Clarifying Achievement Goals and Their Impact
Heidi Grant and Carol S. Dweck
The study of achievement goals has illuminated basic motivational processes, though controversy
surrounds their nature and impact. In 5 studies, including a longitudinal study in a difficult premed
course, the authors show that the impact of learning and performance goals depends on how they are
operationalized. Active learning goals predicted active coping, sustained motivation, and higher achieve-
ment in the face of challenge. Among performance goals, ability-linked goals predicted withdrawal and
poorer performance in the face of challenge (but provided a “boost” to performance when students met
with success); normative goals did not predict decrements in motivation or performance; and outcome
goals (wanting a good grade) were in fact equally related to learning goals and ability goals. Ways in
which the findings address discrepancies in the literature are discussed.
Considerable evidence suggests that much of achievement mo-
tivation (e.g., intrinsic interest, strategy use, and persistence) can
be understood in terms of the different goals individuals bring to
the achievement context (see Ames, 1992; Ames & Archer, 1988;
Butler, 1987, 1993; Dweck & Elliott, 1983; Dweck & Leggett,
1988; Elliott & Dweck, 1988; Harackiewicz, Barron, Carter,
Lehto, & Elliot, 1997; Kaplan & Maehr, 1999; Middleton &
Midgely, 1997; Nicholls, 1984; Pintrich, 2000a; Rawsthorne &
Elliot, 1999; Utman, 1997). However, there are some disagree-
ments and some conflicting findings on the nature of these rela-
tions. Specifically, researchers disagree on how to best define and
operationalize the major classes of goals, and on the precise impact
of these goals on motivation and achievement.
In the original goal models, two classes of goals were identi-
fied—performance goals, where the purpose is to validate one’s
ability or avoid demonstrating a lack of ability, and learning goals,
where the aim is to acquire new knowledge or skills (i.e., to
increase one’s ability; see Dweck & Elliott, 1983). Different
researchers have used different labels for these two classes of
goals—performance goals have also been called ego-involved
goals (e.g., Nicholls, 1984) or ability goals (e.g., Ames, 1992), and
learning goals have also been called mastery goals (e.g., Ames,
1992; Butler, 1993; Elliot & Harackiewicz, 1996; Meece & Holt,
1993) or task goals (e.g., Middleton & Midgely, 1997; Nicholls,
These two classes of goals were then linked to motivation and
performance in achievement situations. Performance goals, with
their emphasis on outcomes as measures of ability, were shown to
produce a vulnerability to helplessness and debilitation after a
setback or negative feedback, particularly in cases where current
perceptions of ability were low (Ames & Archer, 1988; Butler,
1993; Elliott & Dweck, 1988; Jagacinski & Nicholls, 1987; Meece,
Blumenfeld, & Hoyle, 1988). That is, when the goal is to validate
ability and individuals do not believe they can accomplish this,
motivation and performance tend to suffer. Learning goals, with
their emphasis on understanding and growth, were shown to fa-
cilitate persistence and mastery-oriented behaviors in the face of
obstacles, even when perceptions of current ability might be low
(Ames & Archer, 1988; Butler, 1993, Elliott & Dweck, 1988;
Jagacinski & Nicholls, 1987; Utman, 1997).
Performance and learning goals have also been shown to predict
real-world performance, including exam grades, course grades,
and achievement test scores, controlling for past performance
(Dweck & Sorich, 1999; Greene & Miller, 1996; Kaplan & Maehr,
1999; Meece & Holt, 1993; Midgely & Urdan, 1995; Roeser,
Midgely, & Urdan, 1996). In addition, goal effects obtain both
when the goals have been experimentally manipulated (Butler,
1987; Elliott & Dweck, 1988; Graham & Golen, 1991), and when
students’ naturally existing goals have been assessed (Ames &
Archer, 1988; Bouffard, Boisvert, Verzeau, & Larouche, 1995;
Midgely, Anderman, & Hicks, 1995; Miller, Behrens, Greene, &
Newman, 1993; Pintrich & DeGroot, 1990; Pintrich & Garcia,
1991). The fact that induced goals have been found to have strong
impact is important for two reasons. First, it means that goals can
have a causal role in producing achievement patterns. Second, it
means that learning environments can be constructed in ways that
enhance achievement (Ames, 1992; Maehr & Midgley, 1991;
Roeser et al., 1996).
Despite early agreement regarding the effects of performance
and learning goals on motivation and performance, recent research
has revealed a more complicated picture. Some researchers have
questioned whether learning goals affect performance at all, sug-
Heidi Grant and Carol S. Dweck, Department of Psychology, Columbia
This article is based on a doctoral dissertation submitted to Columbia
University by Heidi Grant under the supervision of Carol S. Dweck. It
was supported by National Institute of Mental Health Grant F31-
MH12706-01 to Heidi Grant. We thank Dean Kathleen McDermott and
Professor Leonard Fine and the Department of Chemistry at Columbia
University for their support and assistance with this project. We are also
grateful for the comments and suggestions given by the dissertation com-
mittee: Geraldine Downey, E. Tory Higgins, Harvey Hornstein, and Gab-
riele Oettingen. Finally, we thank Andrew Eliot, Judith Harackiewicz, and
Corwin Senko for their insightful comments on an earlier version of this
Correspondence concerning this article should be addressed to Heidi
Grant, who is now at the Department of Psychology, New York University,
6 Washington Place, 7th Floor, New York, New York 10003. E-mail:
Journal of Personality and Social Psychology Copyright 2003 by the American Psychological Association, Inc.
2003, Vol. 85, No. 3, 541–553 0022-3514/03/$12.00 DOI: 10.1037/0022-35126.96.36.1991
gesting that they chiefly influence intrinsic motivation (e.g., Bar-
ron & Harackiewicz, 2001; Elliot & Church, 1997; Harackiewicz
et al., 1997; Harackiewicz, Barron, Tauer, Carter, & Elliot, 2000).
Some have argued that performance goals predict higher, not
lower, grades, and do not affect intrinsic motivation (e.g., Barron
& Harackiewicz, 2001; Elliot & Church, 1997; cf. Rawsthorne &
We propose that looking at the ways in which performance and
learning goals have been defined or operationalized can help
account for the discrepant findings that have been obtained by
different researchers. To test this proposal, items were created to
measure the different forms of goals that have been prominently
represented in existing research. Five studies explore the relation-
ships among these goals, their ability to predict intrinsic motiva-
tion and performance under highly challenging or difficult circum-
stances, and the mechanisms through which they may bring about
those effects. We begin by describing the important dimensions
along which the operationalizations of performance and learning
goals vary in current achievement goal research, and describing
how each of these dimensions is represented in the following
What Is a Performance Goal and What Is Its Effect?
Achievement goal researchers have already made one important
distinction among performance goals—namely, the distinction be-
tween performance approach goals (where the focus is on attain-
ing success) and performance avoidance goals (where the focus is
on the avoidance of failure; Elliot, 1999; Elliot & Church, 1997;
Elliot & Harackiewicz, 1996; Middleton & Midgely, 1997; Pin-
trich, 2000a). In general, this program of research has suggested
that it is the avoidance form of performance goals that predict
lower intrinsic motivation and performance, with approach goals
often relating positively to performance.
However, as discussed below, the positive and negative effects
of performance approach goals have typically been found when
performance goals are operationalized in particular ways, and the
positive and negative effects of different types of performance
approach goals have not been systematically explored. Thus our
purpose in this article is to distinguish among approach forms of
performance goals, and we propose that they take at least three
distinct forms: (a) goals that are linked to validating an aspect of
self (e.g., one’s ability), (b) goals that are explicitly normative in
nature, and (c) goals that are simply focused on obtaining positive
outcomes (i.e., doing well). It is the first form that was linked to
impairment in the earlier models, but it has tended to be the second
two forms that have been linked to more positive outcomes in
recent work. Let us take a closer look at these different forms of
For some researchers, the essence of a performance goal is
seeking to validate one’s ability (operationalized either by sug-
gesting to participants that their performance on a task measures
the extent to which they possess a valued ability, or by assessing
the extent to which they generally strive to validate their ability).
Debilitation occurs when outcomes indicate a lack of ability, but
performance maintenance or enhancement can occur when success
is expected (Ames, 1992; Elliott & Dweck, 1988; see Dweck &
Leggett, 1988). It should be noted that debilitation here requires
the presence of challenges, setbacks, or failure—an easy task or
course is not expected to produce debilitation, even in the presence
of strong performance goals. To represent this view, we developed
ability goal items (e.g., “It is important to me to validate that I am
For others, the essence of a performance goal is a normative
comparison (i.e., wanting to perform better than others), and a goal
that is nonnormative (e.g., using an absolute standard such as a
perfect score, or tying absolute performance to self-worth) is not
considered to be a performance goal (Elliot, 1999; Elliot &
Church, 1997; Elliot & Harackiewicz, 1996; Maehr & Midgely,
1991; Pintrich, 2000b). Here, performance goals are often opera-
tionalized by informing participants that their performance on a
task will be evaluated normatively, or by measuring their agree-
ment with statements such as “It is important to me to do well
compared to others in this class” (Elliot & Church, 1997).
The issue of whether normative performance goals are empiri-
cally distinct from performance goals that do not contain a nor-
mative standard has not been systematically addressed in the
achievement goal literature. Yet it is an important question, be-
cause to some theorists, as noted, the presence of normative
comparison is the essence of a performance goal (Elliot & Harac-
kiewicz, 1996; see Rawsthorne & Elliot, 1999), and to others, a
potentially interesting but nonessential aspect of a performance
goal (Elliott & Dweck, 1988). It would be interesting to find that
normative and nonnormative performance goals do indeed differ,
particularly if these differences could illuminate discrepancies in
the reported effects of performance goals on motivation and per-
formance. The following studies contain both normative and non-
normative versions of performance goals. An example of an ex-
plicitly normative goal would be the following: “One of my major
goals in school is to feel that I am more intelligent than other
students.” In contrast, the goal item, “It is important to me to
validate that I am intelligent, ” is not explicitly normative.
Sometimes goal items used to measure performance-goal orien-
tation simply ask the participant about wanting to do well on a
task, such as wanting to earn a high grade in a course. For people
who are focused on doing well, negative outcomes do not neces-
sarily indicate a lack of ability (i.e., holding this type of goal does
imply a particular causal attribution for success or failure). We
refer to the goal of wanting to do well on a particular task as an
outcome goal, and it, too, is represented in our studies (e.g., “It is
important to me to get good grades in my classes.”). A closely
related construct is “competence valuation,” or the degree to which
a task is perceived to be important (Elliot & McGregor, 2001),
which has been found to relate positively to intrinsic motivation
and performance (Barron & Harackiewicz, 2001). We find this
type of goal particularly interesting, because “wanting to do well”
can also be an important part of a learning goal framework. In
other words, a person with a learning goal may care very much
about doing well on a task, but perhaps for different reasons (i.e.,
in order to maximize learning, as an indicator of successful learn-
ing, or for instrumental reasons). Later, we address the question of
whether outcome goals are best understood as performance goals.
What Is a Learning Goal? When Is It Helpful?
There is generally less controversy and more agreement with
respect to the nature of learning goals. As noted learning goals,
task goals, and mastery goals have often been regarded as concep-
GRANT AND DWECK
tually equivalent (Ames, 1992; Linnenbrink & Pintrich, 2000).
Yet, potentially important differences among operationalizations
do exist. For some (Ames, 1992; Elliot & Church, 1997; Elliott &
Dweck, 1988; Harackiewicz et al., 1997; Middleton & Midgley,
1997), a learning goal is an active striving toward development and
growth of competence, and is operationalized by emphasizing the
importance and benefits of learning some new knowledge or skill
to the participant, or by asking participants to indicate the extent to
which learning and developing new skills are major academic
goals. However, the terms “task goals” and “mastery goals” do not
put an explicit emphasis on learning; thus, we thought it important
to test the extent to which the desire to learn may be similar or
different from the desire to master challenges. As a result, we
included items measuring two forms of learning goals. An example
of a learning goal without an explicit challenge-mastery compo-
nent is “I strive to constantly learn and improve in my courses.” An
example of an explicit challenge-mastery item is “It is very im-
portant to me to feel that my coursework offers me real
It should be reiterated that, despite the substantial agreement
among researchers with respect to the concept of a learning goal,
the data with respect to the influence of learning goals on moti-
vation and performance are not without inconsistencies. Typically,
those who adopt learning goals are found to engage in deeper,
more self-regulated learning strategies, have higher intrinsic mo-
tivation, and perform better, particularly in the face of challenge or
setbacks (Ames, 1992; Dweck & Leggett, 1988; Kaplan & Midg-
ley, 1997; Pintrich, 2000a; Pintrich & Garcia, 1991; Utman, 1997;
see also Barron & Harackiewicz, 2000). However, recently, sev-
eral studies have failed to find enhanced performance outcomes
resulting from learning goals (although enhanced intrinsic moti-
vation was found; Elliot & Church, 1997; Elliot, McGregor, &
Conditions Under Which Goal Effects Are Tested
The effects of learning and performance goals on motivation
and achievement have been tested under a wide variety of circum-
stances—with students working on interesting “NINA” puzzles
(Elliot & Harackiewicz, 1996), performing a concept-formation
task (Elliott & Dweck, 1988), solving math problems (designed to
be highly challenging in one condition; Barron & Harackiewicz,
2001; cf. Middleton & Midgely, 1997), or taking an intermediate-
level psychology course (Elliot & Church, 1997). Importantly,
these tasks may have varied with respect to the degree of difficulty
or challenge encountered by the participant, and the degree to
which performance on the task had importance or meaning to the
participant. We feel that conditions where the degree of difficulty
is substantial for a large number of participants and the outcome is
highly important are more likely to reveal goal effects on motiva-
tion, coping, and achievement, and have tried to use such condi-
tions in the studies reported here.
In summary, there have been major differences in the ways
goals have been operationalized, and it is not surprising that the
data are inconsistent with respect to how and when performance
and learning goals affect motivation and achievement. In the
following studies, we attempted to illuminate these issues. In three
studies, we developed and tested a set of items to tap different
forms of learning and performance goals. In the fourth study, to
gain an initial sense of the patterns associated with each goal type,
we presented students with scenarios depicting important aca-
demic setbacks and examined how the different goals predicted
intrinsic motivation and coping. In the fifth study, the different
goals were used to predict intrinsic motivation, study strategies,
and performance in an important and challenging course.
Given the number of goals we hoped to measure and compare
(e.g., ability goals, outcome goals, normative outcome goals, nor-
mative ability goals, learning goals, and challenge-mastery goals),
we wanted to use the fewest possible items to measure each type
of goal while still maintaining high reliability. It was felt that using
relatively few items would minimize the frustration and confusion
participants might experience when required to answer many sim-
ilarly worded items. Thus, 10 items for each of type of goal were
created and carefully tested with 560 participants, and the most
reliable three items were selected for each goal. In three prelimi-
nary studies (Studies 1–3) reported below, the items used to
measure each type of goal achieved good reliability and validity, as
demonstrated by the relatively high alphas for each group of items,
the high correlation of each group of items with scales assessing
conceptually similar variables, and the high test–retest correlations
(the Appendix contains the complete list of items). For each goal
item, participants were asked to rate their agreement on a 7-point
Likert-type scale ranging from 1 (strongly disagree)to7(strongly
A total of 451 participants (218 men and 233 women) recruited from the
Columbia University student population were paid $5 for their participa-
tion. Fifty-seven percent of participants were Caucasian, 9% African
American, 19% Asian, 8% Latino, and 7% were “other” or unidentified.
Participants were asked to complete a goal inventory containing the
three items for each of the types of goals along with several unrelated
measures. The goal items were presented in random order. Participants in
this initial study (and all subsequently reported studies) read and signed
consent forms that informed them about the procedure, the information that
would be asked of them, and their rights as research participants. They
were reminded that they were free to leave the study at any time without
Exploratory Factor Analysis
Principal-components analysis, using varimax rotation and eig-
envalues greater than 1, yielded four factors (accounting for 24%,
20%, 16%, and 12% of the total variance, respectively). Factor 1
contains all normative items (both normative ability and normative
outcome). Factor 2 contains both the learning and challenge-
mastery items. Factor 3 contains all of the nonnormative outcome
goal items. Factor 4 contains all nonnormative ability goal items.
CLARIFYING ACHIEVEMENT GOALS
This analysis was repeated using oblimin rotation, which yielded
nearly identical results (see Table 1).
These analyses revealed that, consistent with our expectations,
learning goals (Factor 2), outcome goals (Factor 3), and ability
goals (Factor 4) are distinct constructs. As we will see, learning
and ability goals are the major classes of predictive goals in
subsequent studies. In addition, participants clearly distinguished
between normative and nonnormative forms of performance goals,
suggesting that this distinction is a meaningful and potentially
important one. Normative goals loaded together (Factor 1), sug-
gesting that participants who endorsed them did not distinguish
strongly between normative outcome and normative ability goals.
Similarly, challenge-mastery goals did not distinguish themselves
reliably from learning goals. Thus, four groups or classes of goals
emerged in this study, and we focus on these groups throughout the
remainder of the article: learning (comprised of both learning and
challenge-mastery items), outcome, ability, and normative (com-
prised of both normative outcome and normative ability items).
Although Cronbach’s alpha is dependent on the length of the
measure (i.e., number of items in a scale), our goal measures
nonetheless achieve substantial alphas. The alphas for each of the
four subsets of the goal items (ranging from .81 to .92) had an
average of .86, consistent with unidimensionality for each set of
Confirmatory Factor Analysis (CFA)
CFA was conducted on the achievement goal items using
EQS 5.7 (Bentler & Wu, 1995). Solutions were generated on the
basis of maximum-likelihood estimation. Seven models were
tested. For each model, we calculated multiple indices of fit:
chi-square, comparative fit index (CFI), normed fit index (NFI),
nonnormed fit index (NNFI), root-mean-square error of approxi-
mation (RMSEA), and Akaike Information Criterion (AIC), a
comparison statistic for nonhierarchical models. The results from
these analyses indicated that two models provided a good fit to the
data: Model A,
(123, N ⫽ 451) ⫽ 490.23, CFI ⫽ .93, NFI ⫽ .91,
NNFI ⫽ .92, RMSEA ⫽ .08, AIC ⫽ 244; Model B,
(120, N ⫽
451) ⫽ 395.76, CFI ⫽ .95, NFI ⫽ .93; NNFI ⫽ .94, RMSEA ⫽
.07, AIC ⫽ 155). Although Model B (a six-factor model) does
provide a slightly better fit, Model A—a hierarchical model with
four primary factors (i.e., an ability goal factor, an outcome goal
factor, a normative factor comprised of normative ability and
normative outcome factors, and a learning factor comprised of
learning and challenge-mastery factors)—is consistent with the
results from the two principal-components analyses, as well as
with the pattern of item intercorrelations and scale alphas. Thus, a
four-factor goal model, consisting of learning goals, outcome
goals, ability performance goals, and normative performance
goals, received the most consistent support and provided the best
overall fit to the data.
Correlations Among Classes of Goals
All four goal indices were positively correlated. Outcome goals
(wanting to do well) appear to accompany the valuing of any
achievement goals, whether those goals pertain to learning (r ⫽
.37, p ⬍ .001), to validating one’s ability (r ⫽ .53, p ⬍ .001), or
to outperforming others (r ⫽ .34, p ⬍ .001). Learning goals were
positively related to outcome goals, as noted, as well as ability
goals (r ⫽ .41, p ⬍ .001) and normative goals (r ⫽ .17, p ⬍ .001).
Finally, ability goals and normative goals were strongly correlated
(r ⫽ .52, p ⬍ .001). Although it appears that individuals who value
achievement may value many aspects of it, we will see that clearly
distinct and unique patterns are associated with each type of goal.
Principal-Component Factor Analysis With Item Loadings
Goal item type Item no.
12 3 4
Learning 1 .72 (.74)
2 .70 (.68)
3 .66 (.62) .51 (⫺.44)
Challenge-mastery 1 .84 (.87)
2 .80 (.82)
3 .74 (.75)
Outcome 1 .81 (⫺.79)
2 .74 (⫺.71)
3 .87 (⫺.86)
Ability 1 .50 (⫺.46)
2 .79 (⫺.80)
3 .81 (⫺.83)
Normative outcome 1 .81 (.85)
2 .81 (.86)
3 .83 (.86)
Normative ability 1 .82 (.78)
2 .81 (.77)
3 .85 (.82)
Note. All loadings above .40 are shown. Oblimin rotation values are shown in parentheses.
GRANT AND DWECK
Analysis by Gender
Tests for mean difference in goal ratings by gender revealed
several significant differences, though the pattern of differences
varied across studies. Because these differences did not replicate
across studies, there is little reason to believe that any were
representative of the general population. It is important to note that
there were no interactive effects of goal and gender in any of the
studies. In other words, performance and learning goals exerted the
same effects on both men and women in each study. Therefore, in
the interest of brevity, gender differences will not be discussed for
A total of 54 participants (23 men, 31 women) recruited from the
Columbia University student population were paid $10 for their
Participants completed the goal items as part of a battery of measures,
and then completed the items again in another battery of measures ex-
actly 2 weeks later.
Correlations between Time 1 and Time 2 ratings were calculated
for each goal. The correlations ranged from .69 to .88, and the
average test–retest correlation was .79. Thus, participants’ scores
were substantially consistent over time.
Study 3 was designed to obtain construct validity for the goal
items by relating them to other goal measures. Two commonly
used measures of achievement goal orientation were chosen (i.e.,
Button, Mathieu, & Zajac, 1996, and Elliot & Church, 1997). In
particular, it was important to show that (a) our measure of
learning goals mapped onto other operationalizations of learning
goals, (b) that our measure of normative performance goals was an
accurate representation of how these goals have been measured in
the literature, and (c) that our outcome goals were equally related
to learning and performance goals in other measures, as they had
been in ours. Neither measure taps ability goals as we have defined
Learning and Performance Orientation Scales (Button et al., 1996).
Button et al.’s (1996) inventory is composed of two scales (Learning and
Performance), each containing eight items. In general, Button et al.’s
learning goal items capture the conceptualization proposed by Dweck and
Elliott (1983)—an emphasis on challenge-seeking, use of effort and strat-
egies, and desire to develop and grow. Button et al.’s performance items
emphasize wanting to do well and not make mistakes, though there are two
items that involve social comparison and the opinions of others.
Elliot and Church’s (1997) Achievement Goal Scale. The goal orien-
tation scale used by Elliot and his colleagues in their classroom studies
(Elliot & Church, 1997; Elliot et al., 1999) consists of three subscales with
six items each, of which we focused on two: Mastery and Performance
Approach (the third subscale is Performance Avoidance). The Mastery
items emphasize wanting to learn as much as possible and thoroughly
master new material. Performance Approach items emphasize wanting to
do better than others (i.e., normative items).
A total of 87 participants (37 men, 50 women) were recruited from the
Columbia University student community and paid $5 for their participa-
tion. Sixty percent of participants were Caucasian, 22% African American,
13% Asian, and 5% were “other” or unidentified.
Participants completed our goal items, along with the Learning and
Performance Orientation Scales (Button et al., 1996), and the Achievement
Goals Scale used by Elliot and his colleagues (Elliot & Church, 1997). The
three measures were presented in three different orders across participants.
There were no discernable effects of order.
As expected, the learning goal items were highly positively
correlated with Button et al.’s (1996) Learning scale (r ⫽ .72, p ⬍
.001) and Elliot and Church’s (1997) Mastery scale (r ⫽ .76, p ⬍
.001). This suggests that the items are valid indices of a learning
In Study 1, outcome goal items, with their focus on the value of
doing well, were compatible with learning goals, ability goals, and
normative goals. They were also positively correlated with both
Button et al.’s (1996) Learning (r ⫽ .37, p ⬍ .01) and Performance
(r ⫽ .45, p ⬍ .001) Scales, as well as with Elliot and Church’s
(1997) Mastery (r ⫽ .41, p ⬍ .001) and Performance Approach
(r ⫽ .30, p ⬍ .01) scales. This is further evidence of the hybrid
nature of outcome goals.
Ability goal items were positively correlated with Button et al.’s
(1996) Performance Scale (r ⫽ .45, p ⬍ .001) and Elliot and
Church’s (1997) Performance Approach scale (r ⫽ .46, p ⬍ .001),
but only moderately, because neither of those scales focus on
As predicted, normative goal items were highly correlated with
Elliot and Church’s (1997; normative) Performance Approach
scale (r ⫽ .83). Ability and outcome goal items were significantly
less correlated with this scale (outcome r ⫽ .30 vs. normative r ⫽
.83, t ⫽ 4.01, p ⬍ .001; Ability r ⫽ .46 vs. Normative r ⫽ .83,
t ⫽ 1.98, p ⬍ .05).
In summary, a comparison of these three measures yielded
evidence for the construct validity of our goal items. High corre-
lations with conceptually similar subscales in the Button et al.
(1996) and Elliot and Church (1997) measures can be taken as
evidence that the items are tapping into the right goal constructs.
We believe that it is important to look at goal effects when
individuals experience major setbacks or failure on highly valued
CLARIFYING ACHIEVEMENT GOALS
tasks, because it is under these conditions that we would expect
goal effects on motivation, coping, and achievement to be maxi-
mal. Studies 4 and 5 were designed to look at how each of the
different goals we identified predicts indices of intrinsic motiva-
tion, mastery-oriented coping, and performance, after a significant
or sustained difficulty or setback, by means of hypothetical failure
scenarios (Study 4), reports of habitual coping style (Study 4), and
a very challenging premed college course (Study 5).
We also included measurements of some of the affective and
cognitive variables that comprise the psychological processes that
accompany goal pursuit. Much recent achievement goal work pays
little attention to the psychological concomitants of goals: attribu-
tions, beliefs, and contingency of self-worth (Molden & Dweck,
2000). By including these measures, we hoped to capture a richer
motivational picture of performance and learning goal processes.
In Study 4, two scenarios were generated in which the par-
ticipant encounters failure in an important achievement setting
(adapted from Zhao & Dweck, 1997). The use of hypothetical
scenarios was used here as a first step in relating the different goals
to the variety of cognitive, affective, and behavioral variables
involved in coping with difficulty in achievement situations.
Participants in Study 4 also completed a measure of chronic
coping style (COPE; Carver, Scheier, & Weintraub, 1989), so that
we might look at the relationship between goal orientation and
participants’ own personal history of coping with setbacks. Thus,
the first part of Study 4 asks participants to indicate how they
would respond to a situation if it occurred, and the second part of
Study 4 asks them to reflect on past situations that have actually
A total of 92 participants (40 men, 52 women) were recruited for pay
from the Columbia University community. Sixty-one percent of partici-
pants were Caucasian, 21% African American, 12% Asian, and 6% were
“other” or unidentified. They received $5 for their participation.
Participants completed the goal items, and then, after a 5-min word-
completion filler task, they received one of two randomly assigned sce-
narios, shown in previous work to elicit motivational differences (see Zhao
& Dweck, 1997). The scenario asked them to read about a failure experi-
ence in a college classroom (either getting a bad grade on an important
essay in a key course or doing poorly on the Graduate Record Examination
when they strongly wished to go to graduate school), and to imagine it
happening to them. These two scenarios were vividly written and were
selected to represent situations that they could easily personally relate to
(i.e., doing poorly on an essay in a course in your major, and doing poorly
on a test in science class). Here is an example:
Imagine that during your second semester at Columbia, you take an
important course in your major, in which students are required to read
their essays out loud to the entire class. This happens several times
throughout the semester. The time comes for the first readings. By the
time it’s your turn, most of the students have already presented their
essays. All of them did pretty well, and you know that their essays got
good grades from the professor. But when you read your essay to the
class, the professor and the other students don’t seem to like your
presentation very much, and later you find out that the grade he gave
you was a C⫺.
Participants were then asked to indicate what they would think, how they
would feel, and how they would behave after the failure by rating their
degree of agreement with a series of statements. These statements include
items assessing loss of intrinsic motivation (e.g., “I’d probably feel less
interested in the subject”), help-seeking (e.g., “I would seek help from my
professor or my classmates”), planning (e.g., “I’d start planning how to do
better on the next presentation”), and time and energy withdrawal (e.g., “I
would devote less time and energy to the class”), as well as attributions for
the failure (e.g., “I would feel like I wasn’t smart enough”), loss of
self-worth (e.g., “I would feel like a loser”), and rumination (e.g., “I would
dwell on my failure”). Responses were made by circling a number on a
7-point scale ranging from 1 (not at all true of me) to7(very true of me).
After a second 5-min word-completion filler task, participants were
asked to complete the Ways of Coping Scale (COPE; Carver et al., 1989).
This scale measures the ways in which individuals have coped with
difficulties when they have arisen. Subscales include Active Cop-
ing, Planning, Positive Reinterpretation, Denial, and Behavioral
For each of the analyses conducted, scenario version (1 or 2)
was entered as a predictor, and no effect for scenario version was
found. Therefore, all analyses reported were conducted collapsing
across scenario version. Each of the four goal types (learning,
outcome, ability, and normative) was entered as a predictor in a
series of simultaneous regressions that included all two-way inter-
actions. There were no significant two-way interactions, so these
terms were dropped in subsequent analyses. Thus, the effects of
each goal on the variables of interest control for any effects of the
other three classes of goals. In this way, we could determine what,
if any, were the unique effects of each class of goal on our
Table 2 depicts the unique relationship between each type of
goal and an index of loss of intrinsic motivation, created by adding
together responses from the following three items (
⫽ .89): “I’d
probably feel less interested in the subject,”“I probably wouldn’t
enjoy the course as much as before,” and “I wouldn’t really be
excited about the course anymore.”
As can be seen, learning goals were negatively related to de-
creases in intrinsic motivation, whereas outcome and ability goals
were significantly correlated with decreases in intrinsic motiva-
tion. Of interest, normative goals did not predict loss of intrinsic
motivation. This finding is worth noting, in that the program of
research that has most consistently found that approach forms of
performance goals do not negatively influence intrinsic motivation
has used a normative definition of performance goal (e.g., Elliot &
Church, 1997). Also, Epstein and Harackiewicz (1992) have found
that students high in achievement motivation who were assigned
competitive goals (which are inherently normative) experienced
increased interest in a task when they were given a failure expec-
tancy. This finding suggests that competitive strivings may buffer
individuals when they experience difficulty or failure, in ways that
ability-focused strivings do not.
GRANT AND DWECK
Endorsement of several possible behavioral responses by goal
type is also displayed in Table 2. Consistent with the maintenance
of intrinsic motivation, learning goals predicted planning (one
item: “I’d start planning how to do better on the next presenta-
tion”), and negatively predicted withdrawal of time and energy
(one item: “I would devote less time and energy to the class”).
Ability goals, in contrast, positively predicted withdrawal of time
Outcome goals were the only goals that were positively related
to help-seeking (one item: “I would seek help from my professor
or my classmates”). Help-seeking may be perceived as a good way
to obtain the good grades that those who endorse outcome goals
Turning to the psychological processes that accompany goal
pursuit, learning goals (
⫽ .56, p ⬍ .001) were predictive of
effort-based attributions for failure (one item: “I think that if I
work harder, I can do better”), whereas ability goals (
⫽ .22, p ⬍
.05) and outcome goals (
⫽ .36, p ⬍ .01), in contrast, were
predictive of ability-based attributions (one item: “I feel like I’m
just not good at this subject”). Learning goals were negatively
related to making ability attributions for poor performance (
⫺.37, p ⬍ .01).
These results are consistent with prior research, which found
attributions to low ability to be associated with drops in intrinsic
motivation and helplessness, whereas attributions to effort were
associated with intrinsic motivation maintenance and mastery-
oriented coping (e.g., Mueller & Dweck, 1998).
Again, normative goals were not reliable predictors of negative
ability attributions. Taken together with the finding that these goals
do not reliably predict loss of intrinsic motivation, the data begin
to suggest that normative performance goals may be a hardier form
of performance goal (i.e., one that does not tend to lead to “help-
less” forms of coping and behavior). This is again consistent with
Elliot and colleagues’ findings (see Elliot & Church, 1997; Elliot
et al., 1999) that (normative) performance approach goals do not
lead to lower motivation and performance.
Loss of Self-Worth
Loss of self-worth is akin to a negative ability attribution, but it
is more global. It, too, can often accompany helpless motivational
and behavioral responses to a setback (e.g., Covington, 1992;
Crocker & Wolfe, 2001). A composite index of self-worth loss was
created by adding together responses from the following three
items: “I would feel like a loser,”“I would feel like a failure,” and
“I’d think less of myself as a person” (
⫽ .84). Consistent with
results thus far, ability (
⫽ .56, p ⬍ .001) goals were positively
correlated with loss of self-worth.
The tendency to ruminate on one’s setbacks has been associated
with helplessness. A composite index of ruminating and dwelling
on the failure was created by adding together responses from the
following two items: “I would dwell on how poorly I did” and “I
would replay it all over and over again in my mind” (
Rumination was fairly strongly related to ability goals (
p ⬍ .001). Thus, those goals that tend to lead to ability attributions
and negative self-evaluation also predict dwelling on the negative
outcome and its meaning.
The results from the hypothetical failure scenarios revealed a
consistent pattern among the motivational and coping variables.
Learning goals predicted active, engaged responding, whereas
ability goals predicted self-denigration and withdrawal. Outcome
goals were associated with a hybrid response pattern (i.e., low
ability attributions and decreased intrinsic motivation as well as
help-seeking). Finally, normative goals were not reliable predic-
tors of mastery-oriented or helpless responding.
Chronic Coping Style
We now turn to the question of whether different goals predict
different reported histories of coping with setbacks in past achieve-
ment situations. Different styles of chronic coping were measured
by the Ways of Coping Scale (Carver et al., 1989), which asks
participants to indicate the extent to which they have typically
engaged in various coping strategies.
Consistent with the responses to the failure scenarios, learning
goals predicted active coping (
⫽ .38, p ⬍ .01) and planning
⫽ .33, p ⬍ .01). They were also predictive of positive reinter-
pretation of a setback (
⫽ .30, p ⬍ .05) and negatively related to
⫽⫺.36, p ⬍ .01), behavioral disengagement (
p ⬍ .01), and mental disengagement (
⫽⫺.28, p ⬍ .05).
Ability goals negatively predicted positive reinterpretation of a
⫽⫺.30, p ⬍ .05). Of interest, normative goals were
significant predictors of denial after a setback (
⫽ .25, p ⬍ .05)
and behavioral disengagement (
⫽ .28, p ⬍ .01). The finding for
denial perhaps suggests that competitive striving might keep indi-
Goals and Responses to Failure
Loss of intrinsic
Withdrawal of time
and effort Help-seeking Planning
Learning ⫺.39*** ⫺.40*** .17 .57***
Outcome .29** .00 .36** .03
Ability .40*** .32** ⫺.02 ⫺.02
Normative ⫺.11 ⫺.02 ⫺.16 ⫺.16
Note. Values are standardized regression coefficients.
** p ⬍ .01. *** p ⬍ .001.
CLARIFYING ACHIEVEMENT GOALS
viduals from recognizing a poor performance when they produce
one. This may provide some explanation for the consistent finding
that normative goals did not predict negative cognitive, affective,
and behavioral responding to a hypothetical setback (e.g., loss of
intrinsic motivation, low ability attributions, loss of self-worth,
rumination) as strongly or consistently as nonnormative ability
In summary, learning goals were associated with active coping,
and a wide range of positive, mastery-oriented indicators. Learning
goals appear to be a powerful predictor of behaviors that will
preserve intrinsic motivation and performance in the face of dif-
ficulty. In contrast, ability goals were associated with a loss of
motivation and common indices of helplessness. Outcome goals
(which are related to both learning goals and ability goals) also
predicted a loss of motivation and low ability attributions for
failure, but predicted proactive behaviors as well (e.g., help-
seeking). Taken together, these results suggest that valuing doing
well is not in itself a good predictor of responses to failure—rather,
the goals that accompany valuing doing well (learning or validat-
ing ability) seem responsible for much of the “action.” Normative
goals were not among the performance goals that related strongly
or consistently to the variables measured, suggesting that under
some circumstances, competitive performance goal items may not
predict maladaptive cognitions, affect, or coping when other types
of performance goals (i.e., ability goals) do.
Study 5 differed from Study 4 in several ways. First, the goal
items were used to predict motivation and performance in a “real-
world” setting, specifically for freshman and sophomore under-
graduates taking an important and often career-defining course.
Study 5 also differed from many past course-taking studies in the
level of sustained challenge or difficulty encountered by partici-
pants (and, as explained below, in our special attention to students
who encountered successive setbacks over the course of the se-
mester). For this reason, we would expect to see more facilitative
effects of learning goals on motivation and performance, as well as
the debilitating effects of performance goals.
Aside from being a real-world study, Study 5 differed from
Study 4 in another important way. Study 4 presented students with
a fait accompli—a defined failure—and therefore, perhaps did not
allow us to see the potentially beneficial effects of performance
goals for people experiencing challenge but not failure. Study 5
allowed us to monitor students throughout the semester, by looking
in on students as they began this new, important, and challenging
course. Here we might find that for students who are doing well,
ability goals will provide a boost over time, whereas for students
who are encountering difficulty, ability goals will predict further
impairment. In other words, Study 5 allowed us to see goal effects
as they played out over time—both their facilitative effects and
their detrimental effects.
Most potential premed, engineering, and science majors at Co-
lumbia University enroll in General Chemistry in the Fall of their
freshman year. The permission and support of the Columbia Uni-
versity Provost, Deans of the College of Arts and Sciences, and
General Chemistry instructors were granted to conduct an inten-
sive study of these students throughout the semester. Surveys
tracked students’ intrinsic motivation and performance at several
points throughout the semester, and grades were obtained from the
Chemistry Department with permission of the students.
Participants were 85% freshmen, 50% female and 50% male. The
number of participants in each wave of the study varied between 78 and
128, depending on class/recitation attendance. In the largest sample, par-
ticipants were 59% Caucasian, 7% African American, 26% Asian, and 8%
Latino. The average grade on any exam in this course was a C⫹, suggest-
ing that this was a course in which many participants experienced difficulty
or setbacks. For the smaller samples, we tested to ensure that the partici-
pants were entirely representative of the larger sample and that no system-
atic attrition had occurred. Thus, although attendance (and hence partici-
pation in the study) varied over the waves of the study, no significant
differences among samples at the different time points were found in terms
of gender, ethnicity, goal endorsement, or grades.
General Chemistry is a lecture course that is structured around three
midterms and a final exam. Data were collected from participants at four
points during the semester: twice 2–3 weeks before the first midterm, once
immediately after the first midterm, and again 2 weeks before the final
exam. Data were collected in the last 15–20 min of class or recitation. The
measures were presented (along with other measures in a packet of ques-
tionnaires) in the following sequence:
Session 1 (2–3 weeks before first midterm): goal items, demographic
Session 2 (1 week after Session 1): intrinsic motivation, perception of
Session 3 ( after first midterm): general study strategies (from Elliot
et al., 1999)
Session 4 (before final exam): intrinsic motivation
Consistent with the results of Study 4, we predicted that learning goals
would be positively related to intrinsic motivation and grades (despite the
lack of the influence of learning goals on performance found in previous
studies in what may have been less academically strenuous or personally
relevant contexts). We expected ability goals to be associated with lower
performance after multiple setbacks, as suggested by Dweck and Leggett
(1988), but not necessarily with lower performance overall. In fact, we
expected that students who were doing well in the course might experience
a “boost” from holding strong ability goals.
If different types of goals are systematically related to different
levels of perceived ability, then it is possible that the effects of
goals obtained in this study are simply due to this confounding
factor. To rule out this explanation, perceived level of ability in
chemistry was measured at the beginning of the course (one item:
“Compared to other students in this course, please rate your ability
in chemistry” on a 10-point scale ranging from top 10% to lower
10%). Perception of ability in chemistry was related to overall
course grade (r ⫽ .27, p ⬍ .01). It was also related to intrinsic
GRANT AND DWECK
motivation at the beginning (r ⫽ .26, p ⬍ .01) and at the end (r ⫽
.31, p ⬍ .05) of the course.
Correlations between perceived ability in chemistry and goal
type revealed that normative goals were significantly positively
related to perceived ability (r ⫽ .38, p ⬍ .001). In other words,
people with normative goals tended to believe that their ability was
high relative to others. This could help account for the resilience
(or, better put, lack of negative consequences) associated with
normative goals in Study 4. Outcome goals were also positively
related to perceived ability (r ⫽ .21, p ⬍ .05), whereas learning
and ability goals were unrelated to perceptions of chemistry
In a set of linear regressions, we looked at the relationship
between goal type and intrinsic motivation, measured by enjoy-
ment of and interest in the course (two items,
⫽ .88). These data
were collected at the beginning of the course and again before the
final exam. The regressions controlled for perceived ability
(though an essentially identical pattern emerged when perceived
ability was not included in the analysis) and for the effects of each
of the other three goal indices. We also included gender in our
initial analyses, but gender did not predict intrinsic motivation and
was dropped from subsequent analyses predicting intrinsic
In this highly difficult course, learning goals predicted higher
intrinsic motivation at the beginning (
⫽ .23), t(128) ⫽ 2.34, p ⬍
.05, and at the end (
⫽ .22), t(78) ⫽ 2.02, p ⬍ .05, of the course.
This is consistent with the findings of Elliot, Harackiewicz, and
their colleagues (Elliot & Church, 1997; Elliot & McGregor, 1998;
Harackiewicz & Elliot, 1998) that learning goals positively predict
intrinsic motivation. There were no other significant predictors of
We looked at the relationship between each goal type and
students’ total grades, controlling for Scholastic Aptitude Test
(SAT) score, perceived ability in chemistry, number of prior
courses in chemistry, and gender, as well as the effects of other
goal indices. Gender predicted total grade (
⫽⫺.19), t(126) ⫽
⫺3.03, p ⬍ .01, such that men tended to have higher grades than
women. In addition, we looked at the extent to which each goal
type predicted improvement from Exam 1 to the final exam,
controlling for performance on Exam 1 (the interaction of each
goal with performance on Exam 1 was also included as a
Total Course Grade
Consistent with the pattern of effort attribution and mastery-
oriented coping associated with learning goals in Study 4, learning
goals positively predicted course grade (
⫽ .20), t(120) ⫽ 2.42,
p ⬍ .05. No other goals were significant predictors of course
grade. The fact that learning goals emerged as a significant pre-
dictor of performance supplements the findings of Elliot, Harac-
kiewicz, and their colleagues (e.g., Elliot & Church, 1997;
Harackiewicz et al., 1997), who have suggested that performance
goals, and not learning goals, predict course performance. This
result could imply that when a course involves sustained challenge,
learning goals do positively affect course performance.
Improvement in Grade From Exam 1 to Final Exam
Learning goals also significantly predicted grade improvement
⫽ .25), t(122) ⫽ 2.94, p ⬍ .01, and were the only goals to do
Final Exam Grade
Earlier, we had predicted that ability goals would have a nega-
tive effect on performance for those students who had experienced
prolonged setbacks. To address this question, we looked at how
goals predicted performance on the final exam for those students
who had performed poorly throughout the semester. We simulta-
neously regressed each goal type, the average of students’ Exam 1,
2, and 3 grades (our index of past performance), and the interaction
of goal type with average exam grades, onto final exam grades. We
predicted a significant interaction for ability goals, such that stu-
dents who had done poorly throughout the semester (i.e., those
with low average exam grades) would suffer for holding strong
ability goals, whereas those who had done well throughout the
semester might receive a boost on the final.
As predicted, there was a significant interaction between ability
goals and average grade on Exams 1, 2, and 3 (
t(71) ⫽ 2.23, p ⬍ .05. Figure 1 illustrates this effect. We have
plotted data for participants who were either one standard devia-
tion above or below the mean endorsement of ability goals (see
Jaccard, Turrisi, & Wan, 1990). Participants were further separated
into high- and low-course performance groups (based on a median
split of performance on exams prior to final). As shown, partici-
pants with low prefinal grades score lower on the final exam if
they are high rather than low in ability goals. In contrast, partici-
pants with higher prefinal grades earn better scores on the final
Figure 1. Final exam grade predicted by past performance and ability
CLARIFYING ACHIEVEMENT GOALS
exam if they are high rather than low in ability goals. This finding
suggests that when setbacks are repeated, ability goals predict poor
performance, but may indeed provide a boost when an individual
is doing well (see Elliott & Dweck, 1988).
To further understand the differences we found in performance,
we looked at three study strategies (deep processing, surface
processing, and disorganized processing) that were adapted from a
scale used by Elliot et al. (1999), and were assessed immediately
after students took the first exam. The tendency to engage in deep
processing was significantly correlated with grade in the course
(r ⫽ .29, p ⬍ .01). Disorganized processing was negatively related
to course grade (r ⫽⫺.36, p ⬍ .001). Surface processing was
unrelated to course grade (r ⫽ .08, ns).
Outcome goals predicted surface processing of course material
(r ⫽ .29, p ⬍ .01), and learning goals predicted deeper processing
of course material (r ⫽ .31, p ⬍ .01). In contrast, normative goals
were negatively related to deep processing (r ⫽⫺.21, p ⬍ .05),
suggesting that one drawback associated with a competitive goal
might be the absence of deep analysis of issues or material.
Mediational Analyses for Learning Goal Effects on
The significant correlation between learning goals and deep
processing (r ⫽ .31), as well as the correlation between deep
processing and course grade (r ⫽ .29), suggested processing style
as a possible mediator of the effect of learning goals on course
grade. Consistent with this hypothesis, the relationship between
learning goals and course grade (controlling as we had earlier for
SAT score, perceived ability in chemistry, past chemistry course
experience, and gender), when controlling for extent of deep
processing, is not significant (
⫽⫺.06, ns), whereas deep pro-
cessing remains a significant predictor of course grade (
p ⬍ .05; see Table 3 and Figure 2).
Items measuring different types of performance and learning
goals were created and used in five studies to help to shed light on
several important, unresolved issues in current achievement goal
research. Studies 1–3 yielded evidence for four types of goals:
learning goals, outcome goals (wanting to do well), ability-linked
performance goals, and normative performance goals. Individuals’
responses in these three preliminary studies and two more com-
prehensive studies suggested answers to a number of the funda-
mental questions posed in the literature.
First, are there different types of learning goals? What is the
relationship of learning goals to intrinsic motivation and perfor-
mance? We looked at two types of learning goals: striving to learn
and develop versus seeking to master challenges. These two goals
were highly correlated and loaded together in two principal-
components analyses, so the items were combined into a single
learning goal measure. Although we did not find evidence in our
studies for separating these two types of learning goals, they may
still differ importantly from the “task goals” found in past research
that are often operationalized in ways that contain neither striving
to learn nor challenge-seeking.
Studies 4 and 5 provided evidence for the positive effects of
learning goals on both intrinsic motivation and performance, con-
sistent with the early research on achievement goals (see, e.g.,
Ames, 1992; Ames & Archer, 1988; Butler, 1987; Dweck & Leg-
gett, 1988; Elliott & Dweck, 1988; Meece et al., 1988; Nicholls,
1984). Individuals who endorse learning goals should be more
likely to see negative outcomes as information about ways to
improve the learning process, rather than as indicators of stable
low ability. As expected, in response to a major hypothetical
failure (Study 4), learning goals predicted a wide range of positive,
mastery-oriented indicators—including sustained intrinsic motiva-
tion, planning, and persistence. Participants with strong learning
goals also reported a history of having used more mastery-oriented
coping methods (e.g., active coping, planning) in response to past
In Study 5, in an important and difficult college course, learning
goals predicted better processing of course material, higher intrin-
sic motivation, higher grades, and greater improvement over time.
Further analysis suggested that the relationship between learning
goals and course grades was mediated by the tendency to engage
in deeper processing of course material. The impact of learning
goals on performance may be seen chiefly when a high degree of
challenge is present, when a task is personally important, or when
the processing of complex, difficult material is necessary. A po-
tentially important topic for future research is the role that these
factors play in the presence or absence of learning-goal effects on
Turning to other questions posed earlier, is wanting to do well
different from wanting to prove your ability? When might perfor-
Summary of Learning and Ability Goal Effects From Studies 4 and 5
Goal Study 4 Study 5
Learning No decrease in intrinsic motivation
Less time and effort withdrawal
Seeking positive reinterpretation and growth
Higher intrinsic motivation at beginning and
end of course
Greater improvement over time
Ability Lower intrinsic motivation Lower grades after repeated poor performance
Loss of self-worth Higher grades after repeated good performance
Low ability attributions
Time and effort withdrawal
GRANT AND DWECK
mance goals predict vulnerability, and when might they prove
beneficial to intrinsic motivation and/or performance? Individuals
who endorse ability goals (i.e., seek to validate their ability) should
be more likely to see negative outcomes as indicative of a lack of
ability. Consistent with this prediction, ability goals were associ-
ated with common indices of helplessness after a significant hy-
pothetical failure in Study 4. These goals predicted attributions to
low ability, loss of self-worth, rumination about the setback, and
loss of intrinsic motivation. In Study 5, consistent with the results
of Study 4, after multiple setbacks, ability goals predicted lower
grades. Thus, ability goals tend to predict a pattern of negative
affect and cognition, as well as poorer subsequent performance,
after a significant setback or a series of setbacks. These findings
are also consistent with the early work on achievement goals
(Ames & Archer, 1988; Butler, 1993; Elliott & Dweck, 1988;
Jagacinski & Nicholls, 1987; Meece et al., 1988; see also Midgley,
Kaplan, & Middleton, 2001). However, ability goals do not appear
to have negative effects on performance when one is still “in the
running” (i.e., when success is still possible), or when one is doing
well, and may in these cases sometimes even boost performance
because so much is on the line.
Why do the negative effects of ability goals occur? Dweck and
Leggett (1988) suggested several potential cognitive and affective
mechanisms of debilitation for individuals who hold ability goals
in the face of difficulty. These include the loss of belief in the
efficacy of effort (i.e., “My ability is so low, no amount of effort
could help me”), defensive withdrawal of effort (either as a form
of self-handicapping or as a response to the belief that the need to
put in effort confirms that one has low ability), and interference of
negative affect with concentration or test performance. Another
possibility is that students with ability goals may withdraw effort
strategically when they are doing poorly to redirect the resources
to courses where they have a better chance at getting a good grade.
Although these data do not test specifically for this possibility, the
pattern of negative attributions, rumination, and loss of self-worth
associated with ability goals suggest that withdrawal is not a solely
cool-headed strategic process.
Outcome goals had surprisingly few effects. Although corre-
lated with many key outcomes, these effects were almost always
due to the association of outcome goals with either learning goals
(e.g., for active coping and effort attributions) or ability goals (e.g.,
for loss of self worth and rumination). These effects did not
survive simultaneous regression analyses that controlled for the
influence of learning, ability, and normative goals. Taken together,
these results suggest that those researchers interested in studying
the unique effects of performance goals would do better not to
operationalize them this way, as outcome goals (wanting to do
well) can clearly be as much a part of a learning framework as a
performance framework. In fact, doing well can be a means of
assessing the acquisition and mastery of new skills and knowledge
or of demonstrating ability.
Finally, do normative and nonnormative performance goals
produce different effects? Unlike (nonnormative) ability goals,
normative performance goals did not predict any of the affective,
cognitive, or behavioral variables measured in Study 4, with the
exception of the tendency on the COPE scale (Carver et al., 1989)
to report engaging in denial and behavioral disengagement after
experiencing an academic setback. In other words, wanting to
outperform others might lead you to be reluctant to perceive your
performance as a failure. In Study 5, normative goals, unlike
ability goals, did not predict vulnerable performance, and in fact,
were associated with higher levels of perceived ability. As men-
tioned earlier, the absence of a relationship between competitive
goals and helplessness is worth noting, in that those researchers
who have most consistently found that performance goals do not
negatively influence intrinsic motivation and performance have
used a normative definition of performance goal (e.g., Elliot &
Church, 1997). Further research is warranted to explore the roles
that perceived ability and denial may play in this protective func-
tion. Moreover, it is striking that although deep processing medi-
ates the beneficial effects of learning goals on grades, the negative
relationship between normative goals and deep processing did not
seem to predict poorer grades. If the lower level of deep processing
was not a hindrance in this setting, it is very likely that competitive
zeal could have positive effects in the many settings in which deep
processing is not required (Kanfer & Ackerman, 2000).
Because ability performance goals and normative performance
goals appear to behave so differently, it would seem important for
researchers to include both types of performance goals in future
studies. In this way, we could continue to gain knowledge about
when, why, and for whom each has costs and benefits.
It should be noted that there are several differences between
these studies and many past studies of goal effects. First, the
present studies used a measure of general goal orientation (i.e., the
extent to which students typically felt oriented toward particular
goals in their courses), whereas many past studies have used goal
inventories that were specific to the task at hand, or to the course
the student was currently taking. Although it is not certain how this
difference might have affected our results and their interpretation,
the field of achievement motivation might benefit from research
that addressed differences in general versus specific goal measure-
ment. Next, our participants, attending a highly selective univer-
sity, may have differed somewhat from the typical student in ways
that could increase or decrease the impact of particular goals. Also,
somewhat larger sample sizes in some other studies may have
yielded significant effects for certain performance goals that were
not significant in our studies. Finally, as noted above, the course
that our participants were enrolled in (Study 5) appeared to require
deep processing in order to do well, and it is possible that this
factor heightened the impact of learning goals. Nonetheless, our
findings make sense both in view of much previous research and
in view of the different meanings that various goals have for the
Indeed, in this article we have taken great care to consider the
meaning that particular goals may have for the individual and to
Figure 2. Processing style mediates the effects of learning goals on
course grade. Values are standardized regression coefficients.
ized regression coefficient controlling for Scholastic Aptitude Test, per-
ceived chemistry ability, past chemistry course experience, and gender.
*p ⬍ .05. **p ⬍ .01.
CLARIFYING ACHIEVEMENT GOALS
consider goal effects in that light. For example, in addressing the
effects of ability-linked goals on behavior or performance, we
pointed to the inferences that students with ability-linked goals
draw from setbacks. In addressing the effects of outcome goals, we
noted that wanting to do well, far from being a pure performance
goal, may be equally linked to learning and performance goals. In
thinking about learning goals, we stressed the element of active
striving rather than a simple focus on the task or the absence of
performance goal concerns. Thus, for each goal type, we tried to
spell out the impact it might have in the face of achievement
outcomes and why. We hope that our findings have shown the
importance of conceptualizing the psychological processes that
accompany different types of goals, and of matching operational-
izations to these conceptions. When thought of in this way, it
becomes clearer when and why different goals—even ones that
have typically been classified under the same name—will have
In conclusion, we have found evidence to suggest that a careful
examination of different types of performance and learning goals
can indeed begin to clarify current controversies in the field. These
studies have shown that learning goals do exert a positive influ-
ence on both intrinsic motivation and performance when individ-
uals encounter prolonged challenge or setbacks. In addition, al-
though performance goals that are focused on validating ability
can have beneficial effects on performance when individuals are
meeting with success, these same goals can predict impaired
motivation and performance after setbacks.
Ames, C. (1992). Classrooms: Goals, structures, and student motivation.
Journal of Educational Psychology, 84, 261–271.
Ames, C., & Archer, J. (1988). Achievement goals in the classroom:
Students’ learning strategies and motivation processes. Journal of Edu-
cational Psychology, 80, 260–267.
Barron, K., & Harackiewicz, J. (2000). Achievement goals and optimal
motivation: A multiple goals approach. In C. Sansone & J. Harackiewicz
(Eds.), Intrinsic and extrinsic motivation: The search for optimal moti-
vation and performance (pp. 229–254). San Diego, CA: Academic
Barron, K. E., & Harackiewicz, J. M. (2001). Achievement goals and
optimal motivation: Testing multiple goal models. Journal of Personal-
ity and Social Psychology, 80, 706–722.
Bentler, P. M., & Wu, E. J. C. (1995). EQS for Window’s user’s guide.
Encino, CA: Multivariate Software.
Bouffard, T., Boisvert, J., Verzeau, C., & Larouche, C. (1995). The impact
of goal orientation on self-regulation and performance among college
students. British Journal of Educational Psychology, 65, 317–329.
Butler, R. (1987). Task-involving and ego-involving properties of evalua-
tion: Effects of different feedback conditions on motivational percep-
tions, interest and performance. Journal of Educational Psychology, 79,
Butler, R. (1993). Effects of task- and ego-achievement goals on informa-
tion seeking during task engagement. Journal of Personality and Social
Psychology, 65, 18–31.
Button, S. B., Mathieu, J. E., & Zajac, D. M. (1996). Goal orientation in
organizational research: A conceptual and empirical foundation. Orga-
nizational Behavior and Human Decision Processes, 67, 26–48.
Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping
strategies: A theoretically based approach. Journal of Personality and
Social Psychology, 56, 267–283.
Covington, M. V. (1992). Making the grade: A self-worth perspective on
motivation and school reform. New York: Cambridge University Press.
Crocker, J., & Wolfe, C. T. (2001). Contingencies of self-worth. Psycho-
logical Review, 108, 593–623.
Dweck, C. S., & Elliott, E. S. (1983). Achievement motivation. In P.
Mussen & E. M. Hetherington (Eds.), Handbook of child psychology
(pp. 643–691). New York: Wiley.
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to
motivation and personality. Psychological Review, 95, 256–273.
Dweck, C. S., & Sorich, L. (1999). Mastery-oriented thinking. In C. R.
Snyder (Ed.), Coping (pp. 232–251). New York: Oxford University
Elliot, A. J. (1999). Approach and avoidance motivation and achievement
goals. Educational Psychologist, 34, 169–189.
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., & Harackiewicz, J. M. (1996). Approach and avoidance
achievement goals and intrinsic motivation: A mediational analysis.
Journal of Personality and Social Psychology, 70, 461–475.
Elliot, A. J., & McGregor, H. (1998). Test anxiety and the hierarchical
model of approach and avoidance achievement motivation. Journal of
Personality and Social Psychology, 70, 968–980.
Elliot, A. J., & McGregor, H. (2001). A 2 ⫻ 2 achievement goal frame-
work. Journal of Personality and Social Psychology, 80, 501–519.
Elliot, A. J., McGregor, H., & Gable, S. (1999). Achievement goals, study
strategies, and exam performance: A mediational analysis. Journal of
Experimental Social Psychology, 91, 549–563.
Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation
and achievement. Journal of Personality and Social Psychology, 54,
Epstein, J. A., & Harackiewicz, J. M. (1992). Winning is not enough: The
effects of competition and achievement orientation on intrinsic interest.
Personality and Social Psychology Bulletin, 18, 128–138.
Graham, S., & Golen, S. (1991). Motivational influences on cognition:
Task involvement, ego involvement, and depth of information process-
ing. Journal of Educational Psychology, 83, 187–194.
Greene, B. A., & Miller, R. B. (1996). Influences on achievement: Goals,
perceived ability, and cognitive engagement. Contemporary Educational
Psychology, 21, 181–192.
Harackiewicz, J. M., Barron, K. E., Carter, S. M., Lehto, A. T., & Elliot,
A. J. (1997). Predictors and consequences of achievement goals in the
college classroom: Maintaining interest and making the grade. Journal
of Personality and Social Psychology, 73, 1284–1295.
Harackiewicz, J. M., Barron, K. E., Tauer, J. M., Carter, S. M., & Elliot,
A. J. (2000). Short-term and long-term consequences of achievement
goals: Predicting interest and performance over time. Journal of Edu-
cational Psychology, 92, 316–330.
Harackiewicz, J. M., & Elliot, A. J. (1998). The joint effects of target and
purpose goals on intrinsic motivation: A mediational analysis. Person-
ality and Social Psychology Bulletin, 24, 675–689.
Jaccard, J., Turrisi, R., & Wan, C. K. (1990). Interaction effects in multiple
regression. Newbury Park, CA: Sage.
Jagacinski, C. M., & Nicholls, J. G. (1987). Competence and affect in task
involvement and ego involvement: The impact of social comparison
information. Journal of Educational Psychology, 79, 107–114.
Kanfer, R., & Ackerman, P. L. (2000). Individual differences in work
motivation: Further explorations of a trait framework. Applied Psychol-
ogy: An International Review, 49, 470–482.
Kaplan, A., & Maehr, M. L. (1999). Achievement goals and student
well-being. Contemporary Educational Psychology, 24, 330–358.
Kaplan, A., & Midgley, C. (1997). The effect of achievement goals: Does
level of perceived academic competence make a difference? Contempo-
rary Educational Psychology, 22, 415–435.
GRANT AND DWECK
Linnenbrink, E. A., & Pintrich, P. R. (2000). Multiple pathways to learning
and achievement: The role of goal orientation in fostering adaptive
motivation, affect, and cognition. In C. Sansone & J. M. Harackiewicz
(Eds.), Intrinsic and extrinsic motivation: The search for optimal moti-
vation and performance (pp. 195–227). San Diego, CA: Academic
Maehr, M. L., & Midgley, C. (1991). Enhancing student motivation: A
schoolwide approach. Educational Psychologist, 26, 399–427.
Meece, J. L., Blumenfeld, P. C., & Hoyle, R. H. (1988). Students’ goal
orientations and cognitive engagement in classroom activities. Journal
of Educational Psychology, 80, 514–523.
Meece, J. L., & Holt, K. (1993). A pattern analysis of students’ achieve-
ment goals. Journal of Educational Psychology, 85, 582–590.
Middleton, M. J., & Midgely, C. (1997). Avoiding the demonstration of
lack of ability: An underexplored aspect of goal theory. Journal of
Educational Psychology, 89, 710–718.
Midgley, C., Anderman, E., & Hicks, L. (1995). Differences between
elementary and middle school teachers: A goal theory approach. Journal
of Early Adolescence, 15, 90–113.
Midgely, C., Kaplan, A., & Middleton, M. (2001). Performance-approach
goals: Good for what, for whom, under what circumstances, and at what
cost? Journal of Educational Psychology, 93, 77–86.
Midgley, C., & Urdan, T. (1995). Predictors of middle school students’ use
of self-handicapping strategies. Journal of Early Adolescence, 15, 389–
Miller, R. B., Behrens, J. T., Greene, B. A., & Newman, D. (1993). Goals
and perceived ability: Impact on student valuing, self-regulation, and
persistence. Contemporary Educational Psychology, 18, 2–14.
Molden, D. C., & Dweck, C. S. (2000). Meaning and motivation. In C.
Sansone & J. M. Harackiewicz (Eds.), Intrinsic and extrinsic motivation:
The search for optimal motivation and performance (pp. 131–159). San
Diego, CA: Academic Press.
Mueller, C. M., & Dweck, C. S. (1998). Praise for intelligence can
undermine children’s motivation and performance. Journal of Person-
ality and Social Psychology, 75, 33–52.
Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability,
subjective experience, task choice, and performance. Psychological Re-
view, 91, 328–346.
Pintrich, P. R. (2000a). An achievement goal theory perspective on issues
in motivation terminology, theory, and research. Contemporary Educa-
tional Psychology, 25, 92–104.
Pintrich, P. R. (2000b). The role of goal orientation in self-regulated
learning. In M. Boekaerts, P. Pintrich, & M. Zeidner (Eds.), Handbook
of self-regulation (pp. 451–502). San Diego, CA: Academic Press.
Pintrich, P. R., & DeGroot, E. V. (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. L. Maehr & P. R. Pintrich
(Eds.), Advances in motivation and achievement (Vol. 7, pp. 371–402).
Greenwich, CT: JAI Press.
Rawsthorne, L. J., & Elliot, A, J. (1999). Achievement goals and intrinsic
motivation: A meta-analytic review. Personality and Social Psychology
Review, 3, 326–344.
Roeser, R. W., Midgely, C., & Urdan, T. C. (1996). Perceptions of the
school psychological environment and early adolescents’ psychological
and behavioral functioning in school: The mediating role of goals and
belonging. Journal of Educational Psychology, 88, 408–422.
Utman, C. H. (1997). Performance effects of motivational state: A meta-
analysis. Personality and Social Psychology Review, 1, 170–182.
Zhao, W., & Dweck, C. S. (1997). Implicit theories and vulnerability
to depression-like responses. Unpublished manuscript, Columbia
Achievement Goal Inventory Items
Outcome Goal Items (
It is very important to me to do well in my courses.
I really want to get good grades in my classes.
A major goal I have in my courses is to perform really well.
Ability Goal Items (
It is important to me to confirm my intelligence through my schoolwork.
In school I am focused on demonstrating my intellectual ability.
One of my important goals is to validate my intelligence through my
Normative Goal Items (
It is very important to me to do well in my courses compared to others.
I try to do better in my classes than other students.
A major goal I have in my courses is to get higher grades than the other
It is very important to me to confirm that I am more intelligent than other
When I take a course in school, it is very important for me to validate
that I am smarter than other students.
In school I am focused on demonstrating that I am smarter than other
Learning Goal Items (
I strive to constantly learn and improve in my courses.
In school I am always seeking opportunities to develop new skills and
acquire new knowledge.
In my classes I focus on developing my abilities and acquiring new ones.
I seek out courses that I will find challenging.
I really enjoy facing challenges, and I seek out opportunities to do so in
It is very important to me to feel that my coursework offers me real
Received February 6, 2002
Revision received January 16, 2003
Accepted January 21, 2003 䡲
CLARIFYING ACHIEVEMENT GOALS