Motivating young language learners: A longitudinal model of motivational development in elementary school foreign language classes

Abstract

Promoting intrinsic motivation is often a central concern in teaching foreign languages to elementary school children. Self-determination theory posits that intrinsic motivation develops through the interaction of the person and the environment. The present study investigated how elementary school students’ motivation develops over the course of a school year in Japanese public schools. Five-hundred and fifteen Japanese elementary school children were surveyed over the course of one school year. Self-reported motivation, perceptions of teacher support, need satisfaction, and engagement were measured at different times. External raters observed students’ engagement, while classroom teachers assessed the quality of students’ motivation and learning. Structural equation modeling results indicated a positive, dynamic relationship between motivation, perceptions of the learning environment, and engagement. External raters’ assessments showed significant positive correlations with students’ self-reported engagement. Findings indicate how the instruction offered in these Japanese elementary schools supported students’ foreign language learning motivation.

Full text

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8,
Motivating young language learners: A
longitudinal model of self-determined
motivation in elementary school foreign
language classes
W. L. Quint Oga-Baldwin1,2, Yoshiyuki Nakata3, Philip D.
Parker4, & Richard M. Ryan4,5
,
Cite%as:!
Oga-Baldwin, W. L. Q, Nakata, Y., Parker, P. D., & Ryan, R. M. (2017). Motivating young language
learners: A longitudinal model of motivational development in elementary school foreign
language classes. Contemporary Educational Psychology. doi: 10.1016/j.cedpsych.2017.01.010,
Abstract
Promoting intrinsic motivation is often a central concern in teaching foreign
languages to elementary school children. Self-determination theory posits that
intrinsic motivation develops through the interaction of the person and the
environment. The present study investigated how elementary school students’
motivation develops over the course of a school year in Japanese public schools. Five-
hundred and fifteen Japanese elementary school children were surveyed over the
course of one school year. Self-reported motivation, perceptions of teacher support,
need satisfaction, and engagement were measured at different times. External raters
observed students’ engagement, while classroom teachers assessed the quality of
students’ motivation and learning. Structural equation modeling results indicated a
positive, dynamic relationship between motivation, perceptions of the learning
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
8,D?.?;.#,E*)F21A)/5,;G,>'?@#/);*,
6,H#A2'#,E*)F21A)/5,
I,J;AB)AB#,E*)F21A)/5,
K,L?A/1#&)#*,:#/B;&)@,E*)F21A)/5,
M,E*)F21A)/5,;G,4;@B2A/21,
,

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6,
environment, and engagement. External raters’ assessments showed significant
positive correlations with students’ self-reported engagement. Findings indicate how
the instruction offered in these Japanese elementary schools supported students’
foreign language learning motivation.
Keywords: SDT; longitudinal model; engagement; motivational development;
elementary school; Japan
1. Introduction
For elementary school children, learning a language can often be a process of growth
and discovery. When learners develop positive affect for the foreign language, it can
lead to a lifelong interest. Making the process of foreign language learning attractive
to children is a goal of many instructional programs (Garton, Copland, & Burns,
2011). In these contexts, motivation, and more specifically intrinsic motivation,
becomes a key focus in the classroom process.
Following this trend, the Japanese Ministry of Education, Culture, Sports,
Science, and Technology (MEXT), has emphasized intrinsic motivation (“zest for
life”) in its institution of compulsory (English) foreign language studies for all 5th and
6th grade pupils (MEXT, 2008). Under this Course of Study, students experience
foreign language communication through interaction and games, but are not given
summative assessments due to the potentially damaging motivational consequences
(Berwick & Ross, 1989). Instead, teachers nurture motivation through a low-pressure,
low-stakes learning environment (Ryan & Niemiec, 2009), based on experiential
learning, with no summative assessments and a strong emphasis on enjoyment
(MEXT, 2008). The ultimate goal is to promote motivation through supporting
students’ behavior, interest, and positive attitude toward the foreign language; in other

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
I,
words, their engagement and intrinsic motivation. This leads to the question of how
teachers can effectively support and maintain this type of activity and motive. As
noted by Butler (2015, p. 319), situated research on the learning context is now
needed to find how best to support and maintain young language learners’ motivation.
In this study, we address this call for further research.
Previous longitudinal models of motivational development in first and second
language educational settings have primarily focused on secondary school children
(e.g., Jang, Kim, & Reeve, 2012; Csizér & Dörnyei, 2005). While there have been
cross-sectional studies looking at Japanese elementary students’ language learning
motivation (Carreira, 2011; 2012; etc.), previous studies have not approached this
from a latent-variable, structural equation modeling perspective. Due to numerous
constraints on the use of testing in elementary foreign language classes, previous
models have also not included external assessments. We propose an empirical model
of how motivation to learn a foreign language begins to develop in a public
elementary school setting, including external assessment of learning outcomes.
1.1 Foreign language motivation in elementary schools
In recent years, researchers have given considerable attention to motivation in
elementary schools across East Asia (Butler, 2015). In Japan, the Ministry of
Education currently promotes the ideas of interest and motivation in foreign language
learning through an emphasis on communication and games in elementary
classrooms; the Course of Study specifically refers to promoting interest, behavior,
and motivation (MEXT, 2008). Students learn words and phrases through interest-
building, activity-based classes, without relying on external rewards such as praise
and high-stakes assessment. This paradigm of instruction is consistent with the
motivational ideas put forth in self-determination theory (SDT; Deci & Ryan, 1985),

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
K,
in that both seek to promote a positive motivational climate for language learning
(Oga-Baldwin & Nakata, 2014).
According to SDT, intrinsic motivation is defined as the desire to engage with
a task for its own sake, and is often indexed by personal enjoyment, interest, and
feelings of positive affect. Applied to language education generally, SDT has shown
positive explanatory power for students’ desire to continue learning the new language
(Noels, Clément, Pelletier, & Vallerand, 2000) and interact with foreign cultures
(Noels, Clément, & Pelletier, 2001; Vansteenkiste, Zhou, Lens & Soenens, 2005).
Applying ideas from SDT in a series of cross-sectional studies in Japanese
elementary schools, Carreira (2006; 2011; 2012) found a clear pattern of decreasing
motivation to learn English as a foreign language across school years, subjects, and
genders. Students’ motivation to learn in schools decreased in higher grades, both in
terms of the desire to learn English (Carreira, 2006) and the desire to learn other
subjects (Carreira, 2011). Male students also showed lower levels of adaptive
motivation compared to females. These results are echoed in studies of young learners
in Korea (Kim, 2011).
In a recent study, Carreira and her colleagues (2013) found evidence that
teachers’ support correlated with higher student motivation. Using path analyses, the
results of this study suggested that teachers’ support predicted a sense of more
autonomous motivation for learning a foreign language. While previous work
indicated that the quantity of motivation diminishes as students age (Carreira, 2011),
this study offered the hope that perhaps through effective pedagogy, teachers could
influence the quality of students’ motivation.
Similar findings come from studies in China. Parents from higher socio-
economic backgrounds supported their children’s sense of autonomy and self-

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
M,
determined motivation (Butler, 2014). These higher income parents were also more
likely to provide a less-controlling and more nurturing environment for children
learning a foreign language. Similarly, teachers in Korea attributed decreases in
student motivation to teaching practices (Kim & Seo, 2011). Taken together, these
studies indicate that the decreasing trend in motivation noted in previous studies may
be a partial product of their environment, potentially remedied by providing better
support for students’ motivational needs. If this is so, students with positive
perceptions of their teachers’ support should show a lower decrease in motivation
over time.
According to MEXT (2008), classroom teaching should support positive
interest in and behavior toward language learning. Continuing in the traditions
defined by previous language learning studies (Butler, 2014; Carreira et al., 2013;
Nishida, 2013), we integrate self-determination theory and its minitheories with the
concept of engagement to describe how elementary school learners develop a sense of
positive autonomous motivation.
1.2 Self-determination theory and its minitheories
As a broad theory of human motivation across domains, SDT attempts to
organize the numerous aspects of motivation, including how and why people do what
they do, the effects of the environment, and personal needs and goals (Deci & Ryan,
1985). According to Reeve (2012), SDT is “is a macrotheory of motivation comprised
of five interrelated minitheories” (p. 150). The three minitheories relevant to the
current study are organismic integration theory (OIT), basic psychological need
theory (BPNT), and cognitive evaluation theory (CET) (Ryan, Deci, & Vansteenkiste,
2016). Researchers have tested different combinations of these theories (cf. Carreira et
al., 2013; Jang et al., 2012; Noels et al., 2000), but none have used all of them

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
N,
together in an empirical longitudinal model. In this study, we test all three theories
alongside the concept of engagement to build a motivational model for foreign
language learning.
1.2.1 Reasons why: Regulation of motivation
Self-determination theory posits that learners have a range of motives that can
underpin their efforts at learning. This minitheory, called organismic integration
theory (OIT), describes a set of behavioral regulation patterns, moving from external,
controlled reasons to internalized, autonomous reasons. In broad terms, OIT describes
why learners choose to engage in their schoolwork on a continuum from controlled to
autonomous motivation.
Controlled motivation is comprised of motives whose locus of causality is
outside of the person. It is represented by external and introjected regulations. Under
external regulation, students complete tasks in order to get praise, rewards, or avoid
negative consequences. Extrinsically regulated behavior disappears quickly after the
rewards disappear (Deci, Koestner, & Ryan, 2001). Introjected regulation comes from
a sense of “ought-to,” shame or other social pressure associated with a task. This form
of regulation is brought about by a desire not to seem incapable in the eyes of
classmates, or to receive positive regard from parents or teachers. These two
categories of maladaptive motivation can be used together (Vansteenkiste et al., 2009),
but also may appear as separate and distinct sets of motives (Carreira, 2012; Noels et
al., 2000). Students with more controlled motives generally show less effective time
management and greater anxiety (Senécal, Julien, & Guay, 2003), and ultimately
lower achievement (Soenens & Vansteenkiste, 2005).
The opposite of controlled motivation, autonomous motivation, is defined by
two types of regulation: identified and intrinsic regulation. Prior studies have

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
9,
measured these two regulations together as autonomous motivation (e.g.,
Vansteenkiste et al., 2009). Identified regulation refers to how individuals perceive
personal value in learning. This presents as a desire to learn for tangible or intangible
future gains, such as attaining the skills necessary for a dream job, becoming part of a
desired target community, or other instrumental outcome (Noels, 2013). Intrinsic
regulation is characterized as a belief that the learning task is stimulating, that
accomplishment in and of itself is worthwhile, and that studying and knowing new
things is pleasurable (Noels et al., 2000). Autonomously motivated students use more
deep-level learning strategies (Vansteenkiste et al., 2005) and achieve better grades
(Soenens & Vansteenkiste, 2005).
Studies have replicated the internal to external continuum from autonomous to
controlled motivation presented by organismic integration theory to describe
motivation for learning a foreign language (Noels et al., 2000). Other research using
this aspect of SDT has been conducted in Japanese elementary schools, (Ando, Fuse,
& Kodaira, 2009; Carreira, 2012; Oga-Baldwin & Nakata, in press; Yamauchi &
Tanaka, 2000), showing support for the theory. In some of these studies, students’
responses indicated a three-factor solution, with identified and introjected regulations
factoring together (Ando, Fuse, & Kodaira, 2008; Carreira, 2012), while others
indicated a four-factor solution (Oga-Baldwin & Nakata, in press; Yamauchi &
Tanaka, 2000). Ando and colleagues’ (2008) sample came from a large group of third
through sixth grade students, while the others focused only on students in fifth and
sixth grades. In all of these studies, the patterns of correlation were largely consistent
with OIT, though the slight differences in results and cross-sectional nature of these
studies make it difficult to conclude any specific developmental trends.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
O,
More recently, researchers have used aspects of the continuum to show how
the learning environment positively correlates with Japanese elementary schools
students’ autonomous motivation to study English (Carreira, Ozaki, & Maeda, 2013).
Other cross-sectional studies have shown a positive relationship between engagement
and motivational regulation (Oga-Baldwin & Nakata, in press).
1.2.2 Autonomy, relatedness, and competence needs: Motivational nutrients
Basic psychological needs theory is conceptualized under the assumption that
human beings thrive under situations where their basic psychological needs are met.
Just as people require clean air and water, nutritious food, and sufficient exercise for
optimal physical functioning (Ryan & Deci, 2002), these needs are theorized to
nourish and sustain high quality motivation and psychological wellness. For the sake
of parsimony, SDT recognizes three basic needs: the need to feel a connection to
others (Baumeister & Leary, 1995), referred to as the need for relatedness; the need to
feel capable of influencing the surrounding environment in a meaningful way (White,
1959), titled the need for competence; and the need to feel a sense of personal
causality and volition in one’s actions (deCharms, 1968), referred to as the need for
autonomy. As people are social animals, these needs are necessarily interrelated (and
in most contexts they are highly correlated; Deci & Ryan, 2000) and reciprocally
influence one another. A threat to any one need hinders optimal functioning.
Applied to the field of education, need satisfaction has been used to explain
students’ classroom engagement (Jang, Reeve, Ryan, & Kim, 2009). Longitudinally,
autonomy need satisfaction has been shown to mediate the influence of the classroom
on students’ engagement and achievement (Jang, Kim, & Reeve, 2012). Researchers
have connected need satisfaction to autonomous motivation in various language

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
P,
learning settings (Carreira et al., 2013; Noels, 2013; McEown, Noels, & Saumure,
2014).
1.2.3 Autonomy support, structure, and teaching
According to the final minitheory, cognitive evaluation theory, teachers create a
motivationally supportive environment through the use of interesting activities, timely
feedback, judicious rewards, acceptance of students’ affect, and culturally appropriate
expectations (Reeve, 2012). Teachers may control students toward a single desired
behavioral outcome through rewards and punishments, or may focus on providing
students with the resources to feel initiative and choice in learning by focusing on
autonomy-support (Deci, Koestner, & Ryan, 2001). As described by Reeve (2012),
“autonomy-support is whatever a teacher says and does during instruction to facilitate
students’ perceptions of autonomy and experiences of psychological need satisfaction”
(p. 167). This definition allows for broad interpretation across cultures while retaining
the essential underlying concept that support for the person’s basic psychological
needs is essential to good teaching. Because autonomy refers to the support of
student’s willingness and volition rather than “independence” or separateness,
teachers’ support for learners’ autonomy remains an important factor even in cultural
settings high in collectivism (Jang et al., 2009). In foreign language educational
settings, autonomy-supportive teaching helps to promote autonomous motivation
(Noels, Clément, & Pelletier, 1999; Carreira et al., 2013).
In education, the quality of support for students’ needs is often balanced
against the idea of structure, the form that instruction takes. Structured teaching is
clear, well-organized, appropriately paced, provides feedback, and builds new
knowledge (Jang et al., 2010). Studies have shown that this aspect of instruction is
positively correlated with autonomy support (Sierens et al., 2009), sometimes

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
87,
inextricably so (Oga-Baldwin & Nakata, 2015). Thus, both the form and the quality of
instruction are an integral part of the motivational process in classrooms.
Taken together, the organismic integration, basic psychological needs, and
cognitive evaluation theories explain why people act, what sustains their action, and
how they perceive their environment. However, these theories by themselves do not
account for the action itself. This leads to the question of what students actually do,
feel, and think during their studies; in short, their engagement. Given that behavior
and enjoyment are explicitly mentioned as part of the Course of Study (MEXT, 2008),
engagement should be considered an integral aspect of a dynamic process leading
toward the goal of motivational development.
1.3 Engagement: Energy in action
Recently, researchers have integrated SDT concepts such as need satisfaction,
structure, and support with the idea of engagement (Jang, Reeve, & Deci, 2010; Jang
et al., 2009; Jang et al., 2012; Skinner et al., 2008, etc.) Different from motivation,
engagement represents the temporary state where students are acting, studying, and
doing. Where motivation is the potential and direction of students’ energy,
engagement is that energy being used to learn actively. Research into supportive
teaching has used engagement as a dynamic pivot point in the process of classroom
motivational growth (cf. Skinner et al., 2008; Jang et al., 2012). In these models, the
classroom environment both influences and is influenced by the degree to which
students enjoy their studies, pay attention, and think actively in a virtuous circle.
Likewise, when students feel the classroom exerts a negative impact on them, they are
likely to “turn off” and enter a vicious cycle of maladaptive beliefs and behaviors
with regard to the school environment (Jang et al., 2016; Skinner et al., 2008).
Engagement is a multifaceted concept (Fredricks, Blumenfeld, & Paris, 2004).

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
88,
Students may work hard, pay attention in class, and complete their assignments as a
form of behavioral engagement. They may find the class enjoyable, fun, and
interesting, which represents emotional engagement. They may find their assignments
and activities to be challenging, causing them to think deeply and use their abilities to
the fullest, which represents cognitive engagement.
Together, these forms of engagement have shown meaningful effects on
achievement (Jang et al., 2012; Jang et al., 2016), and further positive reciprocal
effects on teachers’ attitudes and behaviors toward students (Skinner & Belmont,
1993; Skinner et al., 2008). Engagement predicts students’ self-efficacy and goals
over time (Reeve & W. Lee, 2014). Positive engagement has shown direct effects on
students’ learning and achievement (Jang et al., 2012). Teachers may also be more
able to recognize engagement than motivation (W. Lee & Reeve, 2012). This concept
thus represents a crucial element in the process of describing both how students learn
in the classroom and how these selfsame students are perceived by their teachers.
According to recent theory (Reeve, 2012), engagement should positively
support and maintain students’ motivation over time. Oga-Baldwin and Nakata (2015)
demonstrated a predictive relationship of a well-structured and supportive
environment on students’ in-class engagement in Japanese elementary classes. More
recent work has shown that engagement may positively predict more autonomous
motivation, while negatively predicting more controlled motives (Oga-Baldwin &
Nakata, in press). As a direct result of engaging with learning material, students
developed more high-quality autonomous motives. Thus, engagement may act as a
central pivot for the development of motivation (Heckhausen, 1991; Reeve & W. Lee,
2014), mediating the influence of the classroom and prior motivation on future
motivation.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
86,
1.4 Why an amalgamated model is needed
The major concepts of SDT and engagement have been studied within the realm
of general education. At the same time, few empirical studies to date have
demonstrated how elementary students develop motivation to learn a foreign language
longitudinally; prior research in Japan (Carreira et al., 2013; Nishida, 2013) has only
employed cross-sectional models. Oga-Baldwin & Nakata (in press) found a positive
relationship between engagement and different motivational regulations, but did not
measure the classroom environment or need satisfaction. Carreira and colleagues
(2013) showed the most complete cross-sectional model using CEP, BNPS, and
intrinsic motives, but did not include the full continuum of motivational regulation,
nor did they show any connection with engagement. These findings, along with others
from recent longitudinal models (Jang et al., 2012; Jang et al., 2016; Wang & Eccles,
2013) offer the foundations for an amalgamated model of motivational development
across a single school year.
As noted, the Course of Study (MEXT, 2008) makes the development of
interest, desire to learn, and positive behavior through positive learning experiences a
central policy goal; these concepts match with engagement and autonomous
motivation. Organismic integration theory defines the desired quality of motivation as
emphasized in this policy (Ryan & Deci, 2002). Basic psychological needs theory
predicts that autonomy, relatedness, and competence support autonomous motivation
and engagement (Jang et al., 2009). Cognitive evaluation theory provides a
mechanism for how students’ needs may be met by the environment (Skinner et al.,
2008). Engagement represents what students do, think, and feel in a real classroom
setting, and may help explain the development of motivation (Reeve, 2012; Reeve &
W. Lee, 2014). By amalgamating these theoretical elements, we seek to illustrate how
classrooms may help to sustain elementary school students’ language learning

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8I,
motivation (Reeve & Assor, 2011).
2. Research questions and overview
In order to answer the broader question of how teachers can build and maintain
young language learners’ motivation (Butler, 2015), we constructed a longitudinal
model of how motivation develops through the process of learning a foreign language
in a Japanese elementary school classroom. This study represents the first fully latent
longitudinal test of these theories together in a foreign language classroom setting.
While previous tests of theory have used mean-based path analyses (Carreira et al.,
2013), use of fully latent methods allow for inclusion of measurement error from
multiple indicators in the model and therefore to correct estimates for it (Kline, 2011).
All of the constructs investigated are multifaceted (Ryan & Deci, 2002; Fredricks et
al., 2004), and thus should be represented by a latent modeling framework using
multiple indicators rather than reducing the data to mean scores.
Based on the classroom motivation and engagement literature, we constructed a
hypothetical model. We sought to measure both self and environment (Ushioda, 2013).
In this model, the term “prior motivation” refers to the three different motivational
regulations (autonomous, introjected, external) as measured at the beginning of the
school year, while “outcome motivation” refers to the motivational regulation
variables at the end of the school year. We expected students to be high in intrinsic
motivation from the outset, based in the MEXT policies focused on facilitating zest
for learning. We constructed the model to answer the following research questions:
1. To what extent does students’ prior knowledge of English predict
motivation and classroom processes?
Hypothesis 1: Prior vocabulary proficiency was expected to predict all
variables in the model, reflecting the relationship between previous
ability on subsequent on academic and motivational outcomes (Hattie,
2009).
2. To what extent do prior motivations predict perceptions of teacher
support, need satisfaction, and engagement?

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8K,
Hypothesis 2: Prior motivation was expected to predict students’
perceptions of teachers’ support, need satisfaction, and engagement.
Autonomous motives were expected to show a positive effect, and
extrinsic motives expected to have a negative relationship (Cohen & Katz,
2015; Reeve, 2013).
3. To what extent do teacher support and need satisfaction predict
engagement?
Hypothesis 3: Consistent with robust findings across paradigms and
environments (Jang et al., 2012; Skinner et al., 2008; Wang & Eccles,
2013; etc.), teacher support and need satisfaction were expected to
positively predict student engagement.
4. To what extent does prior motivation predict outcome motivation?
Hypothesis 4: Prior motives were expected to predict outcome motives at
the end of the year. Each motive was expected to predict itself, but also
show cross-lagged effects. More autonomous prior motives will
negatively predict more external outcome motives, and more external
prior motives will negatively predict more autonomous outcome motives.
5. To what extent does in-class engagement predict outcome motivation?
Hypothesis 5: We expected that self-reported engagement would
positively predict outcome autonomous motivation, and negatively
predict more external regulations (Oga-Baldwin & Nakata, in press;
Carreira et al., 2013; Reeve, 2012).
6. To what extent do motivation and classroom engagement predict teacher
assessments?
Hypothesis 6: Engagement, prior motivation, and outcome motivation
were expected to predict teachers’ assessments of each individual student,
while external regulation at both times will negatively predict teacher
assessment (Jang et al., 2012).
7. To what extent is self-reported engagement visible to outside observers?
Hypothesis 7: Independent of the structural equation model, external
raters’ assessments of engagement were expected to positively correlate
at greater than .3 with students’ self-reported engagement (Butler & J.
Lee, 2006; Nave et al., 2008).,
3. Methods
3.1. Participants
The current sample came from seven schools in a suburban school district in
western Japan. Five-hundred and fifteen 5th-year students (female n = 253; age 10-11)
in 16 classes at seven schools completed surveys during the 2013-2014 school year.
Several students had absences due to illness at one point during the course of the year.
Surveys were administered at five times during the 2013 school year: once in April,
2013, once in May, once in October, once in January, 2014, and finally in March. The

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8M,
sixteen homeroom teachers attached to each class were given student assessment
surveys at the end of the school year in March.
This research was approved by the [University] Ethics Review Board.
Permission to conduct the research was also provided by the local board of education.
Participating principals volunteered to have their schools join the study, and
individual teachers were approached to obtain agreement. Participants were informed
of the scope and aims of the study before agreeing to participate with signed
permission forms. Fifth-year classes were chosen for the target population as it is the
first year targeted for foreign language study in Japanese elementary schools (MEXT,
2008).
3.2. Instruments
Survey instruments were designed to represent the specific theories describing
the interaction between the person, their actions, and the environment. The survey
instruments used, the theories they represent, and sample items are presented in Table
1.
Prior to taking surveys, students completed a 20-item vocabulary pre-test at
the beginning of April 2013. Students were asked to identify English vocabulary
words from pictures selected from the curriculum. As a small but not insignificant
number of foreign loanwords have been imported into the Japanese language, test
items were selected from English words not commonly used in daily life in Japan.
Students were asked to demonstrate word knowledge productively, and allowed to
write either the Japanese phonetic reading (katakana) or English spelling of the items.
Usage of Japanese phonetic writing and incorrect spelling were not penalized when
they indicated the correct word. Scores ranged from 0 to 20. The overall mean score

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8N,
was roughly 31%, or 6 correct items. This test was administered during the first
weeks of the semester, roughly a week before students completed their first surveys.
Table 1. Theoretical elements and their instrumentations in the model.
Theoretical element
Instrument
Example items
Self-Regulation Questionnaire–
Academic (3 factors)
Autonomous regulation (6 items)
I’m interested in English
I want to be able to use English
in the future
Introjected regulation(3 items)
I want my friends to think I’m
good at English
Organismic integration theory
External regulation (3 items)
If I don’t participate my teacher
will get angry
Cognitive evaluation theory
Teacher support scale (1 factor, 5
items)
My teacher gives clear
explanations
Basic needs theory
Activity Feeling Scales (1 factor, 9
items)
I felt I wanted to learn more
English
I felt I was working together with
my friends
I felt my English was improving
Engagement
Engagement scales (1 factor, 11
items)
I paid attention in today’s class
I was interested in today’s class
I tried to comprehend my
teacher’s English
Prior proficiency
Vocabulary Pre-test (20 items; mean
score)
Concrete words from the
curriculum: ruler, twelve,
globe
Learning outcomes
Teacher assessment (1 factor, 4
items)
Student is interested in English
as a foreign language
Student has good quality
motivation to learn a foreign
language
Student is well behaved in
English classes
Student is well behaved in
English classes
Survey instruments were based on previous research and instrument validations
conducted in the SDT paradigm (Carreira, 2012; Yamauchi & Tanaka, 1998). As
Weeks and colleagues (2007) recommend, items were tested through translation and
back translation, then focus group participants re-wrote items together in groups to
provide the most comprehensible wordings. Wordings were validated through focus
groups with elementary students and teachers at each of the seven participating
schools. These focus groups were designed to elicit the natural wordings that students
and teachers use when discussing the concepts relevant to this study (Devellis, 2012).
We conducted these groups in order to achieve the best emic representations of the
motivational constructs in question (King & McInerney, 2014) and achieve wordings

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
89,
that were most likely to be comprehensible to the larger population of students.
Groups then listened to explanations of the different theoretical factors (e.g.,
autonomy, intrinsic regulation, engagement, etc.), then sorted items into discrete
categories. Wordings and categorizations were deemed acceptable only when more
than half of the participants agreed. This a priori cutoff was used in order to achieve
the greatest consensus and maintain the minimum number of indicators necessary to
create an appropriate latent factor (i.e., 3 indicators; Kline, 2011, pp. 114–115).
The quality of students’ motivation was measured at two time points with a
twelve-item Japanese version of the self-regulation questionnaire (SRQ-A; Ryan &
Connell, 1989; Yamauchi & Tanaka, 2000; Carreira, 2012). Scales showed acceptable
internal reliabilities (> .70; Devellis, 2012; see Table 2). While studies have found
evidence for discriminant validity between intrinsic and identified regulations (Oga-
Baldwin & Nakata, in press), in order to avoid difficulties occurring when predictors
are highly correlated (> .8; Tabachnick & Fidell, 2007), these two variables were
treated as a single latent factor representing autonomous motivation. Prior studies
using the SDT framework have also employed this conceptualization of autonomous
motivation (e.g., Vansteenkiste et al., 2009). Introjected and extrinsic regulations
were treated as separate factors. Scales ranged from 1 (“<50% true for me”) to 5
(“>90% true for me”). A comparative EFA from the pilot study is presented in
Appendix 1.
Students’ perceptions of their teacher were measured using a measure of
supportive teaching (five items, Cronbach’s α = .70; Oga-Baldwin & Nakata, 2015).
Students’ autonomy, relatedness, and competence (ARC) needs were measured using
the nine-item Activity Feeling Scales (AFS; Reeve & Sickenius, 1994; Jang et al.,
2009; Cronbach’s α = .87). As noted in theoretical work (Ryan & Deci, 2002), all

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8O,
three needs are necessary to appropriate psychological functioning. Where previous
models have used means-based path analyses to show differential effects of each
factor, fully latent longitudinal models using the same scales have treated these as a
single latent variable (Jang et al., 2012; Jang et al., 2016). Scales for both supportive
teaching scales and the AFS ranged from 1 (“<50% true for class today”) to 5 (“>90%
true for class today”). Alternative constructions and the justification for the use of a
single latent variable are presented in Appendix 2.
Consistent with prior longitudinal models (Jang et al., 2012; Reeve & W. Lee,
2014), engagement was measured as a single construct, using items representing
cognitive, emotional, and behavioral engagement (eleven items, Cronbach’s α = .91).
All scales were tested previously with independent samples (Oga-Baldwin & Nakata,
2015; in press). As with other measures, this scale ranged from 1 (“<50% true for
class today”) to 5 (“>90% true for class today”).
Homeroom teachers assessed individual students’ in-class interest, behavior,
motivation, and English ability according to the goals outlined by the Ministry of
Education (MEXT, 2008). According to the national curriculum policy, summative
assessment, especially testing, is to be avoided, in part due to the potential negative
impact it may have on motivation. Due to these policy constraints, summative post-
tests were not permitted by schools or boards of education. Outcome measures instead
used teachers’ assessment of students’ language abilities and quality of motivation
(Moore, Lippman, & Ryberg, 2015). This assessment measure should be noted, as no
prior studies of elementary school motivation in Japan have used outcome measures.
Maintaining consistency with other measures, a 5-point scale was used, ranging from
1 (“50% true or less for this student”) to 5 (“90% true or greater for this student”).
This scale was chosen as an attempt to reflect a rating of degree of accuracy rather

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
8P,
than a frequency rating (i.e., “Students often do this”; “Students are rarely like this”),
the latter of which has shown weaker psychometric fit (Mizumoto & Takeuchi, 2009).
Further, ratings reflecting a more precise level of applicability may overcome
previously noted cultural tendencies toward non-committal answers (i.e., marking
toward the center of the scale; Reid, 1990). Internal reliability for this measure was
good (Cronbach’s α = .94).
In order to measure engagement externally, videos were taken of students’ in-
class performance and behavior. Two trained raters were instructed to watch the
whole class and rate activity levels for each minute of the class, leading to roughly 40
observations per class. Using a 5-point rating system, raters documented full class
engagement on a scale ranging from 1 (“all students off-topic, bored, or mindless”) to
5 (“all students working, interested, or thinking”). Raters were selected from a group
of 4th-year university teacher trainees who had completed their teaching practicum
and were preparing to enter the teaching practice in Spring of 2014. The observations
were conducted in the fall and winter of 2013-2014. Rater training was minimal to
allow for naïve assessment as might be made by non-scholarly observers, such as
parents, teachers, and administrators. Inter-rater reliability was calculated using
Pearson’s correlation coefficient, showing acceptable agreement (r = .93, p < .001).
3.3. Analyses
Latent analysis was undertaken using MPlus 7 (Muthén & Muthén, 2012) using
the weighted least squares mean and variance corrected (WLS-MV) estimator for all
structural equation models. As Likert data may be considered ordered categorizations
rather than truly continuous (Carifio & Perla, 2007), we used weighted least squares
estimation for its ability to model non-normal ordered categorical data (Muthén &
Muthén, 2012). Fit cutoffs were set at RMSEA < 0.08, CFI > 0.9, TLI > 0.9 for an

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
67,
acceptable model, with RMSEA < 0.06, CFI > 0.95, TLI > 0.95 deemed to
demonstrate good fit (Kline, 2011). No item error correlations were used.
Missing data due to student absences or non-response was 3.3% of the total
volume of data. Missing data from individuals absent at one time point or from non-
answered questions was handled in MPlus using Full-Information-Maximum-
Likelihood estimation in order to handle the 66 cases with missing data (Schafer &
Graham, 2002).
The nested nature of the data (i.e., participants nested within classes) was
accounted for using cluster-robust standard errors. For this analysis, each individual
class was treated as a cluster. Intraclass correlations ranged from .05 to .16. The ICCs
for each of the variables considered are presented in Table 2. While the intraclass
correlations for the variables were large enough to justify a multilevel approach, the
number of level 2 clusters was potentially small enough to lead to bias (< 50; Maas &
Hox, 2005) and other computational issues (Steenbergen & Jones, 2002); thus we
chose to account for potential nesting issues with cluster-robust standard errors.
Engagement had the strongest ICC, indicating the largest between cluster differences.
We treated this separately through the use of external validation of engagement in
each individual class.
To validate the engagement measures, class self-report means were correlated
with the mean score of both raters’ assessments of collective engagement. Previous
work has indicated that independent ratings of individuals exceeding .3 (as a moderate
effect size; Cohen, 1992) may be useful for explaining behaviors and attitudes (Nave,
Sherman, & Funder, 2008).

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
68,
4. Results
An initial confirmatory factor analysis indicated that all factors were adequately
represented by the indicators. Factor analysis results demonstrated good fit, χ2
(1280) = 1492.254, p < .001, RMSEA = .018 [CI = .014, .022], CFI = .97, TLI = .97.
Standardized factor coefficients ranged from .38 to .98, indicating sufficient internal
factorial validity. Items and their wordings are presented in Appendix 3. Latent
variables had low to moderate correlations (from .00 to .62). Table 2 displays the
factor correlations and descriptive statistics.
Table 2. Latent factor correlations, intraclass correlations, descriptive statistics, and internal
reliabilities
*p ≤ .05, **p ≤ .01, *** p ≤ .001.
The full model with standardized structural regression coefficients is presented
in Figure 1. The structural model fit the data well, χ2 (1329) = 1527.003, p < .001,
RMSEA = .017 [CI = .012, .021], CFI = .97, TLI= .97. The model and its components
will be explained referring to each of the original hypotheses.
Vocabulary proficiency weakly predicted autonomous motivation (
β
= .16, p
< .001) and teacher assessment (
β
= .18, p < .001). The pre-test scores negatively
Latent
ICC
1
2
3
4
5
6
7
8
9
10
1. Autonomous Reg. April
.04
-
.30***
-.56***
.27***
.41***
.32***
.57***
.02
-.48***
.29***
2. Introjected Reg. April
.03
-
.16**
.14**
.18**
.17*
.21***
.57***
.01
.07*
3. External Reg. April
.05
-
-.19***
-.16**
-.18***
-.32***
.08
.51***
-.16***
4. Supportive Teaching June
.14
-
.46***
.59***
.39***
.17***
-.27***
.17***
5. Need Satisfaction October
.08
-
.51***
.41***
.13***
-.27***
.19***
6. Engagement January
.16
-
.60***
.26***
-.38***
.25***
7. Autonomous Reg. March
.08
-
.27***
-.56***
.28***
8. Introjected Reg. March
.04
-
.28***
.05
9. External Reg. March
.07
-
-.20***
10. Teacher Assessment March
-
Mean
3.73
1.94
2.48
4.04
3.64
3.89
3.81
2.06
2.46
3.31
SD
.90
.88
1.06
.71
.78
.75
.88
.87
.96
1.02
95% CI
3.65
3.81
1.86
2.02
2.39
2.57
3.91
4.10
3.57
3.70
3.83
3.96
3.72
3.89
1.98
2.13
2.38,
2.55
3.21
3.40
Cronbach’s Alpha
.87
.72
.73
.70
.87
.91
.89
.78
.73
.94

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
66,
predicted prior introjected regulation (β = -.13, p < .001). No other significant
correlations from Hypothesis 1 were found in the model.
,
Figure 1. Final model results. Latent covariances for external, introjected, and autonomous regulations are not
displayed. Non-significant paths have been removed.
No prior motives significantly predicted supportive teaching or engagement, but
autonomous regulation significantly predicted need satisfaction (
β
= .36, p < .001).
This confirms the theoretical relationship between autonomous motivation and need
satisfaction, but no other relationships from Hypothesis 2.
The model successfully confirmed Hypothesis 3. Supportive teaching predicted
engagement (β = .44, p < .001), partially mediated by need satisfaction (
β
= .27, p
< .001). The structural model explained 44% of the variance for engagement.
Using Cohen’s (1992) criteria, one large and two medium auto-regressive
relationships was found on motivational regulations over time (
β
AUTONOMOUS-AUTONOMOUS
= .41, p < .001,
β
INTROJECTED-INTROJECTED = .67, p < .001,
β
EXTERNAL-EXTERNAL = .32, p
< .001). Autonomous motivation had a medium-sized negative predictive relationship
on introjected regulation (
β
AUTONOMOUS-INTROJECTED = -.37, p < .01), while external

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6I,
regulation had a small relationship (
β
EXTERNAL-INTROJECTED = -.25, p < .05). Autonomous
regulation negatively predicted external regulation (
β
AUTONOMOUS-EXTERNAL = -.23, p
< .01). No other relationship from Hypothesis 4 was found. Differing from previous
findings of statistically significant decrease across years in Japanese elementary
schools (Carreira, 2006; 2011), autonomous motivation showed a slight, non-
significant increase over time (April M = 3.73, March M = 3.80, t = -1.87, p = .06).
Support was also found for Hypothesis 5. Engagement showed a medium-sized
predictive effect on autonomous motives (
β
= .47, p < .001). Engagement further
showed a small positive relationship with introjection (
β
= .24, p < .01) and a small
negative relationship with external regulation (
β
= -.26, p < .001). The model
predicted roughly 52% of the variance on autonomous motivation, 42% of the
variance for introjection and 38% of the variance for external regulation.
Both engagement and prior autonomous regulation weakly predicted teachers’
final assessment of students’ in-class performance (
β
ENGAGEMENT-ASSESSMENT = .16, p
< .01;
β
AUTONOMOUS-ASSESSMENT = .17, p < .001. No other direct effects from students’
self-report data showed a significant effect on teacher assessment. The total model
accounted for roughly 15% of teachers’ assessment, including non-significant effects.
Outside of the structural equation model, observed collective engagement
showed a significant correlation with class mean self-reported engagement (r = .57, p
= .02, 95% CI = .10 ~ .83). Table 3 shows the mean collective engagement score for
each class as rated by observers and students. This correlation is consistent with other
results using external triangulation of self-reports (Butler & J. Lee, 2006; Nave et al.,

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6K,
2008), and confirmed Hypothesis 7. Students’ self-report ratings were generally
higher than the external ratings, Raters’ M = 3.36, Self-reported M = 3.95, t = -6.58, p
< .001.
Table 3. Mean external ratings and self-report ratings for engagement by class.
5. Discussion
In this study, we sought to demonstrate a model of how motivation to learn a
foreign language develops over the course of a school year in a low-stakes, activity
Rater 1
Rater 2
Self-reported Engagement
Class
Mean
(95% CI)
SD
Mean
(95% CI)
SD
Rater
Mean
Mean
(95% CI)
SD
3.77
3.64
3.48
A
(n=40)
(3.52, 4.02)
.81
(3.42, 3.86)
.70
3.71
(3.25, 3.72)
.72
2.77
2.95
3.63
B
(n=39)
(2.59, 2.95)
.57
(2.84, 3.06)
.37
2.86
(3.39, 3.87)
.67
3.58
3.63
3.53
C
(n=39)
(3.4, 3.76)
.58
(3.45, 3.81)
.58
3.61
(3.27, 3.78)
.79
3.04
3.04
3.65
D
(n=39)
(2.8, 3.28)
.77
(2.80, 3.28)
.77
3.04
(3.32, 3.98)
.97
3.42
3.64
4.15
E
(n=29)
(3.27, 3.57)
.50
(3.39, 3.89)
.81
3.53
(3.92, 4.37)
.52
3.42
3.40
3.92
F
(n=29)
(3.27, 3.57)
.50
(3.22, 3.58)
.58
3.41
(3.71, 4.12)
.48
3.25
3.42
3.88
G
(n=29)
(3.08, 3.42)
.55
(3.25, 3.59)
.55
3.34
(3.72, 4.04)
.41
3.00
3.00
3.57
H
(n=29)
(2.87, 3.13)
.41
(2.88, 3.12)
.38
3.00
(3.24, 3.90)
.83
2.64
2.86
3.74
I
(n=33)
(2.49, 2.79)
.49
(2.75, 2.97)
.35
2.75
(3.53, 3.94)
.52
2.67
3.08
3.66
.76
J
(n=36)
(2.45, 2.89)
.70
(2.99, 3.17)
.28
2.88
(3.39, 3.92)
3.53
3.39
4.45
K
(n=21)
(3.35, 3.71)
.57
(3.22, 3.56)
.56
3.46
(4.19, 4.71)
.57
4.00
3.59
4.22
L
(n=34)
(3.77, 4.23)
.73
(3.37, 3.81)
.72
3.80
(4.00, 4.43)
.57
4.30
4.20
4.51
M
(n=34)
(4.01, 4.59)
.95
(3.91, 4.49)
.92
4.25
(4.27, 4.74)
.65
3.26
3.17
4.20
N
(n=33)
(3.01, 3.51)
.80
(2.92, 3.42)
.80
3.22
(3.91, 4.46)
.78
3.37
3.37
4.25
O
(n=25)
(3.13, 3.61)
.79
(3.17, 3.57)
.66
3.37
(4.03, 4.48)
.49
3.77
3.64
4.43
P
(n=25)
(3.52, 4.02)
.81
(3.42, 3.86)
.70
3.71
(3.96, 4.30)
.53

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6M,
oriented learning environment (MEXT, 2008). Consistent with our hypotheses and
self-determination theory (Ryan & Deci, 2002), our results show that motivation
develops in this context at the intersection of the classroom and individual.
Specifically, the quality of students’ motivation at the end of the school year develops
in relation to both their prior motives and their learning experiences. Further, by
engaging students in learning tasks through appropriate support for their needs,
teachers can help to build a sense of autonomous motivation at the end of the school
year. These findings support previous models of motivational development (Carreira
et al., 2013; Jang et al., 2012; Skinner et al., 2008), and help to draw connections
between them. The model provides empirical evidence for Reeve’s (2012) hypothesis
that engagement, developed through a need-supportive environment, may help to
support and maintain autonomous motivation.
5.1 Research questions
RQ 1: To what extent does students’ prior knowledge of English predict motivation
and classroom processes?
Students’ prior vocabulary knowledge showed limited effects on the overall
model. Vocabulary test scores weakly predicted autonomous regulation, and
negatively predicted introjected regulation. This is consistent with SDT, as students
with greater knowledge would likely feel a greater sense of competence (one of the
basic needs) and would be less likely to feel the need to demonstrate their knowledge
to avoid threats to their ego (Ryan & Deci, 2002); they know they have the
knowledge, and want to use it. Teachers showed some recognition of their vocabulary
knowledge as well in their assessments. This may reflect the curriculum in elementary

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6N,
schools, which is primarily focused on teaching new words and phrases through
games (MEXT, 2008).
RQ 2: To what extent do prior motivations predict perceptions of teacher support,
need satisfaction, and engagement?
Students’ prior autonomous motivation predicted their sense of need satisfaction.
Consistent with research showing that more internally regulated students may have a
more positive picture of teachers’ support (Cohen & Katz, 2015; McEown et al.,
2014), more autonomously motivated students were more likely to feel their needs
were being met in class, while lower quality motives (i.e., introjected and external
regulations) showed no significant effects. Students’ existing motivations when they
come into a classroom setting thus may predict their perceptions of what they
experience in the form of their need satisfaction. At the same time, no significant
direct effects from motivation were found on teacher support or engagement. More
autonomously motivated students did not perceive their teachers as more supportive,
indicating that the quality of the instruction itself was independent of students’
perceptions. We interpret the lack of a direct effect between motivation and
engagement to mean that engagement here is situational, and thus more strongly
connected with the classroom environment than students’ existing motivation. Zero-
order correlations corroborate this interpretation. Engagement correlated with the
prior motivation variables, but betas became non-significant after controlling for the
classroom variables.
RQ 3: To what extent do teacher support and need satisfaction predict engagement?
Both supportive teaching and need satisfaction predicted positive engagement,
supporting Hypothesis 3. Teaching showed a slightly stronger relationship, indicating

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
69,
the importance of a supportive classroom environment. By providing a need
satisfying environment, teachers simultaneously helped students to engage in positive
ways. Results are consistent with other SDT research (Jang et al., 2009; Skinner et al.,
2008), further showing that school environments do not always damage students’
motivation, but can also be need supportive (Reeve & Assor, 2011), even in
hierarchically oriented societies like Japan.
While prior autonomous regulation predicted need satisfaction, no prior
motivational factors showed any direct relationship with engagement, indicating that
engagement is tied more strongly to the quality of the classroom environment than
students’ existing motivation. The relatively large effect sizes and coefficients of
determination (R2) values for both need satisfaction and engagement further indicate
the predictive value of a supportive classroom.
RQ 4: To what extent does prior motivation predict outcome motivation?
In support of Hypothesis 4, learners’ motivation showed a range of
autocorrelations over time, showing one strong effect, and two medium sized effects
(Cohen, 1992). Medium sized autolagged effects were found for autonomous and
external regulations, and a strong effect was found for introjected. Results show that
students’ prior motivations predict themselves, indicating how students’ reasons for
studying at the beginning of their studies may affect motivation over the course of the
school year. While the longitudinal correlations between the different regulations
were comparatively weak, they were consistent with organismic integration theory,
showing negative correlations between more and less autonomous motives. Unlike in
previous cross-sectional studies (Carreira, 2006; 2011), students in this sample did not
show decreases in the quality of their motivation; intrinsic motivation remained

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6O,
relatively stable. In light of the previously noted negative trend, this may offer some
hope for foreign language education in the Japanese context.
RQ 5: To what extent does in-class engagement predict outcome motivation?
Engagement, influenced by the classroom environment, had a direct predictive
effect on students’ motivational orientations at the end of the year. This confirms
Hypothesis 5. Based on these results, a crucial step in the process of supporting
students’ long-term motivation is providing an engaging learning environment, as
hypothesized by Reeve (2012). While engagement alone did not predict motivation,
the strong path from engagement to autonomous motivation corroborates previous
findings of the relationship between engagement and motivation (Oga-Baldwin &
Nakata, in press; Reeve & W. Lee, 2014). Thus, while previous models have found
evidence for a direct effect from supportive teaching and need satisfaction to
autonomous motivation (Carreira et al., 2013), the current model suggests that
engagement is an important mediating element in the teaching and learning process.
RQ 6: To what extent do motivation and classroom engagement predict teacher
assessments?
Teachers were somewhat able to understand students’ self-reported motivation
and engagement in order to assess their ability, interest, and behavior. The effects for
each were small, though this is consistent with other findings on the relationship
between engagement, motivation, and assessment in other longitudinal studies of
engagement (Jang et al., 2012; Reeve & W. Lee, 2014). This result confirms
Hypothesis 6, but also differs from the results of Lee and Reeve (2012), who found
self-reported engagement, but not motivation, to be salient and recognizable to
teachers. We interpret the significant relationship between motivation and assessment

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
6P,
to mean that prior motivation may be reflected indirectly throughout the school year,
and teachers may come to recognize this in their subjective assessments.
RQ 7: To what extent is self-reported engagement visible to outside observers?
External raters’ assessments of students’ engagement broadly agreed with self-
reports. The correlation found here between students’ self-report and raters’
assessments (r = .57) is somewhat stronger than those in studies using independent
rating of personality traits (e.g., Nave et al., 2008, etc.), and is consistent with other
correlations found between self-reports and external assessments of on-task behavior
in language learning (Butler & J. Lee, 2006). Engagement is thus visible to outsiders,
providing further validation for the self-report model.
5.2 Implications
This model illustrates how learners may be influenced by their past motives,
while also demonstrating the effects of teachers’ support on students motivation.
While existing motivation significantly, sometimes strongly, predicted itself within
the model, need satisfaction and engagement also had dynamic effects over time.
Motivation showed no direct effects on engagement; teachers’ support did.
Engagement thus functions as the central pivot point for much of the model,
predicting both outcome motivation and teachers’ evaluations. This indicates that
what teachers do and say can have a positive influence on students’ motivation over
time.
Results showed a generally positive trend with regard to the motivational
aspects of the classrooms surveyed. While students’ motivation did not increase
during the year, this may be taken as a positive sign for these school environs; prior

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
I7,
findings have indicated a decrease as students grow older (cf. Carreira, 2011). This
runs counter to the argument that decreases in intrinsic motivation are developmental
(Carreira, 2006), and instead may indicate that these decreases have some relationship
to the teaching and learning environment. This represents how engaging, supportive,
and need-satisfying instruction in a low-pressure, low-stakes environment may
positively influence students’ motivation (Ryan & Niemiec, 2009).
The strong predictive value of prior motivation on outcome motivation is
further evidence for the idea of a stable core of motivation found in previous research
(Nakata, 2006), as well as the presence of a sense of autonomous motivation to learn
beyond simply enjoying English. While motivation may ebb, flow, and change over
time (Dörnyei, Ibrahim, & Muir, 2015), findings here indicate that these changes in
motivation may also center around more stable beliefs that influence students’
behavior indirectly over the course of an academic year.
5.3 Limitations and future directions
While results indicate the relative stability of motivation and the importance of
teachers’ support, how students change as individuals and the minutiae of how
teachers may best provide that support is beyond the scope of the current investigation.
The results display variable-centered but not person-centered statistics. Further, no
qualitative observations of these classes have been discussed. Future studies on this
topic will need to make use of qualitative observations and person-centered analyses.
Qualitative observations will allow a finer discussion of the principles and practices
of the most successful teachers in this cohort, while person-centered analyses may
show the trajectories of the individual students. Testing motivational development
from a person-centered perspective represents a critical direction for further research.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
I8,
Continuing the call from Butler (2015), providing more detailed context to the
learning environment and the changes within the individuals may offer further insight
into how young language learners develop autonomous motivation.
Additionally, this study focused on Japanese elementary classrooms that are
embracing the MEXT (2008) approach that emphasizes variables known to support
autonomy and intrinsic motivation. Thus there was a high level of autonomous
motivation and engagement in evidence. The findings should be generalized with
caution in other L2 settings.
6. Conclusions
The current study details the first study detailing the development of Japanese
elementary school students’ motivation to learn a foreign language using external
assessment under the current Course of Study. This research further represents one of
the first truly longitudinal investigations of the cognitive evaluation, basic needs, and
organismic integration minitheories from self-determination theory. The results
demonstrate the coherence of the three theories, validated with external measures.
Where previous measures of both elementary school language learning (Nishida,
2013) and motivational development (Jang et al., 2012) have modeled for time
periods within a single school term, the current study measured the development of
language learning motivation over the course of an entire school year.
These findings demonstrate how teachers may support students’ motivation
over the course of a school year by providing an engaging classroom experience.
Teachers who provide appropriate need support and structure to their foreign
language lessons help students to engage, which predicts student motivation at the end
of the year. For teachers, this would imply that creating a clear, interesting, and well-

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
I6,
paced learning environment is centrally important for foreign language learners. In
this environment, students feel connected to their peers, capable of the tasks, and
personally invested in their learning, and thus engage in the learning activities.
Engagement in turn leads to more positive assessments from teachers. This rich and
enjoyable language learning environment appears to alleviate the previously noted
motivational declines that occur across school years, thus offering some hope for
teachers wishing to develop motivation, positive affect, and a sense of growth and
discovery for elementary school students learning a foreign language.
7. Compliance with Ethical Standards
Funding: This study was funded by a JSPS KAKENHI grant-in-aid for young
scientists (B) [Funding ID number removed for peer review]
Ethical approval: All procedures performed in studies involving human participants
were in accordance with the ethical standards of the institutional and/or national
research committee and with the 1964 Helsinki declaration and its later amendments
or comparable ethical standards.
Informed consent: Informed consent was obtained from all individual participants
included in the study as well as their legal guardians.
8. References
Ando, F., Fuse, M., & Kodaira, H. (2008). Motivation and elementary school pupils'
positive class participation: Self-determination theory. Japanese Journal of
Educational Psychology, 56(2), 160–170.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
II,
Baumeister, R., & Leary, M. R. (1995). The need to belong: Desire for interpersonal
attachments as a fundamental human motivation. Psychological Bulletin, 117(3),
497–529.
Berwick, R., & Ross, S. (1989). Motivation after matriculation: Are Japanese learners
of English still alive after exam hell? JALT Journal, 11(2), 193–210.
Butler, Y. G. (2014). Parental factors in children’s motivation for learning English: A
case in China. Research Papers in Education, 1–28.
doi:10.1080/02671522.2014.891643
Butler, Y. G. (2015). English language education among young learners in East Asia:
A review of current research (2004–2014). Language Teaching, 48(03), 303–
342. doi:10.1017/S0261444815000105
Butler, Y. G., & Lee, J. (2006). On-task versus off-task self-assessments among
Korean elementary school students studying English. The Modern Language
Journal, 90(4), 506–518. doi:10.1111/j.1540-4781.2006.00463.x
Carifio, J., & Perla, R. J. (2007). Ten common misunderstandings, misconceptions,
persistent myths and urban legends about Likert scales and Likert response
formats and their antidotes. Journal of Social Sciences, 3(3), 106–116.
Carreira, J. M. (2006). Motivation for learning English as a foreign language in
Japanese elementary schools. JALT Journal, 28(2), 135–158.
Carreira, J. M. (2011). Relationship between motivation for learning EFL and
intrinsic motivation for learning in general among Japanese elementary school
students. System, 39(1), 90–102. doi:10.1016/j.system.2011.01.009
Carreira, J. M. (2012). Motivational orienations and psychological needs in EFL
learning among elementary school students in Japan. System, 40(2), 191–202.
doi:10.1016/j.system.2012.02.001

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
IK,
Carreira, J. M., Ozaki, K., & Maeda, T. (2013). Motivational model of English
learning among elementary school students in Japan. System, 41(3), 706–719.
doi:10.1016/j.system.2013.07.017
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159.
doi:10.1037/0033-2909.112.1.155.
Cohen, R., & Katz, I. (2015, August). When I feel good, I see good: Does students'
competence affect the way they perceive their teacher? Paper presented at the
8th SELF Biennial International Conference, Kiel, Germany.
deCharms, R. (1968). Personal causation: The internal affective determinants of
behavior. Academic Press, New York.
Csizér, K., & Dörnyei, Z. (2005). Language learners’ motivational profiles and their
motivated learning behavior. Language Learning, 55(4), 613–659. doi:
10.1111/j.0023-8333.2005.00319.x
Deci, E. L., & Ryan, R.M. (1985). Intrinsic motivation and self-determination in
human behavior. NY: Plenum.
Deci, E. L., & Ryan, R. M. (2000). The “what” and ‘why’ of goal pursuits: Human
needs and the self-determination of behavior. Psychological Inquiry, 11(4),
227–268. doi:10.1207/S15327965PLI1104_01
Deci, E. L., Koestner, R., & Ryan, R. M. (2001). Extrinsic rewards and intrinsic
motivation in education: Reconsidered once again. Review of Educational
Research, 71(1), 1–27. doi:10.3102/00346543071001001
Devellis, R. F. (2012). Scale Development: Theory and Application. (3rd ed.)
Thousand Oaks: Sage.
Dörnyei, Z., Ibrahim, Z., & Muir, C. (2015). 'Directed motivational currents':
Regulating complex dynamic systems through motivational surges. In Z.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
IM,
Dörnyei, P. MacIntyre, & A. Henry, (Eds.), Motivational dynamics in language
learning (pp. 95–105). Bristol: Multilingual Matters.
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
Garton, S., Copland, F., & Burns, A. (2011). Investigating global practices in
teaching english to young learners. British Council ELT Research Paper 11–01.
London: British Council.
Hattie, J. (2009). Visible learning. New York: Routledge.
Heckhausen, H., (1991). Motivation and action. Springer-Verlag: Berlin.
Jang, H., Kim, E. J., & Reeve, J. (2012). Longitudinal test of self-determination
theory's motivation mediation model in a naturally occurring classroom context.
Journal of Educational Psychology, 104(4), 1175–1188. doi:10.1037/a0028089
Jang, H., Kim, E. J., & Reeve, J. (2016). Why students become more engaged or more
disengaged during the semester: A self-determination theory dual-process
model. Learning and Instruction, 1–12. doi:10.1016/j.learninstruc.2016.01.002
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, 103(2), 588–600. doi:10.1037/a0019682
Jang, H., Reeve, J., Ryan, R. M., & Kim, A. (2009). Can self-determination theory
explain what underlies the productive, satisfying learning experiences of
collectivistically oriented Korean students? Journal of Educational Psychology,
101(3), 644–661. doi:10.1037/a0014241
Kim, T.-Y. (2011). Korean elementary school students’ English learning
demotivation: a comparative survey study. Asia Pacific Education Review,

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
IN,
12(1), 1–11. doi:10.1007/s12564-010-9113-1
Kim, T.-Y., & Seo, H.-S. (2012). Elementary school students' foreign language
learning demotivation: A mixed methods study of Korean EFL context. The
Asia-Pacific Education Researcher, 21(1), 160–171.
King, R. B., & McInerney, D. M. (2014). Culture's consequences on student
motivation: Capturing cross-cultural universality and variability through
personal investment theory. Educational Psychologist, 49(3), 175–198.
doi:10.1080/00461520.2014.926813
Kline, R. B. (2011). Principles and practice of structural equation modeling (3rd ed.).
New York: Guilford Press.
Lee, W., & Reeve, J. (2012). Teachers’ estimates of their students’ motivation and
engagement: Being in synch with students. Educational Psychology, 32(6),
727–747. doi:10.1080/01443410.2012.732385
Maas, C. J. M., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling.
Methodology, 1(3), 86–92. doi:10.1027/1614-1881.1.3.86
McEown, M. S., Noels, K. A., & Saumure, K. D. (2014). Self-determined and
integrative orientations and teachers' motivational support in a Japanese as a
foreign language context. System, 1–15. doi:10.1016/j.system.2014.06.001
MEXT. (2008). Shogakkou gakushuu shidou youryou kaisetu: Gaikokugo
katsudouhen [Explanatory commentary for the elementary school curriculum
guidelines: Foreign language activities]. Retrieved 8/26/10 from
http://www.mext.go.jp/component/a_menu/education/micro_detail/__icsFiles/af
ieldfile/2009/06/16/1234931_012.pdf
Mizumoto, A., & Takeuchi, O. (2009). Comparing frequency and trueness scale
descriptors in a Likert scale questionnaire on language learning strategies. JLTA

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
I9,
Journal, 12, 116–136.
Moore, K. A., Lippman, L. H., & Ryberg, R. (2015). Improving outcome measures
other than achievement. AERA Open, 1(2). doi:10.1177/2332858415579676
Muthén, L. K., & Muthén, B. O. (2012). Mplus version 7. Los Angeles, CA: Muthén
& Muthén.
Nakata. (2006). Motivation and experience in foreign language learning. Bern: Peter
Lang.
Nave, C. S., Sherman, R. A., & Funder, D. C. (2008). Beyond self-report in the study
of hedonic and eudaimonic well-being: Correlations with acquaintance reports,
clinician judgments and directly observed social behavior. Journal of Research
in Personality, 42(3), 643–659. doi:10.1016/j.jrp.2007.09.001
Nishida, R. (2013). Empirical studies of affective variables and motivational changes
among Japanese elementary school EFL learners. Tokyo: Kinseido.
Noels, K. A. (2013). Promoting the motivation of Japanese learners of English
through autonomy, competence, and relatedness. In M. T. Apple, D. Da Silva,
& T. Fellner (Eds.) Language learning motivation in Japan (pp.15–34). Bristol,
UK: Multilingual Matters.
Noels, K. A., Clément, R., & Pelletier, L. G. (1999). Perceptions of teachers’
communicative style and students’ intrinsic and extrinsic motivation. The
Modern Language Journal, 83(1), 23–34.
Noels, K. A., Clément, R., & Pelletier, L. G. (2001). Intrinsic, extrinsic, and
integrative orientations of French Canadian leamers of English. Canadian
Modern Language Review, 57(3), 424–442.
Noels, K. A., Pelletier, L. G., Clément, R., & Vallerand, R. J. (2000). Why are you
learning a second language? Motivational orientations and self-determination

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
IO,
theory. Language Learning, 50(1), 57–85.
Oga-Baldwin, W. L. Q., & Nakata, Y. (2014). Supplementing the elementary foreign
language course of study with a self-determination framework. International
Journal of Curriculum Development and Practice, 16(1), 13–26.
Oga-Baldwin, W. L. Q., & Nakata, Y. (2015). Structure also supports autonomy:
Measuring and defining autonomy-supportive teaching in Japanese elementary
foreign language classes. Japanese Psychological Research, 57(3), 167–179.
doi:10.1111/jpr.12077
Oga-Baldwin, W. L. Q. & Nakata, Y. (in press). Engagement, gender, and motivation:
A predictive model for Japanese young language learners. System.
Reeve, J. (2012). A self-determination theory perspective on student engagement. In S.
L. Christenson, A. L. Reschly, & C. Wylie, Handbook of Research on Student
Engagement (pp. 149–172). Boston, MA: Springer US. doi:10.1007/978-1-
4614-2018-7_7
Reeve, J., & Assor, A. (2011). Do social institutions necessarily suppress individuals’
need for autonomy? The possibility of schools as autonomy-promoting contexts
across the globe. In V.I. Chirkov, R.M. Ryan, & K. M. Sheldon (Eds.), Human
autonomy in cross-cultural contexts (pp. 111–132). Dordrecht: Springer
Netherlands. doi: 10.1007/978-90-481-9667-8_6
Reeve, J., & Lee, W. (2014). Students’ classroom engagement produces longitudinal
changes in classroom motivation. Journal of Educational Psychology, 106(2),
527–540. doi:10.1037/a0034934
Reid, J. (1990). The dirty laundry of ESL survey research. TESOL Quarterly, 24(2),
323–338. doi:10.2307/3586913
Ryan, R. M. (1982). Control and information in the intrapersonal sphere: An

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
IP,
extension of cognitive evaluation theory. Journal of Personality and Social
Psychology, 43, 450–461.
Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization:
Examining reasons for acting in two domains. Journal of Personality and Social
Psychology, 57, 749–761.
Ryan, R. M., & Deci, E. L. (2002). An overview of self-determination theory: An
organismic-dialectical perspective. In E. L. Deci & R. M. Ryan (Eds.),
Handbook of self-determination research (pp. 3–33). Rochester, NY: University
of Rochester Press.
Ryan, R. M., & Niemiec, C. P. (2009). Self-determination theory in schools of
education: Can an empirically supported framework also be critical and
liberating? Theory and Research in Education, 7(2), 263–272.
doi:10.1177/1477878509104331
Ryan, R. M., Deci, E. L., & Vansteenkiste, M. (2016). Autonomy and autonomy
disturbances in self-development and psychopathology: Research on motivation,
attachment, and clinical process. Developmental Psychopathology,
doi:10.1002/9781119125556.devpsy109
Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.
Psychological Methods, 7(2), 147–177. doi:10.1037//1082-989X.7.2.147
Senécal, C., Julien, E., & Guay, F. (2003). Role conflict and academic
procrastination: A self-determination perspective. European Journal of Social
Psychology, 33(2), 135–145.
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,

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
K7,
79(1), 57–68. doi:10.1348/000709908X304398
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
Skinner, E. A., Furrer, C. J., 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
Soenens, B., & Vansteenkiste, M. (2005). Antecedents and outcomes of self-
determination in 3 life domains: The role of parents“ and teachers” autonomy
support. Journal of Youth and Adolescence, 34(6), 589–604.
doi:10.1007/s10964-005-8948-y
StataCorp. (2013). Stata statistical software: Release 13. College Station, TX:
StataCorp LP.
Steenbergen, M. R., & Jones, B. J. (2002). Multilevel modeling data structures.
American Journal of Political Science, 46(1), 218–237.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.).
Needham Heights, MA: Allyn & Bacon.
Ushioda, E. (2013) Motivation: L2 learning as a special case? In S. Mercer, S. Ryan,
& M. Williams, (Eds.) Psychology for language learning: Insights from
research, theory and practice (pp. 58–73). Basingstoke: Palgrave Macmillan.
Vansteenkiste, M., Sierens, E., Soenens, B., Luyckx, K., & Lens, W. (2009).
Motivational profiles from a self-determination perspective: The quality of
motivation matters. Journal of Educational Psychology, 101(3), 671–688.
Doi:10.1037/a0015083

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
K8,
Vansteenkiste, M., Zhou, M., Lens, W. & Soenens, B. (2005). Experiences of
Autonomy and Control Among Chinese Learners: Vitalizing or Immobilizing?
Journal of Educational Psychology, 97(3), 468–483 doi:10.1037/0022-
0663.97.3.468
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(C), 12–23.
doi:10.1016/j.learninstruc.2013.04.002
Weeks, A., Swerissen, H., & Belfrage, J. (2007). Issues, challenges, and solutions in
translating study instruments. Evaluation Review, 31(2), 153–165.
doi:10.1177/0193841X06294184
White, R. W. (1959). Motivation reconsidered: The concept of competence.
Psychological Review, 66(5), 297–333.
Yamauchi, H., & Tanaka, K. (1998). Relations of autonomy, self-referenced beliefs,
and self-regulated learning among Japanese children. Psychological Reports, 82,
803–816. doi:10.2466/PR0.82.3.803-816

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
K6,
Appendix 1
Organismic Integration Theory Variables
An exploratory factor analysis using MPlus was conducted on pilot data taken from
an independent sample (n = 470) in April 2012 prior to implementation of this study.
A three-factor solution best represented the data.
Model fit comparisons
df
χ2
CFI
RMSEA
AIC
1 Factor Model
54
754.446
.67
.18 (.16, 18)
14188.688
2 Factor Model
43
204.510
.92
.09 (.08, .11)
13660.752
3 Factor Model
33
77.453
.98
.06 (.04, .07)
13553.696
4 Factor Model
No convergence
Table of factor loadings, with correlation matrix and descriptive statistics
Item
Factor 1
Factor 2
Factor 3
Autonomous 1
.67*
.19*
-.01
Autonomous 2
.67*
.24*
-.01
Autonomous 3
.65*
.25*
-.01
Autonomous 4
.86*
.01
.30*
Autonomous 5
1.00*
-.02
.42*
Autonomous 6
.99*
.01
.35*
Introjected 1
-.00
.64*
.17
Introjected 2
-.06
.80*
.00
Introjected 3
.02
.70*
.06
External 1
-.27*
.05
.48*
External 2
-.01
-.08
.76*
External 3
-.29*
.00
.50*
Factor 1
-
Factor 2
-.06
-
Factor 3
-.62*
.36*
-
Mean (SD)
3.88 (.91)
1.87 (.88)
2.17 (.96)
Cronbach’s α
.88
.79
.75
*p < .05

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
KI,
Appendix%2%
Basic Psychological Needs Variables: Alternative Model Tests
Alternate models using each of the individual Basic Psychological Needs (Autonomy,
Relatedness, and Competence) were tested. Each individual need was first tested,
showing effects very similar to the model selected in Figure 1 of the main document.
We then tested the needs together. Results showed effects from multicollinearity,
represented by inflated standard errors in each of the betas, as well as evidence of
suppression (negative betas despite positive zero-order correlations). Based on the
conceptualization that all three basic needs are necessary for psychological
functioning (Ryan & Deci, 2002), as well as the evidence that within this sample, they
function in similar psychometric fashion, we argue that they should be used together
as a single factor in the full longitudinal model.
Further evident in the model, autonomy showed the strongest mediating effect
between supportive teaching and engagement. This would further indicate that even
within a supposedly collectivist society, spending time satisfying students’ autonomy,
as opposed to competence and relatedness, may have the most positive benefits (cf.
Jang et al., 2009). At the same time, the high correlations between the factors would
indicate that they are likely difficult to separate. While removing theoretical factors
may show a similar pattern of effects, the differences between the effects in these
models and the complete model presented in the main text would indicate that the
three needs should continue to be measured together.
Model of supportive teaching on engagement, mediated by autonomy

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
KK,
Model of supportive teaching on engagement, mediated by relatedness
Model of supportive teaching on engagement, mediated by competence
Model of supportive teaching on engagement, mediated by the three basic needs
individually.

!"#$%#&'()*+,-#.#/#+,0#1.21+,3,45#*+,6789, :;*/2<=;1#15,>'?@#/);*#&,0A5@B;&;"5,
';)C,10.1016/j.cedpsych.2017.01.010,
,
KM,
Appendix%3%
Factor loading coefficients and item wordings for each indicator
Factor
Item wording (English translation)
Coefficient
(Pre) Autonomous 1
English is fun
.76
(Pre) Autonomous 2
I’m interested in English
.82
(Pre) Autonomous 3
English has value
.74
(Pre) Autonomous 4
English will help me in other parts of my life
.72
(Pre) Autonomous 5
I want to be able to use English in the future
.82
(Pre) Autonomous 6
English will help me grow
.77
(Pre) Introjected 1
I want my teacher to like me
.63
(Pre) Introjected 2
I want other people to praise me
.77
(Pre) Introjected 3
I want my friends to think I’m good at English
.83
(Pre) Extrinsic 1
If I don’t participate my teacher will get angry
.89
(Pre) Extrinsic 2
Participating is one of the rules
.50
(Pre) Extrinsic 3
I have no other choice
.83
Supportive Teaching 1
My teacher gives clear explanations
.76
Supportive Teaching 2
The pace of class is appropriate
.69
Supportive Teaching 3
My teacher directs me as to what to do
.71
Supportive Teaching 4
My teacher speaks a lot of English
.36
Supportive Teaching 5
My teacher appeals to my interests
.60
Needs 1 (Autonomy)
I felt I chose what I wanted to do
.63
Needs 2 (Autonomy)
I felt I wanted to learn more English
.74
Needs 3 (Autonomy)
I felt I was learning for myself
.66
Needs 4 (Relatedness)
I felt good working with my friends
.72
Needs 5 (Relatedness)
I felt I was working together with my friends
.74
Needs 6 (Relatedness)
I felt closer to my friends
.69
Needs 7 (Competence)
I felt confident in my English
.75
Needs 8 (Competence)
I felt my English was improving
.79
Needs 9 (Competence)
I felt I could speak English
.63
Engagement 1 (Behavioral)
I participated in today’s English class
.67
Engagement 2 (Behavioral)
I worked on today’s activities until they were complete
.72
Engagement 3 (Behavioral)
I paid attention in today’s class
.74
Engagement 4 (Emotional)
I had fun in class
.80
Engagement 5 (Emotional)
I felt good today
.82
Engagement 6 (Emotional)
I was interested in today’s class
.85
Engagement 7 (Emotional)
I enjoyed learning new things
.75
Engagement 8 (Cognitive)
I tried to express myself in English
.69
Engagement 9 (Cognitive)
I tried to understand my partners
.76
Engagement 10 (Cognitive)
I worked hard to make myself understood in English
.76
Engagement 11 (Cognitive)
I tried to comprehend my teacher’s English
.76
(Post) Autonomous 1
English is fun
.78
(Post) Autonomous 2
I’m interested in English
.82
(Post) Autonomous 3
English has value
.81
(Post) Autonomous 4
English will help me in other parts of my life
.76
(Post) Autonomous 5
I want to be able to use English in the future
.82
(Post) Autonomous 6
English will help me grow
.82
(Post) Introjected 1
I want my teacher to like me
.74
(Post) Introjected 2
I want other people to praise me
.82
(Post) Introjected 3
I want my friends to think I’m good at English
.82
(Post) Extrinsic 1
If I don’t my teacher will get angry
.79
(Post) Extrinsic 2
Participating is one of the rules
.60
(Post) Extrinsic 3
I have no other choice
.82
Teacher Assessment 1
Student is interested in English as a foreign language
.98
Teacher Assessment 2
Student has good quality motivation to learn a foreign language
.96
Teacher Assessment 3
Student is well behaved in English classes
.90
Teacher Assessment 4
Student has good English communication abilities
.85