to practice (pp. 183-194). Rotterdam, The Netherlands: Sense.
Opdenakker, M.-C., & Van Damme, J. (2001). Relationship between school composition and
characteristics of school process and their effect on mathematics achievement. British
Educational Research Journal, 27(4), 407-432. doi: 10.1080/01411920120071434
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 28
Pokorny, S. B., Jason, L. A., Schoeny, M. E., Townsend, S. M., & Curie, C. J. (2001). Do
participation rates change when active consent procedures replace passive consent?
Evaluation Review, 25(5), 567-580. doi: 10.1177/0193841X0102500504
Reeve, J. (2002). Self-determination theory applied to educational settings. Handbook of self-
determination research. (pp. 183-203): University of Rochester Press, Rochester, NY.
Reeve, J. (2006). Extrinsic Rewards and Inner Motivation. In C. M. Evertson & C. S.
Weinstein (Eds.), Handbook of classroom management: Research, practice, and
contemporary issues (pp. 1346-1664). Mahwah, NJ, US: Lawrence Erlbaum
Associates Publishers.
Reeve, J., & Halusic, M. (2009). How K-12 teachers can put self-determination theory
principles into practice. Theory and Research in Education, 7(2), 145-154. doi:
10.1177/1477878509104319
Reeve, J., Jang, H., Carrell, D., Jeon, S., & Barch, J. (2004). Enhancing students' engagement
by increasing teachers' autonomy support. Motivation and Emotion, 28(2), 147-169.
doi: 10.1023/B:MOEM.0000032312.95499.6f
Reyes, M. R., Brackett, M. A., Rivers, S. E., White, M., & Salovey, P. (2012). Classroom
emotional climate, student engagement, and academic achievement. Journal of
Educational Psychology, 104, 700-712. doi:10.1037/a0027268
Sheldon, K.M. & Niemiec, C.P. (2006). It's not just the amount that counts: balanced need
satisfaction also affects well-being. Journal of Personality and Social Psychology, 91,
331-341. doi: 10.1037/0022-3514.91.2.331
Shernoff, D.J. (2013). Optimal Learning Environments to Promote Student Engagement, 77
Advancing Responsible Adolescent Development. New York: Springer. Doi:
10.1007/978-1-4614-7089-2_4
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 29
Sierens, E., Vansteenkiste, M., Goossens, L., Soenens, B., & Dochy, F. (2009). The synergistic
relationship of perceived autonomy support and structure in the prediction of self-
regulated learning. British Journal of Educational Psychology, 79(1), 57-68. doi:
10.1348/000709908x304398
Skinner, E. A., Kindermann, T. A., Connell, J. P., & Wellborn, J. G. (2009). Engagement and
disaffection as organizational constructs in the dynamics of motivational development.
In K. R. Wenzel & A. Wigfield (Eds.), Handbook of motivation at school (pp. 686-
245). New York: Routledge/Taylor & Francis Group.
Skinner, E., Furrer, C., Marchand, G., & Kindermann, T. (2008). Engagement and disaffection
in the classroom: Part of a larger motivational dynamic? Journal of Educational
Psychology, 100(4), 765-781. doi: 10.1037/a0012840
Skinner, E. A., & Belmont, M. J. (1993). Motivation in the classroom: Reciprocal effects of
teacher behavior and student engagement across the school year. Journal of
Educational Psychology, 85(4), 571-581. doi: 10.1037/0022-0663.85.4.571
Snijders, T. A. B., & Bosker, R. J. (1994). Modeled variance in two-level models.
Sociological Methods & Research, 22, 342-364. doi: 10.1177/0049124194022003004
Stroet, K., Opdenakker, M.-C., & Minnaert, A. (2013). Effects of need supportive teaching on
early adolescents’ motivation and engagement: A review of the literature. Educational
Research Review, 9(0), 65-87. doi: 10.1016/j.edurev.2012.11.003
Tucker, C. M., Zayco, R. A., Herman, K. C., Reinke, W. M., Trujillo, M., Carraway, K., . . .
Ivery, P. D. (2002). Teacher and child variables as predictors of academic engagement
among low-income African American children. Psychology in the Schools, 39(4), 477-
488. doi: 10.1002/pits.10038
Vansteenkiste, M., Sierens, E., Goossens, L., Soenens, B., Dochy, F., Mouratidis, A., . . .
Beyers, W. (2012). Identifying configurations of perceived teacher autonomy support
Running head: ARE AUTONOMY SUPPORT AND STRUCTURE EQUALLY IMPORTANT 30
and structure: Associations with self-regulated learning, motivation and problem
behavior. Learning and Instruction, 22(6), 431-439. doi:
10.1016/j.learninstruc.2012.04.002
Vansteenkiste, M., Simons, J., Lens, W., Sheldon, K. M., & Deci, E. L. (2004). Motivating
learning, performance, and persistence: The synergistic effects of intrinsic goal
contents and autonomy-supportive contexts. Journal of Personality and Social
Psychology, 87(2), 246-260. doi: 10.1037/0022-3514.87.2.246
Vezeau, C., Chouinard, R., Bouffard, T., Janosz, M., Bergeron, J., & Bouthillier, C. (2010).
Estimation de l'effet-école et de l'effet classe sur la motivation des élèves du
secondaire [Estimation of school- and classroom-effect on secondary school student
motivation]. Revue des sciences de l'éducation, 36(2), 445-468. doi: 10.7202/044485ar
Wang, M.-T., & Eccles, J. S. (2013). School context, achievement motivation, and academic
engagement: A longitudinal study of school engagement using a multidimensional
perspective. Learning and Instruction, 28(0), 12-23. doi:
10.1016/j.learninstruc.2013.04.002
West, S. G., Cham, H., & Liu, Yu (2014). Causal Inference and Generalization in Field
Settings. Experimental and Quasi-Experimental Designs. In H. T. Reis & Judd, C. M.
(Eds.), Handbook of Research Methods in Social and Personality Psychology (p. 49-
80). New York: Cambridge University Press.
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT
Table 1.
Descriptive statistics and correlation matrix for indicators of engagement, autonomy support and
structure (student level).
M
SD
1
2
3
4
5
6
7
1. Behavioral engagement
2.60
.62
1.00
Cognitive engagement
2. Self-regulation
2.30
.76
.74**
1.00
3. Deep strategies
2.02
.80
.46**
.50**
1.00
Emotional engagement
4. Positive emotions
2.25
.73
.47**
.52**
.44**
1.00
5. Negative emotions
.96
.60
-.25**
-.22**
-.16**
-.34**
1.00
Teacher
6. Autonomy support
2.22
.74
.36**
.33**
.35**
.39**
-.22**
1.00
7. Structure
2.53
.68
.42**
.41**
.32**
.44**
-.27**
.60**
1.00
Note. **p < .01.
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT 32
Table 2.
Results of multilevel analyses for behavioral engagement.
Note. N students = 744; N classrooms = 51.
Model 1
Model 2
Model 3
Coefficient
SE
p
Coefficient
SE
p
Coefficient
SE
p
Intercept
2.80
(0.05)
<.01
2.91
(0.06)
<.01
2.76
(0.05)
<.01
Student level
Gender
-0.26
(0.05)
<.01
-0.22
(0.05)
<.01
-0.22
(0.05)
<.01
Cultural resources
0.25
(0.07)
<.01
0.23
(0.07)
<.01
0.24
(0.08)
<.01
Material resources
0.30
(0.09)
<.01
0.30
(0.09)
<.01
0.29
(0.09)
<.01
Grade retention
-0.07
(0.05)
.12
-0.06
(0.05)
.18
-0.06
(0.04)
.16
Classroom level
1. Classroom
composition
Gender ratio
-0.25
(0.12)
.04
-0.01
(0.10)
.94
Cultural resources
-
-
Material resources
-
-
Ratio of retained
students
-
-
Educational track
-
-
2. Teachers’
practices
Structure
0.15
(0.03)
<.01
Autonomy support
0.05
(0.02)
.06
Structure*
Autonomy support
0.03
(0.03)
.22
∆ Between-classroom
variance
31%
8%
46%
Total of between-
classroom variance
explained
31%
39%
85%
Deviance
1298
1294
1283
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT 33
Table 3.
Results of multilevel analyses for cognitive engagement (use of deep strategies).
Note. N students = 744; N classrooms = 51.
Model 1
Model 2
Model 3
Coefficient
SE
p
Coefficient
SE
p
Coefficient
SE
p
Intercept
2.11
(0.06)
.00
2.27
(0.09)
<.01
2.27
(0.11)
<.01
Student level
Gender
-0.18
(0.06)
<.01
-0.12
(0.07)
.07
-0.13
(0.07)
.07
Cultural resources
0.36
(0.10)
<.01
0.33
(0.09)
<.01
0.35
(0.09)
<.01
Material resources
0.29
(0.12)
.02
0.29
(0.12)
.02
0.28
(0.12)
.03
Grade retention
-0.02
(0.07)
.74
-0.01
(0.07)
.84
-0.02
(0.06)
.69
Classroom level
1. Classroom
composition
Gender ratio
-0.35
(0.16)
.031
-0.26
(0.17)
.13
Cultural resources
-
-
Material resources
-
-
Ratio of retained
students
-
-
Educational track
-
-
2. Teachers’
practices
Structure
0.01
(0.06)
.93
Autonomy support
0.08
(0.07)
.23
Structure*
Autonomy support
-0.05
(0.03)
.09
∆ Between-classroom
variance
25%
8%
33%
Total of between-
classroom variance
explained
25%
33%
66%
Deviance
1718
1713
1714
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT 34
Table 4.
Results of multilevel analyses for cognitive engagement (self-regulation).
Note. N students = 744; N classrooms = 51.
Model 1
Model 2
Model 3
Coefficient
SE
p
Coefficient
SE
p
Coefficient
SE
p
Intercept
2.38
(0.06)
<.01
2.38
(0.06)
<.01
2.37
(0.05)
<.01
Student level
Gender
-0.10
(0.05)
.05
-0.10
(0.05)
.05
-0.07
(0.05)
.19
Cultural resources
0.18
(0.10)
.06
0.18
(0.10)
.06
0.19
(0.10)
.04
Material resources
0.28
(0.11)
<.01
0.28
(0.11)
<.01
0.27
(0.11)
.02
Grade retention
-0.05
(0.07)
.47
-0.05
(0.07)
.47
-0.05
(0.06)
.37
Classroom level
1. Classroom
composition
Gender ratio
-
-
Cultural resources
-
-
Material resources
-
-
Ratio of retained
students
-
-
Educational track
-
-
2. Teachers’
practices
Structure
0.14
(0.04)
<.01
Autonomy support
0.07
(0.05)
.18
Structure*
Autonomy support
-0.01
(0.03)
.93
∆ Between-classroom
variance
10%
0%
60%
Total of between-
classroom variance
explained
10%
0%
70%
Deviance
1661
1661
1646
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT 35
Table 5.
Results of multilevel analyses for emotional engagement (positive emotions).
Note. N students = 744; N classrooms = 51.
Model 1
Model 2
Model 3
Coefficient
SE
p
Coefficient
SE
p
Coefficient
SE
p
Intercept
2.28
(0.07)
<.01
2.19
(0.08)
<.01
2.13
(0.05)
<.01
Student level
Gender
-0.04
(0.05)
.44
-0.04
(0.05)
.36
0.03
(0.04)
.47
Cultural resources
0.23
(0.12)
.04
0.28
(0.12)
.02
0.26
(0.11)
.02
Material resources
0.17
(0.10)
.11
0.19
(0.10)
.06
0.18
(0.10)
.07
Grade retention
0.01
(0.06)
.95
-0.06
(0.06)
.30
-0.06
(0.05)
.27
Classroom level
1. Classroom
composition
Gender ratio
-
-
Cultural resources
-
-
Material resources
-
-
Ratio of retained
students
-
-
Educational track
0.20
(0.09)
.02
0.16
(0.06)
<.01
2. Teachers’
practices
Structure
0.16
(0.04)
<.01
Autonomy support
0.10
(0.04)
.02
Structure*
Autonomy support
0.06
(0.02)
<.01
∆ Between-classroom
variance
0%
14%
71%
Total of between-
classroom variance
explained
0%
14%
85%
Deviance
1634
1630
1608
Running head: CLASSROOM ENVIRONMENT EFFECTS ON ENGAGEMENT 36
Table 6.
Results of multilevel analyses for emotional engagement (negative emotions).
Note. N students = 744; N classrooms = 51.
Model 1
Model 2
Model 3
Coefficient
SE
p
Coefficient
SE
p
Coefficient
SE
p
Intercept
0.97
(0.05)
<.01
0.97
(0.05)
<.01
1.02
(0.05)
<.01
Student level
Gender
-0.02
(0.04)
.68
-0.02
(0.04)
.68
-0.06
(0.04)
.13
Cultural resources
-0.01
(0.08)
.92
-0.01
(0.18)
.92
0.01
(0.08)
.98
Material resources
-0.20
(0.12)
.09
-0.20
(0.12)
.09
-0.19
(0.12)
.10
Grade retention
-0.01
(0.05)
.95
-0.01
(0.05)
.95
-0.01
(0.05)
.81
Classroom level
1. Classroom
composition
Gender ratio
-
-
Cultural resources
-
-
Material resources
-
-
Ratio of retained
students
-
-
Educational track
-
-
2. Teachers’
practices
Structure
-0.06
(0.04)
.15
Autonomy support
-0.09
(0.04)
.03
Structure*
Autonomy support
-0.04
(0.02)
.05
∆ Between-classroom
variance
0%
0%
60%
Total of between-
classroom variance
explained
0%
0%
60%
Deviance
1418
1418
1419
