Access to this full-text is provided by Frontiers.
Content available from Frontiers in Psychology
This content is subject to copyright.
Frontiers in Psychology 01 frontiersin.org
Disentangling motivation and
engagement: Exploring the role
of eort in promoting greater
conceptual and methodological
clarity
Robin P. Nagy , Andrew J. Martin * and Rebecca J. Collie
School of Education, University of New South Wales, Sydney, NSW, Australia
Conflation over motivation and engagement has historically impeded research
and practice. One reason for this is because definition and measurement have
often been too general or diuse—especially in the case of engagement.
Recently conceptual advances aimed at disentangling facets of engagement
and motivation have highlighted a need for better psychometric precision—
particularly in the case of engagement. To the extent that engagement is
inadequately assessed, motivation research involving engagement continues to
behampered. The present study investigates multidimensional eort (a specific
facet of engagement) and how it relates to motivation. In particular, weexamine
the associations between specific positive and negative motivation factors and
dimensions of eort, thereby shedding further insight into how dierent types of
motivation interplay with dierent types of engagement. Drawing on data from
a sample of 946 Australian high school students in 59 mathematics classrooms
at five schools, this study hypothesized a tripartite model of academic eort
in terms of operative, cognitive, and social–emotional dimensions. A novel
nine-item self-report Eort Scale measuring each of the three factors was
developed and tested for internal and external validity—including its relationship
with multidimensional motivation. Multilevel confirmatory factor analyses
were conducted to test the factor structure and validity of multidimensional
eort. Additionally, doubly-latent multilevel structural equation models were
conducted to explore the hypothesized motivation → engagement (eort)
process, and the role of student- and classroom-level background attributes as
predictors of both motivation and eort. Results supported the hypothesized
model of tripartite eort and its distinctiveness from motivation, and showed
that key dimensions of motivation predicted eort at student- and classroom-
levels. This study provides implications and suggestions for future motivation
research and theorizing by (1) establishing evidence for the validity of a
novel engagement framework (multidimensional eort), and (2) supporting
future measurement and practice in academic engagement juxtaposed with
multidimensional motivation—critical for better understanding engagement,
and motivation itself.
KEYWORDS
motivation, eort, engagement, academic development, validity, multilevel
TYPE Original Research
PUBLISHED 13 December 2022
DOI 10.3389/fpsyg.2022.1045717
OPEN ACCESS
EDITED BY
Frédéric Guay,
Laval University,
Canada
REVIEWED BY
Eduardo Aguirre-Dávila,
National University of Colombia, Colombia
Lizeth Guadalupe Parra-Pérez,
Instituto Tecnológico de Sonora (ITSON),
Mexico
*CORRESPONDENCE
Andrew J. Martin
andrew.martin@unsw.edu.au
SPECIALTY SECTION
This article was submitted to
Educational Psychology,
a section of the journal
Frontiers in Psychology
RECEIVED 15 September 2022
ACCEPTED 24 November 2022
PUBLISHED 13 December 2022
CITATION
Nagy RP, Martin AJ and Collie RJ (2022)
Disentangling motivation and engagement:
Exploring the role of eort in promoting
greater conceptual and methodological
clarity.
Front. Psychol. 13:1045717.
doi: 10.3389/fpsyg.2022.1045717
COPYRIGHT
© 2022 Nagy, Martin and Collie. This is an
open-access article distributed under the
terms of the Creative Commons Attribution
License (CC BY). The use, distribution or
reproduction in other forums is permitted,
provided the original author(s) and the
copyright owner(s) are credited and that
the original publication in this journal is
cited, in accordance with accepted
academic practice. No use, distribution or
reproduction is permitted which does not
comply with these terms.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 02 frontiersin.org
Introduction
Motivation and engagement are two intertwined constructs
that have a history of conation by researchers and practitioners.
is has at times impeded advances in theoretical clarity, research,
and practice relevant to both constructs (Reschly and Christenson,
2012). For example, it has been highlighted that inappropriately
conating motivation and engagement can create theoretical
ambiguity, introduce validity challenges for measurement and
research, and lay a shaky foundation for educational intervention
(Martin, 2012; Reschly and Christenson, 2012; Martin etal.,
2017). In recent years, much theorizing and research has been
conducted into the multidimensionality of motivation (e.g.,
Martin, 1999–2022) and engagement (e.g., Fredricks etal., 2004).
However, the basic demarcation of motivation as intent, and
engagement as action, has thus far limited a more nuanced
understanding of unique associations between their various
dimensions, particularly with respect to non-observable
dimensions of engagement. Whereas reliable scales have been
developed and extensively tested for the measurement of
motivation’s key dimensions, there has been much less focus on
theoretically-informed measurement of multidimensional
engagement, especially its internal aspects. By harnessing such a
measurement scale, motivation and engagement can befurther
disentangled by examining the relation between adaptive and
maladaptive motivation factors and specic dimensions of eort,
thereby shedding further insight into the interplay of dierent
motivation and engagement types.
Eort (as a specic form of engagement) is an illustrative case
in point of the blurred conceptual and empirical terrain relevant
to engagement. Despite appearing ubiquitously throughout the
engagement literature, it is as yet a largely untapped and undened
construct that warrants further attention and denition. In this
study, wetherefore closely considered eort from a conceptual
perspective and harnessed this conceptual foundation to develop
a multidimensional measure of it—the Eort Scale. In particular,
it was anticipated that this novel tool would enable demarcation
between individual motivation dimensions and their unique
associations with dierent types of eort. Mathematics was chosen
as a specic subject area of focus, due to well-documented declines
in motivation and engagement highlighted by recent research (see
Collie etal., 2019), together with continued declines in students’
mathematics achievement, especially in Australia (e.g., omson
etal., 2016, 2019).
Utilizing a multilevel approach, wetested the measurement
properties of the Eort Scale at student- and classroom-levels in
mathematics to determine its psychometric properties and its
associations with multidimensional motivation via bivariate
correlations at both levels. Wethen employed structural equation
modelling to examine the role of multidimensional motivation in
predicting multidimensional eort at student- and classroom-
levels (as shown in Figure 1). rough these conceptual and
empirical processes, weshed further light on the unique and
shared variance between motivation and engagement (by way of
eort) and provide a foundation for greater clarity and coherence
for educational researchers and practitioners in their future work
aimed at optimizing students’ academic outcomes.
Motivation and engagement
To foreground our study of motivation and engagement,
werst briey summarize some key features of motivation and
engagement, some broad dimensions that distinguish them, and
the multidimensional motivation framework we harness as a
means to better understand how motivation and engagement
interrelate.
Where have webeen? Where are
wenow? Where are wegoing?
Motivation and engagement are signicant areas of interest in
educational psychology, seen as drivers of proximal and long-term
academic (and other) success and accomplishment (Reschly and
Christenson, 2012). Out of the two, motivation has received far
more focused theorizing and research, as indicated by the
numerous major theories that have been developed in the past ve
decades (e.g., social cognitive theory, Bandura, 2012; goal theory,
Elliot, 2005; need achievement theory, McClelland, 1961; self-
worth theory, Covington, 2000; self-determination theory, Ryan
and Deci, 2017; [situated] expectancy-value theory, Wigeld and
Eccles, 2000; Eccles and Wigeld, 2020; etc.). By contrast, there are
few theories about engagement, and relatively little work on
clarifying its measurement and theoretical grounding. However,
in the past 2 decades, there has been an uptick in attention being
given to engagement.
ere is now broad consensus that engagement is
multidimensional, comprising components of behavior, cognition,
and emotion/aect (Fredricks etal., 2004), but there remain
diering ideas about how these dimensions are dened and where
they reside within an overarching “engagement” construct
(Christenson etal., 2012). In their review of student engagement,
Reschly and Christenson (2012) identied three main channels of
engagement literature: one driven by reducing school dropout,
one emanating from a school reform perspective, and one
emerging from motivation theory and research. Especially in
relation to the latter, there has been conation with motivation
theory, denitions, and measurement. Given the lack of consensus
on denitions of engagement and its association with/distinction
from motivation, Reschly and Christenson (2012) argued there is
a need for theoretical and psychometric advancement of
engagement that can then beimplemented in motivation research
in order to better understand the two. ey encouraged new
expositions of engagement to advance the eld, and to test the
convergent and divergent validity of these expositions in relation
to motivation. With a focus on multidimensional eort, the
present study oers one approach toward a new exposition of
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 03 frontiersin.org
multidimensional engagement, and its alignments and dierences
from motivation. e envisaged yields are 2-fold: better
understanding and measurement of engagement (via a novel
multidimensional eort framework and measurement tool) that
aords a better understanding of the unique associations between
motivation’s and engagement’s individual dimensions.
Dierentiating motivation and
engagement
As noted, in the past 2 decades, researchers have attended
more closely to the distinctions and alignments between
motivation and engagement. In his commentary on major
researchers’ perspectives on motivation and engagement, Martin
(2012) (see also Martin, 2022) observed that at a fundamental
level, motivation and engagement may be demarcated into
internal and external dimensions. For example: Reeve (2012)
suggested motivation comprises “private, unobservable,
psychological, neural, and biological” factors, while engagement
constitutes “publicly observable behavior” (p.151); Ainley (2012)
identied motivation in terms of inner psychological factors,
whereas engagement reected more outward involvement; and,
Voelkl (2012) suggested that motivation aligns with internal
aective states and engagement with behavioral factors. All this
being the case, motivation has been dened as the inclination,
energy, emotion, and drive to learn, work eectively, and
achieve—and engagement as the more externally-evident factors
reecting the internal motivational phenomena (e.g., Martin etal.,
2017). However, although helpful in clearly dierentiating
between these two constructs, this basic demarcation of
motivation as internal, versus engagement as external, is not
intended as a prescriptive or denitive distinction.
Many researchers illustrate the blurred edges to this internal/
external classication, referring to internal facets of engagement,
typically characterized by cognitive and aective/emotional
dimensions (e.g., Fredricks etal., 2004; Appleton et al., 2006;
Cleary and Zimmerman, 2012; Wang and Eccles, 2012; Morgan
etal., 2022). Indeed, Reschly and Christenson (2012) specically
highlight conation over these internal facets of engagement and
aspects of motivation, such as self-regulation. ey point out that
dening motivation as intent (internal), and engagement as action
(external), implies that engagement is always behavioral and so
observable, whereas it is clear that cognitive and aective
engagement are largely internal processes, and so apparently
indistinguishable from motivation using this distinction.
As such, we draw on the denitions of motivation and
engagement in Martin etal. (2017) and extend them for this study,
with motivation being the inclination and drive to learn, work
eectively, and achieve—and engagement as the expression of this
inclination and drive to learn via either external (e.g., behaviors)
or internal (e.g., cognitive and aective) processes. In this study,
FIGURE1
Hypothesized full process model (processes are estimated at student- and classroom-levels). SES, Social-economic status indicator; NESB, Non-
English speaking background indicator.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 04 frontiersin.org
weaim to expand this distinction between the two constructs by
capturing a more comprehensive engagement characterization
(targeting eort as a specic active form of engagement). Our
study encompasses both internal and external dimensions of
engagement, and specically distinguishes between its internal
aspects (e.g., cognitive and social–emotional) and motivation.
In addition to motivation being considered as an internal
process, and engagement as both internal and external, there is
tentative agreement about the ordering of the process in which
they manifest, with motivation generally considered to lead to
engagement. For example, Schunk and Mullen (2012) used social-
cognitive theory as the basis for explaining how motivation and
engagement inter-relate, with self-ecacy (a motivation factor)
inuencing behavioral engagement. In another conceptualization,
Cleary and Zimmerman (2012) employed self-regulation theory
to describe how self-ecacy (motivation) leads to changes in
engagement (encompassing strategizing and self-regulatory
processes). ere is thus some agreement that “motivation is a
basis for subsequent engagement” (Martin, 2012, p.305). is
hypothesized ordering of motivation and engagement is important
in the present study as it is a means to examine how motivation
and engagement inter-relate and is thus a way to better understand
both constructs. Specically, weinvestigated the extent to which
multidimensional motivation predicted a novel engagement
(eort) construct (see Figure1).
Multidimensional motivation
e Motivation and Engagement Wheel (Martin, 1999–2022)
has been developed to capture multidimensional motivation as
proposed by seminal motivation theorizing. It is the framework
harnessed in the present study as the means to better understand
how motivation and engagement (by way of multidimensional
eort) interrelate. e Wheel comprises six (rst order) motivation
factors that can also beintegrated to form two higher-order factors
(positive/adaptive and negative/maladaptive motivation). Positive
motivation consists of: self-ecacy (the belief and condence in
one’s ability to learn), valuing (the belief in the importance,
usefulness, and relevance of one’s academic work), and mastery
orientation (the orientation to develop one’s learning and task
mastery). Negative motivation comprises: anxiety (the tendency
to feel anxious about one’s academic work), failure avoidance (the
inclination to work in order to avoid doing poorly), and uncertain
control (the lack of agency in eecting positive academic
outcomes). Positive motivation factors reect students’ positive
attitudes and orientations to academic learning, whereas negative
motivation factors represent students’ attitudes and orientations
that inhibit learning. As noted, these six factors emanate from
foundational motivation theories. Self-ecacy is very much based
on the work of Bandura (2001) and reects students’ task-specic
competence beliefs. Valuing draws on (situated) expectancy-value
theory (Wigeld and Eccles, 2000; Eccles and Wigeld, 2020) that
underscores the motivational boost students experience if they
value a task in one or more ways (e.g., in terms of utility and
importance). Mastery orientation is underpinned by goal theory
(Elliot, 2005), which reects students’ goal orientation towards
achieving academic success via eort, skill development, and
learning. Anxiety and failure avoidance draw from need
achievement and self-worth theories (McClelland, 1961;
Covington, 2000) that oer perspectives on students’ fear of failure
(failure avoidance is also implicated in goal theory by way of
performance avoidance goals; Elliot, 2005). Finally, uncertain
control is informed by attribution theory (Weiner, 2010) which
describes how the dimensions of stability, locus, and control
inuence students’ motivation to learn.
e factors in the Wheel are assessed via an accompanying
assessment tool, the Motivation and Engagement Scale (MES;
Martin, 1999–2022). e MES has been extensively employed and
tested in a variety of research studies (see Liem and Martin, 2012
for a review). e MES assesses not only the six positive and
negative motivation factors described above, but also three
positive engagement factors (planning and monitoring, task
management, and persistence) and two negative engagement
factors (self-handicapping and disengagement). With an expanded
denition of engagement (by way of eort) that includes internal
as well as external factors, the present study extends the
operationalization of engagement in the MES to engagement
factors outside it. To our knowledge, only one study has
investigated the predictive links between the MES motivation and
engagement factors, tentatively suggesting that motivation
predicts engagement (Martin etal., 2017). e present study’s
focus on multidimensional eort (as an active, energetic form of
positive engagement), and how the six motivation factors predict
it (Figure1), is an opportunity to incorporate a new measure of
engagement into the evidence base.
Multidimensional eort: A means to
better understand engagement and
motivation
Researchers are increasingly focusing on students’ engagement
at school as a predictor of academic success (Lei etal., 2018).
Fredricks etal. (2004) provided a seminal review of research and
theorizing on engagement, describing it in terms of a
(multidimensional) tripartite model, with behavioral, cognitive,
and emotional engagement as constituent factors of an overarching
engagement construct. Fredricks etal. (2004) described behavioral
engagement in terms of student involvement and participation in
school activities (in both academic and non-academic arenas).
ey described cognitive engagement in terms of a willingness
and thoughtfulness to invest eort to comprehend academic
concepts. Emotional engagement was described as encompassing
(both positive and negative) reactions to teachers, peers, and the
school environment (thus, also reecting a social–emotional
element), that in turn inuences students’ willingness to invest
eort. Engagement is thus now generally acknowledged to bea
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 05 frontiersin.org
multidimensional construct, typically considered as tripartite with
behavioral, cognitive, and aective (or emotional) components
(e.g., Christenson etal., 2012; Lei etal., 2018).
Tripartite eort
Of particular relevance to the present investigation, the meta-
analysis of engagement and achievement conducted by Lei etal.
(2018), framed students’ tripartite engagement in terms of being
actively involved in learning tasks and learning processes. Active
engagement implies the investment of energy and eort in learning
tasks, as opposed to a more passive involvement in class (such as
passively watching a video, or listening/paying attention in class
but not making any eort to participate or play an active role in
discussions). is emphasis on active (as opposed to passive)
engagement implies eortful engagement in each of the constituent
tripartite engagement dimensions. is being the case, wepropose
that academic eort sits under the umbrella construct of (positive)
engagement and comprises similar components, namely:
behavioral, cognitive, and aective/emotional dimensions—with a
higher-order eort factor that represents the theoretical and
empirical conuence of these rst-order dimensions.
e few researchers who have sought to more explicitly
account for both engagement and eort have emphasized the
importance of distinguishing between them. For example, it is
clear from Fredricks etal. (2004) that whereas school engagement
encompasses positive and negative (e.g., disengagement) academic
and non-academic dimensions, academic eort is a
sub-component of school engagement that specically relates to
the academic arena and involves positive engagement (not
disengagement) that is an active, volitional expenditure of energy
in the domains of behavior, cognition, and social–emotional
interactions. Fredricks etal. (2004) acknowledged that although
engagement has received substantial empirical attention, it is
theoretically messy and overlaps considerably with other
constructs. According to them, the broad umbrella term of
“engagement” is problematic as “it can result in a proliferation of
constructs, denitions, and measures of constructs that dier
slightly, thereby doing little to improve conceptual clarity” (p.60).
Of relevance to the present study, Fredricks etal. identied eort
(a construct incorporated under engagement) as a particular
example of this, and an avenue requiring further clarication and
then investigation in this space. Indeed, Fredricks etal. (2004),
Nagy (2016, 2017), and Carbonaro (2005) have all underscored
the importance of eort and its multidimensional nature,
comprising behavioral (or operative), cognitive, and social–
emotional factors.
Following Fredricks etal. (2004), Nagy (2016, 2017), and
Carbonaro (2005), the behavioral dimension of eort in the
present study is focused on the notion of “doing” and “outcomes-
completion”—referred to herein as operative eort and dened as
active, purposeful, and energetic action-based application to
learning. Operative eort is typied by the application of
behavioral energy in the production and completion of
schoolwork. Cognitive eort is dened as active, purposeful, and
energetic mental/psychological application to learning. It is
typied by concentration, attention, and focus directed toward
understanding, comprehension, and mastery of schoolwork.
Social–emotional eort is dened as active, purposeful, and
energetic interpersonal/aective application to learning. It is
typied by appropriate and respectful classroom social–emotional
interactions that involve self-control and sensitivity to the social
context of learning, conducive to completing schoolwork.
Measurement of tripartite eort
Building on this tripartite framing of eort in terms of its
operative, cognitive, and social–emotional dimensions, a
multidimensional eort scale (hereaer, the Eort Scale) was
developed for implementation in the present study. is Eort
Scale is designed to capture the three distinct aspects of eort, and
also to represent a hypothesized overarching eort factor reecting
appropriate weighting (or loading) of each of the three constituent
factors onto the whole—enabling both specicity (in the case of a
rst-order structure) and broader application (in the case of a
higher-order structure) as appropriate to the research purpose. It
is this tool that will represent an approach to multidimensional
engagement (i.e., via eort) and bethe basis of analyses with
multidimensional motivation in the present study. It is described
more fully in the section Materials and methods, below.
Context and background attributes
relevant to motivation and engagement
In line with major motivation theories (e.g., Bandura, 2001;
Ryan and Deci, 2017; Eccles and Wigeld, 2020), we also
accounted for contextual and background attributes known to
beimplicated in motivation and engagement. Wedid so in two
ways: by employing multilevel modelling to extend the typical
student-level analyses of motivation and engagement to analyses
at the classroom-level, and by including numerous pertinent
student- and classroom-level background attributes as predictors
of motivation and engagement (see Figure1). e former enabled
us to disentangle student- and classroom-level motivation and
engagement. e latter enabled us to determine the unique
association of motivation predicting eort, by controlling for
variance attributable to pertinent student- and classroom-level
background factors.
Aims of the present study
Historical conation over the intertwined constructs of
motivation and engagement has impeded advances in theory and
practice relating to students’ academic development. As the eld
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 06 frontiersin.org
of engagement has progressed over the past decade, advances in
its theoretical conceptualizing have led to new opportunities to
better understand the interface between motivation and
engagement. Questions can now beposed that further unpick the
distinctiveness of these two intertwined constructs, such as how
positive and negative motivation factors uniquely predict
individual dimensions of engagement. is study seeks to bring
clarity to this space through a purposeful investigation using the
hitherto untapped and undened construct of eort
(representing a specic active form of multidimensional
engagement) and to investigate the unique associations of its
respective dimensions with multidimensional motivation.
We aimed to closely consider eort from a conceptual
perspective, hypothesizing a tripartite model of academic eort
in terms of operative, cognitive, and social–emotional
dimensions—and then developing a practical multidimensional
eort measure—the Eort Scale, incorporating each of the three
component factors.
We adopted a construct validation approach to explore
motivation and this novel eort framework (e.g., Marsh, 1997,
2002; Martin and Marsh, 2008). Such an approach considers
assessment of the validity of both within-network (“internal
validity”) and between-network (“external validity”). Wepursued
this construct validation by rst testing the measurement
properties of the Eort Scale at student- and classroom-levels
(internal validity), then testing the association between motivation
and eort via bivariate correlations at both levels (external
validity), and then examining the role of motivation predicting
eort at student- and classroom-level (external validity),
appropriately controlling for pertinent student- and classroom-
level background attributes (see Figure1). For the purposes of this
study, weare particularly interested in convergent (the extent to
which motivation is associated with eort in theoretically plausible
ways) and discriminant (the extent to which there remains
sucient unshared variance to indicate their distinctiveness)
aspects of the constructs’ external validity. Wehypothesized that
motivation and eort would beassociated with each other (by way
of correlations and predictive parameters) but le as an open
empirical question the precise nature and strength of associations
between their dierent dimensions.
Materials and methods
Participants and procedure
The sample for this study comprised 946 Australian high
school students nested within 59 mathematics classrooms
from five schools. The sample was chosen, within the given
constraints and practicalities of data collection, to be as
diverse as possible in terms of gender, academic ability, age,
and school gender profile (viz. single-sex or coeducational)
and therefore as representative as possible of potentially
influential covariate attributes. All schools were non-
academically selective in intake, in the independent school
sector and located in and around a major capital city of New
South Wales (NSW) on the east coast of Australia. Of the five
schools, three were coeducational, one was a single-sex boys’
school, and one was a single-sex girls’ school. Just over half
(53%) of students were boys. Students were in the first 4 years
of high school in Australia and comprised: Year 7 (8%),
Year 8 (41%), Year 9 (34%), and Year 10 (17%). The average
age was 14.70 years (SD = 0.98 years). Non-English-speaking
background (NESB) students accounted for 16% of the sample.
Students typically came from higher socio-economic status
(SES) postal districts (M = 1,084, range from 846 to 1,179,
SD = 64) than the Australian average (M = 1,000, SD = 100)
based on the Australian Bureau of Statistics (ABS) index of
relative socio-economic advantage and disadvantage
classification (SEIFA; Australian Bureau of Statistics, 2016).
Of the 59 classrooms, class size varied from 7 to 29 (M = 21,
SD = 5), with participation rates ranging from 31% to 100%
(M = 74%, SD = 17%). Human ethics approval was received
from the lead researcher’s university, and school principals
then provided approval for their school’s participation in the
study. Following this, parents/careers and students provided
consent. An online survey was administered to students, in a
regular mathematics lesson, in the final term of 2020.
Materials
e measures included in the survey comprised the
substantive factors of motivation and eort. We also assessed
student and classroom background attributes as covariates.
Motivation
Motivation was measured using six self-rated items from the
brief form of the Motivation and Engagement Scale—High
School [MES-HS-Short; Martin, 1999–2022]. e items captured
three positive motivation constructs (self-ecacy, valuing, and
mastery orientation) and three negative motivation constructs
(anxiety, failure avoidance, and uncertain control). e items
(e.g., for self-ecacy, “I believe Ican do well in this subject”) were
rated using a seven-point Likert scale (1 = strongly disagree to
7 = strongly agree). As each motivation factor was represented by
a single-item, wecould not estimate them as latent variables, and
so wemodelled each factor as error-adjusted mean scores so that
our analyses could correct for unreliability (as latent modeling
would do). e following equation was used to calculate the
error-adjusted mean score: σ2 * (1−ω), where σ2 is the estimated
variance of the substantive factor and ω is the reliability estimate
of this factor (Hayduk, 1987; see also Cole and Preacher, 2014).
e reliabilities (omega total; McNeish, 2018) and variances were
taken from a prior research program using the full (multi-item)
MES-HS in mathematics (Martin and Marsh, 2005; Marsh etal.,
2008). Descriptive statistics for the present study are presented
in Results.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 07 frontiersin.org
Academic eort
Student-rated eort was measured using a three-factor, nine-
item scale (the Eort Scale) emanating from work by Nagy (2016,
2017, 2022). Operative eort was measured via three items [e.g., “In
mathematics, Itry hard on schoolwork (e.g., in class or at home etc.)
given to me”]; cognitive eort was measured via three items (e.g., “I
amfocused in mathematics class”), and social–emotional eort was
measured via three items [e.g., “I show self-control in mathematics
lessons (e.g., Iwait my turn, do not interrupt, and do not talk over
other students etc.)”]. As described in the Introduction, the
hypothesized eort framework comprises three rst-order factors
and also an overarching higher-order eort factor. All eort items
were rated using a seven-point Likert scale (1 = strongly disagree to
7 = strongly agree) and are detailed in Supplementary material
(Supplementary Table S1). For completeness, also presented in
Supplementary material is a brief form of the Eort Scale (the Eort
Scale—Short [ES-S]; one item for each of the three dimensions, thus
a three-item measure)—and its psychometric properties and
correlations with motivation. Descriptive, reliability, and factor
analytic ndings for rst- and higher-order Eort Scale factors are
presented in the Results section below.
Student and classroom background attributes
Our hypothesized process model (Figure1) was designed to
assess the unique associations between motivation and eort
beyond the roles of student and classroom background attributes.
It was therefore important to account for notable student and
classroom background attributes. Student background factors
were: age (in years); gender (0 = female, 1 = male); socio-economic
status (SES), home language background (NESB; 0 = English,
1 = non-English speaking background), and mathematics ability.
e SES score was derived from self-rated postcode and/or
suburb, using the Australian Bureau of Statistics Index of Relative
Socio-Economic Advantage and Disadvantage classication
(SEIFA; Australian Bureau of Statistics, 2016), with higher values
representative of areas of greater socio-economic advantage.
Mathematics ability was assessed via a 10-item mathematics
assessment, the High School Mathematics Competency scale
(HSMC; Nagy, 2021; and evidence of validation demonstrated in
Martin etal., 2020), developed to test the underlying mathematical
competencies of students. Assessment items were graduated in
diculty but accessible to all students in years 7–10 without the
need for stage-specic subject knowledge. Items were mapped
against the New South Wales (NSW) and Australian national
curriculums (Australian Curriculum Assessment and Reporting
Authority, n.d.; NSW Education Standards Authority, 2019).
Example items from this assessment that reected the curriculum
domains of Time, Patterns and Algebra, and Ratios and Rates,
were respectively, [Time]: “What time will it be 75 min aer
11:15 am? [(A) 11:30 am, (B) 12:30 am, (C) 11:30 pm, and (D)
12:30 pm]”; (Patterns and Algebra): “Find the next number in the
pattern: 8, 11, 14, 17, [(A) 20, (B) 21, (C) 22, and (D) 23]”; (Ratio
and Rates): “If the ratio of boys: girls in a class is 4:5, what fraction
of the class is boys? [(A) 1/4, (B) 1/5, (C) 4/5, and (D) 4/9].” A
mathematics ability score was calculated for each student
(corresponding to the total number of correct responses out of 10)
and then standardized by year group. ree classroom covariates
were also included: class size, class-average age, and class-average
ability (using the mean mathematics ability score for
each classroom).
Data analyses
Data collected from school students that relates to their
learning is typically part of a multilevel structure, with students
clustered into classrooms. Within these classrooms, there is
generally greater similarity among students than between students
of dierent classrooms, due to factors such as how classroom
groupings are chosen (e.g., streaming by ability-level) and unique
classroom culture (e.g., due to the unique combination of teacher
expectations and classroom climate). Typically, it is statistically
invalid to analyze clustered data at a single-level, as it can violate
statistical assumptions and give rise to Type 1 errors (Marsh etal.,
2008). Furthermore, in measuring and analyzing constructs at
either student-level or classroom-level, the interpretation of
results may bedierent and yield dierent practical implications.
It is now well established that accounting for these realities
requires multilevel modelling that accommodates the clustering
of students within classrooms and distinguishes between student-
level eects and classroom-level eects. Indeed, dierences in
motivation and engagement (specically, eort, in this study) may
beinuenced by both individual and classroom factors, and it is
therefore appropriate to use a multilevel approach in bringing
conceptual and empirical clarity to their association. e central
analyses therefore consisted of multilevel conrmatory factor
analysis (MCFA) and doubly-latent multilevel structural equation
modelling (MSEM).
Analyses were carried out in Mplus version 8 using maximum
likelihood estimation with robust standard errors (MLR; Mplus
RRID:SCR_015578; Muthén and Muthén, 1998–2022), which
accounts for non-normality of the sample. Missing data (4%) were
handled using the Mplus full information maximum likelihood
(FIML) default. All multilevel modelling included Level 1 (L1;
student-level) and Level 2 (L2; classroom-level) variables. To
determine model t, a Comparative Fit Index (CFI) greater than
0.90 and Root Mean Square Error of Approximation (RMSEA)
less than 0.08 were used as thresholds for acceptable t (Hu and
Bentler, 1999), and a CFI greater than 0.95 and RMSEA less than
0.05 as thresholds for excellent t. Prior to conducting multilevel
analyses, measurement invariance tests as a function of key
sub-groups (e.g., age and gender) were conducted for the eort
factors and demonstrated relative invariance across all sub-groups
tested. Full details of these tests can be found in
Supplementary material in the section titled “Invariance Tests”
and in Supplementary Tables S4–S6.
Multilevel descriptive analyses comprised student-level (L1) and
classroom-level (L2) scale means, standard deviations, skewness,
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 08 frontiersin.org
kurtosis, reliability, and intra-class correlations (ICCs). To test factor
structure, two MCFAs were rst conducted using the Eort Scale
(one involving only rst-order eort factors, and the other including
a higher-order eort factor). en, these two MCFAs were
re-estimated but with the motivation factors also included. ese
latter MCFAs enabled a test of t for models where motivation and
eort were represented as distinct factors and an assessment of
correlations between motivation and eort. In MCFAs, L1 and L2
parallel latent factor loadings for eort (but not for motivation as
these were single-item factors—see Materials) were constrained to
beequal (i.e., isomorphism) and L2 residuals were constrained to
begreater than zero to ensure a more parsimonious model with
greater accuracy in parameter estimation at both levels (e.g., Morin
et al., 2014). e hypothesized process model of motivation
predicting eort was tested with two doubly-latent MSEMs (one for
rst-order eort and one for higher-order eort; Figure 1) that
included controls for student- and classroom-level background
attributes (as predictors of motivation and eort). In the MSEMs, all
background covariates were correlated, motivation predictors were
correlated, and eort outcomes were correlated.
Results
Preliminary descriptive statistics
Means and standard deviations (SDs) for eort factors at L1
(student-level) and L2 (classroom-level) are shown in
Table1A. Skewness and kurtosis values are also in Table1A and
are within indicative guidelines for approximately normal
distributions (Kim, 2013). Descriptive statistics at L1 and L2 for
motivation are displayed in Table1B, with skewness and kurtosis
values also reecting approximately normal distributions.
Fit and dimensionality of motivation and
eort
As described in the Introduction, it is vital to have sound
measurement of engagement (by way of eort in this study) in order
to eectively explore the distinctiveness of motivation and
engagement. erefore, we rst conducted MCFAs to test the
hypothesized eort dimensions, operationalized via the Eort Scale
[see Supplementary material for a summary of single-level (student)
CFAs of the Eort Scale]. e rst-order eort structure yielded an
excellent t to the data [
2
c
(54) = 177.284, p < 0.001, RMSEA = 0.049,
CFI = 0.966], as did the higher-order eort structure [
2
c
(56) = 176.868, p < 0.001, RMSEA = 0.048, CFI = 0.966]. As Table1A
demonstrates, mean MCFA loadings on the rst-order eort factors
ranged from 0.71 to 0.90 (L1) and 0.97 to 1.00 (L2), with a grand
mean of 0.81 (L1) and 0.98 (L2). e mean MCFA loadings on the
higher-order eort factor were 0.85 (L1) and 0.87 (L2). All factor
loadings were therefore within an acceptable range (Byrne, 2012).
Reliability estimates for rst-order eort factors ranged from ω = 0.75
to 0.93 (L1) with a mean of 0.84, and ω = 0.98 to 1.00 (L2) with a
mean of 0.99, indicating acceptable internal consistency. Reliability
for the higher-order eort factor was ω = 0.89 (L1) and 0.91 (L2) and
so also indicated acceptable internal consistency. Table1A shows
intra-class correlations (ICCs) which ranged from 0.09 to 0.15 for
rst-order eort factors and was 0.15 for the higher-order eort
factor. e grand mean ICC (0.12) was above the 10% threshold
recommended by Byrne (2012) and provided justication for our
multilevel approach in this study.
Having established the dimensionality and measurement
properties of eort, wethen included motivation in the MCFAs to
ascertain its dimensionality and distinctiveness relative to eort. Two
models were run1 that both yielded excellent t to the data: one for
rst-order eort [
2
c
(126) = 296.852, p < 0.001, RMSEA = 0.038,
CFI = 0.970] and one for higher-order eort [
2
c
(152) = 332.056,
p < 0.001, RMSEA = 0.035, CFI = 0.969]. us, when modelled as
separate factors, there is excellent t, signaling distinct dimensionality
between motivation and eort. Table1B shows motivation factor
loadings, determined from a fully-saturated MCFA that only
included motivation items, which ranged from 0.76 to 0.89 (L1) with
a mean of 0.82 and from 0.94 to 0.98 (L2) with a mean of 0.97. ICCs
ranged from 0.07 to 0.18 with a mean of 0.13 indicating that the
variance attributable to motivation at the classroom-level was above
the recommended threshold (Byrne, 2012) justifying modelling
motivation at L1 and L2.
Multilevel correlations between
motivation and eort
e MCFAs involving both motivation and eort also
generated latent correlations that were a further means of assessing
their distinctiveness. All correlations are summarized in Table2,
with the correlations between the target substantive factors of
motivation and eort displayed in bold font for clarity. At both L1
(student-level) and L2 (classroom-level), there were signicant
positive correlations between all three positive motivation and
rst- and higher-order eort factors; for operative eort (L1:
r = 0.50 to 0.53, mean r = 0.52, p < 0.001; L2: r = 0.81 to 0.86, mean
r = 0.84, p < 0.001), cognitive eort (L1: r = 0.46 to 0.50, mean
r = 0.49, p < 0.001; L2: r = 0.70 to 0.74, mean r = 0.72, p < 0.001),
social–emotional eort (L1: r = 0.34 to 0.43, mean r = 0.39,
p < 0.001; L2: r = 0.60 to 0.66, mean r = 0.64, p < 0.001), and higher-
order eort (L1: r = 0.53 to 0.56, mean r = 0.55, p < 0.001; L2:
r = 0.79 to 0.84, mean r = 0.82, p < 0.001). ere were also
1 For completeness wealso tested a model where the six motivation
factors and three first-order eort factors loaded onto a single higher-
order factor (thus, a model where motivation and eort were not
dierentiated as separate constructs). This yielded a significantly poorer
fit to the data relative to the excellent fit of the MCFAs separating motivation
and eort as distinct constructs: (
2
c
[188] = 775.512, p < 0.001,
RMSEA = 0.057, CFI = 0.897).
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 09 frontiersin.org
signicant negative correlations between the negative motivation
factor of uncertain control and all eort factors: for operative
eort (L1: r = −0.30, p < 0.001; L2: r = −0.57, p < 0.001), for
cognitive eort (L1: r = −0.28, p < 0.001; L2: r = −0.54, p < 0.001),
for social–emotional eort (L1: r = −0.23, p < 0.001; L2: r = −0.35,
p < 0.05), and for higher-order eort (L1: r = −0.32, p < 0.001; L2:
r = −0.57, p < 0.001). ere were no signicant correlations
between the negative motivation factors of anxiety and failure
avoidance and any of the eort factors at either level. Taken
together, the bivariate associations between motivation and eort
demonstrated signicant alignments, but at the same time
suciently sized unshared variance to support their distinctiveness.
Multilevel structural equation modelling
of motivation predicting eort
e multilevel process model (Figure 1) of motivation
predicting eort was then tested using doubly-latent MSEM. Two
MSEMs were conducted, the rst (MSEM
1
) examined motivation
TABLE1A Multilevel descriptive statistics and CFAs of first-order and higher-order eort.
Vari a ble
Statistics
MSD Skew Kurtosis ωCFA loadings
(min., max., mean)
ICC
Level 1 (Student)
First-order eort factors
Operative eort 5.934 0.955 −1.362 2.755 0.853 0.687, 0.870, 0.808 -
Cognitive eort 5.779 1.051 −1.471 3.225 0.925 0.832, 0.933, 0.896 -
Social-emotional eort 6.243 0.713 −1.102 1.915 0.754 0.660, 0.742, 0.711 -
Second-order eort factor
Higher-order eort 5.985 0.797 −1.093 1.733 0.888 0.698, 0.958, 0.847 -
Level 2 (Classroom)
First-order eort factors
Operative eort 5.893 0.405 −0.806 0.946 0.985 0.953, 1.000, 0.978 0.087
Cognitive eort 5.739 0.448 −0.637 -0.031 0.999 0.998, 1.000, 0.999 0.104
Social-emotional eort 6.227 0.306 −0.521 -0.281 0.977 0.904, 1.000, 0.966 0.147
Second-order eort factor
Higher-order eort 5.953 0.350 −0.361 -0.445 0.912 0.657, 1.000, 0.870 0.149
ω = reliability (omega total; McNeish, 2018); ICC = Intra Class Correlation; CFA Loadings = Conrmatory factor analysis standardized factor loadings; M, SD, Skew and Kurtosis are
calculated from unit-weighted scale scores of raw items.
TABLE1B Multilevel descriptive statistics of motivation items.
Vari a ble
Statistics
MSD Skew Kurtosis ωaCFA loading ICC
Level 1 (Student)
Self-ecacy (positive motivation) 5.822 1.333 −1.683 2.950 0.771 0.852 -
Valuing (positive motivation) 5.542 1.384 −1.120 1.018 0.770 0.847 -
Mastery orientation (positive motivation) 5.590 1.283 −1.145 1.329 0.806 0.888 -
Anxiety (negative motivation) 5.251 1.754 −0.923 −0.129 0.771 0.759 -
Failure avoidance (negative motivation) 4.715 1.819 −0.474 −0.862 0.766 0.765 -
Uncertain control (negative motivation) 3.012 1.665 0.679 −0.439 0.788 0.821 -
Level 2 (Classroom)
Self-ecacy (positive motivation) 5.738 0.635 −0.986 0.582 0.777 0.978 0.183
Valuing (positive motivation) 5.507 0.619 −0.578 0.871 0.789 0.971 0.165
Mastery orientation (positive motivation) 5.570 0.445 −0.709 0.595 0.840 0.969 0.106
Anxiety (negative motivation) 5.210 0.493 −0.260 −0.203 0.779 0.936 0.072
Failure avoidance (negative motivation) 4.746 0.605 −0.143 0.428 0.842 0.963 0.098
Uncertain control (negative motivation) 3.120 0.680 0.095 −0.201 0.876 0.980 0.141
ω = reliability (omega total; McNeish, 2018); ICC = Intra Class Correlation; CFA Loadings = Conrmatory factor analysis standardized factor loadings; M, SD, Skew and Kurtosis are
calculated from raw items. a Motivation items are modelled as error-adjusted scores using established reliability and variance measures from a prior research program (ω and σ2 values
were derived from data used in: Martin and Marsh, 2005; Marsh et al., 2008).
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 10 frontiersin.org
TABLE2 Multilevel correlation matrix within and between motivation and eort factors.
Variables Operative
eort
Cognitive
eort
Social–
emotional
eort
Self-ecacy Valuing Mastery
orientation Anxiety Failure
avoidance
Uncertain
control
Level 1 (Student)
Eort factors
Cognitive eort 0.848***
Social–emotional eort 0.611*** 0.657***
Motivation factors
Self-ecacy (positive) 0.533*** 0.503*** 0.344***
Valuing (positive) 0.533*** 0.501*** 0.407*** 0.675***
Mastery orientation (positive) 0.501*** 0.461*** 0.428*** 0.407*** 0.532***
Anxiety (negative) 0.079 0.019 0.071 −0.158** −0.058 −0.002
Failure avoidance (negative) −0.058 −0.055 −0.074 −0.176*** −0.191*** −0.030 0.537***
Uncertain control (negative) −0.301*** −0.281*** −0.228*** −0.497*** −0.329*** −0.181*** 0.370*** 0.387***
Level 2 (Classroom)
Eort factors
Cognitive eort 0.942***
Social–emotional eort 0.721*** 0.782***
Motivation factors
Self-ecacy (positive) 0.861*** 0.744*** 0.598***
Valuing (positive) 0.851*** 0.720*** 0.652*** 0.770***
Mastery orientation (positive) 0.805*** 0.699*** 0.658*** 0.631*** 0.683***
Anxiety (negative) 0.149 0.085 0.217 0.078 0.044 0.016
Failure avoidance (negative) 0.021 0.005 0.011 −0.014 −0.043 0.180 0.332**
Uncertain control (negative) −0.571*** −0.544*** −0.354*−0.697*** −0.527*** −0.243 0.062 0.301*
aLevel 1 Higher order eort 0.557*** 0.563*** 0.529*** 0.055 −0.065 −0.316***
aLevel 2 Higher order eort 0.836*** 0.826*** 0.792*** 0.144 0.014 −0.566***
Motivation items are modelled as error-adjusted scores; All correlations taken from the rst-order CFA model with the exception of ahigher-order eort correlations which are taken from the higher-order CFA model; values in bold highlight correlations
between the substantive factors; *p < 0.05. **p < 0.01. ***p < 0.001.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 11 frontiersin.org
predicting rst-order eort factors, and the second (MSEM2)
investigated motivation predicting higher-order eort. To
appropriately ascertain the unique associations between
motivation and eort (beyond student and classroom background
attributes), the MSEMs included controls for a range of student
covariates (age, gender, SES, NESB, and mathematics ability) and
classroom-level attributes (class-average ability, class size, and
class-average age)—with these covariates predicting motivation
and eort in the MSEMs. Both models yielded an excellent t to
the data [rst-order eort model MSEM1:
2
c
(174) = 365.398,
p < 0.001, RMSEA = 0.034, CFI = 0.971; higher-order eort model
MSEM2:
2
c
(216) = 458.662, p < 0.001, RMSEA = 0.034,
CFI = 0.964]. In the summary of substantive ndings described
here, only signicant L1 and L2 standardized paths (β) between
the substantive factors and notable results involving covariates are
presented (and these are shown in Figures2 and 3 for the rst-
order eort and higher-order eort models respectively). All
signicant and non-signicant standardized substantive and
covariate paths are reported in Table3.2
At L1, student-level self-ecacy signicantly positively predicted
student-level operative eort (β = 0.27, p < 0.001), cognitive eort
(β = 0.25, p < 0.01), and higher-order eort (β = 0.27, p < 0.001).
Student-level valuing signicantly positively predicted student-level
operative eort (β = 0.18, p < 0.05), cognitive eort (β = 0.18, p < 0.05),
social–emotional eort (β = 0.18, p < 0.05), and higher-order eort
(β = 0.20, p < 0.01). Student-level mastery orientation signicantly
2 Full results of the higher-order eort model (MSEM
2
) can beseen in
Supplementary Material Table S7.
Self-efficacy
Operative effort
Cognitive effort
Anxiety
Failure
avoidance
Uncertain
control
Valuing
Mastery
Social-emotional
effort
Operative effort
Cognitive effort
Social-emotional
effort
Valuing
Mastery
Level 1 (Student)
Level 2 (Classroom)
FIGURE2
Significant substantive paths in central multilevel analysis—First order eort factors (MSEM1). Only significant paths are shown and are labeled with
standardized betas (β); *p < 0.05. **p < 0.01. ***p < 0.001; All paths controlled for variance attributed to covariates (Level 1: age, gender, social-
economic status, non-English speaking background, mathematics ability; Level 2: class-average ability, class size and class-average age). See
Table3 for all covariate associations and all non-significant paths.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 12 frontiersin.org
Level 1 (Student)
Level 2 (Classroom)
Self-efficacy
Higher-order effort
Anxiety
Valuing
Mastery
Failure
avoidance
Uncertain
control
Higher-order effort
.42**
Mastery
FIGURE3
Significant substantive paths in central multilevel analysis—Higher Order Eort Factor (MSEM2). Only significant paths are shown and are labeled
with standardized betas (β); *p < 0.05. **p < 0.01. ***p < 0.001; All paths controlled for variance attributed to covariates (Level 1: age, gender, social-
economic status, non-English speaking background, and mathematics ability; Level 2: class-average ability, class size, and class-average age). See
Table3 for all covariate associations and all non-significant paths.
positively predicted student-level operative eort (β = 0.28, p < 0.001),
cognitive eort (β = 0.25, p < 0.001), social–emotional eort (β = 0.27,
p < 0.001), and higher-order eort (β = 0.30, p < 0.001). Examining
the negative motivation factors at L1, student-level uncertain control
signicantly negatively predicted student-level operative eort
(β = −0.11, p < 0.05) and higher-order eort (β = −0.11, p < 0.05).
Interestingly, student-level anxiety signicantly positively predicted
student-level operative eort (β = 0.16, p < 0.01), s ocial–emotional
eort (β = 0.12, p < 0.05), and higher-order eort (β = 0.14, p < 0.01).
At L2, the signicant paths found between the classroom-level
motivation and eort factors were in relation to valuing, which
positively predicted social–emotional eort (β = 0.47, p < 0.05), and
mastery orientation, which positively predicted operative eort
(β = 0.34, p < 0.01), cognitive eort (β = 0.49, p < 0.01), and higher-
order eort (β = 0.42, p < 0.01).
Although not the substantive focus of the study, for
completeness wereport here noteworthy patterns of covariate
associations where a given L1 (student-level) or L2 (classroom-
level) covariate signicantly predicted both motivation and eort
(see Table 3 for all covariate associations). At L1, gender
signicantly predicted student-level motivation and eort.
Specically, being male was positively associated with self-ecacy
(β = 0.16, p < 0.001), and valuing (β = 0.10, p < 0.01), and negatively
associated with anxiety (β = −0.16, p < 0.001), uncertain control
(β = −0.09, p < 0.01), operative eort (β = −0.10, p < 0.05), so cial–
emotional eort (β = −0.24, p < 0.001), and higher-order eort
(β = −0.09, p < 0.05). Student-level mathematics ability
signicantly positively predicted self-ecacy (β = 0.21, p < 0.001),
valuing (β = 0.16, p < 0.001), and higher-order eort (β = 0.06,
p < 0.05), and negatively predicted failure avoidance (β = −0.12,
p < 0.01), and uncertain control (β = −0.14, p < 0.01). At L2, class-
average ability signicantly positively predicted self-ecacy
(β = 0.37, p < 0.01), valuing (β = 0.44, p < 0.01), and s ocial–
emotional eort (β = 0.35, p < 0.05), and negatively predicted
failure avoidance (β = −0.55, p < 0.001), and uncertain control
(β = −0.63, p < 0.001).
Discussion
e present study sought to bring new insights to the
alignment and distinctiveness of motivation and engagement
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 13 frontiersin.org
TABLE3 Multilevel structural equation process model: Standardized beta coecients.
Variables
Outcomes (MSEM1 using rst-order eort factors) MSEM2
Self-ecacy Valuing Mastery
orientation Anxiety Failure
avoidance
Uncertain
control
Operative
eort
Cognitive
eort
Social–
emotional
eort
Higher-order
eort
Level 1 (Student)
L1 Covariates
SES 0.056 0.035 0.007 0.063 0.012 0.012 0.023 −0.018 −0.083* −0.010
Age 0.022 0.034 0.028 −0.032 −0.025 −0.014 −0.010 −0.018 −0.015 −0.016
Gender (male) 0.158*** 0.101** 0.041 −0.162*** −0.034 −0.086** −0.099* −0.035 −0.236*** −0.087*
NESB −0.045 0.052 −0.002 −0.025 0.061 0.070 −0.046 −0.022 −0.035 −0.036
Mathematics ability 0.213*** 0.160*** 0.044 −0.070 −0.123** −0.142** 0.056 0.052 0.059 0.061*
L1 Motivation factors
Self-ecacy (positive) 0.270*** 0.252** 0.099 0.272***
Valuing (positive) 0.179*0.177*0.175*0.198**
Mastery orientation (positive) 0.278*** 0.248*** 0.274*** 0.295***
Anxiety (negative) 0.162** 0.087 0.121*0.138**
Failure avoidance (negative) −0.005 0.025 −0.037 0.005
Uncertain control (negative) −0.114* −0.089 −0.107 −0.113*
Level 2 (Classroom)
L2 Covariates
Class-average ability 0.368** 0.437** −0.127 −0.097 −0.550*** −0.629*** 0.132 0.337 0.351*0.238
Class size 0.298* −0.008 0.379*0.292 0.407* −0.142 −0.093 −0.259 −0.269 −0.171
Class-average age −0.153 −0.361*** −0.287** 0.204 −0.036 −0.101 −0.178 −0.102 0.113 −0.135
L2 Motivation factors
Self-ecacy (positive) 0.319 0.236 0.177 0.292
Valuing (positive) 0.193 0.000 0.465*0.149
Mastery orientation (positive) 0.340** 0.490** 0.390 0.423**
Anxiety (negative) 0.101 0.082 0.015 0.101
Failure avoidance (negative) 0.014 0.061 0.006 0.027
Uncertain control (negative) −0.158 −0.197 0.112 −0.148
Motivation items modelled as error-adjusted scores. SES, Social-economic status indicator (positive is higher SES); NESB, Non-English speaking background; *p < 0.05. **p < 0.01. ***p < 0.001.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 14 frontiersin.org
(operationalized as eort). Beginning with a conceptual review to
clarify denitional parameters of both motivation and engagement
(with specic focus on the relatively neglected construct of eort),
wetested a hypothesized multidimensional eort structure and
then its empirical juxtaposition with a well-established motivation
framework. Multilevel (student- and classroom-level) ndings
supported the reliability and validity of multidimensional eort
(by way of the Eort Scale) and the distinctiveness of eort from
motivation by way of multidimensional model t and latent
bivariate multilevel correlations. en, MSEM explored the
“classic” motivation → engagement (eort) process. is revealed
signicant associations between student- and classroom-level
motivation and student- and classroom-level eort—as well as
some noteworthy patterns of covariates predicting both
motivation and eort at student- and classroom-levels. ese
ndings and their implications for motivation and engagement
theory, research, and practice are now discussed.
Findings of note
is study has not only reinforced well-established
understanding of motivation and engagement as two inter-related
constructs (Martin, 2009; Martin etal., 2017), it has also shed new
light on some of the precise ways in which individual motivation
factors interplay with specic multidimensional engagement
factors. MCFA ndings showed multidimensional motivation and
multidimensional eort to have distinct factor structures, with
signicant and theoretically plausible bivariate correlations
between rst-order motivation and rst- and higher-order eort
factors. MSEM further supported this via unique predictive
associations between rst-order motivation, and rst- and higher-
order eort factors. In this study wewere especially interested in
the extent to which motivation is associated with eort in
theoretically plausible ways (convergent validity) and also the
extent to which there remained sucient unshared variance to
indicate their distinctiveness (discriminant validity). Our ndings
garner strong evidence for both convergent (signicant
associations) and discriminant validity (noteworthy unshared
variance) between motivation and eort.
e MSEM provided a particularly nuanced insight into how
multidimensional motivation and eort are aligned and distinct,
bringing greater psychometric clarity to developments in theorizing,
and aording a better understanding of the distinctiveness and
interface of motivation and engagement (by way of our novel eort
framework). Positive motivation factors were found to
overwhelmingly predict eort at the student-level. Specically (aer
controlling for student-level background attributes—discussed
below), mastery orientation and valuing uniquely predicted all three
eort factors (operative, cognitive, and social–emotional), and self-
ecacy predicted both operative and cognitive eort. All three
positive motivation factors predicted higher-order eort. In
explaining the salient role of mastery orientation, it is worth noting
central tenets of goal theory (Elliot, 2005) that posits eort as a means
by which students’ mastery orientation is operationalized. Indeed,
classroom-level mastery orientation also predicted classroom-level
eort, which is in line with the role of classroom motivational climates
in classroom-level engagement under goal theory (Ames, 1992;
Wentzel, 2012; Wentzel etal., 2017). ere are thus strong theoretical
roots underpinning the role of mastery orientation in predicting
eort at both student- and classroom-levels.
Valuing was also predictive of all three eort dimensions at
the student-level and of social–emotional eort at the classroom-
level. us, when an individual student believes in the importance
and relevance of their academic work to learning, they are more
likely to try harder in their application to that learning. is
nding aligns with major psycho-educational perspectives—
particularly, expectancy-value theory—contending that “students’
subjective task values predict both intentions and actual decisions
to persist at dierent activities” (Wigeld and Cambria, 2010,
p.21). In addition, Wigeld and Cambria (2010) highlight that
students’ values are socio-culturally situated which may well
explain why, at the classroom-level, valuing predicted social–
emotional eort. Indeed, Eccles and Wigeld (2020) recently
updated their conceptual framework to “situated expectancy-value
theory” to reect the situated nature of motivation and
motivational processes. In the case of our study, classrooms
comprising students who view academic tasks as more important
(higher classroom-average valuing) seemed to be contexts
conducive to greater class-average extension of interpersonal
respect and self-control (higher classroom-average social–
emotional eort). It is interesting that class-average valuing did
not signicantly predict either operative or cognitive eort at the
classroom-level. e reason for this is not clear, but there may
be something about classroom-level valuing that lends to
classroom-level interpersonal prosocial behavior (in the form of
social–emotional eort) but not classroom-level intrapersonal
behavior (in the forms of operative and cognitive eort) that
requires further investigation (see Warrington and Younger, 2011
for an example of related research identifying the role of peer
group inclusion and exclusion in school).
It was also interesting to note that student-level self-ecacy,
although predictive of students’ operative and cognitive eort, did
not signicantly predict students’ social–emotional eort. is
suggests that the belief and condence that students have about
their own ability is reected more towards the eort they invest in
their own personal application and cognition rather than towards
their inter-personal self-regulation and demonstration of respect
for others. is conrms that self-ecacy as a motivational driver
is associated more with what Bandura (2001) described as direct
personal agency, than to other-oriented agency.
Another result warranting further consideration is that of
student-level anxiety (a negative motivation factor) positively
predicting operative, social–emotional, and higher-order eort. One
could beforgiven for expecting that anxious students would bemore
avoidant or debilitated in their eort/engagement (Yang etal., 2021;
Quintero etal., 2022). However, our results indicate that anxiety is a
potentially arousing (rather than debilitating) factor—in line with
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 15 frontiersin.org
classic cognitive appraisal theories where task demands can
be perceived as challenges more than threats (e.g., Lazarus and
Folkman, 1984). Of course, another interpretation is that students
responded to their anxiety with greater eort so they could avoid the
poor performance they are anxious about (see Covington, 2000;
Martin etal., 2003). However, failure avoidance did not predict eort
at either student-or classroom-levels and so we believe we can
discount this possibility.
Notwithstanding mastery orientation and valuing, our ndings
showed that the link between motivation and eort is predominantly
manifested between students rather than between classrooms. is
is consistent with ndings of other studies demonstrating that the
majority of variance in motivation and engagement occurs at the
student-level (e.g., Martin and Marsh, 2005). At the same time,
however, there was a more consistent pattern of classroom-level
background attributes that predicted motivation and eort—and in
fact, more so for motivation than for eort. Specically, our ndings
indicated that: classroom-average ability was associated with higher
positive motivation and lower negative motivation, in line with prior
motivation research (see Elliot, 2005); classroom-average age was
negatively associated with positive motivation factors, consistent
with well-documented developmental declines in motivation (e.g.,
Jacobs etal., 2002; Gottfried etal., 2007); and, class size positively
predicted positive motivation, but also positively predicted failure
avoidance, potentially reecting the somewhat equivocal results in
class size research over the past ve decades (e.g., see Glass and
Smith, 1979; Blatchford, 2011).
Turning to the student-level background attributes, gender
was the only factor predicting both motivation and eort.
Interestingly, despite having higher positive motivation and lower
negative motivation, boys were also signicantly less likely than
girls to invest this motivation in academic eort. Indeed, other
research has also suggested that boys are higher than girls in some
aspects of motivation (perceived competence) but lower in eort
(Wilkie, 2019). Why this is the case requires further investigation,
but wesuspect answers may lie in gender-specic constructions
of eort. For example, research has shown that being seen to put
eort into academic work may not t with culturally prescribed
representations of masculinity (Connell and Messerschmidt,
2005) or what is considered “cool” for boys to do (Martino, 1999,
2000; Jackson and Dempster, 2009). Perhaps in some support of
this, our results showed that it was the more visible and observable
aspects of eort (operative and social–emotional) where boys
scored lower, not the internal (cognitive) aspect of eort. Taken
together, these ndings have highlighted some of the student and
context background attributes that are important to include in
research seeking to better understand the salient alignments and
distinctions between motivation and engagement.
Implications for theory and research
In line with the call for new expositions of engagement to
advance the eld (Reschly and Christenson 2012), we sought to
bring greater lucidity to the motivation and engagement space
through a purposeful focus on eort (a specic active form of
engagement) and how it relates to multidimensional motivation.
In this way, our ndings build on recent developments in concepts
and theory, helping to further understand the distinctiveness of
motivation and engagement, the interface between them, and the
interplay between their individual dimensions. For example, it
supported theorized distinctions between internal and external
aspects of motivation and engagement (Martin, 2012, 2022) in
that there was clear measurement and correlational distinction
between the study’s motivation and eort factors. As noted above,
ndings also shed light on what aspects of major motivational
theories [e.g., goal theory regarding mastery orientation, Elliot,
2005; (situated) expectancy-value theory regarding valuing, Eccles
and Wigeld, 2020] are associated with distinct aspects of
engagement. By introducing a novel engagement framework by
way of multidimensional eort, our ndings extend claims made
by these theories with respect to motivation and its
academic eects.
e study also oers measurement yields. To capture our
hypothesized multidimensional eort framework, wedeveloped
and established evidence for the validity of a novel instrument—
the Eort Scale—that assessed three distinct aspects of eort
(operative, cognitive, and social–emotional) in line with its
overarching umbrella construct, tripartite engagement (Fredricks
et al., 2004). is study therefore oers future researchers a
feasible new method of studying eort (as a pertinent example of
active classroom engagement). In addition to the Eort Scale, in
Supplementary material, wealso established evidence for the
validity of a parallel three-item version (the Eort Scale—Short)
that may beuseful in research where longer forms are not feasible
(e.g., in real-time research, intensive longitudinal work, etc. see
Gogol etal., 2014; Martin etal., 2020).
Implications for practice
e dominant pattern of ndings suggests the importance of
targeting self-ecacy, valuing, and mastery orientation—as these
were the main predictors of eort. Martin (2007) gives some
practical examples to develop each of these facets; for instance, the
restructuring of learning to maximize opportunities for success
may boost students’ self-ecacy, as might enhancing students’
beliefs about themselves and their academic capabilities, and
developing their skills in eective goal-setting to boost
competence. Providing students with relevance and meaning in
their learning is one way of improving valuing (Martin, 2007),
which is further enhanced by teachers modeling positive attitudes
in valuing what they teach (Eccles and Wigeld, 2002). Mastery
orientation can beenhanced by focusing students on the task at
hand more than on the assessment grade associated with it
(Martin, 2007), and also on students’ own personal learning and
progress more than how they compare and compete with other
students (Martin and Elliot, 2016).
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 16 frontiersin.org
Alongside motivational intervention as a means of enhancing
eort, it is also important to boost eort directly. is is where the
multidimensional perspective on eort is especially useful, as it
enables targeted and specic educational action. Operative eort
may besupported by teachers encouraging students to complete
schoolwork by the given deadline, emphasizing the importance of
students’ active investment of time and energy in the completion
and quality of their academic work. Teachers who regularly check
students’ work are best placed to assess their operative eort, and
in doing so, actively encourage such eort by commending
students for trying hard where applicable. Teachers can also
suitably acknowledge students’ proactive academic output that is
additional to the minimum specied task requirements,
encouraging students to engage in supplementary practice, where
appropriate, to cement understanding and techniques.
Cognitive eort may be targeted by encouraging students to
develop ‘active listening’ and attentional skills (e.g., presenting positive
indications of concentration and focus during instruction, such as
eye-contact), commending students for their focus, and reective
thinking in their comments and academic work, where appropriate.
Another strategy that can be adopted is for teachers to explicitly
promote cerebral challenge (or “brain burn”; for specic examples
appropriate to the mathematics classroom, such as the metaphoric
“brain gym” see Nagy, 2013). Teachers can promote cognitive eort by
allowing students sucient processing time before eliciting responses
to questions that arise in class discussions, aording students more
opportunity to think about questions, and formulate proactive
contributions to class discussions to clarify their developing schema.
Students should beencouraged, where appropriate, to extend their
learning by engaging in mentally challenging tasks, and to use
cognitive strategies such as visualization and self-talk. A further
strategy to improve cognitive eort is for students to increase the
duration and frequency of their quiet task-focus time, in class and at
home, including turning o mobile phones and social
media notications.
Teachers can enhance students’ social–emotional eort by
developing clear classroom expectations of mutual support and
respect and being explicit about the behaviors they want
sustained, such as interest in others’ classroom contributions,
support for others’ participation, management of impulsivity,
proactive self-regulation, and contribution to positive classroom
culture. At the same time, teachers might seek to eliminate
behaviors that are not acceptable, such as derision of others’
contributions and achievements, shouting out, talking over
others, not taking turns, and so on—so that they foster a social–
emotional classroom that is a safe environment in which to
explore and test ideas and critical thinking.
Not only does the study suggest direction on the
motivation and effort factors to target, it also provides
direction on the students and classrooms for whom boosting
motivation and effort is particularly important. For example,
the findings suggest the need to target boys’ operative and
social–emotional effort. It is also evident that girls may need
help to reduce anxiety and uncertain control, alongside
support to boost their self-efficacy and valuing. The study also
suggests improving social–emotional effort among students in
low ability classrooms and reducing failure avoidance in larger
classes. When considering these students and classrooms for
applied focus, it is worth remembering that the present study
was conducted in the mathematics domain and it is known
that this is an area where, for example, there are gender
differences in motivation (Meece etal., 2006; Watt etal., 2012)
and also motivation and engagement differences as a function
of ability (Wang and Eccles, 2013). Indeed, as discussed in
Limitations below, the extent to which this study’s findings
and practical advice apply to other subject domains remains
to beinvestigated.
Limitations and future directions
ere are some limitations to consider when interpreting
our results and which provide potential direction for future
research. First, the correlational approach in this study cannot
beinterpreted as supporting causal conclusions. Experimental
and longitudinal designs are required to establish the causal
ordering of motivation and engagement (eort) implied in our
research. Second, as noted above, our study targeted motivation
and engagement in mathematics. Further research is needed to
verify the extent to which results are replicated in other subject
areas. ird, data were collected via self-reports, reecting
students’ perceptions of their motivation and eort. Recent
research (e.g., Collie and Martin, 2017) has highlighted the
importance of garnering perspectives from multiple informants
(e.g., in this case, the students and their teachers) in order to
obtain a more comprehensive understanding of a target
construct like eort, that comprises both internal and external
facets. Fourth, although weestablished illuminating associations
between motivation and eort, an extension of the present study
might investigate a fuller process, such as including achievement
in our hypothesized process as a consequence of eort. Fih,
motivation was assessed using single-item indicators, modelled
as error-adjusted scores using established reliability and variance
measures from a prior research program. Further studies might
consider using multiple-item latent motivation measures to
ensure greater measurement accuracy. Sixth, our analyses were
based on variable-centered techniques (MCFA, MSEM) which
highlight associations between variables at a whole-sample level,
but may mask important ndings that are pertinent to
subpopulations within the sample. Person-centered techniques
such as latent prole analysis may identify eort proles among
particular subpopulations of students that are not evident in
variable-centered approaches. Seventh, data collection took
place at the end of 2020, in the rst year of the COVID-19
pandemic. ere was some minor disruption to learning earlier
in the year, but Australian schools had returned to face-to-face
learning for a period of 6 months prior to data collection, and as
such wedo not expect this to have signicantly impacted results.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 17 frontiersin.org
Finally, our sample comprised Year 7–10 Australian high-school
students from independent schools. It is important to expand the
age-range, national context, and type of school sampled in future
studies to establish the generality of the present ndings.
Conclusion
is study sought to shed further light on the unique and
shared variance between motivation and engagement (by way of
eort). e ndings have provided several avenues of focus for
subsequent motivation research and theorizing. ey have also
established evidence for the validity of a novel engagement
framework (multidimensional eort) that may support future
measurement and practice in academic engagement. In so doing,
the research presented here oers greater clarity and coherence for
educational researchers and practitioners in their approaches to
optimizing students’ academic development.
Data availability statement
e datasets presented in this article are not readily available
because consent from participants to share the dataset is not
available. Summative data are available and can be requested.
Requests to access the datasets should be directed to AM; andrew.
martin@unsw.edu.au.
Ethics statement
e studies involving human participants were reviewed and
approved by UNSW Human Ethics Committee (Approval
#HC200273). Written informed consent to participate in this
study was provided by the participants’ legal guardian/next of kin.
Author contributions
RN shared in the development of the research and materials
design and led data analysis and report writing. AMand RC
shared in the development of the research and materials design
and assisted with data analysis and report writing. All authors
contributed to the article and approved the submitted version.
Acknowledgments
We acknowledge the Australian Government RTP
Scholarship and NSW Department of Education Waratah
Scholarship in support of RN’s Ph.D research into high school
students’ eort.
Conflict of interest
One of the measures (the MES) in the study was developed by
the second author, AM, and is a published instrument attracting
a small fee, of which a part is put towards its ongoing development
and administration, and part of which is also donated to
UNICEF. However, for this study, there was no fee involved for
its use.
e remaining authors declare that the research was conducted
in the absence of any commercial or nancial relationships that
could be construed as a potential conict of interest
e handling editor FG declared a past co-authorship with the
authors AM and RC.
Publisher’s note
All claims expressed in this article are solely those of the
authors and do not necessarily represent those of their aliated
organizations, or those of the publisher, the editors and the
reviewers. Any product that may be evaluated in this article, or
claim that may be made by its manufacturer, is not guaranteed or
endorsed by the publisher.
Supplementary material
e Supplementary material for this article can befound online
at: https://www.frontiersin.org/articles/10.3389/fpsyg.2022.1045717/
full#supplementary-material
References
Ainley, M. (2012). “Students’ interest and engagement in classroom activities,” in
Handbook of Research on Student Engagement. eds. S. L. Christenson, A. L. Reschly
and C. Wylie (New York, NY: Springer), 283–302.
Ames, C. (1992). Classrooms: goals, structures, and student motivation. J. Educ.
Psychol. 84, 261–271. doi: 10.1037/0022-0663.84.3.261
Appleton, J. J., Christenson, S. L., Kim, D., and Reschly, A. L. (2006). Measuring
cognitive and psychological engagement: validation of the student engagement
instrument. J. Sch. Psychol. 44, 427–445. doi: 10.1016/j.jsp.2006.04.002
Australian Bureau of Statistics (2016). Census of population and housing: Socio-
Economic Indexes for Areas (SEIFA), Australia, 2016. Available at: https://www.abs.
gov.au/websitedbs/censushome.nsf/home/seifa
Australian Curriculum Assessment and Reporting Authority (n.d.). Foundation
to year 10 curriculum: Mathematics. Available at: https://www.australiancurriculum.
edu.au/f-10-curriculum/mathematics/
Bandura, A. (2001). Social cognitive theory: an agentic perspective. Annu. Rev.
Psychol. 52, 1–26. doi: 10.1146/annurev.psych.52.1.1
Bandura, A. (2012). Social cognitive theory in Handbook of eories of Social
Psychology 1. (eds.) P. A. M. Van Lange, A. W. Kruglanski and E. T. Higgins (London:
SAGE Publications Ltd.), 349–374
Blatchford, P. (2011). “Three generations of research on class-size
effects,” in APA Educational Psychology Handbook, Vol 2: Individual
Differences and Cultural and Contextual Factors, eds. Sandra Graham, J. M.
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 18 frontiersin.org
Royer, and M. Zeidner (Washington DC: American Psychological Association),
530–554.
Byrne, B. M. (2012). Structural Equation Modeling With Mplus: Basic Concepts,
Applications, and Programming Routledge.
Carbonaro, W. (2005). Tracking, students’ eort, and academic achievement.
Sociol. Educ. 78, 27–49. doi: 10.1177/003804070507800102
Christenson, S. L., Reschly, A. L., and Wylie, C. (eds.) (2012). Handbook of
Research on Student Engagement. New York, NY: Springer
Cleary, T. J., and Zimmerman, B. J. (2012). “A cyclical self-regulatory account of
student engagement: theoretical foundations and applications,” in Handbook of
Research on Student Engagement. eds. S. L. Christenson, A. L. Reschly and C. Wylie
(New York, NY: Springer), 237–257.
Cole, D. A., and Preacher, K. J. (2014). Manifest variable path analysis: potentially
serious and misleading consequences due to uncorrected measurement error.
Psychol. Methods 19, 300–315. doi: 10.1037/a0033805
Collie, R. J., and Martin, A. J. (2017). Students’ adaptability in mathematics:
examining self-reports and teachers’ reports and links with engagement and
achievement outcomes. Contemp. Educ. Psychol. 49, 355–366. doi: 10.1016/j.
cedpsych.2017.04.001
Collie, R. J., Martin, A. J., Bobis, J., Way, J., and Anderson, J. (2019). How students
switch on and switch o in mathematics: exploring patterns and predictors of (dis)
engagement across middle school and high school. Educ. Psychol. 39, 489–509. doi:
10.1080/01443410.2018.1537480
Connell, R. W., and Messerschmidt, J. W. (2005). “Hegemonic masculinity
rethinking the concept, Gender and Society, 19, 829–859. doi: 10.1177/
0891243205278639
Covington, M. V. (2000). Goal theory, motivation, and school achievement: an
integrative review. Annu. Rev. Psychol. 51, 171–200. doi: 10.1146/annurev.
psych.51.1.171
Eccles, J. S., and Wigeld, A. (2002). Motivational beliefs, values, and goals. Annu.
Rev. Psychol. 53, 109–132. doi: 10.1146/annurev.psych.53.100901.135153
Eccles, J. S., and Wigeld, A. (2020). From expectancy-value theory to situated
expectancy-value theory: a developmental, social cognitive, and sociocultural
perspective on motivation. Contemp. Educ. Psychol. 61:101859. doi: 10.1016/j.
cedpsych.2020.101859
Elliot, A. J. (2005). “A conceptual history of the achievement goal construct,” in
Handbook of Competence and Motivation. eds. A. J. Elliot and C. S. Dweck. 1st ed
(New York, NY: e Guilford Press), 52–72.
Fredricks, J. A., Blumenfeld, P. C., and Paris, A. H. (2004). School engagement:
potential of the concept, state of the evidence. Rev. Educ. Res. 74, 59–109. doi:
10.3102/00346543074001059
Glass, G. V., and Smith, M. L. (1979). Meta-analysis of research on class size and
achievement. Educ. Eval. Policy Anal. 1, 2–16. doi: 10.3102/01623737001001002
Gogol, K., Brunner, M., Goetz, T., Martin, R., Ugen, S., Keller, U., et al. (2014). My
questionnaire is too long! e assessments of motivational-aective constructs with
three-item and single-item measures. Contemp. Educ. Psychol. 39, 188–205. doi:
10.1016/j.cedpsych.2014.04.002
Gottfried, A. E., Marcoulides, G. A., Gottfried, A. W., Oliver, P. H., and
Guerin, D. W. (2007). Multivariate latent change modeling of developmental
decline in academic intrinsic math motivation and achievement: childhood
through adolescence. Int. J. Behav. Dev. 31, 317–327. doi: 10.1177/
0165025407077752
Hayduk, L. A. (1987). Structural Equation Modelling With LISREL: Essentials and
Advances, Baltimore, MD: e Johns Hopkins University Press.
Hu, L. T., and Bentler, P. M. (1999). Cuto criteria for t indexes in covariance
structure analysis: conventional criteria versus new alternatives. Struct. Equ. Model.
6, 1–55. doi: 10.1080/10705519909540118
Jackson, C., and Dempster, S. (2009). I sat back on my computer … with a bottle
of whisky next to me: constructing ‘cool’ masculinity through ‘eortless’ achievement
in secondary and higher education. J. Gend. Stud. 18, 341–356. doi:
10.1080/09589230903260019
Jacobs, J. E., Lanza, S., Osgood, D. W., Eccles, J. S., and Wigeld, A. (2002). Changes
in children’s self-competence and values: gender and domain dierences across
grades one through twelve. Child Dev. 73, 509–527. doi: 10.1111/1467-8624.00421
Kim, H.-Y. (2013). Statistical notes for clinical researchers: assessing normal
distribution (2) using skewness and kurtosis. Restor. Dentis. Endodon. 38:52. doi:
10.5395/rde.2013.38.1.52
Lazarus, R. S., and Folkman, S. (1984). Stress, Appraisal, and Coping New York,
NY: Springer Publishing Company.
Lei, H., Cui, Y., and Zhou, W. (2018). Relationships between student engagement
and academic achievement: a meta-analysis. Soc. Behav. Pers. 46, 517–528. doi:
10.2224/sbp.7054
Liem, G. A. D., and Martin, A. J. (2012). e motivation and engagement scale:
theoretical framework, psychometric properties, and applied yields. Aust. Psychol.
47, 3–13. doi: 10.1111/j.1742-9544.2011.00049.x
Marsh, H. W. (1997). “e measurement of physical self-concept: a construct
validation approach,” in e Physical Self: From Motivation to Well-Being. ed. K. R.
Fox (Champaign, IL: Human Kinetics), 27–58.
Marsh, H. W. (2002). A multidimensional physical self-concept: a construct
validity approach to theory, measurement and research. Psychology 9, 459–493. doi:
10.12681/psy_hps.24074
Marsh, H. W., Martin, A. J., and Cheng, J. H. S. (2008). A multilevel
perspective on gender in classroom motivation and climate: potential benefits
of male teachers for boys? J. Educ. Psychol. 100, 78–95. doi:
10.1037/0022-0663.100.1.78
Martin, A. J. (1999–2022). e motivation and engagement scale. Lifelong
Achievement Group. Available at: www.lifelongachievement.com
Martin, A. J. (2007). Examining a multidimensional model of student motivation
and engagement using a construct validation approach. Br. J. Educ. Psychol. 77,
413–440. doi: 10.1348/000709906X118036
Martin, A. J. (2009). Motivation and engagement across the academic life span.
Educ. Psychol. Meas. 69, 794–824. doi: 10.1177/0013164409332214
Martin, A. J. (2012). “Part II commentary: motivation and engagement:
conceptual, operational, and empirical clarity,” in Handbook of Research on Student
Engagement. eds. S. L. Christenson, A. L. Reschly and C. Wylie (New York, NY:
Springer), 303–311.
Martin, A. J. (2022). “e role of academic engagement in mediating the link
between instruction and academic outcomes: new insights from load reduction
instruction and the 4M academic engagement framework,” in Handbook of Research
on Student Engagement. eds. S. Christenson and A. Reschly. 2nd ed (Cham,
Switzerland: Springer).
Martin, A. J., and Elliot, A. J. (2016). e role of personal best (PB) goal setting in
students’ academic achievement gains. Learn. Individ. Dier. 45, 222–227. doi:
10.1016/j.lindif.2015.12.014
Martin, A. J., Ginns, P., and Papworth, B. (2017). Motivation and engagement:
same or dierent? Does it matter? Learn. Individ. Dier. 55, 150–162. doi: 10.1016/j.
lindif.2017.03.013
Martin, A. J., Mansour, M., and Malmberg, L.-E. (2020). What factors inuence
students’ real-time motivation and engagement? An experience sampling study of
high school students using mobile technology. Educ. Psychol. 40, 1113–1135. doi:
10.1080/01443410.2018.1545997
Martin, A. J., and Marsh, H. (2005). Motivating boys and motivating girls: does
teacher gender really make a dierence? Aust. J. Educ. 49, 320–334. doi:
10.1177/000494410504900308
Martin, A. J., and Marsh, H. W. (2008). Workplace and academic buoyancy:
psychometric assessment and construct validity amongst school personnel and
students. J. Psychoeduc. Assess. 26, 168–184. doi: 10.1177/0734282907313767
Martin, A. J., Marsh, H. W., and Debus, R. L. (2003). Self-handicapping and
defensive pessimism: a model of self-protection from a longitudinal perspective.
Contemp. Educ. Psychol. 28, 1–36. doi: 10.1016/S0361-476X(02)00008-5
Martino, W. (1999). “Cool boys”, “party animals”, “squids” and “pooers”:
interrogating the dynamics and politics of adolescent masculinities in school. Br. J.
Sociol. Educ. 20, 239–263. doi: 10.1080/01425699995434
Martino, W. (2000). Mucking around in class, giving crap, and acting cool:
adolescent boys enacting masculinites at school. Can. J. Educ. 25, 102–112.
McClelland, D. C. (1961). e Achieving Society, New York, NY: Free Press.
McNeish, D. (2018). anks coecient alpha, We’ll take it from here. Psychol.
Methods 23, 412–433. doi: 10.1037/met0000144
Meece, J. L., Glienke, B. B., and Burg, S. (2006). Gender and motivation. J. Sch.
Psychol. 44, 351–373. doi: 10.1016/j.jsp.2006.04.004
Morgan, L., Close, S., Siller, M., Kushner, E., and Brasher, S. (2022). Teachers’
experiences: social emotional engagement–knowledge and skills. Educ. Res. 64,
41–59. doi: 10.1080/00131881.2021.1988865
Morin, A. J. S., Marsh, H. W., Nagengast, B., and Scalas, L. F. (2014). Doubly latent
multilevel analyses of classroom climate: an illustration. J. Exp. Educ. 82, 143–167.
doi: 10.1080/00220973.2013.769412
Muthén, L. K., and Muthén, B. O. (1998–2022). Mplus user’s guide. Available at:
www.StatModel.com
Nagy, R. P. (2013). Building a solid foundation from which to launch our future
mathematicians. Aust. Math. Teach. 69, 20–25.
Nagy, R. P. (2016). Tracking and visualizing student eort: evolution of a practical
analytics tool for sta and student engagement. J. Learn. Analy. 3, 165–193. doi:
10.18608/jla.2016.32.8
Nagy et al. 10.3389/fpsyg.2022.1045717
Frontiers in Psychology 19 frontiersin.org
Nagy, R. P. (2017). “Tracking and visualising student eort: a practical analytics tool
for student engagement.” in Excellence in Professional Practice Conference 2017: Case
Studies of Practice. (ed.) L. Rosman, Australian Council for Educational Research,
53–62. Available at: https://www.acer.org/les/EPPC_Case_Study_Digital.pdf
Nagy, R. P. (2021). “e high school mathematics competency scale,” in e role
of eort in high school students’ academic development: A longitudinal perspective
in mathematics and English. Doctoral thesis in preparation. University of New
South Wales. doi: 10.17605/OSF.IO/2GKAF
Nagy, R. P. (2022). e role of eort in high school students’ academic
development: A longitudinal perspective in mathematics and English. Doctoral
thesis in preparation. University of New South Wales.
NSW Education Standards Authority (2019). “NSW syllabus for the Australian
curriculum: mathematics K-10.” in NSW Syllabus for the Australian Curriculum.
Available at: https://educationstandards.nsw.edu.au/wps/portal/nesa/k-10/learning-
areas/mathematics/mathematics-k-10
Quintero, M., Hasty, L., Li, T., Song, S., and Wang, Z. (2022). A multidimensional
examination of math anxiety and engagement on math achievement. Br. J. Educ.
Psychol. 92, 955–973. doi: 10.1111/bjep.12482
Reeve, J. (2012). “A self-determination theory perspective on student engagement,”
in Handbook of Research on Student Engagement. eds. S. L. Christenson, A. L.
Reschly and C. Wylie (New York, NY: Springer), 149–172.
Reschly, A. L., and Christenson, S. L. (2012). “Jingle, jangle, and conceptual
haziness: evolution and future directions of the engagement construct,” in Handbook
of Research on Student Engagement. eds. S. L. Christenson, A. L. Reschly and C.
Wylie (New York, NY: Springer), 3–19.
Ryan, R. M., and Deci, E. L. (2017). Self-Determination eory: Basic Psychological
Needs in Motivation, Development, and Wellness, New York, NY: Guilford Publications.
Schunk, D. H., and Mullen, C. A. (2012). “Self-ecacy as an engaged learner,” in
Handbook of Research on Student Engagement. eds. S. L. Christenson, A. L. Reschly
and C. Wylie (New York, NY: Springer), 219–235.
omson, S., de Bortoli, L., and Underwood, C. (2016). PISA 2015: a rst look at
Australia’s results. Australian Council for Educational Research. Available at: https://
research.acer.edu.au/ozpisa/21/
omson, S., de Bortoli, L., Underwood, C., and Schmid, M. (2019). PISA 2018:
reporting Australia’s results. Volume Istudent performance. Available at: https://
research.acer.edu.au/ozpisa/35/
Voelkl, K. E. (2012). “School identication,” in Handbook of Research on Student
Engagement. eds. S. L. Christenson, A. L. Reschly and C. Wylie (New York, NY:
Springer), 193–218.
Wang, M.-T., and Eccles, J. S. (2012). Adolescent behavioral, emotional, and
cognitive engagement trajectories in school and their dierential relations to
educational success. J. Res. Adolesc. 22, 31–39. doi: 10.1111/j.1532-7795.2011.
00753.x
Wang, M.-T., and Eccles, J. S. (2013). School context, achievement motivation, and
academic engagement: a longitudinal study of school engagement using a
multidimensional perspective. Learn. Instr. 28, 12–23. doi: 10.1016/j.learninstruc.
2013.04.002
Warrington, M., and Younger, M. (2011). Life is a tightrope: reections on peer
group inclusion and exclusion amongst adolescent girls and boys. Gend. Educ. 23,
153–168. doi: 10.1080/09540251003674121
Watt, H. M. G., Shapka, J. D., Morris, Z. A., Durik, A. M., Keating, D. P., and
Eccles, J. S. (2012). Gendered motivational processes aecting high school
mathematics participation, educational aspirations, and career plans: a comparison
of samples from Australia, Canada, and the United States. Dev. Psychol. 48,
1594–1611. doi: 10.1037/a0027838
Weiner, B. (2010). e development of an attribution-based theory of motivation:
a history of ideas. Educ. Psychol. 45, 28–36. doi: 10.1080/00461520903433596
Wentzel, K. R. (2012). “Part III commentary: socio-cultural contexts, social
competence, and engagement at school,” in Handbook of Research on Student
Engagement. eds. S. L. Christenson, A. L. Reschly and C. Wylie (New York, NY:
Springer), 479–488.
Wentzel, K. R., Muenks, K., McNeish, D., and Russell, S. (2017). Peer and teacher
supports in relation to motivation and eort: a multi-level study. Contemp. Educ.
Psychol. 49, 32–45. doi: 10.1016/j.cedpsych.2016.11.002
Wigeld, A., and Cambria, J. (2010). Students’ achievement values, goal
orientations, and interest: denitions, development, and relations to achievement
outcomes. In. Dev. Rev. 30, 1–35. doi: 10.1016/j.dr.2009.12.001
Wigeld, A., and Eccles, J. S. (2000). Expectancy–value theory of achievement
motivation. Contemp. Educ. Psychol. 25, 68–81. doi: 10.1006/ceps.1999.1015
Wilkie, K. J. (2019). “Middle school boys’ and girls’ own expressions of aspirations
for their mathematics learning,” in Aect and Mathematics Education: Fresh
Perspectives on Motivation, Engagement, and Identity. eds. M. S. Hannula, G. C.
Leder, F. Morselli, M. Vollstedt and Q. Zhang (Cham, Switzerland: Springer),
165–194.
Yang, X., Zhang, M., Kong, L., Wang, Q., and Hong, J.-C. (2021). e eects of
scientic self-ecacy and cognitive anxiety on science engagement with the
“question-observation-doing-explanation” model during school disruption in
COVID-19 pandemic. J. Sci. Educ. Technol. 30, 380–393. doi: 10.1007/
s10956-020-09877-x
Available via license: CC BY
Content may be subject to copyright.
Content uploaded by Andrew J Martin
Author content
All content in this area was uploaded by Andrew J Martin on Dec 13, 2022
Content may be subject to copyright.
