Motivation and Engagement Across the Academic Life SpanA Developmental Construct Validity Study of Elementary School, High School, and University/College Students

Abstract

From a developmental construct validity perspective, this study examines motivation and engagement across elementary school, high school, and university/college, with particular focus on the Motivation and Engagement Scale (comprising adaptive, impeding/maladaptive, and maladaptive factors). Findings demonstrated developmental construct validity across the three distinct educational stages in terms of goodfitting first- and higher order factors, invariance of factor structure across gender and age, and a pattern of correlations with cognate constructs (e.g., homework completion, academic buoyancy, class participation) consistent with predictions. Notwithstanding the predominantly parallel findings, there was also notable distinctiveness, primarily in terms of mean-level effects, such that elementary school students were generally more motivated and engaged than university/college students who in turn were more motivated and engaged than high school students. Implications for motivation and engagement measurement and theory, research in the psychoeducational domain, and the subsequent potential for performance profiling across the academic life span are discussed.

Full text

0
Martin, A.J. (2009). Motivation and engagement across the academic lifespan: A developmental construct
validity study of elementary school, high school, and university/college students. Educational and
Psychological Measurement, 69, 794-824. DOI: 10.1177/0013164409332214.
This article may not exactly replicate the authoritative document published in the journal. It is not the copy
of record. The exact copy of record can be accessed via the DOI: 10.1177/0013164409332214.

1
Motivation and Engagement Across the Academic Lifespan:
A Developmental Construct Validity Study of Elementary School, High School, and
University/College Students
Andrew J. Martin
KEYWORDS: construct validity; developmental; measurement; motivation; engagement; students
Author Note. This article was in part prepared while the first author was Visiting Senior Research
Fellow in the Department of Education at the University of Oxford. Requests for further
information about this investigation can be made to Associate Professor Andrew J. Martin, Faculty
of Education and Social Work, A35 – Education Building, University of Sydney, NSW 2006,
AUSTRALIA. E-Mail: a.martin@edfac.usyd.edu.au.

2
Motivation and Engagement Across the Academic Lifespan:
A Developmental Construct Validity Study of Elementary School, High School, and
University/College Students
Abstract
From a ‘developmental construct validity’ perspective, the present study examines motivation and
engagement across elementary school, high school, and university/college with particular focus on
the Motivation and Engagement Scale (comprising adaptive, impeding/maladaptive, and
maladaptive factors). Findings demonstrated developmental construct validity across the three
distinct educational stages in terms of good-fitting first and higher order factors, invariance of factor
structure across gender and age, and a pattern of correlations with cognate constructs (e.g.,
homework completion, academic buoyancy, class participation) consistent with predictions.
Notwithstanding the predominantly parallel findings, there was also notable distinctiveness,
primarily in terms of mean-level effects such that elementary school students were generally more
motivated and engaged than university/college students who in turn were more motivated and
engaged than high school students. Implications for motivation and engagement measurement and
theory, research in the psycho-educational domain, and the subsequent potential for performance
profiling across the academic lifespan are discussed.

1
Motivation and Engagement Across the Academic Lifespan:
A Developmental Construct Validity Study of Elementary School, High School, and
University/College Students
Students within elementary school, high school, and university/college share a great deal in
common. In each context students are required to apply themselves over a sustained period of time
to develop their academic skills, engage with key performance demands, negotiate the rigors of
competition, deal with setback and adversity, cope with possible self-doubt and uncertainty, and
develop psychological and behavioral skills to effectively manage the ups and downs in the
ordinary course of academic life. Given these congruencies across distinct educational stages, it is
feasible to propose that there will be core and common constructs relevant and meaningful across
the academic lifespan. The present study seeks to assess this issue in the context of academic
motivation and engagement and, more specifically, the validity of recently developed academic
motivation and engagement instrumentation in the context of students from elementary school, high
school, and university/college. Analyses conducted in the present investigation across these three
distinct educational stages are proposed as a ‘developmental construct validity’ study of academic
motivation and engagement.
Substantive Background: An Integrative Framework for Motivation and Engagement and
Implications for Measurement
The substantive background to the study centers on academic motivation and engagement and
the need for more pragmatic and integrative approaches to their measurement and theorizing. In
critical reviews of motivation and engagement research, it has been suggested that such research
oftentimes yields limited practical implications and applications and that there is a need to devise
research that advances scientific understanding but which also has applied utility. Hence, there have
been calls to give greater attention to use-inspired basic research in education and psychology
contexts (Stokes, 1997; see also Greeno, 1998; Pintrich, 2000, 2003). Critical reviews of motivation
and engagement research also point to the fact that such research is diverse and fragmented. As a

2
result, there have also been calls for more integrative approaches to its research and theorizing
(Bong, 1996; Murphy & Alexander, 2000; Pintrich, 2003). It is in this context that the Motivation
and Engagement Wheel (Martin, 2001, 2002, 2007a) was developed. The Wheel is presented in
Figure 1.
As Figure 1 shows, there are two levels at which the Wheel has been conceptualized: the
integrative higher order level comprising four factors and the lower (or first) order level comprising
eleven factors. As discussed fully in Martin (2007a, 2008a, 2008b), higher (and first) order factors
are adaptive cognitions (self-efficacy, valuing, mastery orientation), adaptive behaviors (planning,
task management, persistence), impeding/maladaptive cognitions (anxiety, failure avoidance,
uncertain control), and maladaptive behaviors (self-handicapping, disengagement). Initially this
Wheel was developed to better understand motivation and engagement amongst high school
students; however, in the present study its application to elementary school and university students
is assessed from a developmental construct validity perspective (described below).
Higher Order Dimensions of Motivation and Engagement
Martin (2007a, 2008a, 2008b) proposed that over the past four decades a number of
psychological theories and models have been developed that explain the nature of human cognition
and behavior. He demonstrated that there are significant commonalities across these theories and
models and which provide direction as to fundamental (higher order) dimensions of motivation and
engagement. These commonalities operate at three levels.
The first delineates cognitive and behavioral elements, including work encompassing
cognitive and behavioral orientations in learning strategies (Pintrich & DeGroot, 1990; Pintrich &
Garcia, 1991), cognitive antecedents of behavioral strategies used to negotiate environmental
demands (Buss & Cantor, 1989), cognitive-behavioral approaches to engagement and behavior
change (Beck, 1995), and cognitive-affective and behavioral dimensions to academic engagement
(Miller et al, 1996; Miserandino, 1996). The second demonstrates the differential empirical strength of
distinct aspects of motivation and engagement—for example, self-efficacy reflects highly adaptive
motivation (Bandura, 1997; Pajares, 1996), anxiety impedes individuals’ engagement (Sarason &

3
Sarason, 1990; Spielberger, 1985), and behaviors such as self-handicapping reflect quite maladaptive
engagement (Martin, Marsh, & Debus, 2001a, 2001b, 2003; Martin, Marsh, Williamson, & Debus,
2003). The third informs the structure of motivation and engagement frameworks, such as those
hypothesizing and empirically demonstrating hierarchical models of human cognition and behavior
that encompass specific factors under more global characterizations (e.g., Elliot & Church, 1997;
Marsh & Shavelson, 1985; Shavelson, Hubner, & Stanton, 1976).
Taken together and in consideration of the joint issues of: motivational and behavioral
orientations; cognitive-behavioral frameworks; differing empirical levels of adaptive, impeding, and
maladaptive dimensions in applied settings; and, hierarchical models of cognition and behavior, Martin
(2007a, 2008a, 2008b) proposed that motivation can be characterized in terms of four higher order
dimensions: (a) adaptive cognition, (b) adaptive behavior, (c) impeding/maladaptive cognition, and (d)
maladaptive behavior. These dimensions and their component first order factors have been synthesized
under the Motivation and Engagement Wheel (Martin, 2001, 2003a, 2003c, 2007a, 2008b) presented in
Figure 1.
First Order Dimensions of Motivation and Engagement
Pintrich (2003) identified core substantive questions for the development of a motivational
science. Taken together, these questions underscore the importance of considering, conceptualizing,
and articulating a model of motivation from salient and seminal theorizing related to: self-efficacy,
control, valuing, goal orientation, need achievement, self-worth, and self-regulation. These, it is
suggested, provide a useful heuristic for the identification of first order constructs for
operationalizing the Motivation and Engagement Wheel.
As discussed fully in Martin (2001, 2002, 2003c, 2007a), (a) self-efficacy theory (e.g.,
Bandura, 1997) is reflected in the self-efficacy dimension of the Wheel, (b) attributions and control
are reflected in the uncertain control dimension (tapping the controllability element of
attributions—see Connell, 1985; Weiner, 1994), (c) valuing (e.g., Eccles, 1983; Wigfield & Tonks,
2002) is reflected in a valuing dimension, (d) self-determination (in terms of intrinsic motivation—
see Ryan & Deci, 2000) and motivation orientation (see Dweck, 1986; Martin & Debus, 1998;

4
Nicholls, 1989) are reflected in a mastery orientation dimension, (e) self-regulation (e.g., Martin,
2001, 2002, 2003c, 2007a; Martin et al., 2001a, 2001b, 2003; Zimmerman, 2002) is reflected in
planning, task management, and persistence dimensions, and (f) need achievement and self-worth
(e.g., Atkinson 1957; Covington, 1992; Martin & Marsh, 203; McClelland, 1965) are reflected in
failure avoidance, anxiety, self-handicapping, and disengagement dimensions, and Hence, the
Wheel comprises eleven lower or first-order dimensions – see Figure 1.
Measurement and the Motivation and Engagement Scale
Alongside the Motivation and Engagement Wheel is its accompanying instrumentation the
Motivation and Engagement Scale (MES). Typically administered to high school students, the
Motivation and Engagement Scale – High School (MES-HS; Martin, 2001, 2003c, 2007a, 2007b,
2008a) demonstrates a strong factor structure that is invariant across gender and age (but there are
mean-level differences such that females generally report higher levels of motivation than males
and middle high school students report lower motivation than junior and senior high school
students) and is reliable and normally-distributed. It has also been found to predict a variety of
educational outcomes such as enjoyment of school, classroom participation, educational aspirations
as well as achievement-related outcomes such as school grades. To extend this line of research, the
present investigation assesses a parallel form of the MES using the Motivation and Engagement
Scale – Junior School (MES-JS) and Motivation and Engagement Scale – University/College
(MES-UC).
Over the past few years, there has been growing research around the Motivation and
Engagement Wheel and its accompanying instrumentation, the Motivation and Engagement Scale.
The MES is robust in high school (Martin, 2007a), workplace (Martin, in press b; see also Martin
2005b, 2005c), music (Martin, 2008b), sport (Martin, 2008b), and physical activity domains
(Martin, Tipler and colleagues, 2006). The Wheel and MES are useful as bases for educational
intervention (Martin, 2005a, 2008b). The Wheel and MES are helpful foundations for assessing
group-level (climate) effects (Martin & Marsh, 2005). Finally, the Wheel and MES are useful in

5
addressing more specific educational issues such as domain specificity (Green, Martin, & Marsh,
2007), teacher effects (Martin & Marsh, 2005), and the role of parents and teachers in the
motivation and engagement process (Martin, 2003b, 2006). However, to date, there has been no
thoroughgoing and detailed scoping of the Wheel and MES across the span of education – that is,
across elementary school, high school, and universities samples (but see Martin, in press b, for brief
research in the context of sport, music, work, and daily life motivation and engagement). The
present study does so from a proposed developmental construct validity perspective.
Methodological Background: A Developmental Construct Validity Perspective
Researchers in psychology and education have increasingly emphasized the need to develop
and evaluate instruments within a construct validation framework (e.g., see Marsh, 2002; Marsh &
Hau, 2007). Investigations that adopt a construct validation approach can be classified as within-
network or between-network studies. Moreover, it is proposed here that when construct validity is
assessed across distinct educational stages it constitutes something of a developmental construct
validity perspective. Specifically, it is proposed that a dual within- and between-network approach
across elementary school, high school, and university represents a developmental construct validity
approach to assessing the generality of motivation and engagement across the academic lifespan.
Within-network Validity
Beginning with a logical analysis of internal consistency of the construct definition,
measurement instruments, and generation of predictions, within-network studies typically employ
empirical techniques such as exploratory factor analysis (EFA), confirmatory factor analysis (CFA),
and reliability analysis. The present study conducts within-network analyses across the three
samples using confirmatory factor analysis to test the multidimensional motivation and engagement
framework and reliability analysis to test the internal consistency of scores. Consistent with
previous studies of high school students (e.g., Green et al, 2007; Martin, 2001, 2003c, 2007a) and
across diverse performance settings such as music and sport (Martin, 2008b), it is hypothesized that
at each educational stage (elementary school, high school, and university), the motivation and

6
framework instrumentation (MES) will evince a sound first and higher order factor structure and
comprise reliable scores.
Between-network Validity
Between-network research explores relationships between a target central framework and a
set of factors external to the framework. It typically does so through statistical procedures such as
correlation, regression, or structural equation modeling (SEM) analyses to examine relationships
between measures and instruments. The present study conducts between-network analyses across
the three samples by assessing: (a) the invariance of factor structure across gender, age groups, and
educational stages (elementary school, high school, university/college), (b) mean-level differences
across educational stages, and (c) the empirical links between the hypothesized first and higher
order factors and a set of cognate between-network measures (enjoyment of school/university, class
participation, positive intentions, academic buoyancy, homework/assignment completion). Each of
these between-network techniques is described in turn.
Factorial invariance in the structure of motivation and engagement. As described in Martin
(2007a, 2008b), insufficient attention is given to analyses of factor structure of motivation and
engagement and the extent to which a given motivation and engagement instrument and its
components are invariant across different groups. Such concerns about factor structure invariance
are most appropriately evaluated using CFA to determine whether—and how—the structure of
motivation and engagement vary according to key sub-populations (see Hattie, 1992; Marsh, 1993).
Martin (2004, 2007a) has previously shown the MES factor structure (factor loadings,
uniquenesses, correlations/variances) to be invariant across early-, mid-, and late-adolescent
samples and also across gender. The present study is an opportunity to assess invariance across
gender and age within elementary school and university. It is also an opportunity to assess
invariance across elementary school, high school, and university samples. Consistent with previous
studies of high school students (e.g., Green et al, 2007; Martin, 2001, 2003c, 2007a) and across
diverse performance settings such as music and sport (Martin, 2008b), it is hypothesized that factor

7
structure (including loadings, correlations/variances, and uniquenesses) across gender, age, and
educational stage will evince relative invariance.
Mean-level educational stage effects. Very little research has assessed mean levels of
motivation and engagement across the academic lifespan: elementary school, high school, and
university. The transition from elementary to middle school has been found to pose difficulties and
challenges unique to that time (Anderman & Midgley, 1997; Roeser, Eccles, & Sameroff, 2000) and
a decline in student motivation and engagement is typically found to emerge after this transition
(see Martin, 2001, 2003c, 2004, 2007a; Wigfield & Tonks, 2002) including changes in subjective
task value (Wigfield, Eccles, Mac Iver, Reuman, & Midgley, 1991). As students move on to
university/college some research has found them to be more confident in the quantity and quality of
their abilities whereas other research finds it a difficult transition with less support and structure and
a major challenge in asserting one’s identity amongst highly capable peers (Martin, Marsh,
Williamson, & Debus, 2003). Increasingly, universities and colleges are recognizing the stresses
and strains of undergraduate life and the difficulties in making a successful transition from high
school (see Martin, Milne-Home, Barrett, & Spalding, 1997; Martin, Milne-Home, Barrett,
Spalding, & Jones, 2000). Indeed, Martin and colleagues (Martin, Marsh, Williamson, & Debus,
2003) have found university to present distinct challenges that instill doubts and uncertainties that in
some cases lead to self-handicapping, poorer academic performance, and eventual dropout. Taken
together, then, it is hypothesized that elementary school students will evince relatively higher mean
levels of motivation and engagement than high school and university samples, however, no
predictions are made regarding the relative mean levels of the latter two groups.
Motivation, engagement, and cognate correlates. Consistent with the construct validity
approach, it is proposed that five between-network constructs provide a theoretically relevant basis
for examining the external validity of the MES across the academic lifespan: positive intentions,
class participation, enjoyment of school, academic buoyancy, and homework/assignment
completion. In terms of positive intentions, several researchers have shown that students higher in
motivation and engagement are more likely to take advanced or optional courses and also more

8
likely to report future course enrolment intentions (Meece, Wigfield, & Eccles, 1990). In addition to
positive intentions, class participation is deemed a feasible between-network construct. Learning
environments that foster student participation are found to enhance students’ commitment to
learning (Richter & Tjosvold, 1980) while a lack of participation is found to lead to unsuccessful
educational outcomes such as emotional withdrawal and poor identification with the school (Finn,
1989). Enjoyment of school is another feasible between-network construct. Elliot and Sheldon
(1997), for example, included enjoyment as one of the five key variables in their study of goal
pursuit. Even research in higher education finds that enjoyment is a key factor in students’
engagement at university (Lee, Sheldon & Turban, 2003). Martin and Marsh (2006, 2008a, 2008b)
have shown academic buoyancy to be a factor relevant to students’ ability to deal with academic
setback in the ordinary course of academic life and also shown a variety of motivation and
engagement factors to be significantly associated with such buoyancy. It is also proposed that in
addition to these four intra-psychic measures, there is a need for more behavioral measures (Green
et al., 2007) that in the present study takes the form of homework/assignment completion.
Consistent with previous studies of high school students (e.g., Green et al, 2007; Martin, 2001,
2003c, 2007a) and across diverse performance settings such as music and sport (Martin, 2008b, in
press a, in press b), it is hypothesized that the adaptive dimensions will be positively (to a modest or
strong degree) associated with these correlates, the impeding/maladaptive dimensions will be
associated at near-zero or negatively (to a weak or modest degree), whilst maladaptive dimensions
will be more markedly negatively (to a modest or strong degree) associated with these correlates.
Aims of the Present Study
The overarching aim of the present study is to examine the developmental construct validity
of motivation and engagement across elementary school, high school, and university samples. More
specifically, the present study assesses a recently developed integrative motivation and engagement
instrumentation across the academic lifespan with a view to assessing: (a) within-network validity
in terms of first and higher order factor structure and reliability and (b) between-network validity in

9
terms of invariance of factor structure across groups (gender, age, educational stage), mean-level
differences across educational stage, and associations with cognate correlates.
Method
Elementary School Sample and Procedure
The elementary school sample comprised 624 upper-age elementary students in five schools.
All schools were located in urban areas drawing from two capital cities in Australia. Students were
aged 9 to 11.5 years (N=114, 56% females and 44% males) and 11.5 years to 13 years (N=510, 38%
females and 62% males). The mean age of students was 11.13 (SD=.69) years. Teachers read the
Motivation and Engagement Scale – Junior School (MES-JS; Martin, 2007b) items aloud to
students during class or pastoral care/tutorial groups. The rating scale was first explained and
sample items were presented. Students were then asked to complete the instrument as the teacher
read out each item in turn and to return the completed form to the teacher at the end of class or
pastoral care/tutorial group. Previous work has been conducted in a smaller urban/rural elementary
school sample (Martin, Craven, & Munns, 2006), however, this work only comprised a factor
analysis of the MES-JS with no invariance testing, mean-level analyses, analyses in the context of
the academic lifespan, and external validity checks. The present study, then, is a significant
progression on previous work.
High School Archive Sample and Procedure
The high school sample comprised data collected from 21,579 high school students from 58
Australian schools. Thirty-six schools were government and 22 schools were independent, from
urban and regional areas across most states in Australia. Students were aged 12 years to 13 years
(N=6,640, 49% females and 51% males), 14 years to 15 years (N=7,894, 43% females and 57%
males), and 16 years to 18 years (N=7,045, 44% females and 56% males). The mean age of students
was 14.52 (SD=1.57) years.
The high school sample is something of an archive sample that has been compiled over
recent years across numerous research projects. Portions of the data have been reported on

10
elsewhere with a more substantial construct validity study by Martin (2007a) assessing the
Motivation and Engagement Scale – High School (MES-HS) amongst 12,237 high school students,
all of whom are included as part of the present archive sample of 21,579 students. The reader is
urged to consult Martin (2007a; see also Martin, 2008b, in press a, in press b) for these academic
motivation and engagement data in the context of other performance domains such as sport, music,
and work) as the first substantial large-sample investigation into the MES-HS. The archive dataset
represents the integration of data collected over the previous five years and so can be considered to
be relatively current. Teachers administered the MES-HS (Martin, 2001, 2003c, 2007a, 2007b) to
students during class or pastoral care/tutorial groups. The rating scale was first explained and
sample items were presented. Students were then asked to complete the instrument on their own and
to return the completed form to the teacher at the end of class or pastoral care.
University Sample and Procedure
University (college) respondents were 420 undergraduate students from two Australian
universities. One university is well-established and one of the oldest in the country (68% of
sample). The other is a more recently established institution (32%). Most respondents were female
(80%), with 20% male. Most students were enrolled in education (66%), with other students
enrolled in arts (18%), psychology/social science (8%), social work (3%), science (3%), and
communications (2%). Most were full-time students (96%), with 4% part-time. Most were in their
first year of study (65%), with 25% in second year, 7% in third year, and 3% in fourth or fifth year.
The mean age of students was 21.47 (SD=6.62) years, with 60% under 20 years of age and 40% 20
years and over. Students completed the instrument in lecture or tutorial time. Students were asked to
complete the Motivation and Engagement Scale – University/College (MES-UC; Martin, 2007b) on
their own and return the completed instrument at the end of the lecture or tutorial they were
attending at the time.
Materials
Motivation and Engagement Scale

11
General overview. The Motivation and Engagement Scale – Junior School (MES-JS; Martin,
2007b), Motivation and Engagement Scale – High School (MES-HS; Martin, 2001, 2003c, 2007a,
2007b), and Motivation and Engagement Scale – University/College (MES-UC, Martin, 2007b) are
instruments that measure elementary, high school, and university students’ motivation and
engagement respectively. Adapted from the MES-HS, the MES-JS and MES-UC assess motivation
and engagement through three adaptive cognitive dimensions (self-efficacy, valuing, mastery
orientation), three adaptive behavioral dimensions (persistence, planning, task management), three
impeding/maladaptive cognitive dimensions (anxiety, failure avoidance, uncertain control), and two
maladaptive behavioral dimensions (self-handicapping, disengagement).
Each of the eleven factors comprises four items—hence the MES is a 44-item instrument.
The MES-JS and MES-UC comprise the same number of items (44) and the same number of first
order (11) and higher order (4) factors as the original high school instrument (MES-HS). As much
as possible, item adaptation aimed to make simple and transparent word and terminology changes in
order to remain very parallel to the high school form. In the Appendix a sample item from the MES-
HS is presented along with its MES-JS and MES-UC adaptations (see Martin, 2007a for a full
account of the origins of and rationale for the scale and item development). To simplify the survey
for younger students the MES-JS asks students to rate themselves on a shorter scale of 1 (‘Strongly
Disagree’) to 5 (‘Strongly Agree’) whereas for the MES-HS and MES-UC, students rate themselves
on a scale of 1 (‘Strongly Disagree’) to 7 (‘Strongly Agree’). In most studies using the MES (e.g.,
Martin, 2007a, 2008a, 2008b, in press a), the 7-point rating scale is typically used. However, the
elementary school sample posed a distinct challenge in that a simpler survey form was desirable:
pilot work indicated students had difficulty teasing apart the finer-grained rating points on the 7-
point scale.
Adaptive cognitive and behavioral dimensions. Each adaptive dimension falls into one of
two groups: cognitions and behaviors. Adaptive cognitions include self-efficacy, mastery
orientation, and valuing. Adaptive behaviors include persistence, planning, and task management.
Self-efficacy is students’ belief and confidence in their ability to understand or to do well in their

12
school/university work, to meet challenges they face, and to perform to the best of their ability.
Valuing of school/university is how much students believe what they do and learn at
school/university is useful, important, and relevant to them. Mastery orientation entails being
focused on understanding, learning, solving problems, and developing skills. Planning is how much
students plan their work and how much they keep track of their progress as they are doing it. Task
management refers to the way students use their time, organize their timetable, and choose and
arrange where they prepare for school/university and school/university tasks. Persistence reflects
students’ capacity to persist in situations that are challenging and at times when they find it difficult
to do what is required.
Impeding and maladaptive cognitive and behavioral dimensions. Impeding/maladaptive
cognitive dimensions are anxiety, failure avoidance, and uncertain control. Anxiety has two parts:
feeling nervous and worrying. Feeling nervous is the uneasy or sick feeling students get when they
think about their school/university work or school/university tasks. Worrying is their fear of not
doing very well in their school/university work. Failure avoidance occurs when the main reason
students try at school/university is to avoid doing poorly or to avoid being seen to do poorly.
Uncertain control assesses students’ uncertainty about how to do well or how to avoid doing
poorly. Maladaptive behavioral dimensions are self-handicapping and disengagement. Self-
handicapping occurs when students reduce their chances of success at school/university. Examples
are engaging in other activities while they are meant to be doing their school/university work or
preparing for upcoming school/university work tasks. Disengagement occurs when students give up
or are at risk of giving up at school/university or in particular school/university activities.
Between-network Correlates
Students were also administered items that explored their enjoyment of school/university (4
items; e.g., elementary school item: “I like school”, Cronbach’s  = .94; high school item: “I like
school”,  = .91; university item: “I like university”,  = .91), class participation (4 items; e.g.,
elementary school item: “I get involved in things we do in class”,  = .90; high school item: “I get

13
involved in things we do in class”,  = .90; university item: “I get involved in things we do in
class”,  = .93), positive intentions (4 items; e.g., high school item: “I intend to complete school”, 
= .82; university item: “I intend to complete university”,  = .72), and academic buoyancy (4 items;
e.g., elementary school item: “I think I’m good at dealing with schoolwork pressures”,  = .78; high
school item: “I think I’m good at dealing with schoolwork pressures”, = .80; university item: “I
think I’m good at dealing with university pressures”,  = .84). These measures were rated on a 1
(‘Strongly Disagree’) to 7 (‘Strongly Agree’) scale and were adapted directly from Martin (2007a,
2008b; see also Martin & Marsh, 2006, 2008a, 2008b) who has shown them to be reliable, a good
fit to the data in confirmatory factor analysis, and significantly associated with motivation and
engagement in other performance domains such as sport and music. Homework/assignment
completion (“How often do you do and complete your assignments?”) was a single item assessed on
a 1 (‘Never’) to 5 (‘Always’) rating scale.
Confirmatory Factor Analysis and Structural Equation Modeling
Confirmatory factor analysis (CFA) and structural equation modeling (SEM), performed
with LISREL 8.80 (Jöreskog & Sörbom, 2006), were used to test the hypothesized models. In CFA
and SEM, the researcher posits an a priori structure and tests the ability of a solution based on this
structure to fit the data by demonstrating that: (a) the solution is well defined, (b) parameter
estimates are consistent with theory and a priori predictions, and (c) the subjective indices of fit are
reasonable (McDonald & Marsh, 1990). Maximum likelihood was the method of estimation used
for the models. In evaluating goodness of fit of alternative models, the root mean square error of
approximation (RMSEA) is emphasized as are the comparative fit index (CFI), the non-normed fit
index (NNFI), and an evaluation of parameter estimates. For RMSEAs, values at or less than .05
and .08 are taken to reflect a close and reasonable fit respectively (see Jöreskog & Sörbom, 1993).
The CFI and NNFI vary along a 0 to 1 continuum in which values at or greater than .90 and .95 are
typically taken to reflect acceptable and excellent fits to the data respectively (McDonald & Marsh,

14
1990). The CFI contains no penalty for a lack of parsimony whereas the RMSEA contains penalties
for a lack of parsimony.
Missing Data
For large-scale studies, the inevitable missing data is a potentially important problem,
particularly when the amount of missing data exceeds 5% (e.g., Graham & Hoffer, 2000). A
growing body of research has emphasized potential problems with traditional pairwise, listwise, and
mean substitution approaches to missing data (e.g., Graham & Hoffer, 2000), leading to the
implementation of the Expectation Maximization Algorithm, the most widely recommended
approach to imputation for missing data that are missing at random, as operationalized using
missing value analysis in LISREL. In fact, less than 5% of the MES data were missing in each of
the elementary school, high school, and university samples and so the EM Algorithm was
implemented for all samples. Also explored were alternative approaches to this problem which
showed that results based on the EM algorithm used here were very similar to those based on the
traditional pairwise deletion methods for missing data—as would be expected to be the case when
there was so little missing data.
Multi-group CFA and Tests of Invariance
Two broad sets of invariance tests were conducted. The first assessed invariance within
samples. The second assessed invariance between samples. For the within-sample invariance tests,
for each of elementary school, high school, and university, multi-group CFAs were conducted to
assess invariance across gender and age. For the between-sample invariance tests, three invariance
analyses were conducted – that between high school and university on the original 7-point rating
scale, that between elementary school, high school, and university using a common 5-point rating
scale (reliabilities for the transformed 5-point variables: high school  range = .75 - .81; university
 range = .66 to .86), and that between elementary school and university on a common 5-point
rating scale (the common 5-point rating scale was derived by aggregating the first and last two
points of the 7-point rating scale). Although the chi-square difference test is the most

15
straightforward means of assessing differences between nested models, problems associated with
such tests exist (e.g., see McDonald & Marsh, 1990; Tabachnick & Fidell, 1996). Hence, in
formally assessing differences in models, emphasis is given to differences in fit indices (Cheung &
Rensvold, 2002).
Multiple-Indicator-Multiple-Cause (MIMIC) Models
Notwithstanding the importance of testing for invariance in factor structure, there is also
reason to investigate the mean-level developmental effects on the eleven facets of the MES-JS,
MES-HS, and MES-UC. Kaplan (2000) suggested the multiple-indicator multiple-cause (MIMIC)
approach, which is similar to a regression model in which latent variables (e.g., multiple
dimensions of motivation and engagement) are ‘caused’ by discrete grouping variables (e.g.,
educational stage) that are represented by single indicators. This MIMIC model assessed the role of
educational stage (elementary school, high school, university) as a predictor of motivation and
engagement. Being a multinomial predictor and using high school as the reference point,
educational stage was represented by two dummy variables: high school (0) vs elementary school
(1) and high school (0) vs university (1) – hence, positive beta weights for both dummy variables
indicate higher scores for elementary school and university students compared to high school
students and negative beta weights for both dummy variables indicate lower scores for elementary
school and university students compared with high school students.
Results
First and Higher Order Confirmatory Factor Analysis (CFA)
In the first instance, an 11-factor model was examined using CFA. The CFA yielded a very
good fit to the data for elementary school (

2 = 1,881.10, df = 847, p < .001, CFI = .98, NNFI = .97,
RMSEA = .04), high school (

2 = 28,217.75, df = 847, p < .001, CFI = .98, NNFI = .98, RMSEA =
.04), and university (

2 = 1,697.75, df = 847, p < .001, CFI = .96, NNFI = .95, RMSEA = .05).
Factor loading ranges and means are presented in Table 1. Taken together, for all three samples the
loadings are acceptable. This is supported by the acceptable reliability coefficients (e.g., see

16
Henson, 2001) also presented in Table 1. Correlations for the sample are presented in Table 2.
Predictably, for the three samples all adaptive dimensions were strongly (significantly) positively
correlated and correlated strongly (significantly) negatively with maladaptive dimensions and
slightly (but significantly) negatively or at near-zero with impeding/maladaptive dimensions.
Maladaptive dimensions were markedly (significantly) positively correlated as were
impeding/maladaptive dimensions. For the three samples, all correlations indicate lower levels of
shared variance between factor groupings than within factor groupings.
In addition to the first order dimensions constituting the eleven facets of the Motivation and
Engagement Wheel, there is also hypothesized a higher order structure delineated by adaptive
cognitive dimensions, adaptive behavioral dimensions, impeding/maladaptive cognitive dimensions,
and maladaptive behavioral dimensions. In higher order models, correlations between first order
dimensions are constrained to be zero and relations among these first order dimensions are
explained in terms of higher order dimensions. For each of elementary school, high school, and
university samples, the higher order CFAs comprised the 44 items, the 11 first order dimensions,
and the four higher order dimensions. The higher order elementary school structure fit the data very
well (

2 = 2,155.87, df = 886, p < .001, CFI = .97, NNFI = .97, RMSEA = .05), as did the higher
model for high school students (

2 = 36,732.07, df = 886, p < .001, CFI = .98, NNFI = .98, RMSEA
= .04) and university students (

2 = 1,968.82, df = 886, p < .001, CFI = .95, NNFI = .94, RMSEA =
.05). Table 2 presents higher order correlations – which broadly confirm cluster correlations in the
first order model.
Multi-group Confirmatory Factor Analysis and Invariance Tests
Eight models were tested in each of the multi-group CFAs assessing invariance of factor
structure across gender, age, and educational stage. The initial five models related to the first order
factor structure. The first model allowed all factor loadings, uniquenesses, and correlations to be
freely estimated; the second held first order factor loadings invariant across groups; the third held
first order factor loadings and correlations/variances invariant; the fourth held first order factor

17
loadings and uniquenesses invariant, and the fifth held first order factor loadings, uniquenesses, and
correlations/variances invariant. The final three models focused on invariance of higher order
loadings and correlations/variances: the sixth freely estimated the higher order loadings and
correlations/variances, the seventh held higher order loadings invariant, and the eighth held higher
order loadings and correlations/variances invariant.
Within sample invariance tests. For elementary school, results in Table 3 indicate that when
successive elements of the first and higher order factor structure are held invariant across groups,
the fit indices are predominantly comparable across (Table 3 also indicates

2, df, and p values): (a)
males and females (ranges: CFIs=.97 for first order and .96 for higher order solutions; NNFIs=.98
for first order and .97 for higher order solutions; RMSEAs=.05 for first and higher order solutions)
and (b) younger (9-11.5 years) and older (11.5-13 years) students (ranges: CFIs=.97 for first order
and .96 for higher order solutions; NNFIs=.96 for first and higher order solutions; RMSEAs=.05 for
first and higher order solutions).
For high school, the fit indices are predominantly comparable across: (a) males and females
(ranges: CFIs=.98 for first order and .97 for higher order solution; NNFIs=.98 for first order and .97
for higher order solution; RMSEAs=.04 for first order and higher order solutions) and (b) early-
(12-13 years), mid- (14-15 years), and late- (16-18 years) adolescence (ranges: CFIs=.98 for first
order and .97 for higher order solution; NNFIs=.98 for first order and .97 for higher order solution;
RMSEAs=.04 for first order and higher order solutions).
For university, the fit indices are predominantly comparable across: (a) males and females
(ranges: CFI=.93 to .94 for first order and .92 to .93 for higher order solution; NNFIs=.93 for first
order and .92 for higher order solution; RMSEAs=.06 for first order and higher order solutions) and
(b) younger (17-19 years) and older (20+ years) students (ranges: CFIs=.94 for first order and .92
for higher order solution; NNFIs=.93 for first order and .92 for higher order solution; RMSEA=.05
to .06 for first order and .06 for higher order solutions). For all three samples, the application of

18
recommended criteria for evidence of lack of invariance (i.e., a change of 0.01 in fit indices—see
Cheung & Rensvold, 2002) indicates that there is invariance across groups.
Between sample invariance tests. The final set of invariance tests assessed first and higher
order factor structure across elementary school, high school, and university samples. This is a direct
assessment of the generalizability of the framework and measurement across diverse settings. Fit
indices in Table 4 (Table 4 also indicates

2, df, and p values) show that when successive elements
of the factor structure are held invariant across high school and university samples on the original 7-
point rating scale (ranges: CFIs and NNFIs=.98 for first order and higher order solutions;
RMSEAs=.04 for first order and higher order solutions), there is invariance across all first order and
higher order parameters. In terms of elementary school, high school, and university samples on a
common 5-point scale (the common 5-point rating scale was derived by aggregating the first and
last two points of the 7-point rating scale), there is also invariance across the three samples (ranges:
CFIs and NNFIs=.98 for first order and higher order solutions; RMSEAs=.04 for first order and
higher order solutions). Finally, when assessing invariance between elementary school and
university samples (thereby omitting the extremely large high school sample which could bias
invariance findings), there is also evidence of invariance when aspects of factor structure (loadings,
correlations/variances, uniquenesses) are systematically constrained to be equal (ranges: CFI = .96
to.97 for first order and .96 for higher order solution; NNFI = .96 to.97 for first order and .96 for
higher order solution; RMSEAs = .05 for first order and higher order solutions). For each of these
three sets of between-sample invariance tests, the application of recommended criteria for evidence
of lack of invariance (i.e., a change of 0.01 in fit indices) indicates that there is invariance across
elementary school, high school, and university domains.
Multiple-Indicator Multiple-Cause (MIMIC) Modeling
The previous analyses explored possible differences in factor structure as a function of
educational stage. It was also of interest to explore possible mean-level differences in motivation
and engagement as a function of educational stage (elementary school, high school, university).
Multiple-indicator multiple-cause (MIMIC) modeling was the analytical method used to examine

19
this and involved structural equation models in which educational stage was used as a predictor of
the first and higher order factors of the Wheel. The first order model yielded a good fit to the data
(

2 = 39,347.85, df = 914, p < .001, CFI = .95, NNFI = .94, RMSEA = .04) as did the higher order
model (

2 = 45,508.66, df = 966, p < .001, p < .001, CFI = .95, NNFI = .94, RMSEA = .05). Beta
coefficients are presented in Table 1 along with the main effects for educational stage. Results show
that there are significant stage differences on all motivation and engagement factors. Compared
with high school students, elementary school and university students are significantly higher on all
adaptive dimensions. Also, compared with high school students, elementary school and university
students are significantly lower in uncertain control, self-handicapping, and disengagement.
However, compared to high school students, elementary school and university students are
significantly higher on anxiety and failure avoidance. As a general finding, there is a greater
difference between elementary and high school students than between high school and university
students. Again, however, to note is that the high school and university 1-7 rating continuum was
transformed to a 1-5 rating continuum to place them on the same scale of measurement as
elementary school – hence, caution is advised when interpreting these findings. Due to the large
high school sample, caution is also advised when interpreting the significance of the MIMIC results
– and this being the case, greater emphasis is given to findings in relation to self-efficacy, mastery
orientation, valuing of school, planning, task management, persistence, uncertain control, and self-
handicapping that yielded standardized beta values greater than .30.
Motivation, Engagement, and Between-network Cognate Correlates
As indicated earlier, consistent with the between-network construct validity approach, it was
of interest to explore the nature of relationships between each facet of motivation and a set of key
between-network correlates across the three educational stages. To this end, the three samples were
also administered items that explored enjoyment of school/university (elementary school, high
school, university), class participation (elementary school, high school, university), positive
academic intentions (high school, university), academic buoyancy (elementary school, high school,

20
university), and homework completion (high school, university). For each of the three samples, first
and higher order CFAs were conducted.
The first order elementary school CFA yielded a very good fit to the data (

2 = 2,915.33, df
= 1393, p < .001, CFI = .98, NNFI = .98, RMSEA = .04) and showed that: (a) adaptive dimensions
are significantly positively associated with these between-network constructs and (b)
impeding/maladaptive and maladaptive dimensions (particularly uncertain control, self-
handicapping, and disengagement) are negatively correlated with these constructs. Table 5 presents
findings. These first order findings were broadly supported in the high school sample (

2 = 52,112,
df = 1650, p < .001, CFI = .98, NNFI = .98, RMSEA = .04) and the university sample (

2 =
3,251.39, df = 1650, p < .001, CFI = .96, NNFI = .96, RMSEA = .05). Interestingly – and consistent
with Martin (2007; see also Martin & Marsh, 2006, 2008a, 2008b) – academic buoyancy is a
notable exception in being more markedly correlated with impeding/maladaptive cognitions than
maladaptive behaviors – largely a function of its very high correlation with anxiety (discussed fully
in Martin & Marsh, 2006, 2008a, 2008b). Again, however, due to the large high school sample
caution is advised when interpreting the correlations – emphasis is given to the size and direction of
the correlation coefficients themselves rather than their significance levels.
The higher order factor analysis for elementary school (

2 = 3,361.64, df = 1453, p < .001,
CFI = .97, NNFI = .97, RMSEA = .05) provides general support for the first order findings. Higher
order correlations are also presented in Table 5 (again, due to the large samples involved, emphasis
is given to the size and direction of the correlation coefficients themselves rather than their
significance levels). Consistent with the elementary school findings, the higher order factor analysis
for high school (

2 = 67,868.55, df = 1724, p < .001, CFI = .98, NNFI = .98, RMSEA = .04)
provides support for the first order findings as did the higher order model for the university sample
(

2 = 3,683.58, df = 1724, p < .001, CFI = .95, NNFI = .95, RMSEA = .05).
Discussion

21
Through the integration of multivariate measurement and the hypothesized motivation and
engagement framework, the study supports the developmental construct validity of motivation and
engagement at elementary school, high school, and university/college levels. From this
developmental construct validity perspective, perhaps the most significant yield of the present study
is the predominantly comparable findings across three very distinct educational stages. The data
confirm the hypothesized generality of the Wheel and its accompanying instrumentation amongst
very young students in elementary school through to mature age students in university. In some
ways the most revealing tests were the multi-group invariance analyses across the elementary
school, high school, and university samples. These analyses directly addressed the question posed at
the outset of the study regarding the generality of the proposed motivation and engagement
framework in diverse educational settings. The invariance data suggested that there is generality –
and developmental validity – of the framework across the academic lifespan.
Notwithstanding the important consistencies across the three educational stages, findings
also suggest issues distinct to each academic setting. For example, the data showed that elementary
school students reflect higher levels of motivation and engagement and this is consistent with prior
work showing declines between elementary and middle/high school (e.g., Anderman & Midgley,
1997; Roeser et al, 2000; Wigfield et al, 1991; Wigfield & Tonks, 2002). In terms of university
students, there was some question as to their level of motivation relative to school students with
some research recognizing the challenges they face in higher education and other research reporting
on their confidence in their abilities (e.g., see Martin, Marsh, Williamson, & Debus, 2003; Pitts,
2005). The present data shed light on these competing views by showing that, notwithstanding
equivalence in factor structure, university students reflect higher mean levels of motivation and
engagement than their high school counterparts. In the case of all MIMIC analyses, however, due to
the large samples involved emphasis is given to the size and direction of the standardized beta
coefficients rather than the attained significance levels.
Because the constructs within the Wheel have a theoretical basis, researchers are able to
draw on theory to provide direction for intervention aimed at addressing facets within the Wheel.

22
Research shows that targeted intervention is more effective than intervention that does not focus on
specific target behaviors (O’Mara, Marsh, Craven, & Debus, 2006) and so it is proposed that
intervention programs seeking to build specific academic skills and competencies need to provide
targeted support that can do this. The Wheel provides a basis for doing so. Martin (2007a; see also
Martin, 2008b, in press a, for strategy in sport and music settings) has proposed specific classroom
strategy that targets each of these dimensions and this strategy incorporated into intervention work
has demonstrated significant yields for students (Martin, 2005a, 2008a). In addition to what Martin
(2007a) suggests in terms of specific classroom strategy there are other approaches to intervention
that have more of a measurement basis to them. One such approach that Martin (2008b) has
previously proposed in relation to motivation and engagement involves performance profiling.
Performance profiling (Butler & Hardy, 1992) has very direct synergies with the Wheel both
in form and substance – indeed, Martin (2008b) has demonstrated how performance profiling can
be conducted with the Wheel in sport and music domains. Performance profiling provides a means
by which to effectively and parsimoniously contextualize individuals’ profile in reference to a set of
psychological and behavioral criteria. Although there are various ways and levels to profile under a
performance profiling schedule, the example in the present study is the mean-level profile (rounded)
for the high school sample as a whole (N=21,579). In Figure 2, the traditional performance profiling
format (see Butler & Hardy, 1992; see also Martin, 2008b; Weinberg & Gould, 2001) has been
adapted to interface with the Motivation and Engagement Wheel. Obviously at the individual level
it would reflect the student’s mean scores on each dimension. Or, it could be readily employed at a
class or school level (and bringing into focus the issue of multilevel models of motivation and
engagement – see Marsh, Martin, & Cheng, 2008; Martin & Marsh, 2005, for multilevel research
along these lines).
Limitations, Future Directions, and Conclusion
The present study provides an enhanced understanding of the validity of motivation and
engagement in the context of three educational stages: elementary school, high school, and
university. There are, however, a number of potential limitations important to consider when

23
interpreting findings. Firstly, although the large sample involved in the study is a distinct strength of
the research, it posed some challenges when interpreting data, with the need to emphasize the
practical significance of findings as much as or more than the statistical significance of findings. It
is also important to recognize that the data presented in this study are all self-reported. Although
this is a logical and defensible methodology in its own right given the substantive focus, it is
important to conduct research that examines the same constructs using data derived from additional
sources such as, for example, achievement and that from teachers and parents. Just as important as
the self-report nature of findings is the fact that the data presented in the study are cross-sectional.
Tracking the same students over time and assessing factor structure and inter-relationships from a
longitudinal perspective would shed further light on the developmental processes relevant to
motivation and engagement. Additionally, examining reliability and stability of the scores over time
and the causal ordering of motivation and engagement in relation to the cognate constructs assessed
here are other issues of interest in longitudinal work.
The nature of quantitative survey-based methods also warrants some further comment.
Although Martin, Marsh, Williamson, and Debus (2003) conducted qualitative work amongst
university samples, future research might encompass qualitative work that can more fully scope the
detailed nature and extent of motivation and engagement across the academic lifespan. Alongside
this qualitative work, there may also be yields in multi-level approaches to developmental construct
validity in motivation and engagement. Advances in statistical software enable researchers to more
accurately assess the relative influence of individual-, class-, and school-level factors using multi-
level modeling (see Goldstein, 2003) and so future research can readily explore the influence of
class- and school-level motivation climates relative to individual-level variation in motivation and
engagement as relevant to developmental construct validity.
To conclude, the research presented here supports the developmental construct validity of
the Motivation and Engagement Wheel and its accompanying instrumentation, the Motivation and
Engagement Scale, across the academic lifespan. The findings of the present investigation hold
implications for researchers studying issues relevant to motivation and engagement across the

24
academic lifespan. The findings also present new insights and opportunities for educators seeking to
enhance the educational outcomes of their students – outcomes that are affected by motivation and
engagement and the extent to which educators can effectively measure and enhance them.

25
References
Aiken, L.S., & West, S.G. (1991). Multiple regression: Testing and interpreting interactions.
London: Sage.
Anderman, E.A., & Midgley, C. (1997). Changes in personal achievement goals and the perceived
classroom goal structures across the transition to middle level schools. Contemporary
Educational Psychology, 22, 269-298.
Atkinson, J.W. (1957). Motivational determinants of risk-taking. Psychological Review, 64, 359-
372.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman & Co.
Beck, A.T. (1995). Cognitive therapy: Basics and beyond. New York: Guilford.
Bong, M. (1996). Problems in academic motivation research and advantages and disadvantages of
their solutions. Contemporary Educational Psychology, 21, 149-165.
Buss, D.W., & Cantor, N. (1989). Personality psychology: Recent trends and emerging directions.
New York: Springer-Verlag.
Butler, R.J., & Hardy, L. (1992). The performance profile: Theory and application. The Sport
Psychologist, 6, 253-264.
Cheung, G.W., & Rensvold, R.B. (2002). Evaluating goodness-of-fit indexes for testing
measurement invariance. Structural Equation Modeling, 9, 233-255.
Connell, J.P. (1985). A new multidimensional measure of children’s perceptions of control. Child
Development, 56, 1018-1041.
Covington, M.V. (1992). Making the grade: A self-worth perspective on motivation and school
reform. Cambridge: Cambridge University Press.
Dweck, C.S. (1986). Motivational processes affecting learning. American Psychologist, 41, 1040-
1048.
Eccles, J. (1983). Expectancies, values, and academic behaviors. In J. Spence (Ed.). Achievement
and achievement motivation. San Francisco: Freeman.
Elliot, A.J., & Church, M.A. (1997). A hierarchical model of approach and avoidance achievement
motivation. Journal of Personality and Social Psychology, 72, 218-232.
Elliot, A. J., & Sheldon, K. M. (1997). Avoidance achievement motivation: A personal goals
analysis. Journal of Personality and Social Psychology, 73, 171-185.
Finn, J. D. (1989). Withdrawing from school. Review of Educational Research, 59, 117-142.
Goldstein, H. (2003). Multilevel statistical models. London: Edward Arnold.
Graham, J. W., & Hoffer, S. M. (2000). Multiple imputation in multivariate research. In T. D. Little,
K. U. Schnable, & J. Baumert (Eds.), Modeling longitudinal and multilevel data: Practical

26
issues, applied approaches, and specific examples (pp. 201–218). Mahwah, NJ: Erlbaum.
Green, J., Martin, A.J., & Marsh, H.W. (2007). Motivation and engagement in English,
mathematics and science high school subjects: Towards an understanding of multidimensional
domain specificity. Learning and Individual Differences, 17, 269-279.
Greeno, J. (1998). The situativity of knowing, learning, and research. American Psychologist, 53, 5-
26.
Hattie, J. (1992). Self-concept. Hillsdale, NJ: Erlbaum.
Henson, R. K. (2001). Understanding internal consistency reliability estimates: A conceptual primer
on coefficient alpha. Measurement and Evaluation in Counseling and Development, 34, 177-
189.
Jöreskog, K.G. & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS
command language. Chicago: Scientific Software International.
Jöreskog, K.G. & Sörbom, D. (2006). LISREL 8.80. Chicago: Scientific Software International.
Kaplan, D. (2000). Structural Equation Modeling: Foundations and Extensions. Newbury Park,
CA: Sage.
Lee, F. K., Sheldon, K. M., & Turban, D. B. (2003). Personality and goal-striving process: The
influence of achievement goal patterns, goal level, and mental focus on performance and
enjoyment. Journal of Applied Psychology, 88, 256-265.
Marsh, H. W. (1993). The multidimensional structure of academic self-concept: Invariance over
gender and age. American Educational Research Journal, 30, 841-860.
Marsh, H. W. (2002). A multidimensional physical self-concept: A construct validity approach to
theory, measurement, and research. Psychology: The Journal of the Hellenic Psychological
Society, 9, 459-493.
Marsh, H. W., & Hau, K-T. (2007). Applications of latent-variable models in educational
psychology: The need for methodological-substantive synergies. Contemporary Educational
Psychology, 32, 151-171.
Marsh, H.W., Martin, A.J., & Cheng, J. (2008). A multilevel perspective on gender in classroom
motivation and climate: Potential benefits of male teachers for boys? Journal of Educational
Psychology, 100, 78-95.
Marsh, H. W., & Shavelson, R. J. (1985). Self-concept: Its multifaceted, hierarchical structure.
Educational Psychologist, 20,107-125.
Martin, A.J. (2001). The Student Motivation Scale: A tool for measuring and enhancing motivation.
Australian Journal of Guidance and Counselling, 11, 1-20.

27
Martin, A.J. (2002). Motivation and academic resilience: Developing a model of student
enhancement. Australian Journal of Education. 14, 34-49.
Martin, A.J. (2003a). How to motivate your child for school and beyond. Sydney: Bantam.
Martin, A.J. (2003b). The relationship between parents’ enjoyment of parenting and children’s
school motivation. Australian Journal of Guidance and Counselling, 13, 115-132.
Martin, A.J. (2003c). The Student Motivation Scale: Further testing of an instrument that measures
school students’ motivation. Australian Journal of Education, 47, 88-106.
Martin, A.J. (2004). School motivation of boys and girls: Differences of degree, differences of kind,
or both? Australian Journal of Psychology, 56, 133-146.
Martin, A.J. (2005a). Exploring the effects of a youth enrichment program on academic motivation
and engagement. Social Psychology of Education, 8, 179-206.
Martin, A.J. (2005b). Perplexity and passion: Further consideration of the role of positive
psychology in the workplace. Journal of Organizational Behavior Management, 24, 203-205.
Martin, A.J. (2005c). The role of positive psychology in enhancing satisfaction, motivation, and
productivity in the workplace. Journal of Organizational Behavior Management, 24, 113-133.
Martin, A.J. (2006). The relationship between teachers’ perceptions of student motivation and
engagement and teachers’ enjoyment of and confidence in teaching. Asia-Pacific Journal of
Teacher Education, 34, 73-93.
Martin, A.J. (2007a). Examining a multidimensional model of student motivation and engagement
using a construct validation approach. British Journal of Educational Psychology, 77, 413-
440.
Martin, A.J. (2007b). The Motivation and Engagement Scale. Sydney, Australia: Lifelong
Achievement Group (www.lifelongachievement.com).
Martin, A.J. (2008a). Enhancing student motivation and engagement: The effects of a
multidimensional intervention. Contemporary Educational Psychology, 33, 239-269.
Martin, A.J. (2008b). Motivation and engagement in music and sport: Testing a multidimensional
framework in diverse performance settings. Journal of Personality, 76, 135-170.
Martin, A.J. (in press a). How domain specific are motivation and engagement across school, sport,
and music? A substantive-methodological synergy assessing young sportspeople and
musicians. Contemporary Educational Psychology.
Martin, A.J. (in press b). Motivation and engagement in diverse performance domains: Testing their
generality across school, university/college, work, sport, music, and daily life. Journal of
Research in Personality.
Martin, A.J., Craven, R.G., & Munns, G. (2006). Motivation and engagement in young children:
How well does a high school conceptualization generalize to junior school? In R.G. Craven.,
J. S. Eccles, & T. M. Ha (Eds). Self-concept, Motivation, Social and Personal Identity for the

28
21st Century. Proceedings of the Fourth International Biennial SELF Research Conference,
Ann Arbor, University of Michigan (ISBN 1 74108 148 3).
Martin, A.J., & Debus, R.L. (1998). Self-reports of mathematics self-concept and educational
outcomes: The roles of ego-dimensions and self-consciousness. British Journal of
Educational Psychology, 68, 517-535.
Martin, A.J. & Marsh, H.W. (2003). Fear of failure: Friend or foe? Australian Psychologist, 38, 31-
38.
Martin, A.J., & Marsh, H.W. (2005). Motivating boys and motivating girls: Does teacher gender
really make a difference? Australian Journal of Education, 49, 320-334.
Martin, A.J., & Marsh, H.W. (2006). Academic resilience and its psychological and educational
correlates: A construct validity approach. Psychology in the Schools, 43, 267-282.
Martin, A.J., & Marsh, H.W. (2008a). Academic buoyancy: Towards an understanding of students’
everyday academic resilience. Journal of School Psychology, 46, 53-83.
Martin, A.J., & Marsh, H.W. (2008b). Workplace and academic buoyancy: Psychometric
assessment and construct validity amongst school personnel and students. Journal of
Psychoeducational Assessment, 26, 168-184.
Martin, A. J. Marsh, H.W., & Debus, R. L. (2001a). A quadripolar need achievement
representation of self-handicapping and defensive pessimism. American Educational
Research Journal, 38, 583-610.
Martin, A.J. Marsh, H.W., & Debus, R.L. (2001b). Self-handicapping and defensive pessimism:
Exploring a model of predictors and outcomes from a self-protection perspective. Journal of
Educational Psychology, 93, 87-102.
Martin, A.J. Marsh, H.W., & Debus, R.L. (2003). Self-handicapping and defensive pessimism: A
model of self-protection from a longitudinal perspective. Contemporary Educational
Psychology, 28, 1-36.
Martin, A.J., Marsh, H.W., Williamson, A., & Debus, R.L. (2003). Self-handicapping, defensive
pessimism, and goal orientation: A qualitative study of university students. Journal of
Educational Psychology, 95, 617-628.
Martin, A.J., Milne-Home, J., Barrett, J., & Spalding, E. (1997). Stakeholder perceptions of the
institution: To agree or not to agree. Journal of Institutional Research in Australasia, 6, 53-
67.
Martin, A.J., Milne-Home, J., Barrett, J., Spalding, E., & Jones, G. (2000). Graduate satisfaction
with university and perceived employment preparation. Journal of Education and Work, 13,
199-214.

29
Martin, A.J., Tipler, D.V., Marsh, H.W., Richards, G.E., & Williams. M.R. (2006). Assessing
multidimensional physical activity motivation: A construct validity study of high-school
students. Journal of Sport and Exercise Psychology, 28, 171-192.
McClelland, D.C. (1965). Toward a theory of motive acquisition. American Psychologist, 20, 321-
333.
McDonald, R. P., & Marsh, H. W. (1990). Choosing a multivariate model: Noncentrality and
goodness-of-fit. Psychological Bulletin, 107, 247-255.
Meece, J.L., Wigfield, A., & Eccles, J.S. (1990). Predictors of mathematics anxiety and its
influence on young adolescents’ course enrolment intentions and performance in
mathematics. Journal of Educational Psychology, 82, 60-70.
Miller, R.B., Greene, B.A., Montalvo, G.P., Ravindran, B., & Nichols, J.D. (1996). Engagement in
academic work: The role of learning goals, future consequences, pleasing others, and
perceived ability. Contemporary Educational Psychology, 21, 388-422.
Miserandino, M. (1996). Children who do well in school: Individual differences in perceived
competence and autonomy in above-average children. Journal of Educational Psychology, 88,
203-214.
Murphy, P.K., & Alexander, P.A. (2000). A motivated exploration of motivation terminology.
Contemporary Educational Psychology, 25, 3-53.
Nicholls, J.G. (1989). The competitive ethos and democratic education. Cambridge: Harvard
University Press.
O’Mara, A. J., Marsh H. W., Craven, R. G., & Debus, R. (2006). Do self-concept interventions
make a difference? A synergistic blend of construct validation and meta-analysis. Educational
Psychologist, 41, 181-206.
Pajares, F. (1996). Self-efficacy beliefs in achievement settings. Review of Educational Research.
66, 543-578.
Pintrich, P.R. (2000). Educational psychology at the millennium: A look back and a look forward.
Educational Psychologist, 35, 221-226.
Pintrich, P.R. (2003). A motivational science perspective on the role of student motivation in
learning and teaching contexts. Journal of Educational Psychology, 95, 667-686.
Pintrich, P.R., & DeGroot, E. (1990). Motivational and self-regulated learning components of
classroom academic performance. Journal of Educational Psychology, 82, 33-40.
Pintrich, P.R., & Garcia, T. (1991). Student goal orientation and self-regulation in the college
classroom. In M. Maehr & P.R. Pintrich (Eds.). Advances in motivation and achievement:
Goals and self-regulatory processes (Vol. 7 pp. 371-402). Greenwich, CT: JAI Press.

30
Pitts, S. (2005). Valuing musical participation. Hants, UK: Ashgate Publishing.
Richter, F.D., & Tjosvold, D. (1980). Effects of student participation in classroom decision making
on attitudes, peer interaction, motivation and learning. Journal of Applied Psychology, 65, 74-
80.
Roeser, R., Eccles, J.S., & Sameroff, A.J. (2000). School as a context of early adolescents’
academic and social-emotional development: A summary of research findings. Elementary
School Journal, 100, 443-471.
Ryan, R.M., & Deci, E.L. (2000). Self-determination theory and the facilitation of intrinsic
motivation, social development, and well-being. American Psychologist, 55, 68-78.
Sarason, I.G., & Sarason, B.R. (1990). Test anxiety. In H. Leitenberg (Ed.). Handbook of social and
evaluation anxiety. (pp. 475-495). New York: Plenum Press.
Shavelson, R. J., Hubner, J. J., & Stanton, G. (1976). Self-concept: Validation of construct
interpretations. Review of Educational Research, 46, 407-441.
Spielberger, C.D. (1985). Assessment of state and trait anxiety: Conceptual and methodological
issues. Southern Psychologist, 2, 6-16.
Stokes, D. (1997). Pasteur’s quadrant: Basic science and technological innovation. Washington,
DC: Brookings Institution Press.
Tabachnick, B.G., & Fidell, L.A. (2007). Using multivariate statistics. Boston: Pearson/Allyn &
Bacon
Weinberg, R.S., & Gould, D. (1999). Foundations of sport and exercise psychology. Champaign,
IL: Human Kinetics.
Weiner, B. (1994). Integrating social and personal theories of achievement striving. Review of
Educational Research, 64, 557-573.
Wigfield, A., Eccles, J., Mac Iver, D., Reuman, D., & Midgley, C. (1991). Transitions at early
adolescence: Changes in children’s domain-specific self-perceptions and general self-esteem
across the transition to junior high school. Developmental Psychology, 27, 552-565.
Wigfield, A., & Tonks, S. (2002). Adolescents’ expectancies for success and achievement task
values during middle and high school years. In F. Pajares & T. Urdan (Eds). Academic
motivation of adolescents. Connecticut: Information Age Publishing.
Zimmerman, B.J. (2002). Achieving self-regulation: The trial and triumph of adolescence. In F.
Pajares & T. Urdan (Eds). Academic motivation of adolescents. Connecticut: Information Age
Publishing.

31
Anxiety
Uncertain
control
Self-
handicapping
Disengagement
Valuing
Planning
Persistence
Self-
efficacy
Mastery
orientation
Task
management
Failure
avoidance
ADAPTIVE
COGNITION
ADAPTIVE
BEHAVIOR
IMPEDING/MALADAPTIV
E COGNITION
MALADAPTIVE
BEHAVIOR
Figure 1
Motivation and Engagement Wheel – adapted from Martin (2003a)

32
Table 1. Cronbach’s Alphas, CFA Factor Loadings, and MIMIC Modeling Standardized Betas
Cronbach’s 
ES / HS / UNI
CFA Loadings
Range (Mean)
ES / HS / UNI
HS (0)
vs
ES (1)
HS (0)
vs
UNI (1)
ADAPTIVE COGNITION
Self-efficacy
.76 / .77 / .71
.60-.72 (.67) / .63-.75 (.69) / .54-.71 (.62)
.24*** (ES>HS)
.15*** (U>HS)
Mastery orient
.82 / .81 / .82
.69-.79 (.73) / .65-.78 (.72) / .63-.82 (.73)
.30*** (ES>HS)
.31*** (U>HS)
Valuing
.74 / .77 / .70
.49-.77 (.65) / .55-.76 (.68) / .49-.70 (.61)
.50*** (ES>HS)
.38*** (U>HS)
Higher Order
ES: Range = .84-.90, Mean = .87; HS: Range = .84-.92,
Mean = .87; UNI: Range = .75-.89, Mean = .80
.45*** (ES>HS)
.36*** (U>HS)
ADAPTIVE BEHAVIOR
Planning
.87 / .77 / .73
.73-.89 (.80) / .57-.79 (.70) / .33-.91 (.66)
.33*** (ES>HS)
.26*** (U>HS)
Task management
.86 / .82 / .82
.61-.88 (.78) / .71-.85 (.76) / .62-.87 (.74)
.26*** (ES>HS)
.24*** (U>HS)
Persistence
.79 / .81 / .75
.63-.79 (.70) / .60-.79 (.71) / .59-.75 (.66)
.25*** (ES>HS)
.24*** (U>HS)
Higher Order
ES: Range = .72-.80, Mean = .76; HS: Range = .84-.88,
Mean = .86; UNI: Range = .59-.90, Mean = .74
.35*** (ES>HS)
.30*** (U>HS)
IMPEDING/MALADAPTIVE COGNITION
Anxiety
.75 / .77 / .78
.52-.74 (.65) / .61-.74 (.68) / .55-.82 (.69)
.04*** (ES>HS)
.22*** (U>HS)
Failure avoidance
.84 / .79 / .85
.61-.85 (.76) / .65-.84 (.70) / .71-.83 (.77)
-.18*** (HS>ES)
-.14*** (HS>U)
Uncertain control
.78 / .79 / .80
.65-.73 (.69) / .62-.75 (.69) / .62-.82 (.72)
-.50*** (HS>ES)
-.28*** (HS>U)
Higher Order
ES: Range = .51-.87, Mean = .69; HS: Range = .56-.83,
Mean = .69; UNI: Range = .51-.74, Mean = .65
-.47*** (HS>ES)
-.24*** (HS>U)
MALADAPTIVE BEHAVIOR
Self-handicapping
.82 / .81 / .87
.68-.77 (.73) / .61-.78 (.72) / .72-.84 (.79)
-.47*** (HS>ES)
-.26*** (HS>U)
Disengagement
.70 / .81 / .72
.33-.85 (.63) / .65-.84 (.74) / .50-.79 (.65)
-.31*** (HS>ES)
-.13*** (HS>U)
Higher Order
ES: Range = .72-.89, Mean = .81; HS: Range = .70-.87,
Mean = .79; UNI: Range = .64-.80, Mean = .72
-.49*** (HS>ES)
-.24*** (HS>U)
ES=Elementary school; HS=High school; UNI=University
Note 1. Means, SDs, skewness, and kurtosis available from the author upon request
Note 2. HS results are bolded to assist readability

33
Table 2. Inter-Scale Correlations in CFA: First and Higher Order Solutions
FIRST ORDER CORRELATIONS
Elementary School / High School / University
SE
MO
V
PLN
TM
P
A
FA
UC
SH
SE
-
MO
.78 / .73 / .60
-
V
.75 / .76 / .61
.72 / .78 / .71
-
PLN
.60 / .55 / .41
.56 / .54 / .42
.51 / .57 / .43
-
TM
.57 / .58 / .25
.50 / .56 / .42
.52 / .58 / .39
.63 / .79 / .57
-
P
.71 / .68 / .64
.58 / .59 / .48
.52 / .65 / .64
.59 / .74 / .65
.63 / .66 / .46
-
A
-.08 / .03 / -.08
.03 / .21 / .17
.04 / .14 / .08
-.19 / .11 / .13
-.11 / .15 / .09
-.19 / .07 / .08
-
FA
-.24 / -.16 / -.24
-.15 / -.05 / -.11
-.22 / -.11 / -.28
-.23 / -.02 / -.15
-.20 / -.02 / -.10
-.29 / -.09 / -.31
.50 / .43 / .39
-
UC
-.54 / -.34 / -.50
-.38 / -.10 / -.12
-.42 / -.17 / -.13
-.39 / -.17 / -.21
-.35 / -.15 / -.10
-.52 / -.27 / -.38
.40 / .49 / .47
.57 / .53 / .45
-
SH
-.47 / -.37 / -.30
-.37 / -.26 / -.26
-.49 / -.32 / -.32
-.36 / -.33 / -.30
-.36 / -.32 / -.24
-.45 / -.40 / -.45
.26 / .19 / .17
.50 / .45 / .53
.62 / .53 / .36
-
D
-.59 / -.62 / -.47
-.59 / -.56 / -.36
-.75 / -.71 / -.63
-.45 / -.51 / -.26
-.48 / -.51 / -.26
-.59 / -.60 / -.54
.11 / .06 / .10
.36 / .32 / .40
.51 / .43 / .39
.65 / .59 / .51
HIGHER ORDER CORRELATIONS
Elementary School / High School / University
AC
AB
IMC
MB
AC
-
AB
.86 / .78 / .77
-
IMC
-.46 / -.16 / -.29
-.56 / -.14 / -.33
-
MB
-.79 / -.75 / -.69
-.74 / -.68 / -.66
.70 / .61 / .73
-
Note 1. Elementary school r > .07 significant at p<0.05; High school r > .02 significant at p<0.05 (but note large sample); University r > .10 significant at p<0.05
Note 2. Self-efficacy (SE); Mastery orientation (MO); Valuing (V); Planning (PLN); Task management (TM); Persistence (P); Anxiety (A); Failure avoid (FA); Uncertain control
(UC); Self-handicapping (SH); Disengagement (D); Adaptive cognitions (AC); Adaptive behaviors (AB); Impeding/maladaptive cognitions (IMC); Maladaptive behaviors (MB)
Note 3. HS results are bolded to assist readability

34
Table 3. Invariance Tests across Gender and Age Group
Chi Square
DF
CFI
NNFI
RMSEA
Elementary School / High School / University
Invariance across Males and Females
First order parameters are free (Model 1: no invariance)
2947 / 28707 / 2720
1694 / 1694 / 1694
.97 / .98 / .94
.97 / .98 / .93
.05 / .04 / .05
FIRST ORDER FACTOR LOADINGS invariant (Model 2)
3084 / 28859 / 2761
1727 / 1727 / 1727
.97 / .98 / .94
.97 / .98 / .93
.05 / .04 / .05
Model 2 + CORRELATIONS/VARIANCES invariant
3165 / 29343 / 2923
1793 / 1793 / 1793
.97 / .98 / .94
.97 / .98 / .93
.05 / .04 / .06
Model 2 + UNIQUENESSES invariant
3108 / 31109 / 2983
1771 / 1771 / 1771
.97 / .98 / .94
.97 / .98 / .93
.05 / .04 / .06
Model 2 + CORRELATIONS/VARIANCES, UNIQUENESSES invariant
3269 / 31759 / 3162
1837 / 1837 / 1837
.97 / .98 / .93
.96 / .98 / .93
.05 / .04 / .06
Higher order parameters free
3409 / 39563 / 3285
1849 / 1849 / 1849
.96 / .97 / .93
.96 / .97 / .92
.05 / .04 / .06
HIGHER ORDER FACTOR LOADINGS invariant (Model 3)
3413 / 39595 / 3320
1855 / 1855 / 1855
.96 / .97 / .93
.96 / .97 / .92
.05 / .04 / .06
Model 3 + CORRELATIONS/VARIANCES invariant
3558 / 40077 / 3455
1876 / 1876 / 1876
.96 / .97 / .92
.96 / .97 / .92
.05 / .04 / .06
Invariance across Age Groups
First order parameters are free (Model 1: no invariance)
3011 / 30639 / 2728
1694 / 2541 / 1694
.97 / .98 / .94
.96 / .98 / .93
.05 / .04 / .05
FIRST ORDER FACTOR LOADINGS invariant (Model 2)
3036 / 31021 / 2792
1727 / 2607 / 1727
.97 / .98 / .94
.96 / .98 / .93
.05 / .04 / .05
Model 2 + CORRELATIONS/VARIANCES invariant
3156 / 32005 / 2924
1793 / 2739 / 1793
.97 / .98 / .94
.96 / .98 / .93
.05 / .04 / .06
Model 2 + UNIQUENESSES invariant
2993 / 32800 / 2875
1771 / 2695 / 1771
.97 / .98 / .94
.96 / .98 / .93
.05 / .04 / .06
Model 2 + CORRELATIONS/VARIANCES, UNIQUENESSES invariant
3091 / 33857 / 3004
1837 / 2827 / 1837
.96 / .98 / .94
.96 / .98 / .93
.05 / .04 / .06
Higher order parameters free
3320 / 41931 / 3208
1849 / 2812 / 1849
.96 / .97 / .93
.96 / .97 / .92
.05 / .04 / .06
HIGHER ORDER FACTOR LOADINGS invariant (Model 3)
3325 / 42050 / 3217
1855 / 2824 / 1855
.96 / .97 / .93
.96 / .97 / .93
.05 / .04 / .06
Model 3 + CORRELATIONS/VARIANCES invariant
3364 / 42582 / 3261
1876 / 2866 / 1876
.96 / .97 / .92
.96 / .97 / .92
.05 / .04 / .06
Note 1. HS results are bolded to assist readability; Note 2. All chi square values significant at p<0.001; Note 3. Maximum 90% Confidence Interval range for all first order RMSEAs
= .04-.06. Note 4. Maximum 90% Confidence Interval range for all higher order RMSEAs = .04-.07

35
Table 4. Invariance Tests across Samples
Chi
Square
DF
CFI
NNFI
RMSEA
Invariance High School and University (7-point scale)
First order parameters are free (Model 1: no invariance)
28875
1694
.98
.98
.04
FIRST ORDER FACTOR LOADINGS invariant (Model 2)
29002
1727
.98
.98
.04
Model 2 + CORRELATIONS/VARIANCES invariant
29249
1793
.98
.98
.04
Model 2 + UNIQUENESSES invariant
29110
1771
.98
.98
.04
Model 2 + CORRELATIONS/VARIANCES, UNIQUE invariant
29291
1837
.98
.98
.04
Higher order parameters free
37548
1849
.98
.98
.04
HIGHER ORDER FACTOR LOADINGS invariant (Model 3)
37563
1855
.98
.98
.04
Model 3 + CORRELATIONS/VARIANCES invariant
37609
1876
.98
.98
.04
Invariance Elementary, High School, University (5-point scale)
First order parameters are free (Model 1: no invariance)
26203
2541
.98
.98
.04
FIRST ORDER FACTOR LOADINGS invariant (Model 2)
26645
2607
.98
.98
.04
Model 2 + CORRELATIONS/VARIANCES invariant
27480
2739
.98
.98
.04
Model 2 + UNIQUENESSES invariant
26878
2695
.98
.98
.04
Model 2 + CORRELATIONS/VARIANCES, UNIQUE invariant
27550
2827
.98
.98
.04
Higher order parameters free
34745
2812
.98
.98
.04
HIGHER ORDER FACTOR LOADINGS invariant (Model 3)
34823
2824
.98
.98
.04
Model 3 + CORRELATIONS/VARIANCES invariant
35171
2866
.98
.98
.04
Invariance Elementary and University (5-point scale)
First order parameters are free (Model 1: no invariance)
3472
1694
.97
.97
.05
FIRST ORDER FACTOR LOADINGS invariant (Model 2)
3657
1727
.97
.97
.05
Model 2 + CORRELATIONS/VARIANCES invariant
3931
1793
.97
.96
.05
Model 2 + UNIQUENESSES invariant
3895
1771
.97
.96
.05
Model 2 + CORRELATIONS/VARIANCES, UNIQUE invariant
4181
1837
.96
.96
.05
Higher order parameters free
4403
1849
.96
.96
.05
HIGHER ORDER FACTOR LOADINGS invariant (Model 3)
4419
1855
.96
.96
.05
Model 3 + CORRELATIONS/VARIANCES invariant
4561
1876
.96
.96
.05
Note 1. All chi square values significant at p<0.001
Note 2. Maximum 90% Confidence Interval range for all first order RMSEAs = .03-.05
Note 3. Maximum 90% Confidence Interval range for all higher order RMSEAs = .04-.05

36
Table 5. First and Higher Order Correlations with Between-Network Constructs
Enjoyment
Participation
Buoyancy
Positive Intent
Homework
Completion
FIRST ORDER CORRELATIONS
Elementary School / High School / University
ADAPTIVE COGNITIONS
Self-efficacy
.43 / .57 / .45
.44 / .51 / .45
.42 / .38 / .41
– / .67 / .68
– / .35 / .05
Mastery orientation
.57 / .55 / .37
.48 / .45 / .36
.35 / .20 / .16
– / .56 / .56
– / .34 / .01
Valuing
.55 / .63 / .51
.42 / .46 / .44
.31 / .25 / .26
– / .68 / .72
– / .39 / .15
ADAPTIVE BEHAVIORS
Planning
.40 / .49 / .21
.40 / .46 / .34
.47 / .35 / .19
– / .49 / .30
– / .42 / .12
Task management
.40 / .48 / .11
.33 / .41 / .26
.39 / .27 / .13
– / .50 / .23
– / .40 / .11
Persistence
.46 / .54 / .35
.51 / .48 / .41
.53 / .37 / .27
– / .60 / .53
– / .48 / .13
IMPEDING / MALADAPTIVE COGNITIONS
Anxiety
-.11 / -.04 / -.23
-.16 / -.08 / -.15
-.62 / -.69 / -.74
– / .02 / -.10
– / .04 / -.06
Failure avoidance
-.16 / -.17 / -.33
-.24 / -.15 / -.19
-.34 / -.31 / -.39
– / -.18 / -.35
– / -.14 / -.11
Uncertain control
-.28 / -.26 / -.29
-.40 / -.25 / -.24
-.52 / -.47 / -.54
– / -.32 / -.31
– / -.24 / .05
MALADAPTIVE BEHAVIORS
Self-handicapping
-.32 / -.34 / -.28
-.40 / -.30 / -.30
-.29 / -.25 / -.25
– / -.40 / -.34
– / -.37 / -.19
Disengagement
-.67 / -.68 / -.57
-.49 / -.46 / -.33
-.29 / -.29 / -.23
– / -.68 / -.67
– / -.47 / -.19
HIGHER ORDER CORRELATIONS
Elementary School / High School / University
Adaptive cognitions
.59 / .67 / .55
.52 / .54 / .52
.41 / .31 / .34
– / .73 / .81
– / .42 / .10
Adaptive behaviors
.55 / .59 / .32
.55 / .53 / .46
.61 / .39 / .28
– / .62 / .50
– / .50 / .16
Impeding/mal cognitions
-.28 / -.20 / -.38
-.41 / -.21 / -.26
-.66 / -.74 / -.87
– / -.19 / -.29
– / -.12 / -.06
Maladaptive behaviors
-.66 / -.71 / -.62
-.54 / -.50 / -.41
-.33 / -.33 / -.30
– / -.72 / -.73
– / -.53 / -.11
Note 1. Elementary school r > +/-.07 significant at p<0.05; High school r > +/-.02 significant at p<0.05 (but note large sample); University r > +/-.12 significant at p<0.05
Note 2. HS results are bolded to assist readability

37
Figure 2
Performance Profile for Motivation and Engagement (adapted from Butler & Hardy, 1992; Martin,
in press b, Weinberg & Gould, 2001) – Reflecting Mean-level /7 (rounded to nearest 0.5) Profile for
High School Sample (N=21,579)
Mastery
orientation
1
2
3
4
5
6
Self-
efficacy
Valuing
Planning
Task
management
Persistence
Anxiety
Failure
avoidance
Uncertain
control
Self
handicapping
Disengagement
IMPEDING/
MALADAPTIVE
COGNITION
MALADAPTIVE
BEHAVIOR
ADAPTIVE
BEHAVIOR
ADAPTIVE
COGNITION
7

38
Appendix. Sample MES items
MES-JS
MES-HS
MES-UC
Self-efficacy
“If I try hard, I believe I can do my schoolwork
well”
“If I try hard, I believe I can do my
schoolwork well”
“If I try hard, I believe I can do my university work
well”
Valuing
“Learning at school is important”
“Learning at school is important”
“Learning at university is important”
Mastery orientation
“I feel very pleased with myself when I really
understand what I’m taught at school”
“I feel very pleased with myself when I really
understand what I’m taught at school”
“I feel very pleased with myself when I really
understand what I’m taught at university”
Planning
“Before I start a project, I plan out how I am
going to do it”
“Before I start an assignment, I plan out how
I am going to do it”
“Before I start an assignment, I plan out how I am
going to do it”
Task management
“I usually do my homework in places where I
can concentrate”
“When I study, I usually study in places
where I can concentrate”
“When I study, I usually study in places where I can
concentrate”
Persistence
“If I can’t understand my schoolwork, I keep
going over it until I do”
“If I can’t understand my schoolwork at first,
I keep going over it until I do”
“If I can’t understand my university work at first, I
keep going over it until I do”
Anxiety
“When I have a project to do, I worry about it a
lot”
“When exams and assignments are coming
up, I worry a lot”
“When exams and assignments are coming up, I
worry a lot”
Failure avoidance
“The main reason I try at school is because I
don’t want to disappoint my parents”
“Often the main reason I work at school is
because I don’t want to disappoint my
parents”
“Often the main reason I work at university is because
I don’t want to disappoint others”
Uncertain control
“When I get a bad mark I don’t know how to
stop that happening again”
“When I get a bad mark I’m often unsure how
I’m going to avoid getting that mark again”
“When I get a bad mark I’m often unsure how I’m
going to avoid getting that mark again”
Self-handicapping
“I sometimes don’t work very hard at school so
I can have a reason if I don’t do well”
“I sometimes don’t study very hard before
exams so I have an excuse if I don’t do as
well as I hoped”
“I sometimes don’t study very hard before exams so I
have an excuse if I don’t do as well as I hoped”
Disengagement
“I’ve given up being interested in school”
“I’ve pretty much given up being involved in
things at school”
“I’ve pretty much given up being involved in things at
university”