AN EXAMINATION OF PSYCHOMETRIC PROPERTIES OF BANDURA'S MULTIDIMENSIONAL SCALES OF PERCEIVED SELF-EFFICACY
By Janice Williams Miller; William T. Coombs and Dale R. Fuqua
This study examined selected psychometric characteristics of Bandura's (1989) Multidimensional Scales of Perceived Self-Efficacy, including reliability and factor structure. Interrelationships among the factors and implications of the findings are discussed as well as suggestions for future inquiry.
Bandura's (1986) social-cognitive theory of self-efficacy guided the development of the Multidimensional Scales of Perceived Self-Efficacy (MSPSE; Bandura, 1989). According to this theory, self-efficacy perceptions refer to "beliefs in one's capabilities to organize and execute the courses of action required to produce given attainments" (Bandura, 1997, p. 3). Bandura (1997) proposed that individuals who perceive themselves as capable tend to attempt and successfully execute tasks or activities, Self-efficacy studies in education have clarified and extended the role of efficacy beliefs as one mechanism underlying learning strategy use (Pintrich & DeGroot, 1990), goal setting (Locke & Latham, 1990), persistence (Gorrell & Capron, 1988), and academic success (Schunk, 1985; Williams, 1996). Precise and detailed measurement of efficacy judgments are typically highly related to subsequent school performance (Schunk, 1991). Therefore, the ability to accurately assess efficacy perceptions in educational settings seems to warrant systematic investigation.
The MSPSE were developed in response to the theoretical and applied importance of the self-efficacy construct. Bandura's (1986) social-cognitive theory of perceived self-efficacy specified the origins and structure of efficacy beliefs. Items on the scales were tailored to academic domains of functioning. The utility of such measures is based on the integrity of the scores produced by the instrument. Studies using this assessment tool have recently been cited in the literature (e.g., Bryant & Fuqua, 1997). Furthermore, researchers have begun administering separate subscales from the instrument (see Williams, 1996; Zimmerman, Bandura, & Martinez-Pons, 1992). However, there is very little psychometric data available on this important measure. Research results indicate that perceptions of academic efficacy are more predictive of academic achievement then the widely used traditional measures of self-concept of ability (Bandura, Barbaranelli, Caprata, & Pastorelli, 1996). The refinement of assessment tools that increase the explanatory and predictive power of self-efficacy constructs may advance the understanding of social-cognitive processes. The purpose of this study was to examine psychometric properties (e,g., reliability, construct validity) of the MSPSE, which is now being used to measure student self-efficacy in applied settings. Specifically, the following research questions are addressed:
The participants (N = 500; 248 girls, 202 boys, and 50 who did not report gender) were primarily junior (43%) or senior (50%) public high school students. Students came from urban, suburban, and rural backgrounds and they were predominantly (96%) middle-class and White. All students were participants in one of nine single-day university-sponsored ACT preparation workshops. Students voluntarily attended the workshop on a Saturday for various reasons, including personal choice, parental desire, or on the advice of their school counselor. The workshop coordinator facilitated completion of the self-administered self-efficacy instrument among the students who had agreed to participate in the study and who had returned parental or guardian permission slips.
The MSPSE is a 57-item self-report measure with nine subscales. Each subscale comprises items rated on a 7-point Likert-type scale (1 = not well at all, 3 = not too well, 5 = pretty well. 7 = very well). Larger student scores indicate higher levels of self-efficacy beliefs. Table 1 shows the names and the number of items for each subscale. Sample items include the following: "How well can you organize your schoolwork?" (Self-Efficacy for Self-Regulated Learning), "How well can you control your temper?" (Self-Regulatory Efficacy), and "How well can you stand up for yourself when you are being treated unfairly?" (Self-Assertive Efficacy). Internal consistency reliability (alphas ranging from .63 to .87 with an overall coefficient of .92) has been reported with a college-aged sample (Bryant & Fuqua, 1997).
Table 1 shows the means and standard deviations for each of the nine MSPSE subscales. To place descriptive statistics in the same metric as the original Likert-type scale, subscale scores for each participant were divided by the number of items on that subscale before computing the mean and standard deviation for that subscale. For example, the Self-Efficacy in Enlisting Social Resources subscale score for each individual was divided by 4 before computing the mean (5.3) and standard deviation (.84) because there are four items on that subscale. The subscale means, ranging from 5.1 to 6.0, indicated that those in this college-bound sample tended to have positive attitudes about their capabilities to produce desired performance levels.
Also shown in Table 1 as diagonal elements are Cronbach's (1951) alpha reliability coefficients, which ranged from 0.60 to 0.87 for the Self-Efficacy in Enlisting Social Resources (ESR) and the Self-Efficacy for Self-Regulated Learning (SRL) subscale scores, respectively. These values demonstrate varying degrees of response consistency adequacy, with eight of the nine scales associated with values that ranged between .70 and .87. Furthermore, although reliability indices are typically a function of the number of subscale items, three of the four 4-item subscales were associated with alphas ranging from .70 to .84. Therefore, the ESR 4-item subscale (alpha = .60) may be an outlier. It is clear that continued use of this subscale would require improvement in reliability. It might be recommended that additional items be added to this subscale to increase their respective reliabilities.
Correlations between subscale pairs are shown as off-diagonal elements in Table 1. These coefficients were relatively dispersed, ranging from 0.12 (Self-Regulatory Efficacy--Self-Efficacy for Leisure Time Skills and Extracurricular Activities) to 0.56 (Self-Efficacy for Self-Regulated Learning-Self-Efficacy to Meet Other's Expectations Subscales). Hence, the proportion of variance in one scale accounted for by another scale (i.e., r2 ranges between 1% and 31%. The magnitude of some of these inter-subscale correlations suggests that higher-order factors may exist.
To ascertain the structure of the 57-item MSPSE, (common) factors were extracted using (a) principal factor analysis, (b) squared multiple correlations as initial commonality estimates, (c) the raw score data set, and (d) PROC FACTOR in SAS. A PROMAX rotation (Hendrickson & White, 1964) was selected with VARIMAX (Kaiser, 1958) prerotation and where the power of k = 4 was chosen to compute the target pattern. This resulted in eight factors with eigenvalues greater than 1, accounting for approximately 89% of the total variance. By rotating nine factors, the total percentage of variance accounted for increased to 92%. An examination of a scree test was inconclusive between an eight- and nine-factor solution. However, three observations strongly encouraged the rotation of nine factors rather than eight. First, the most important of these was that the ninth factor, as will be seen, offered some real conceptual and theoretical value. Second, the original theory on which the instrument is based had proposed nine factors. Third. the 3 percentage points of total variance gained by including the ninth factor supported the decision.
The nine factors in combination accounted for approximately 92% of the total variance of the item set. For interpretational clarity, a salient loading (Gorsuch, 1983, p. 208) of 0.40 was selected as one that is sufficiently high to assume the existence of an item-factor relationship. (Reader's Note. A table of the rotated oblique first-order factor structure ]item-factor correlations], rotated oblique first-order factor pattern [standardized regression coefficients], and percent of variance explained by each first-order factor ignoring other first-order factors may be obtained from the authors.) We considered the nature of the factors and labeled them.
The first factor, which accounted for 18% of the variance, was labeled Self-Efficacy in Basic Study Skills (BSS). This factor included 14 items such as "How well can you plan your school work?" Accounting for 10% of the variance, the second factor, Self-Efficacy in Resisting Peer Pressure (RPP), contained 10 items such as "resist pressure to drink beer, wine, or liquor?" The third factor, which accounted for 13% of the variance, was labeled Self-Assertive Efficacy (SAE). This factor included 10 items such as "stand up for yourself when you feel you are being treated unfairly?" The fourth factor, Social Self-Efficacy (SOC), accounted for 12% of the variance, with 8 items such as "carry on conversations with others?" Also accounting for 12% of the variance, the fifth factor, Self-Efficacy in Seeking Help (SHE), consisted of 8 items such as "get your parent(s) to help you with a problem?" Self-Efficacy for Academic Achievement (AAE), the sixth factor, contained 10 items, which accounted for 12% of the variance (e.g., "learn reading and writing language skills?"). The seventh factor, Self-Efficacy for Extracurricular Activities (EXT, 8 items like "do the kinds of things needed to take part in school plays?"), accounted for 9% of the variance. Accounting for 8% of the variance, Self-Efficacy for Team Athletics (ATH) included 4 items such as "learn skills needed for team sports?" The ninth factor, Self-Efficacy in Mathematics Achievement (MTH, 4 items, e.g., "learn algebra?"), accounted for 6% of the variance. Two of the items on the MSPSE ("get another student to help you when you get stuck on homework?" and "learn to use computers?") failed to load on any empirical factor and were associated with correspondingly low estimates of commonality, .23 and .19, respectively.
Descriptive statistics presented in Table 2 are means and standard deviations for each of the nine empirical factors. Because factor scores have a standardized metric, the interpretation of summary statistics differs from those examined for the original MSPSE subscales. In this case, a positive factor score for an individual would indicate higher self-reported efficacy on that factor, whereas a negative factor score would be interpreted as lower self-perceived efficacy.
Shown as diagonal elements in Table 2, alpha coefficients--computed using only the items with an item-factor correlation of at least 0.45--ranged from .70 to .89 for the MTH and BSS, respectively. The increase in reliability (relative to the original MSPSE) is obvious.
Correlations between factor pairs are shown as off-diagonal elements in Table 2. These coefficients were dispersed, ranging from -.01 (MTH-EXT) to 0.57 (SOCSAE). Hence, anywhere from 0% to 32% of the variance is shared between factors. This measurement overlap between some combinations of the factors suggested the potential presence of higher order factors.
Table 3 shows correlations between the original MSPSE subscale and empirical first-order factor pairs. It is clear that factors related fairly well to the subscale structure. For example, seven of the nine correlations were at least 0.82. Five of these correlations were at least 0.90. However, two of the original MSPSE four-item subscales (Self-Efficacy to Meet Others' Expectations [EXP] and Self-Efficacy in Enlisting Social Resources [ESR]) were somewhat incongruent with the empirical model. The EXP subscale seemed to be less concentrated, which correlated moderately with five of the empirical factors (BSS, RPP, SAE, SOC, SHE). The ESR subscale had low reliability (.60, see Table 1) and only moderate correlations with two of the empirical factors (SOC. SHE). The Self-Efficacy for Leisure Time Skills and Extracurricular Activities (LEI) subscale splintered into two conceptual empirical factors--an extracurricular factor (EXT) and a team sports factor (ATH). Similarly, the MSPSE academic achievement factor was redefined as two factors--a general academic achievement factor (AAE) and a specific mathematical factor (MTH). A noticeable feature of Table 3 (see also Table 1 and Table 2) is the nonindependence of the subscales. Several of the MSPSE subscales load on several empirical factors and vice versa. Higher order factors need to be examined.
To examine potential higher order relationships among the first-order empirical factors, second-order factors were extracted using principal axis factor analysis with an oblique (PROMAX, k = 2) rotation. Consideration of previous empirical findings. the theoretical nature of the construct, as well as inspection of eigenvalues, a scree test, and the percentage for which total variance was accounted suggested a three-factor solution. More specifically, the selection of this three-factor solution was heavily influenced by the fact that Bandura et al. (1996) had already reported a three-factor solution. Furthermore, the theoretical implications of the nature of these three factors were reasonable, and we consider this in the Discussion section. Although there was a minor break between the third and fourth factors on the scree plot, it was much less convincing than the consistency of the three-factor solution with existing research and theory.
Table 4 shows the (a) rotated oblique second-order factor structure (first-order factor and second-order factor correlations), (b) rotated oblique second-order factor pattern (standardized regression coefficients), (c) percentage of variance explained by each second-order factor ignoring other second-order factors (i.e., the sum of the squared elements of the factor structure corresponding to each factor), and (d) estimates of commonality. The factors made sense conceptually, and we labeled them Social Efficacy, Task Management Efficacy, Academic Efficacy.
The second-order Social Efficacy factor comprises first-order factors (and their corresponding items) that either directly (SOC, SAE) or indirectly (ATH, EXT) require social skills. It is clear, for example, that there is a social component to the EXT factor. Similarly, the second-order Task Management factor comprised first-order factors that were more product-oriented than process-oriented (e.g., BSS, RPP, SHE). The second-order Academic Efficacy factor was defined by first-order factors (AAE, MTH) and corresponding items that were conceptually consistent with the selected label. First-order factors that loaded on more than one second-order factor did so in a logical manner. For example, one would expect basic study skills, which involve the management of tasks, to be related to academic efficacy. These three empirical second-order factors (BSS, Task Management Efficacy, and Academic Efficacy) are similar to those identified (Social Efficacy, Self-Regulatory Efficacy, and Academic Efficacy) in a prior study by Bandura et al. (1996).
Correlations between second-order factor pairs were 0.50 (Social Efficacy-Task Management Efficacy), 0.45 (Social Efficacy-Academic Efficacy), and 0.65 (Task Management Efficacy-Academic Efficacy). It is intuitively attractive that the smallest correlation corresponds to the social-academic relationship, and the largest correlation corresponds to the academic-task management relationship.
The theoretical implications of the MSPSE seem to be important given the prominent role of self-efficacy beliefs in diverse lines of research, for example in human development and adaptation (Schwarzer, 1992), persistence and success in academic course work (Lent, Brown, & Hackett, 1994), and goal aspiration (Locke & Latham, 1990). More specifically, the relationship of either the nine scales or the nine first-order factors could be used in conjunction with the three higher order factors to build a hierarchical model that could be subjected to confirmatory factor analysis. This could provide a more complete hierarchical model than first-order factors alone.
The similarity of the factors found in the current sample to the nine scales of the MSPSE (see Table 3) was expected, because the original scales resulted from factor analytic study with self-efficacy purported to be a multidimensional construct (Bandura, 1997). In the current study, efficacy perceptions seemed to vary across academic domains of functioning and represented more than an undifferentiated disposition. Previous research documents the value of specificity of self-efficacy assessments in explaining different facets of academic performance (Pajares & Miller, 1995). Given that the MSPSE subscales were empirically derived, their theoretical fit with the construct deserves serious attention. We did not find the relationship of the subscale structure meaningful in any theoretical or construct schema we could identify. The higher order factors identified here could be interpreted in several ways. For example, the task-management factor might represent basic cognitive skills applied in either social (higher order Factor 1) or academic (higher order Factor 3) situations. Of course alternative schema could be proposed. The point is that the relation between the instrument and its structures to well organized and clear theoretical dimensions is elusive. Careful theoretical analysis and further study of the relationship of MSPSE dimensions to external variables might be enlightening regarding the nature of the construct and this instrument.
The generalizability of the study is limited by the inherent subjectivity of the factor-analytic process (e.g., selection of the number of factors, choice of rotation). Other researchers might provide different labels for the obtained factors. Furthermore, the demographic characteristics of the sample may be a limiting factor. For example, the current study used a homogeneous, college-bound sample with few students scoring at the lower ranges of the self-efficacy scale. The factor structure presented here may differ somewhat for less academically oriented students. In addition, pattern differences due to gender or cultural groupings were unexplored and may exist. The specific situation under which perceived self-efficacy perceptions are reported may also alter the factor-analytic structure. For example, recent studies have explored self-efficacy across several skill and anxiety domains (see Schunk, 1987). In light of the nonacademic factors uncovered here, the predictive validity of self-efficacy within various domains might also be of interest.
Interpretation of subscale scores for an individual who has taken the MSPSE must be approached cautiously. The basic factor structure of the instrument is a reasonable representation of the model on which it is based. However, there is considerable variability in the degree to which specific subscales are either reliable or valid. At the two extremes are the SRL and ESR subscales, SRL is an 11-item subscale with relatively high reliability (alpha = .87), displaying reasonable content validity (i.e., the items are generally conceptually consistent with the subscale label), The subscale also demonstrates a high degree of construct validity, as evidenced by the strong relationship (r = .96) between the subscale and the first-order empirically based BSS factor. Conversely, ESR is a four-item subscale with low reliability (alpha = .60), content deficiency due to the limited number of items included in the subscale, and an item structure that failed to replicate empirically in the form of a single factor.
The instrument seems to have been constructed with the theoretical assumption that a general academic self-efficacy factor would emerge. If limited to visual inspection and consideration of the items (content validity) and the reliability of the instrument as a single scale (alpha = .92), one might concede the existence of this overall factor. The similarity in factor structure between Bandura et al.'s [1996) final conceptualization of the MSPSE and the empirical factor structure found in this study provides compelling evidence against the existence of a general academic construct. The disproportional per subscale item frequencies suggest that the model selected to guide final subscale construction was empirically, rather than theoretically, driven. The instrument would benefit if additional items were developed to complement what seems to have (initially) been a theoretical model that underdefined what is apparently a complex self-efficacy construct structure.
Finally, further research is needed to confirm the replicability of our higher order factor structure findings (e.g., confirmatory factor analysis), to assess the stability of these factors across different samples and groups within academic settings, and to evaluate self-efficacy under a broader range of situationally specific stimuli (e.g., verbal, athletic). It is conceivable, for example, that an entirely different factor structure might materialize if participants were asked "How well can you..." in a scenario that was not so clearly academic in orientation. There are numerous personality dimensions (e.g., sex role identity, confidence, locus of control) that, if related to the second-order Social, Task Management, and Academic Efficacy factors, would provide insight into the theoretical underpinnings of the self-efficacy construct. Does intelligence. for example, relate in a conceptually attractive manner? Conversely, how do dimensions that are theoretically unrelated to these three factors covary? That is, to what degree is there divergent validity? Answers to these questions would be a useful extension of the self-efficacy literature.
Legend for Chart: A - Subscale B - ESR C - AAE D - SRL E - LEI F - SRE G - EXP H - SOC I - SAE J - EPS K - M L - SD A B C D E F G H I J K L ESR (.60)[a] -- -- -- -- -- -- -- -- 5.3 0.84 AAE .24[*] (.74) -- -- -- -- -- -- -- 5.2 0.73 SRL .37[*] .47[*] (.87) -- -- -- -- -- -- 5.1 0.86 LEI .31[*] .24[*] .32[*] (.74) -- -- -- -- -- 5.2 0.85 SRE .22[*] .22[*] .42[*] .12[*] (.79) -- -- -- -- 6.0 0.89 EXP .41[*] .34[*] .56[*] .30[*] .44[*] (.73) -- -- -- 5.3 0.91 SOC .44[*] .21[*] .33[*] .45[*] .15[*] .44[*] (.83) -- -- 5.9 0.92 SAE .32[*] .23[*] .25[*] .43[*] .13[*] .40[*] .53[*] (.84) -- 5.6 1.03 EPS .48[*] .13[*] .33[*] .31[*] .27[*] .39[*] .34[*] .26[*] (.70) 5.1 1.14
Note. MSPSE = Multidimensional Scales of Self-Efficacy; ESR = Self-Efficacy in Enlisting Social Resources (4 items); AAE = Self-Efficacy for Academic Achievement (9 items); SRL = Self-Efficacy for Self-Regulated Learning (11 items); LEI = Self-Efficacy for Leisure Time Skills and Extracurricular Activities (8 items); SRE = Self-Regulatory Efficacy (9 items); EXP = Self-Efficacy to Meet Others' Expectations (4 items); SOC = Social Self-Efficacy (4 items); SAE = Self-Assertive Efficacy (4 items); EPS = Self-Efficacy for Enlisting Parental and Community Support (4 items).
a Statistics reported in parentheses in the main diagonal are internal consistency estimates using coefficient alpha.
b Subscale scores for each participant were divided by the number of items on that subscale before computing the means and standard deviation for that subscale.
* Significant correlation p < .01.
Legend for Chart: A - Subscale B - BSS C - RPP D - SAE E - SOC F - SHE G - AAE H - EXT I - ATH J - MTH K - M L - SD A B C D E F G H I J K L BSS (.89) -- -- -- -- -- -- -- -- .00 .96 RPP .50[*] (.81) -- -- -- -- -- -- -- .00 .93 SAE .34[*] .20[*] (.85) -- -- -- -- -- -- .00 .94 SOC .32[*] .15[*] .57[*] (.82) -- -- -- -- -- .00 .93 SHE .55[*] .37[*] .48[*] .42[*] (.78) -- -- -- -- .00 .91 AAE .52[*] .23[*] .46[*] .29[*] .31[*] (.76) -- -- -- .00 .92 EXT .34[*] .17[*] .33[*] .47[*] .23[*] .42[*] (.76) -- -- .00 .90 ATH .06 .01 .46[*] .42[*] .26[*] .10 .23[*] (.86) -- .00 .94 MTH .28[*] .13[*] .25[*] .01 .21[*] .41[*] -.01 .10 (.70) .00 .90
Note. BSS = Self-Efficacy in Basic Study Skills; RPP = Self-Efficacy in Resisting Peer Pressure; SHE = Self-Efficacy in Seeking Help; EXT = Self-Efficacy for Team Athletics; MTH = Self-Efficacy in Mathematics Achievement. See Table 1 Note.
See Table 1 specific and probability notes.
Legend for Chart: A - MSPSE Subscale B - First-Order Empirical Factors, BSS C - First-Order Empirical Factors, RPP D - First-Order Empirical Factors, SAE E - First-Order Empirical Factors, SOC F - First-Order Empirical Factors, SHE G - First-Order Empirical Factors, AAE H - First-Order Empirical Factors, EXT I - First-Order Empirical Factors, ATH J - First-Order Empirical Factors, MTH A B C D E F G H I J ESR .38 .19 .38 .61[a] .67[a] .26 .33 .17 .26 AAE .44[a] .22 .34 .19 .23 .90[b] .29 .10 .68[b] SRL .96[b] .39 .37 .33 .50[a] .57[a] .41 .09 .31 LEI .28 .13 .48[a] .48[a] .34 .34 .82[b] .67[b] .07 SRE .52[a] .97[b] .19 .13 .42[a] .25 .16 -.03 .11 EXP .67[a] .46[a] .57[a] .46[a] .63[a] .37 .17 .24 .31 SOC .31 .16 .62 .94[b] .39 .33 .41[a] .41[a] .01 SAE .22 .15 .94[b] .52[a] .34 .34 .33 .32 .12 EPS .36 .25 .34 .37 .88[b] .21 .24 .25 .06
Note. See Tables 1 and 2 Notes.
a r greater than or equal to .40. b Largest subscale-factor correlation in each column.
Legend for Chart: A - Factor B - Social Self-Efficacy C - Task Management Efficacy D - Academic Efficacy E - Commonalities h2 A B C D E BSS -- .79 (.71) .47 .64 RPP -- .57 (.59) -- .33 SAE .72 (.63) -- .46 .60 SOC .74 (.71) -- -- .57 SHE .47 .61 (.48) -- .45 AAE -- .52 .66 (.51) .52 EXT .51 (.44) -- -- .32 ATH .57 (.61) -- -- .35 MTH -- -- .57 (.59) .33 Percent of Variance 38 36 25 --
Note. See Tables 1 and 2 Notes. Only factor structure and pattern coefficients whose magnitude is greater than 0.40 are presented. Factor pattern coefficients are given in parentheses.
Bandura A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura A. (1989). The Multidimensional Self-Efficacy Scales. Unpublished test, Stanford University, Stanford, CA.
Bandura A. (1997). Self-efficacy: The exercise of control. New York: Freeman.
Bandura, A., Barbaranelli, C., Caprara, G. V., & Pastorelli, C. (1996). Multifaceted impact of self-efficacy beliefs on academic functioning. Child Development, 67, 1206-1222.
Bryant, N. C., & Fuqua, D. R. (1997. March). A structural analysis of the relationship between self-efficacy and sex role orientation. Paper presented at the meeting of the American Educational Research Association. Chicago, IL.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika. 16, 297-334.
Gorrell, J., & Capron, E. W. (1988). Effects of instructional type and feedback on prospective teachers' self-efficacy beliefs. Journal of Experimented Education, 57, 231-244.
Gorsuch, R. L. (1983). Factor analysts. Hillsdale, NJ: Erlbaum.
Hendrickson, A. E., & White, P.O. (1964). A quick method for rotation to oblique simple structure. British Journal of Statistical Psychology, 17, 65-70.
Kaiser, H. F. (1958). The varimax criterion for analytic rotation in factor analysis. Psychometrika, 23, 187-200.
Lent, R. W., Brown, S. D., & Hackett, G. (1994). Toward a unifying social cognitive theory of career and academic interest, choice, and performance. Journal of Vocational Behavior, 45, 79-122.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task performance. Englewood Cliffs, NJ: Prentice-Hall.
Pajares, F., & Miller, M.D. (1995). Mathematics self-efficacy and mathematics performances: The need for specificity of assessment. Journal of Counseling Psychology, 42, 190-198.
Pintrich, P. R., & DeGroot, E. V. (1990). Motivational and self-regulated learning components of classroom academic performance. Journal of Educational Psychology, 82. 33-40.
Schunk, D. (1985). Self-efficacy and school learning. Psychology in the Schools. 22, 208-223.
Schunk, D. (1987,April). Domain-specific measurement of students' self-regulated learning processes. Paper presented at the meeting of the American Educational Research Association, Washington, DC.
Schunk, D. (1991). Self-efficacy and academic motivation. Educational Psychologist, 26, 207-231.
Schwarzer, R. (Ed.). (1992). Self-efficacy: Thought control of action. Washington, DC: Hemisphere.
Williams, J. E. (1996). The relationship between efficacy for self-regulated learning and domain-specific academic performance, controlling for test anxiety. Journal of Research and Development in Education, 29, 77-80.
Zimmerman, B. J., Bandura, A., & Martinez-Pons, M. (1992). Self-motivation for academic attainment: The role of self-efficacy beliefs and personal goal setting. American Educational Research Journal, 29, 663-676.
By Janice Williams Miller; William T. Coombs and Dale R. Fuqua
Janice Williams Miller is an associate professor, William T. Coombs is an associate professor, and Dale R. Fuqua is a professor, all in the School of Educational Studies at Oklahoma State University, Stillwater. Correspondence regarding this article should be sent to Janice Williams Miller, 313 Willard Hall, Oklahoma State University, Stillwater, OK 74078 (e-mail: email@example.com).