Longitudinal Analysis of Intrinsic Motivation and Competence Beliefs : Is There a Relation over Time ?

Intrinsic motivation and competence beliefs
1



Running Head: INTRINSIC MOTIVATION AND COMEPETENCE BELIEFS






Longitudinal Analysis of Intrinsic Motivation and Competence Beliefs: Is There a Relation
over Time?


Birgit Spinath & Ricarda Steinmayr
University of Heidelberg, Germany


Date of Submission: February 20th 2007
1. Revision submitted: August 10th 2007
2. Revision submitted: November 16th 2007
3. Revision submitted: December 16th 2007

---

Intrinsic motivation and competence beliefs
2
Abstract
The present study explored whether competence beliefs and intrinsic motivation for different
school domains show reciprocal effects over time. A sample of N = 670 German elementary
school pupils (M = 8.8 years, SD = 0.51) was followed over one year. At four measurement
occasions, children completed self-reports on their intrinsic motivation and competence
beliefs for math, German, and school in general. Latent growth models revealed that intrinsic
motivation and competence beliefs decreased over time. Comparing correlational and cross-
lagged structural equation models yielded only weak evidence for cross-lagged influences
between the two constructs. Results suggest that the developmental curves of competence
beliefs and intrinsic motivation might be less inextricably interwoven than frequently
assumed.

---

Intrinsic motivation and competence beliefs
3
Longitudinal Analysis of Intrinsic Motivation and Competence Beliefs: Is There a Relation
over Time?
Maybe more than anything else, to be well-equipped for life-long learning, individuals
need high, sustainable motivation to learn. In sharp contrast to this need for life-long learning
stands the observation that important prerequisites for learning, such as intrinsic motivation
for school-related learning, diminish both during an individual’s development (e.g., Gottfried,
1990; Gottfried, Fleming, & Gottfried, 2001; Spinath & Spinath, 2005b) and from one
generation of pupils to the next (Cocodia et al., 2003; Howard, 2001). Therefore, there is a
vital interest among researchers and educators to understand why intrinsic motivation for
learning weakens and to find ways to preserve or reactivate the initially high intrinsic
motivation for learning among young children. The present study investigated whether the
well-established decline of competence beliefs and intrinsic motivation for learning during the
first school years (i.e. in 6 to 10 year old children) are not only parallel developments but
mutually influencing each other. Specifically, we investigated the assumption that intrinsic
motivation to learn declines as a consequence of children developing more realistic and
therefore less positive ability self-concepts.
We chose this particular age for investigation because of the changes in children’s
concepts about ability between 6 and 10 years of age (e.g., Nicholls, 1978; Nicholls & Miller,
1984) and the co-occurring changes in the school environment (i.e. increasingly normative
feedback). These developments should make potential reciprocal effects between competence
beliefs and intrinsic motivation especially well observable. As it is common practice in most
German elementary schools, children received their first grades during the investigated time.
Usually, up until third grade only verbal evaluations are given that allow for less social
comparisons among pupils. Therefore, if there is any relation between competence beliefs and
intrinsic values over time, they should be very pronounced at this phase during school
trajectory.

---

Intrinsic motivation and competence beliefs
4
This study sets out to overcome some of the methodological short-comings of prior
research on related issues, in that its design follows established guidelines for investigating
reciprocal effects in general (Marsh, Byrne, & Yeung, 1999) and relations between
competence beliefs and intrinsic values specifically (Eccles, 2005). We realized a one-year
longitudinal assessment of competence beliefs and intrinsic motivation in three domains
(math, German, and school in general) with four measurement occasions. Using structural
equation models, it was tested whether cross-lagged paths between competence beliefs and
intrinsic motivation at different measurement occasions were necessary to describe the data
adequately (cross-lagged models) or whether more parsimonious correlational models
provided equally good descriptions. Only if competence beliefs and intrinsic values were to
show substantial cross-relation over time, would there be reason to assume that intrinsic
motivation might drop as a consequence of less positive ability self-perceptions.
Defining the concepts of intrinsic motivation and competence beliefs
The concept of intrinsic motivation plays a vital role in different motivation theories
(e.g., Deci & Ryan, 1985; Wigfield & Eccles, 2000). In the present paper, we focus on the
concept of intrinsic task values as included in the expectance-value theory of motivation by
Jacquelynne Eccles and her associates (cf. Wigfield & Eccles, 2000). Intrinsic task values
denote the degree of positive affective evaluation of an activity, i.e. liking, enjoyment, for
reasons that lie within the activity itself rather than its consequences. Although intrinsic task
values are not the only reason for learning, task enjoyment can be considered as the most
desirable state for learners, because learning comes as a by-product of engaging in a
pleasurable activity. Moreover it has been shown that intrinsic task values are the most
important reasons for task engagement in elementary school children who do not yet
differentiate other task values such as utility or importance (e.g., Eccles, Wigfield, Harold, &
Blumenfeld, 1993; Wigfield & Eccles, 1992).

---

Intrinsic motivation and competence beliefs
5
The second motivational construct we focus on in the present work is the concept of
competence beliefs. In the expectancy-value model by Eccles and her colleagues (Eccles et
al., 1983; Wigfield & Eccles, 2000), task values and expectations of future success are the
two most proximal determinants of achievement-related choices. Expectancies for future
success are usually operationalized by domain-specific ability self-perceptions. In school
contexts, students are asked how good they think they are at certain tasks or in certain
domains (e.g., Wigfield et al., 1997). It has been shown that children distinguish very early
between expectancies and values for certain school domains (e.g., Eccles et al., 1993), so that
in elementary school, children’s ability self-perceptions and intrinsic values are distinct
characteristics.
Why should intrinsic motivation and competence beliefs be related?
The assumption that intrinsic motivation for learning and perceptions of one’s
competences are related is derived from both empirical observations and motivation theories.
Experienced teachers as well as developmental researchers observe that not only intrinsic
motivation for learning declines from elementary school age through adolescence but that the
same is true for children’s competence beliefs (e.g., Bouffard, Marcoux, Vezeau, &
Bordeleau, 2003; Gottfried et al., 2001; Spinath & Spinath, 2005b; Wigfield et al., 1997).
Whereas pre-school children view their own competences in an over-optimistic way (cf.
Stipek & MacIver, 1989), older children perceive themselves more realistically and
sometimes even develop over-pessimistic self-perceptions (e.g., Newman, 1984; Spinath &
Spinath, 2005a; Stipek & Daniels, 1988; Wigfield et al., 1997). This parallel decline nourishes
the plausible assumption that children’s competence beliefs and their intrinsic values are
causally linked: When children believe less in their competences they are bound to lose their
enjoyment of task engagement. Yet, whether the decline of competence beliefs and intrinsic
motivation at this specific age are actually mutually influencing or only parallel but
independent developments still needs to be discovered.

---

Intrinsic motivation and competence beliefs
6
Reasons for declining competence beliefs. What are the reasons for children between 6
and 10 years old to perceive their competencies more realistic and therefore, in most cases,
less positive? During this time, children become increasingly able to differentiate between
effort and ability as causes of achievement outcomes, and attain an understanding of ability as
a capacity of the person (Nicholls, 1978; Nicholls & Miller, 1984). Once children attain this
understanding about ability, they are more likely to use social comparison information to
evaluate their own ability (e.g., Ruble, Boggiano, Feldman, & Loebl, 1980; Ruble, Parsons, &
Ross, 1976). In addition, children learn to differentiate more strongly between ability domains
such as physical, social, and academic abilities (Harter, 1982). Taken together, these
developments contribute to children’s growing competence to perceive their own ability
accurately, that is, more in line with valid criteria of their actual ability (e.g., Harter, 1982;
Newman, 1984; Spinath & Spinath, 2005a). Coming from over-optimistic self-perceptions,
for most children these changes mean a downward correction of their ability self-perceptions.
It is probably no coincidence that in most school systems, children are confronted with
more social comparisons and normative feedback at a time when they develop a normative
concept of ability. For example, in most German schools, children receive their first grades
when they are about 9 years old, i.e. in grade three. Up to that point they would have had only
verbal evaluations that are meant to focus on children’s individual development rather than
social comparisons among pupils. Thus, the age group of 9 year olds that we investigated is at
a special phase in their school trajectory where normative evaluations of competence become
extremely salient (e.g., Eccles et al., 1993). We reasoned that if there are mutual influences
between competence beliefs and intrinsic values over time, they should be very pronounced at
this phase during school trajectory.
Theoretical grounds on which to assume a causal link between ability beliefs and
intrinsic motivation. A theoretical background for the assumption that intrinsic motivation
depends on competence perceptions is provided by White’s (1959) theorizing about

---

Intrinsic motivation and competence beliefs
7
effectance motivation. According to White, individuals have an innate desire to feel
competent. This effectance motivation drives individuals to engage in tasks in which mastery
is at stake. Tasks that generate feelings of efficacy should be experienced as enjoyable or
intrinsically rewarding and, as a by-product, prompt learning. Thus, in White’s theory, feeling
competent and being intrinsically motivated are the same. One has to note, however, that
White’s concept of feelings of efficacy or competence is not the same as the ability beliefs we
focus on in the present investigation. Competence beliefs as operationalized by ability self-
perceptions are cognitive representations about the level of ability one perceives to have.
Nevertheless, White’s theorizing provided the basis on which many motivation researchers
grounded the assumption of a causal relation between competence beliefs and intrinsic
motivation.
Building up on White’s (1959) more general theorizing, Harter (1981a) refined and
extended White’s assumptions in an effort to provide testable hypotheses under a
developmental perspective on competence beliefs. Harter distinguishes between competence
beliefs as cognitive representations of the level of one’s ability and motivational orientations
(i.e. intrinsic and extrinsic motivation). In Harter’s effectance motivation model (1981a,
p.38), intrinsic pleasure in task engagement is a function of successful mastery of challenging
tasks and perceived competence. Conversely, failure and perceived lack of competence should
result in anxiety in mastery situations, an emotion known to be incompatible with enjoyable
feelings. In postulating this sequential process, Harter disentangles perceived competence and
motivational orientation and predicts that higher levels of perceived competence entail higher
levels of intrinsic motivation. Some of the most prolific modern motivation theories share the
assumption that more positive ability self-perceptions should generate more intrinsic
motivation for a given task (e.g., Deci & Ryan, 1985; Wigfield & Eccles, 2000).
It is important to note that all theoretical reasoning about the relation between
competence beliefs and intrinsic motivation assumes that prior competence beliefs influence

---

Intrinsic motivation and competence beliefs
8
subsequent intrinsic motivation. Although one could argue that prior intrinsic motivation
could also have an influence on subsequent competence beliefs, there is no elaborated
theoretical background making a strong point for such directional effects. Therefore, in the
present investigation, we focus on the question whether there is evidence that competence
beliefs might influence intrinsic motivation.
Why could competence beliefs and intrinsic motivation develop in parallel but without
mutual influences? While there is strong theoretical reasoning on which to ground the
assumption that competence beliefs are important for intrinsic motivation, it is also
worthwhile to look for theoretical reasons why such a link might not exist or be only weak.
One could argue that it is not the absolute level of self-perceived ability that is important for
the degree of task enjoyment but other kinds of ability related self-perceptions. One line of
reasoning can be drawn from Harter’s (1981a) specifications to White’s (1959) theory of
effectance motivation. This specification concerns the relation between success and feelings
of efficacy. Harter pointed out that feelings of efficacy, and therefore intrinsic motivation for
task engagement, are maximized not by the mere quantity of success but by succeeding at
optimally challenging tasks. In other words, intrinsic motivation should depend on an optimal
fit between task difficulty and own competence. Such an optimal fit can be given at different
levels of competence, so that not only the most competent individuals experience intrinsic
motivation. Therefore, the degree of intrinsic motivation might not so much depend on the
absolute level of normatively-based ability perceptions but rather on perceiving oneself as
successful at tasks that fit the level of one’s competence. If this was true, we would expect no
or only weak indications of normatively-based ability self-perceptions influencing intrinsic
motivation.
Taken together, developmentally focused motivational theories often assume a causal
influence of competence beliefs on intrinsic motivation. Because competence beliefs undergo
specific changes in the early school years that lead most children to less positive ability self-

---

Intrinsic motivation and competence beliefs
9
perception, declines in intrinsic motivation seem to be an inevitable consequence from this
point of view. If, however, it could be shown that the development of intrinsic motivation is
not or only weakly dependent on the development of competence beliefs, then the trajectories
of the two might be disentangled by means of interventions.
Methodological guidelines for investigating reciprocal effects over time.
After theorizing about the nature of the relation between competence beliefs and
intrinsic motivation, we take a look at the empirical literature. There is ample evidence on the
positive correlation among intrinsic motivation and competence beliefs within domains (e.g.,
Deci & Ryan, 1985; Harter, 1981b; Wigfield, 1994; Wigfield et al., 1997). Most of these
studies are cross-sectional, correlational and allow no conclusions about a potentially causal
relationship between the two constructs over time. In order to investigate whether prior
competence beliefs might be influential for subsequent intrinsic motivation, longitudinal
designs are needed.
In an effort to develop stronger methodological approaches to test for reciprocal
effects between concepts in longitudinal designs, Marsh (1990) and Marsh et al. (1999, p.
156) provided valuable guidelines. These authors state that in order to detect causal relations,
first, two constructs need to be shown to have a significant statistical relation (i.e., substantial
path coefficients in structural equation models). Second, clear time precedence needs to be
established in longitudinal studies with at least two and preferably more measurement
occasions. Third, theoretical models must be tested by means of statistical techniques such as
Structural Equation Modelling (SEM) with all latent constructs inferred on the basis of
multiple indicators. The most adequate description of the empirical data is derived by fitting
the data to different theoretically plausible models among which to choose the best fitting
one.
The present study realizes all of these recommendations to examine the relationship
between competence beliefs and intrinsic motivation. We are not aware of another study that

---

Intrinsic motivation and competence beliefs
10
conducted a one-year longitudinal assessment of competence beliefs and intrinsic motivation
in three domains (math, German, and school in general) with four measurement occasions.
The two constructs were measured on a latent basis by three manifest indicators each. This
design allows for comparing different theoretical models that differ in whether or not they
assume causal relations between competence beliefs and intrinsic motivation. Specifically, it
was tested whether cross-lagged paths between competence beliefs and intrinsic motivation at
different measurement occasions were necessary to describe the data adequately (cross-lagged
models) or whether more parsimonious correlational models provided equally good
descriptions. According to a basic scientific principle, in the case of two competing models
which describe a phenomenon equally well, the more parsimonious model or theory should be
preferred over the more complex one. Although the preference of the more parsimonious
model in the case of an equal fit to the data does not rule out the existence of cross-lagged
effects, such effects would be rather weak.
Of course, longitudinal designs cannot provide unequivocal evidence for causal
influences. They are, however, very useful to test for necessary prerequisites of causal
influences and give an impression of the how strong such influences might be. Concerning the
present investigation, this means that in case of no cross-lagged relations between prior
competence beliefs and later intrinsic motivation, there is no basis for further investigating
causal influences. Furthermore, the present design provides an estimation of the upper limit of
the strength of potential causal influences. If there are only weak relations over time between
competence beliefs and intrinsic motivation, this indicates that causal influences between the
two might at best be weak.
Previous longitudinal studies investigating the link between intrinsic motivation and
competence beliefs

---