Vous êtes sur la page 1sur 9

Journal of Counseling Psychology

1995, Vol. 42, No. 2, 190-198

Copyright 1995 by the American Psychological Association, Inc.


0022-0167/95/S3.00

Mathematics Self-Efficacy and Mathematics Performances: The Need for


Specificity of Assessment

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Frank Pajares
Emory University

M. David Miller
University of Florida

In this study, 391 students were asked to provide 3 types of mathematics self-efficacy
judgments: confidence to solve mathematics problems, confidence to succeed in math-related
courses, and confidence to perform math-related tasks. Criterial tasks were solution of math
problems and choice of math-related majors. As hypothesized, students' reported confidence
to solve the problems they were later asked to solve was a more powerful predictor of that
performance than was either their confidence to perform math-related tasks or to succeed in
math-related courses. Similarly, confidence to succeed in math-related courses was a stronger
predictor of choice of math-related majors than was either confidence to solve problems or
to perform math-related tasks. Results support A. Bandura's (1986) contention that, because
judgments of self-efficacy are task specific, measures of self-efficacy should be tailored
to the criterial task being assessed and the domain of functioning being analyzed to
increase prediction.

Social cognitive theorists contend that self-efficacy beliefs, or "people's judgments of their capabilities to organize
and execute courses of action required to attain designated
types of performances" (Bandura, 1986, p. 391), strongly
influence the choices people make, the effort they expend,
the strength of their perseverance in the face of adversity,
and the degree of anxiety they experience. In part, these
self-perceptions can be better predictors of behavior than
actual capability because such self-beliefs are instrumental
in determining what individuals do with the knowledge and
skills they have. The mediational role these beliefs play also
helps explain why people's performance attainments may
differ even when they have similar knowledge and skills.
Bandura (1986) cautioned that because judgments of selfefficacy are task and domain specific, "ill-defined global
measures of perceived self-efficacy or defective assessments of performance will yield discordances" (p. 397). For
this reason, measures of self-efficacy should be specifically
tailored to the criterial task being assessed and the domain
of functioning being analyzed. This caution has also been
voiced by other social cognitive researchers (e.g., Lent &
Hackett, 1987; Multon, Brown, & Lent, 1991) but has often
gone unheeded in educational research, where assessments
of self-efficacy frequently bear little resemblance to the
criterial task with which they are compared but, instead,
reflect generalized, or sometimes even unrelated, attitudes
about capabilities. Although findings have generally sup-

ported the predictive role of self-efficacy (Maddux, Norton,


& Stoltenberg, 1986), many studies are plagued by assessments of self-efficacy that do not conform to Bandura's
guidelines regarding specificity of measurement (Multon et
al., 1991). This mismatch between self-efficacy and criterial
task assessment is a recurring theme in educational research,
often producing confounded relationships and ambiguous
findings.
When the efficacy beliefs assessed do not reflect with any
sense of specificity the criterial task with which they are
compared, their predictive value is diminished. Multon et al.
(1991) found only 68 published and unpublished papers
written between 1977 and 1988 on the relationship between
self-efficacy and academic performance or persistence. Of
these, only 36 studies with 38 samples, encompassing 4,998
participants, assessed academic performance and met their
criteria for inclusion in a meta-analysis, which were that the
study contain a measure of self-efficacy and academic performance and provide sufficient information to calculate
effect size estimates. They computed that efficacy beliefs
were related to performance (r u = .38) and accounted for
approximately 14% of the variance in students' academic
performance. However, effect sizes depended on specific
characteristics of the studies, notably on the types of efficacy and performance measures used.
Multon et al. (1991) discovered that the strongest effects
were obtained by researchers who compared specific efficacy judgments with basic skills measures of performances,
who developed highly concordant self-efficacy and performance indices, and who administered them at the same
time. Recall that Bandura (1986) cautioned that self-efficacy judgments be specifically rather than globally assessed, correspond directly to the criterial task, and be
measured as closely as possible in time to that task. Significant relationships are obtained even with generalized self-

Frank Pajares, Department of Educational Studies, Emory University; M. David Miller, Department of Foundations, University
of Florida.
Correspondence concerning this article should be addressed to
Frank Pajares, Department of Educational Studies, Emory University, Atlanta, Georgia 30322. Electronic mail may be sent via
Internet to mpajare@unix.cc.emory.edu.
190

MATHEMATICS SELF-EFFICACY
efficacy indices, a phenomenon that Multon et al. described
as reinforcing "the theoretical and practical value of attending to students' self-efficacy beliefs" (p. 35) but that also
may produce confounded and misleading results. In fact, if
global and generalized self-efficacy assessments can predict
performances that are not specifically related, the relationship between properly assessed self-efficacy and performance should certainly increase.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Overview of Mathematics Self-Efficacy Findings


Self-efficacy research in academic settings has focused
primarily on two major areas. The first has investigated the
relationships among efficacy beliefs, related psychological
constructs, and academic motivation and achievement (see
Schunk, 1989, 1991). The second has attempted to explore
the link between efficacy beliefs and college major and
career choices (see Lent & Hackett, 1987). Studies in this
second area have important implications for counseling and
vocational psychology theory and practice, given that findings have provided insights into the career development of
young men and women and can be used to develop appropriate career intervention strategies (Lent & Hackett, 1987;
Multon et al., 1991).
Early studies investigated confidence in learning mathematics, a conceptual forerunner to math self-efficacy that
has consistently been found to predict math-related behavior
and performance (Reyes, 1984). This confidence was globally assessed by asking students general questions about
their perceived mathematics capabilities. Mathematics selfefficacy has more recently been assessed in terms of individuals' judgments of their capabilities to solve specific
math problems, to perform math-related tasks, and to succeed in math-related courses (Betz & Hackett, 1983).
Dowling (1978) was the first researcher to create a confidence measure to specifically correspond with a performance assessment in which students were asked to solve the
same or similar math problems on which their confidence
was based. Drawing from math problems created for the
National Longitudinal Study of Mathematical Abilities
(NLSMA), she developed the Mathematics Confidence
Scale (MCS). Students were first asked to provide judgments of confidence to solve these math problems and later
were asked to solve an alternate-forms test of the problems
on which their confidence was assessed. Dowling obtained
a correlation of .54 between self-efficacy and performance,
a significantly higher figure than that reported by most
previous researchers assessing math confidence globally
and that is consistent with the results of Multon et al.
(1991).
Subsequently, Betz and Hackett (1983) created the Mathematics Self-Efficacy Scale (MSES). They incorporated as
one subscale a measure similar to Dowling's (1978) MCS
and added two additional subscales, one to assess students'
confidence to perform certain math-related tasks and another to assess their confidence to earn an A or B in certain
math-related courses. Betz and Hackett created these subscales in the belief that math self-efficacy required "detailed

191

specification of the domain of math-related behavior" (p.


331) and that these domains included solution of math
problems, math behaviors used in daily life, and, of particular relevance to college students, satisfactory performance
in a college course. They operationalized mathematics selfefficacy as the composite score of the three subscales.
Since its creation, the MSES has been used in studies
focusing both on academic performance and on major and
career choice. For example, Hackett (1985) used the MSES
in a path analysis to discover the mediational role of selfefficacy in predicting undergraduate students' choice of
math-related majors. With the composite score of the three
MSES subscales as the self-efficacy assessment, Hackett
reported a correlation of .50 between perceived efficacy and
choice of major, as well as a moderate direct effect of
self-efficacy on choice (.24). Hackett and Betz (1989) later
administered the MSES to undergraduates and reported that
a hierarchical regression analysis demonstrated that the fullscale MSES score was strongly predictive of a choice of
college major between science and nonscience fields (jB =
.45).
Hackett and Betz (1989) also investigated the relationship
between math self-efficacy and mathematical problem solving. They obtained a correlation of .44 between undergraduates' full-scale MSES scores and performance scores on an
alternate-forms test of the MSES Problems scalea correlation the researchers noted was not as large as they had
expected. In contrast, Pajares and Miller (1994) obtained a
correlation of .70 when they compared scores on one of the
forms of Dowling's (1978) MCS with performance scores
on the same test on which self-efficacy was assessed. Pajares and Miller also found thatwith math self-concept,
perceived usefulness of mathematics, math background, and
gender as part of a path analysis modelself-efficacy had
the strongest direct effect on problem-solving performance
0 = .545).
Other researchers have used full-scale MSES scores as a
generalized math self-efficacy assessment. Randhawa,
Beamer, and Lundberg (1993) adapted the MSES for use
with high school students and used LISREL procedures to
find that the full-scale, generalized self-efficacy score mediated the effect of various math attitudes on math problem
solving. The criterial task, the solving of math problems,
was conceptually related only to the Problems subscale.
Moreover, the problems on the performance measure differed from those presented on the self-efficacy assessment.
Consequently, although the math attitude measures had a
strong direct effect on self-efficacy (.64), they also had a
stronger direct effect on performance (.44) than did generalized self-efficacy (.32).
Studies that report a lack of relationship between math
self-efficacy and performance often suffer from similar conceptual or measurement flaws. Benson (1989) found that the
path from what she called math self-efficacy to performance
was not significant, whereas that between math self-concept
and performance was. However, self-efficacy was assessed
with three global items dealing with expected success in a
statistics class (e.g., "No matter how hard I study, I will not
do well in this class"). Self-concept was assessed with seven

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

192

FRANK PAJARES AND M. DAVID MILLER

items specific to feelings of math self-worth (e.g., "I feel


insecure in a math class"). Performance was the midterm
exam grade in a statistics course. Benson concluded that
"additional studies need to be conducted to verify the relationship between math self-concept and self-efficacy and to
explore why self-efficacy did not influence [performance]
nor statistical test anxiety" (p. 259). Perhaps the answer is
that a comparison of general confidence to succeed in a
class with a statistics midterm grade is not likely to produce
the sort of correspondence that social cognitive theory
hypothesizes.
Other studies are more problematic. Smith, Arnkoff, and
Wright (1990) used hierarchical multiple regression to test
the predictive power of three theoretical modelscognitive-attentional, cognitive-skills, and social learningon
academic performance. The researchers concluded that variables within each model predicted performance to some
degree but that self-efficacy was a weak predictor. Selfefficacy, however, was operationalized as study skills or
test-taking self-perceptions and measured with items such
as "Rate how certain you are that you can study at a time
and place where you won't get distracted." These selfperceptions were then compared with academic outcomes
such as course and exam grades and grade point average.
Cooper and Robinson (1991) found a low but significant
correlation between math self-efficacy and performance,
but a regression model with math anxiety, the quantitative
score on the American College Test, and prior math experience revealed that self-efficacy did not account for a
significant portion of the variance in math performance.
Here again, self-efficacy and performance assessments were
mismatchedscores from the Courses subscale of the
MSES were compared with scores on a performance measure that consisted of solving problems from the Missouri
Mathematics Placement Test.

Purpose of the Study


Hackett and Betz (1989) defined mathematics self-efficacy as "a situational or problem-specific assessment of an
individual's confidence in her or his ability to successfully
perform or accomplish a particular [mathematics] task or
problem" (p. 262). In line with Bandura's (1986) guidelines
regarding consistency of self-efficacy and performance assessment, we hypothesized that individuals' judgments to
solve math problems, to perform math-related tasks, or to
succeed in math-related courses entail substantively different judgments of mathematics capability. Although they all
are judgments about certain mathematical abilities, their
predictive value regarding these abilities should largely
depend on the nature of the specific capabilities with which
they are compared. Consequently, students' judgments to
solve math problems should be more strongly predictive of
their ability to solve those problems than should their confidence to perform other math-related tasks or to succeed in
math-related courses. Similarly, their judgments to succeed
in math-related courses should be more strongly predictive
of their choice to enroll in such courses than should their
confidence to solve specific problems or perform mathematics tasks.
In essence, our argument is that, although researchers
have tended to use generalized measures or multiple-scale
math self-efficacy scores to predict math-related outcomes,
the appropriateness of the match between self-efficacy assessment and the outcome is crucial to optimal prediction of
that outcome. Figure 1 illustrates the traditional method of
using multiple-scale math self-efficacy assessments with
dependent measures such as solving math problems or
choosing math-related majors. However, the more parsimonious model in Figure 2 more accurately reflects the role of
the same self-efficacy assessments in predicting those
outcomes.

Mathematics Problems
Self-Efficacy
Mathematics
Problem-solving
Performance
Math-Related Tasks
Self-Efficacy
Selection of
Math-Related
Major
Math-Related Courses
Self-Efficacy
Figure 1. Multiple mathematics self-efficacy assessments to predict mathematics problems and
selecting math-related majors.

193

MATHEMATICS SELF-EFFICACY

Mathematics Problems
Self-Efficacy
Mathematics
Problem-solving
Performance
Math-Related Tasks ^ \
Self-Efficacy
)

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Selection of
Math-Related
Major
Math-Related Courses
Self-Efficacy

Figure 2. Mathematics self-efficacy assessment matched to related outcomes.

In this study, we asked students to provide judgments of


their confidence to solve specific math problems, to perform
math-related tasks, and to succeed in math-related courses.
We compared these judgments of confidence with two
outcome measures: their ability to solve the problems on
which their confidence was assessed and the math-relatedness of their academic majors. Social cognitive theory holds
that judgments of capability to solve specific problems
should be more predictive of capability to solve those problems. Similarly, judgments of confidence to succeed in
math-related courses should be more predictive of choice of
math-related majors that involve enrolling in such courses.
If this is the case, findings have important implications for
future assessment of self-efficacy both in the area of mathematics and other academic areas.
We acknowledge from the outset that confidence to succeed in mathematics courses does not relate as directly to
choice of major as confidence to solve math problems does
to actual problem solving, and we expected the latter to
provide stronger results. Clearly, however, the choice of a
math-related major is predicated on the need to succeed in
certain mathematics courses. Consequently, there is a strong
logical correspondence between judgments of confidence to
succeed in math courses and choice to pursue a major that
requires enrolling in such courses. This correspondence is
stronger than that between choice of major and either tasks
or problems self-efficacy. One might argue that a better
criterial task could have been selected to compare with
judgments to succeed in math-related courses. The logical
task would be a measure of success in such courses. However, most undergraduates do not take most of the courses
on the efficacy instrumentlow-efficacy students, for example, avoid such courses. Lent, Lopez, and Bieschke
(1993) resolved this issue by comparing courses self-efficacy scores with three outcomes: intention to take the
courses listed on the instrument, grades obtained in mathrelated courses that students took during the subsequent
term, and interest in the math courses listed on the instrument. All three seemed to us reasonable choices. Choice of

math-related major is equally as reasonable, as this choice


provides a logical reflection of both interest and intention.
In some ways, choice of major has greater practical utility to
the counseling and career literature (hence, its use by Betz,
Hackett, and their colleagues). It is interesting that our
results in this study are consistent with those of Lent at al.
(1993).
Method

Participants and Procedure


Participants for the study consisted of 391 undergraduates (247
women and 144 men; 26 freshmen, 74 sophomores, 172 juniors,
and 119 seniors) enrolled in various upper division courses at three
large public universities, two in the South (ns = 212 and 92) and
one in the Southwest (n = 87). The students represented over 30
majors throughout their respective universities. Testing procedures
were carried out in individual, upper division classes in colleges of
education and of arts and sciences at the three universities. In all
cases, colleagues volunteered the use of their classes at our request. Participation by the students was voluntary and no remuneration was provided, but most instructors provided course credit.
Appropriate informed consent was obtained, and students were
assured anonymity. Students were first explained the procedures.
Subsequently, they were asked to complete a descriptive questionnaire and the MSESRevised (MSES-R). Students were informed that they would be asked to solve the problems on which
their problems self-efficacy was assessed. However, they were
asked not to attempt to solve the problems as they provided
confidence judgments, and these directions were also clearly
printed on the efficacy instrument. The test administrator monitored the session to ensure that students followed directions
carefully.
After these instruments were collected, students were administered a test that asked them to solve the problems on which their
self-efficacy had been assessed. Bandura (1986) suggested that
efficacy and performance be assessed within as close a time period
as possible and that efficacy assessment precede performance
assessment. Therefore, all instruments were administered during
one class period. This procedure also ensured that student absences

194

FRANK PAJARES AND M. DAVID MILLER

would not create a situation whereby some students completed one


measure but not another. Pilot testing had demonstrated that the
50-min periods provided ample time in which to complete all
measures.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Instrumentation
Mathematics Self-Efficacy ScaleRevised. Bandura (1977)
wrote that, because the definition of self-efficacy is straightforward, efficacy questionnaires in academic areas are adequate operational measures of "the conviction that one can successfully
execute the behavior required to produce the outcomes" (p. 93).
Norwich (1986) also found that traditional self-report assessments
of self-efficacy had enough reliability and validity to warrant their
use. As we noted earlier, the MSES was developed by Betz and
Hackett (1983) to assess the math self-efficacy of college students.
The instrument has 52 items and three subscales representing three
domains of math-related behavior: solution of math problems,
completion of math tasks used in everyday life (e.g., balancing a
checkbook, computing income taxes), and satisfactory performance in college courses that require knowledge and mastery of
mathematics (e.g., calculus, statistics, and biochemistry).
The Problems subscale of the MSES was adopted from the
mathematics confidence scales created by Dowling (1978), who
selected midrange difficulty items from the NLSMA and developed an instrument to specifically assess the math confidence of
college students. The problems she selected represent three components of mathematics (arithmetic, algebra, and geometry), three
levels of cognitive demand (computation, comprehension, and
application), and two problem contexts (real and abstract). Betz
and Hackett (1983) reported item-total score correlations ranging
from .29 to .63 for tasks, .33 to .73 for courses, and .24 to .66 for
problems. Coefficient alphas for the three subscales were .90 for
Tasks, .93 for Courses, and .92 for Problems. Betz and Hackett
also reported a correlation of .66 between the MSES and the Math
Confidence subscale of the Fennema-Sherman (1976) scales.
Hackett and Betz (1989) reported on unpublished raw data by
Hackett and O'Halloran (1985) that demonstrated moderate
2-week test-retest reliabilities for the total scale (.88) and for each
of the three subscales (.79 for Tasks, .91 for Courses, and .82 for
Problems). Since its creation, the MSES or one of its scales has
been used in a number of studies (e.g., Betz & Hackett, 1983;
Hackett, 1985; Hackett & Betz, 1989; Langenfeld & Pajares, 1993;
Lapan, Boggs, & Morrill, 1989; Lent, Lopez, & Bieschke, 1991,
1993; Pajares & Miller, 1994; Randhawa et al., 1993).
We made two alterations to the MSES for use in this study.
When Dowling (1978) constructed two final forms of math confidence and performance, she constructed and tested preliminary
scales, selecting items according to predetermined criteria. Her
final forms reflected the results of this process. Betz and Hackett
(1983) selected a preliminary scale as the Problems subscale of the
MSES. We selected one of Dowling's final forms as the Problems
subscale of the MSESR. Dowling reported a correlation of .57
between the MCS and the Confidence subscale of Fennema and
Sherman's (1976) Mathematics Attitude Scales. Langenfeld and
Pajares (1993) conducted a factor-analytic study with 520 undergraduates using the MSES-R with a 5-point Likert scale and
reported that the instrument suffered no loss of internal consistency (alpha coefficients = .94 for Tasks, .91 for Courses, and .91
for Problems). Pajares and Kranzler (1994) reported Cronbach's
alpha coefficient of .92 using a version of the Problems subscale
with 329 high school students. We also used a 5-point scale and
obtained alpha coefficients of .90 for Problems, .92 for Courses,
and .91 for Tasks in the present study.

Mathematics outcome measures. Self-efficacy is assessed by


asking individuals to provide a judgment of whether they can
accomplish a specific task and subsequently asking them to accomplish the same, or a similar, task (see Bandura & Schunk,
1981; Schunk, 1981, 1982a, 1982b). In this study, the tasks on
which confidence was assessed were consistent with those on
which our two performance outcomes were measured.
The first outcome measure asked students to solve the same
problems on which their math self-efficacy was assessed. The
problems performance measure used in this study, like the math
problems efficacy scale, was one of the two final forms developed
by Dowling (1978). This was an 18-item, multiple-choice instrument constructed with midrange difficulty items from the
NLSMA, developed specifically for use with college undergraduates, and composed of three nonorthogonal subscales consisting of
mathematics components, cognitive demand, and problem context.
Dowling reported K-R 20 coefficient of reliability of .79 and mean
item difficulty of .29. Dowling constructed the final efficacy and
performance measures from four preliminary forms. For an item to
be included in the final instrument, several criteria had to be met:
(a) percentage correct between .30 and .70, (b) point-biserial
correlation coefficient greater than .50, (c) discrimination index
greater than .40, and (d) significant corrected phi coefficient (p <
.01). Pajares and Kranzler (1994) reported a K-R 20 of .83 on a
similar instrument with a high school sample. We obtained a K-R
20 of .86 in the present study.
The second performance measure consisted of the students'
declared majors, which were rated along Goldman and Hewitt's
(1976) science-nonscience continuum adapted for mathematics.
The rating on this continuum provides a 5-point continuous scale
that includes majors in the fine arts (1), humanities (2), social
sciences (3), natural-biological sciences and lower order mathematics (4), and physical sciences and higher order mathematics
(5). This scale is the same used by Hackett (1985), Betz and
Hackett (1981, 1983), and Lent et al. (1991, 1993) in their studies
of self-efficacy and choice of major. To test the underlying continuum of this variable, we subjected the scale to trend analysis,
full model F(l, 389) = 41.95, p < .001. Results showed a
significant linear trend, F(2, 388) = 8.19, p < .0003.

Data Analysis
The hypotheses of this study were that there would be a stronger
relationship between scores on the math problems self-efficacy
scale and those on the problem-solving performance measure than
between scores on the outcome measure and those of either the
math tasks or courses self-efficacy scales, and that there would be
a stronger relationship between scores on the courses self-efficacy
scale and choice of majors than between that outcome measure and
those of either the math problems or math tasks scales. Four
analyses were performed for each hypothesis: Zero-order correlations were obtained to discover the strength of relationships between variables, the Williams T 2 statistic was used to determine
whether the correlations were significantly different, and a multiple regression analysis was conducted to further inform the relationships. Finally, the difference between the models in Figure 1
and Figure 2 were tested using LISREL 7 (Joreskog & Sorbom,
1989). This procedure provides a joint test for the significance of
the four paths missing in Figure 2.

Results
Table 1 contains the means, standard deviations, and
Pearson product-moment correlations for the three

195

MATHEMATICS SELF-EFFICACY

Table 1
Correlations Between Math Self-Efficacy Scales and Math Performances
Variable
Problems self-efficacy
Tasks self-efficacy
Courses self-efficacy
Full-scale math selfefficacy
5. Problem-solving
performance
6. Choice of math-related
major
' p < .0001.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

1.
2.
3.
4.

SD

73.8
72.1
54.0

10.3
11.4
11.8

.66*
.67*

.66*

199.9

29.6

.87*

.88*

.89*

14.0

2.8

.69*

.45*

.53*

.63*

3.0

0.9

.39*

.34*

.52*

.48*

.28*

MSES-R scales, the total scale, and the two outcome measures (problem solving and choice of majors). Problems
self-efficacy scores ranged from 36 to 90. Possible scores on
this 18-item subscale ranged from 18 to 90; hence, although
some students expressed maximum confidence in their
problem-solving ability, even the least confident student
averaged 2 on the 5-point Likert scale. Also, the mean score
of 4.1 (much confidence) per item suggested that students
were largely confident about their ability to solve the problems. The same was true for the 18-item Tasks self-efficacy
subscalescores ranged from 30 to 90. When asked
whether they could complete a math-related course with a
final grade of B or A, however, students showed less confidence. Scores on this 16-item scale ranged from 22 to 80,
with a mean score of 3.4 per item. Total math self-efficacy
scores ranged from 108 (minimum possible = 52) to 260
(maximum possible), with a mean score of 3.8 per item. On
the whole, students were confident of their math self-efficacy. Scores on the 18-item math performance outcome
measure ranged from 4 to 18, with a mean of 14.0. These
results are consistent with those of Hackett and Betz (1989)
and Pajares and Miller (1994).
As expected, the efficacy subscales were strongly related
with each other. Also, as hypothesized, problems self-efficacy had a stronger relationship with problem-solving performance than did either tasks or courses self-efficacy,
although all were significant. We used Williams's (1959) T 2
modification of Hotelling's T1 to discover whether the differences between these correlations were significant. Critical value for the Williams T 2 was f(388, a < .05) = 1.96.
Results showed that the correlation between problem-solving performance and problems self-efficacy was significantly stronger than that between either performance and
courses self-efficacy, T 2 = 5.00, or tasks self-efficacy, T 2 =
7.76.
A similar phenomenon occurred between the three scales
and choice of majors, whose strongest correlation was with
courses self-efficacy. Again, results of Williams T 2 showed
that the correlation between courses self-efficacy and choice
of majors was significantly stronger than that between either
choice of majors and problems self-efficacy, T 2 = 3.51, or
tasks self-efficacy, T 2 = 4.80.
Results of two multiple regression analyses with the three
subscales as independent variables and math problem-solv-

ing performance as the dependent variable in one analysis


and choice of majors in another, as shown in Tables 2 and
3, help clarify the relationships. Both models were significant. The model with problem-solving performance accounted for 48% of the variance, F(3, 387) = 120.89, p <
.0001, adjusted R2 = .48; the model with choice of majors
accounted for 27%, F(3, 387) = 47.66, p < .001, adjusted
R2 = .26. Although problems self-efficacy and courses
self-efficacy both achieved significance in the problemsolving performance model, the difference between them
was pronounced. Students' confidence in their capability to
solve the problems had a much stronger influence on their
problem-solving performance than did their confidence to
succeed in math-related courses. Confidence to perform
math-related tasks was nonsignificant in both models.
Finally, a test for the differences between the models
presented in Figures 1 and 2 yielded a nonsignificant difference, ^ ( 4 , N = 391) = 3.58, suggesting that the four
additional paths in Figure 1 do not help in the explanation of
the data. Because of these results and those of the Williams
T 2 analysis, we concluded that it is not only theoretically
sound but empirically warranted to use math problems
self-efficacy as the appropriate self-efficacy assessment
when problem-solving performance is the criterial task and,
similarly, to use courses self-efficacy when a related outcome variable such as choice of math-related majors is the
criterial task.
Discussion
Results from this study confirm that Bandura's (1986)
warnings regarding self-efficacy and performance assess-

Table 2
Multiple Regression Analysis of the Effect
of the Math Self-Efficacy Subscales on Math
Problem-Solving Performance
Variable
Problems self-efficacy
Tasks self-efficacy
Courses self-efficacy

Parameter Standard
Probability
estimate
error
t
> \t\
0.170
0.015
11.70
.0001
-0.019
0.013 -1.46
.1460
0.041
0.013
3.19
.0015

196

FRANK PAJARES AND M. DAVID MILLER

Table 3

Multiple Regression Analysis of the Effect of


the Math Self-Efficacy Subscales on Choice of
Undergraduate Major
Variable

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Problems self-efficacy
Tasks self-efficacy
Courses self-efficacy

Parameter Standard
estimate
error

0.009
-0.003
0.036

0.006
0.005
0.005

Probability
> \t\

1.56
-0.57
7.45

.1205
.5698
.0001

merit are well founded. There are different ways of assessing self-efficacy, but the most theoretically appropriate and
empirically warranted is one in which the self-efficacy
measure assesses the same or similar skills required for the
performance task. Students' confidence to solve mathematics problems was a more powerful predictor of their ability
to solve those problems than was their confidence to perform math-related tasks or their confidence to earn As or Bs
in math-related courses. Similarly, their confidence to succeed in such courses was more predictive of their choice of
majors that required them to take many of the math-related
courses on which they expressed that confidence. As we
outlined in the introduction, the self-efficacy-performance
correspondence between confidence to succeed in mathrelated courses and choice of major is not as specific as that
between confidence to solve problems and actual capability
to solve those problems. Consequently, we expected the
latter relationship to be stronger, and indeed it was. This
finding also supports Bandura's (1986) theoretical argument
and our empirical hypothesis.
Researchers have observed that even generalized or less
closely related indexes will correlate significantly with specific outcomes, and we found this phenomenon as well (see
Multon et al., 1991). That is, each subscale, as well as the
full scale, correlated significantly with each outcome. As we
found, however, prediction is enhanced as self-efficacy and
performance correspondence more closely matches. The
clear implication of these results is that researchers are well
advised to consider the nature of the performance task when
choosing or developing an appropriate self-efficacy measure. As findings showed, this advice is both theoretically
and empirically sound.
It is not altogether easy to see what value full-scale scores
provided by multiple-scale instruments such as the MSES or
the MSES-R may have if one wishes to predict specific
mathematics outcomes. Such global scores decontextualize
efficacy beliefs and transform the construct of self-efficacy
into a generalized personality trait rather than the contextspecific judgment Bandura (1986) suggested it is (pp.
5-12). If, as Bandura argued, self-efficacy assessment must
conform to the criterial task to be useful and predictive,
what criterial task can be compared with a global score that
comprises judgments of confidence to succeed in math
courses, to complete math-related tasks, and to solve specific math problems? Moreover, one might question the
practical utility of administering a 52-item instrument when
greater prediction may be had from a shorter instrument

more closely matching the performance task. Hackett and


Betz (1989) suggested that teachers and counselors would
be well advised to pay as much attention to students' selfefficacy beliefs as they presently pay to actual test scores,
for it is these beliefs that may better reveal the choices and
decisions students will make about matters such as choice of
college courses or majors and directions for career development. This is sound advice, if it is tempered by the
warning that professionals would also be well served by
focusing on the self-efficacy beliefs most predictive of the
outcomes they wish to predict.
Some researchers have noted the need to explore the
generality of self-efficacy beliefsthat is, the extent to
which they relate to, or transfer across, different performance tasks or domains (Lent & Hackett, 1987; Multon et
al., 1991). Lent and Hackett (1987) wrestled with this issue
when they argued that the level of specificity of an efficacy
assessment should ultimately depend on the complexity of
the performance criteria with which it is compared (pp.
365-366). Math-related outcomes, particularly in investigations of career choices and decisions, are seldom as precise
and specific as one's capability to solve mathematics problems. Lent and Hackett rightly observed that specificity and
precision are often purchased at the expense of external
validity and practical relevance.
Our findings suggest that research questions should be
formulated with an eye to enhancing the correspondence
between self-efficacy and performance assessments, especially if the aim is to help predict or explain academic
outcomes. Bandura (1986) cautioned against the use of
omnibus tests to assess "global dispositions" (p. 396). It
may be argued that the use of the composite score of a
multiscale test that provides a general judgment of capability for an entire academic area is, in essence, assaying a
global disposition (in this case, one's generalized mathematics confidence) with an omnibus test. Nonetheless, we
agree that full-scale scores from reliable, multiscale academic self-efficacy instruments tap into a general factor of
academic or subject-specific capability. This general selfperception provides teachers and counselors with information regarding students' general mathematics confidence
(see Kranzler & Pajares, 1994), and results may be useful in
predicting complex math-related outcomes that do not easily lend themselves to microanalytic analysis.
As we have argued regarding the full-scale MSES, we
admit that we see limited value in a scale that asks students
to provide confidence judgments to perform 18 varied,
math-related tasks. Each judgment is logically related only
to the task on which that judgment is made. Computing
one's income taxes for a year or figuring out how much
lumber one needs to buy in order to build a set of bookshelves, two of the tasks in the Tasks self-efficacy subscale,
clearly have "figuring" in common. However, it is difficult
to see what one's judgments of confidence to accomplish
these two tasks may predict other than one's ability to
accomplish the specific tasks themselves. Neither judgment
may prove particularly useful in understanding a myriad of
other math-related tasks, choosing a math-related major, or
predicting other mathematical capabilities. Completing cer-

197

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

MATHEMATICS SELF-EFFICACY

tain IRS forms, for example, may tax the minds and wills of
even our finest psychometricians and nuclear scientists.
We expect that the implications from our findings will not
be viewed as limited to investigations of mathematics selfefficacy. The mismatch between self-efficacy and performance assessment is also prevalent in investigations in other
academic areas. Shell, Murphy, and Bruning (1989) reported significant correlations between measures of reading
and writing self-efficacy and related outcomes. Their selfefficacy instrument was composed of two scalesone measured students' confidence that they could perform specific
reading or writing tasks and another that they possessed
specific skills. The criteria for assessing reading and writing
performance were similar only to the skills on which the
skills self-efficacy scale were based. Although the researchers reported that the skills subscale scores had stronger
correlations with performance than did the task subscale's
scores, they used the composite score as the self-efficacy
assessment. Using procedures and instruments similar to
those of Shell et al., Pajares and Johnson (1994) also found
that writing skills self-efficacy was a stronger predictor of
essay scores holistically assessed in terms of writing skills
than was writing tasks self-efficacy. Consequently, we suggest that investigators in academic areas other than mathematics also think carefully about the match between the
appropriateness of self-efficacy and outcome assessment.
Our findings also have implications for the type of mathematics research that focuses on international and multicultural comparisons of math confidence. For example, results
from the Second Study of Mathematics conducted by the
International Association for the Evaluation of Educational
Achievement (Robetaille & Garden, 1988) revealed that
although Japanese students outperformed American students in most areas of mathematics, American students
exhibited greater confidence in their mathematical ability.
This confidence, however, was assessed with global, decontextualized items of the sort that have plagued self-efficacy
research (e.g., "I am not so good at mathematics"). Robetaille and Garden (1988) posited that the negative responses
of Japanese students may have been due to a culturally
based reluctance to boast about their academic ability but
argued that a genuine lack of self-confidence was a more
likely explanation. Regardless of which explanation is more
likely, it would be instructive to assess mathematics selfefficacy in international and multicultural settings in a contextual manner consistent with the guidelines of social cognitive theory and with the findings from this investigation.
If, for example, the criterial task in such international studies involves the solving of math problems, the confidence
assessment should consist of students' judgments of their
confidence to solve those problems rather than of global
confidence statements infused with personal judgments of
self-worth. The results would be a much less culturally
biased assessment, a clearer and more specific judgment of
perceived capability, and a better understanding of the
relationship between self-efficacy beliefs and academic
outcomes.

References
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of
behavioral change. Psychological Review, 84, 191-215.
Bandura, A. (1986). Social foundations of thought and action: A
social cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A., & Schunk, D. H. (1981). Cultivating competence,
self-efficacy, and intrinsic interest through proximal self-motivation. Journal of Personality and Social Psychology, 41, 586598.
Benson, J. (1989). Structural components of statistical test anxiety
in adults: An exploratory study. Journal of Experimental Education, 57, 247-261.
Betz, N. E., & Hackett, G. (1981). The relationship of careerrelated self-efficacy expectations to perceived career options in
college women and men. Journal of Counseling Psychology, 28,
399-410.
Betz, N. E., & Hackett, G. (1983). The relationship of mathematics
self-efficacy expectations to the selection of science-based college majors. Journal of Vocational Behavior, 23, 329-345.
Cooper, S. E., & Robinson, D. A. G. (1991). The relationship of
mathematics self-efficacy beliefs to mathematics anxiety and
performance. Measurement and Evaluation in Counseling and
Development, 24, 4-11.
Dowling, D. M. (1978). The development of a mathematics confidence scale and its application in the study of confidence in
women college students. Unpublished doctoral dissertation,
Ohio State University, Columbus.
Fennema, E., & Sherman, J. A. (1976). Fennema-Sherman Mathematics Attitude Scales: Instruments designed to measure attitudes toward the learning of mathematics by females and males.
JSAS Catalog of Selected Documents in Psychology (No. 1225),
6, 31.
Goldman, R. D., & Hewitt, B. N. (1976). The scholastic aptitude
test "explains" why college men major in science more often
than college women. Journal of Counseling Psychology, 23,
50-54.
Hackett, G. (1985). The role of mathematics self-efficacy in the
choice of math-related majors of college women and men: A
path analysis. Journal of Counseling Psychology, 32, 47-56.
Hackett, G., & Betz, N. E. (1989). An exploration of the mathematics self-efficacy/mathematics performance correspondence.
Journal for Research in Mathematics Education, 20, 261-273.
Hackett, G., & O'Halloran, S. (1985). [Test-retest reliability of
career self-efficacy measures]. Unpublished raw data.
Joreskog, K. G., & Sorbom, D. (1989). LISREL 7 user's guide.
Mooresville, IN: Scientific Software.
Kranzler, J., & Pajares, F. (1994). An exploratory factor analysis
of the Mathematics Self-Efficacy ScaleRevised. Manuscript
submitted for publication.
Langenfeld, T. E., & Pajares, M. F. (1993, April). The Mathematics Self-Efficacy Scale (MSES): Refining the construct. Paper
presented at the meeting of the American Educational Research
Association, Atlanta, GA.
Lapan, R. T., Boggs, K. R., & Morrill, W. H. (1989). Self-efficacy
as a mediator of investigative and realistic general occupational
themes on the Strong-Campbell Interest Inventory. Journal of
Counseling Psychology, 36, 176-182.
Lent, R. W., & Hackett, G. (1987). Career self-efficacy: Empirical
status and future directions. Journal of Vocational Behavior, 30,
347-382.
Lent, R. W., Lopez, F. G., & Bieschke, K. J. (1991). Mathematics
self-efficacy: Sources and relation to science-based career
choice. Journal of Counseling Psychology, 38, 424-430.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

198

FRANK PAJARES AND M. DAVID MILLER

Lent, R. W., Lopez, F. G., & Bieschke, K. J. (1993). Predicting


mathematics-related choice and success behaviors: Test of an
expanded social cognitive model. Journal of Vocational Behavior, 42, 223-236.
Maddux, J. E., Norton, L. W., & Stoltenberg, C. D. (1986).
Self-efficacy expectancy, outcome expectancy, and outcome
value: Relative effects on behavioral intentions. Journal of Personality and Social Psychology, 51, 783-789.
Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of
self-efficacy beliefs to academic outcomes: A meta-analytic
investigation. Journal of Counseling Psychology, 38, 30-38.
Norwich, B. (1986). Assessing perceived self efficacy in relation
to mathematics tasks: A study of the reliability and validity of
assessment. British Journal of Educational Psychology, 56,
180-189.
Pajares, F., & Johnson, M. J. (1994). Confidence and competence
in writing: The role of self-efficacy, outcome expectancy, and
apprehension. Research in the Teaching of English, 29, 313
331.
Pajares, F., & Kranzler, J. (1994). Role of self-efficacy, selfconcept, and general mental ability in mathematical problemsolving. Manuscript submitted for publication.
Pajares, F., & Miller, M. D. (1994). The role of self-efficacy and
self-concept beliefs in mathematical problem-solving: A path
analysis. Journal of Educational Psychology, 86, 193-203.
Randhawa, B. S., Beamer, J. E., & Lundberg, I. (1993). Role of
mathematics self-efficacy in the structural model of mathematics
achievement. Journal of Educational Psychology, 85, 41-48.
Reyes, L. H. (1984). Affective variables and mathematics education. The Elementary School Journal, 84, 558-581.

Robetaille, D. F., & Garden, R. A. (Eds.). (1988). The IEA study of


mathematics: 2. Contexts and outcomes of school mathematics.
Oxford, England: Pergamon Press.
Schunk, D. H. (1981). Modeling and attributional effects on children's achievement: A self-efficacy analysis. Journal of Educational Psychology, 73, 93-105.
Schunk, D. H. (1982a). Effects of effort attributional feedback on
children's perceived self-efficacy and achievement. Journal of
Educational Psychology, 74, 548-556.
Schunk, D. H. (1982b). Verbal self-regulation as a facilitator of
children's achievement and self-efficacy. Human Learning, 1,
265-277.
Schunk, D. H. (1989). Self-efficacy and achievement behaviors.
Educational Psychology Review, 1, 173-208.
Schunk, D. H. (1991). Self-efficacy and academic motivation.
Educational Psychologist, 26, 207-231.
Shell, D. F., Murphy, C. C , & Bruning, R. H. (1989). Self-efficacy
and outcome expectancy mechanisms in reading and writing
achievement. Journal of Educational Psychology, 81, 91-100.
Smith, R. J., Arnkoff, D. B., & Wright, T. L. (1990). Test anxiety
and academic competence: A comparison of alternative models.
Journal of Counseling Psychology, 37, 313-321.
Williams, E. J. (1959). The comparison of regression variables.
Journal of the Royal Statistical Society (Series B), 21, 396-399.

Received March 28, 1994


Revision received July 18, 1994
Accepted August 1, 1994

Vous aimerez peut-être aussi