A question of calibration: a review of the self-efficacy beliefs of students with learning disabilities.

Abstract. This article reviews the literature on the self-efficacy beliefs of students with learning disabilities (LD). Motivational and metacognitive difficulties of students with LD are briefly discussed, followed by a synopsis of Bandura's self-efficacy theory, with special attention to the issue of calibration. From the literature search, 22 studies met the criteria of (a) using a measure of self-efficacy, and (b) including a sample of students identified as having learning disabilities. The resulting body of literature is summarized and analyzed in terms of the nature of the sample, the performance task or domain, the self-efficacy measure used, the research question and outcomes, and the accuracy of calibration between perceived self-efficacy and task outcome. The results from this review suggest that in specific contexts--in the writing performance of students with specific writing difficulties, in particular--students appear to optimistically miscalibrate their self-efficacy. Methodological problems found in some of the research, such as "conceptual blurring," are discussed. Finally, implications for practice are considered, and suggestions are made for future research.

**********

According to Bandura, "Among the mechanisms of personal agency, none is more central or pervasive than people's beliefs about their capabilities to exercise control over events that affect their lives" (1989, p. 1175). Self-efficacy has been defined as "beliefs in one's capabilities to organize and execute the courses of action required to produce given attainments" (Bandura, 1997, p. 2). These perceptions of self-capabilities or self-efficacy have been identified as key factors affecting thought patterns and performance in a wide variety of tasks. For example, self-efficacy perceptions influence choice of activity, task perseverance, level of effort expended, and ultimately, degree of success achieved. Inaccurate estimates of self-efficacy may develop from faulty task analysis or from a lack of self-knowledge (Bandura & Schunk, 1981)--two problems shown to be prevalent in students with learning disabilities' (Butler, 1999; Meltzer, Roditi, Houser, & Perlman, 1998; Swanson, 1989).

The issue of calibration of self-efficacy beliefs with ensuing performance has been described by self-efficacy researchers (e.g., Bandura, 1997; Pajares, 1996; Sawyer, Graham, & Harris, 1992), who maintain that optimistic self-efficacy beliefs are instrumental to the successful completion of challenging tasks. If, as suggested by various researchers (e.g., Butler, 1998a; Meltzer et al., 1998), students with learning disabilities (LD) struggle with self-knowledge and task awareness, their ability to judge their efficacy--the way in which they calibrate their ability and performance--may be affected. Although a number of investigators have addressed the self-efficacy beliefs of students with learning disabilities, this work has not been synthesized, and the issue of calibration of efficacy beliefs has not been comprehensively addressed.

The past 20 years have seen considerable research of a variety of "self" constructs and their relation to students with learning disabilities (for example, see review of the self-concept of children with LD by Chapman, 1988). A great deal of research on various aspects of self-efficacy beliefs has also been conducted, with a number of studies including students with learning disabilities. To date, however, there has been no comprehensive and critical review of the role self-efficacy beliefs play in the academic functioning of individuals with learning disabilities.

This review examines the literature on the self-efficacy beliefs of students with learning disabilities. After briefly addressing motivational and metacognitive problems of students with learning disabilities, the self-efficacy component of Bandura's social cognitive theory is discussed. The significance of the issue of calibration --the degree of congruence between efficacy beliefs and actual performance--is then reviewed. A certain degree of optimism or positive bias in one's calibration is thought to be advantageous (Bandura, 1997) but students with LD struggle with self-knowledge and self-assessment (Butler, 1999), and these deficits may adversely influence their accuracy of calibration. Following the discussion of calibration issues, I summarize and critically review the empirical studies that explore the self-efficacy beliefs of LD students. I then approach the extant literature with the following questions: How accurately calibrated are the self-efficacy measurements of students with learning disabilities? Why might some students with learning disabilities overestimate their beliefs? What is the potential impact of miscalibration of efficacy beliefs? What are some of the problems with the way in which self-efficacy is measured in these studies? Finally, implications for practice and suggestions for future research are presented based on the findings in this review.

Motivation, Metacognition and Learning Disabilities

Motivational beliefs influence task approaches and affect the development of metacognitive skills. Well-developed skills in metacognition--awareness of one's cognitive processes, cognitive strengths and weaknesses, and self-regulation (Flavell, 1976)--are necessary for successful academic functioning, but LD students have been shown to display deficiencies in evaluating their skills and progress (e.g., Swanson, 1989; Wong, 1987, 1991). Specifically, students with learning disabilities experience difficulties analyzing task requirements, selecting and implementing strategies, and monitoring and adjusting performance (Butler, 1998a). Butler (1999) found 76% of post-secondary LD students to struggle with task analysis. In writing tasks, students with LD have been found to focus on lower-order processes, like spelling or grammar, while ignoring higher-order demands such as writing to an audience or organizing ideas (Wong, Butler, Ficzere, & Kuperis, 1996). Further, in mathematics, students with LD struggle with persistent skill deficits, which adversely affect their ability to decipher the demands of higher-level mathematics problems (Jones, Wilson, & Bhojwani, 1997). When compared with their typically achieving peers, LD students are in general less metacognitively aware and tend to focus on the concrete demands of tasks, rather than on the more obscure evaluative and self-awareness skills demanded by metacognitive processes (Butler, 1998a).

An understanding of task demands--what Borkowski (1992) calls the "sizing up" of a task--is essential for successful academic performance. Students with LD may mis-analyze tasks because of misconceptions held about the nature of the task, or because of a lack of awareness that analyzing tasks is an important step in learning (Butler, 1996). The level of effort and persistence expended by a learner is at least partly the result of awareness of task demands and personal capabilities (Butler & Winne, 1995). To sum up, students with learning disabilities struggle with various aspects of metacognition, one component of which may be assessment or evaluation of the nature of the task encountered.

Self-Efficacy Theory

Self-efficacy beliefs are context-specific evaluations of the ability to successfully complete a task. They are formed through mastery experiences, vicarious experiences (observation of others), social/verbal persuasion, and interpretations of physiological and emotional states (Bandura, 1995). These beliefs contribute to prediction of academic outcomes beyond the contributions offered by ability, previous attainments, knowledge and skill alone. Students need more than ability and skills to perform successfully; they also need a sense of efficacy to use their skills well and to regulate their learning (Bandura, 1993). Self-efficacy beliefs differ from related constructs such as competence beliefs and self-concept in that they are more task-specific, and are established through normative criteria rather than comparison with others (Zimmerman, 1995). Perceptions of efficacy play a part in managing motivation in expectancy-value theory, which asserts that individuals evaluate courses of behavior for their value or potential to produce certain outcomes. Shell, Murphy, and Bruning (1989) found that adding a self-efficacy component significantly increased the predictiveness of expectancy-value constructs. Self-efficacy beliefs, then, consist of the degree to which individuals believe they can control their level of performance and their environment in specific contexts (Bandura, 1997).

Calibration

Calibration addresses the accuracy of one's beliefs about potential functioning. In the measurement of self-efficacy for academic functioning, students might be asked to rate how confident they are that they can perform a certain academic task. For example, Graham and Harris (1989b) measured students' self-efficacy for writing essays using a 5-item, 10-point scale that asked about "perceived ability to write an essay with a `good' beginning, that gave three reasons to support the premise, and had a `good' ending" (p. 206). Students then responded on a scale ranging from "not sure" (10) through "maybe" (40), and "pretty sure" (70) to "real sure" (100), in 10-point increments.

The calibration of beliefs with performance is assessed by comparing mean efficacy ratings with task performance. In this example, the students displayed a low level of essay-writing performance, although measured self-efficacy beliefs showed a mean of close to 70, or "pretty sure" (i.e., "pretty sure" I can write a good essay). These assessments of self-efficacy can be viewed as a function of metacognitive knowledge, and are derived from the understandings students build about themselves (Butler, 1998a).

Optimistic estimates of one's efficacy are hypothesized to increase effort and persistence, and to promote accomplishment in challenging circumstances (Bandura, 1986, 1997). In an academic setting, optimistic efficacy beliefs are necessary for attempting novel tasks or learning new material. Evidence suggests that most typically achieving students are somewhat overconfident when asked to rate their academic abilities (e.g., Pajares & Kranzler, 1995; Pajares & Miller, 1994). Although moderate overconfidence is purported to promote achievement (Bandura, 1997), significant incongruence between efficacy beliefs and subsequent performance may not be so benign: Naive optimism or "gross miscalculation (between efficacy judgments and performance) can create problems" (Bandura, 1989, p. 1177).

One of the aims of this review was to examine how students with learning disabilities--who have been shown to display poor task analysis and metacognitive skills--calibrate their efficacy beliefs with criterial tasks.

METHODS

In this review of the literature on the self-efficacy beliefs of students with learning disabilities, the decision was made to restrict inclusion to empirical studies that specifically used the terms "self-efficacy" and "learning disabilities," resulting in the exclusion of studies that used comparable but differently named constructs. For example, the term "learning disability" is typically operationally defined in research through the presence of an IQ-achievement discrepancy. Studies investigating "underachievement" (e.g., Carr, Borkowski, & Maxwell, 1991) use a similar operational definition, but the two terms are conceptually different with regard to hypothesized etiology. In a similar fashion, ability self-concept (e.g., Wigfield, Eccles, MacIver, Reuman, & Midgley, 1991) is related to self-efficacy, but differs in its theoretical underpinnings. The present study was restricted to experimental and correlational research specifically investigating the perceived self-efficacy of students with learning disabilities in an educational context.

Search Procedure

To locate the pertinent literature, the following procedure was followed. First, the time period searched (1977-2000) was restricted to the period following Bandura's 1977 publication of Self-Efficacy: Toward a Unifying Theory of Behavioral Change, which was thought to signal the advent of self-efficacy research. Next, two on-line databases--Educational Resources Information Center (ERIC) and PsycINFO--were searched using the following key words found in the abstracts: learning disab* and self-efficacy.

The abstracts of the 28 "hits" from ERIC and the 37 from PsycINFO were scanned (there was some overlap between the two databases). Twenty-two studies were deemed to meet the stated criteria--the inclusion of students with learning disabilities and the use of a measure of self-efficacy. Of the 22 studies included in this review, 13 were published after 1990, whereas nine were published between 1985 and 1989.

Analysis

The 22 studies were summarized and analyzed in accordance with the following questions:

* What is the nature of the sample (participants) and what academic domains are included in the study?

* Are gender differences in efficacy beliefs addressed?

* What is the main research question in the study and what are the outcomes in terms of self-efficacy?

* Are the efficacy beliefs of students with learning disabilities compared with the beliefs of typically achieving or low-achieving students, and if so, what are the differences?

* Is calibration addressed, and if so, what conclusions are drawn?

RESULTS

The studies included in this review are listed and summarized in Table 1. The columns include the following: (a) author and date published; (b) number of students included in the study, with age and grade span; (c) domain and nature of performance task; (d) brief description of the self-efficacy measure used; (e) primary research question or intervention used; and (f) self-efficacy outcomes.

Participants and Domain

As can be seen in column 2 of Table 1, the number of participants in each study ranged from 3, in the case of Graham and Harris (1989b), to 336, in the study conducted by Gresham, Evans, and Elliott (1988). The mean sample size was 66.8. The youngest students were the 6- to 8-year-olds in the Omizo, Cubberly, and Cubberly (1985) study; four of the studies included students in a college or university setting. In terms of educational setting, nine studies included students in elementary schools, five examined the efficacy beliefs of students in high schools, four studies took place in college or university settings, and four included students from an age range spanning elementary and high school. Since research on reading disabilities dominates the LD field (Shaywitz, Fletcher, Holahan, & Shaywitz, 1992), it might have been predicted that research situated in the field of learning disabilities would have predominantly focused on the reading efficacy beliefs of students. This was not the case, however.

Column 3 displays the domain(s) investigated in each study. The performance domains ranged as follows: eight of the studies investigated writing as the area of interest; five examined efficacy for various math skills; one study examined reading; one study looked at career and vocational interests; and in seven studies the domain explored was either general academic functioning or "mixed" academic functioning. In this last category--mixed or general--some of the studies incorporated measures of global academic efficacy beliefs (Baum & Owen, 1988; Gresham et al., 1988; Hampton, 1998; Saracoglu, Minden, & Wilchesky, 1989). The two studies by Butler (1995, 1998b) investigated self-efficacy in a variety of academic domains, depending on student need. In the Slemon and Shafrir (1997) article, students estimated their efficacy for performance on standardized intelligence and academic achievement tests.

Gender

With non-LD populations, significant differences in self-efficacy beliefs due to gender have been found in a variety of academic topics (e.g., Pajares & Johnson, 1996; Pintrich & DeGroot, 1990). Gender as a variable was included in a minority (5 of 22) of the studies in this review. In some cases, the low proportion of girls in the sample of children with learning disabilities made it difficult to perform meaningful statistical analyses of gender differences. Gresham et al. (1988) found no significant gender interaction in their multivariate analysis of variance. It should be noted that although their sample of 336 students included some students with learning disabilities, the majority (287) were either gifted or "nonhandicapped." Furthermore, the LD students were included in a mixed group of "mildly handicapped" students, which consisted of unspecified numbers having learning disabilities, mild mental handicaps, and behavior disorders. Panagos and DuBois (1999) included gender as a variable in their multiple regression investigating career interest, and found no significant contribution made by gender to career interest. They did not, however, investigate any gender differences in the students' perceived self-efficacy. In the study of the social and academic self-efficacy beliefs of university students with LD, Saracoglu et al. (1989) found that female students reported lower self-esteem than males, but not significantly lower self-efficacy beliefs. In this study, post-secondary students were asked to rate their efficacy beliefs on a scale that measured their "general and social self-efficacy expectancies" (p. 591). As will be discussed below, measurements of global self-efficacy have been criticized: "It is no more informative to speak of self-efficacy in global terms than to speak of nonspecific social behavior" (Bandura, 1986, p. 411). Schunk (1985) included sex as a variable in his study on goal setting in arithmetic, but found no significant differences or interactions involving sex of child. Likewise, Schunk and Cox (1986) found no significant effects for sex in their study on strategy training and attributional feedback.

Interventions and Outcomes

Most of the studies (13 of 22) were designed to assess the effectiveness of various interventions in a wide variety of settings and domains. Column 5 summarizes the intervention being investigated, while column 6 examines the outcomes in terms of changes in self-efficacy beliefs. In general, these studies found that increases in performance skills following the intervention were mirrored by increases in measured self-efficacy beliefs. However, there were some anomalous findings and exceptions to this generalization.

In Butler's (1998b) study of her intervention method--strategic content learning--task-specific self-efficacy beliefs increased along with academic skills, but global efficacy beliefs did not significantly improve. Graham, MacArthur, Schwartz, and Page-Voth (1992) found efficacy beliefs to decrease after the treatment in three of four cases. The authors remarked that pretest efficacy beliefs were overestimated and became more realistic after treatment. In the study investigating goal setting and strategy use in writing, Page-Voth and Graham (1999) found no significant increases in self-efficacy beliefs as a result of intervention. Finally, Wong, Butler, Ficzere, and Kuperis (1997) did not note increased self-efficacy beliefs in conjunction with their strategy intervention for writing compare-and-contrast essays; however, self-efficacy did increase along with performance in their previous study (Wong et al., 1996) involving the writing of opinion essays.

Students with LD Compared to NA (Normally Achieving) Students

Of the studies (9 of 22) not investigating a specific academic intervention, several were designed to explore differences in efficacy beliefs between LD and typically achieving populations. None of the studies reviewed used low-achieving students as control groups. Column 2 lists the numbers of LD students included in each study and makes note of the studies that included comparisons using control groups of normally achieving (NA) students.

Baum and Owen (1988) found that the efficacy beliefs of high-ability/LD students were lower than those of either the LD/average or high-ability (non-LD) groups. In answer to their research question, "Why do bright, learning disabled students have such a poor sense of self-efficacy when they posses greater intellectual and creative potential?" (p. 325), they suggested that bright LD students do not view their classroom accomplishments as meaningful or meeting their own elevated internal standards. In their comparison of LD and typically achieving students in the domain of writing, Graham, Schwartz, and MacArthur (1993) found no differences in self-efficacy between the two groups, even though the LD students were reported to struggle with writing tasks. The authors concluded that more research is needed to explore why LD students tend to miscalibrate and overestimate their writing abilities. (Overestimation of efficacy beliefs can be defined as either average task performance coupled with overly optimistic self-efficacy ratings, or average efficacy ratings coupled with low task performance. In the Graham et al. [1993] study, the self-efficacy ratings of the two groups were not significantly different, but the level of performance was much lower for the LD students.)

Gresham et al. (1988) reported that the mainstreamed MH (mildly handicapped--including children with LD) students displayed lower levels of academic and social self-efficacy than did the nonhandicapped and gifted students. As mentioned, the results of this study need to be viewed in the light of the nature of the sample of the MH students: The proportion of LD students in the MH group was not known to the researchers, which makes generalization to specific groups difficult. Hampton's (1998) investigation of the sources of LD found that LD adolescents rated each of the four sources of academic self-efficacy--past performance, vicarious learning, social persuasion, and physical arousal--lower than did the non-LD control group. In the second section of this article, the results from a regression analysis showed that for LD students only past performance and vicarious learning contributed significantly to the predictive equation (non-LD students were not included in this analysis). Pintrich, Anderman, and Klobucar (1994) compared students with and without learning disabilities on a number of motivation and cognitive variables. Although self-efficacy levels were not significantly different between the two groups, metacognition and reading comprehension were lower for the LD group. In other words, whereas both groups felt about the same in terms of reading confidence, the LD group performed at a significantly lower level on the reading task. University students with and without LD were compared in the study conducted by Saracoglu et al. (1989). On the global self-efficacy scale used, no differences were found between the two groups in either of general or social self-efficacy.

Slemon and Shafrir (1997) included LD and typically achieving college students in their study of calibration of efficacy and performance and compared the two groups for accuracy of prediction (calibration) of standardized achievement (WRAT-R) and intelligence test (WAIS-R) subtest scores. The authors found the LD students to underestimate, overestimate, and accurately estimate their performance on the various subtests, but to generally show more accuracy or less optimism than the typically achieving students. For example, on the achievement test, the WRAT-R, the LD students accurately estimated their reading score, underestimated their spelling score, and overestimated their arithmetic score. The NA also showed variation in the accuracy of their estimation, but displayed a different pattern. These students overestimated their reading and accurately estimated spelling and arithmetic. Little difference between the two groups was seen on the Performance subtests of the WAIS-R. However, on the Verbal subtests, the LD group estimated three subtests accurately, underestimated one subtest, and overestimated another. By comparison, the NA group overestimated four subtests on the Verbal Scale, and accurately estimated the remaining subtest. The authors concluded that "although students with LD are motivated ... they nonetheless tend to lack the optimistic beliefs about ability of the NA students" (p. 11).

Calibration

In eight of the studies reviewed, the authors addressed the apparent miscalibration of efficacy beliefs with the subsequent performance tasks (see Table 2 for scale used, degree of calibration, and self-efficacy ratings). In the eight studies addressing accuracy of self-beliefs, the authors themselves determined whether or not the self-beliefs measured were underestimates, generally accurate, reasonably optimistic, or serious overestimates. It is not always easy to label self-efficacy beliefs as "accurate" or as "overestimates"; in most of the studies (for example, see the above description of Graham & Harris, 1989b), the self-efficacy measures were not directly connected with the criterial task. As an example, a finding of a mean self-efficacy rating of 65 on a 10- to 100-point efficacy scale might be difficult to interpret in terms of accuracy of perceived efficacy for writing essays. In a few of the studies (e.g., Alvarez & Adelman, 1986; Schunk, 1985; Schunk & Cox, 1986), the domain investigated--arithmetic --and the methodology used made the assessment of the accuracy of the judgments quite clear. In the domain of arithmetic, it is possible to briefly display the actual task to the student for purposes of efficacy measurement, and then follow up with the same task used as a performance measure. Of the eight studies in which overestimates of efficacy beliefs are discussed, five investigated the domain of writing, one explored efficacy for reading, one examined arithmetic, and the final study looked at all three of these academic areas plus cognitive functioning in the context of a standardized achievement and intelligence tests.

Alvarez and Adelman (1986) investigated the nature of miscalibrations in students with "psychoeducational problems" (including students with LD and dyslexia). The researchers showed students a page of pairs of increasingly difficult math items and asked them to judge their confidence to complete each pair of questions. After asking the students for efficacy judgments, the researchers asked them to complete as many of the items as they could. Alvarez and Adelman found that 30% of judgments to complete specific math items were overestimates, with just 2% underestimating ability to complete the item (Table 2); the remainder (68%) of the judgments were reasonably accurate. Perhaps most interesting, the authors noted that the students apparently were capable of accurately judging their efficacy when the task items were judged as either obviously easy or obviously difficult. Students typically miscalibrated on items that were viewed as potentially being within their capability. This phenomenon of overstatement of efficacy for realistically challenging tasks was attributed to a "self-protective" function, in which students erected a "facade of competence" to hide their academic difficulties.

Graham and Harris (1989a), investigating self-instructional strategy training for LD students with writing deficiencies, found that "LD students consistently overestimated their composition abilities" (p. 360). They postulated that the unrealistic pretask expectancies of the students in their study might be due to "comprehension deficiencies, use of a self-protective coping strategy, or a developmental delay in the ability to match task demands to ability level" (p. 360). In their case study of three LD students exposed to strategy training, Graham and Harris (1989b) again found that the students consistently overestimated their writing skills. In another study, Graham et al. (1992) noted that students with learning disabilities overestimated self-efficacy for writing at pretest, but displayed more realistic confidence beliefs after goal-setting instruction. The authors had predicted that the students' efficacy beliefs would become more realistic because "the process of goal setting has been shown to facilitate self-evaluation" (p. 332). In other words, the students' self-knowledge was enhanced by the intervention, resulting in more accurate self-appraisals. Again in the domain of writing, Graham et al. (1993) and Sawyer et al. (1992) found students with learning disabilities to express relatively high confidence for composition tasks despite poor performance on pretest writing measures.

In the only study reviewed that specifically examined perceived efficacy for reading tasks, Pintrich et al. (1994) found that LD students expressed levels of reading confidence equal to those of the more capable readers without learning disabilities. That is, the LD students averaged a rating of 5.37 on a 7-point scale, whereas the non-LD students reached a mean of 5.71 on the same scale--a nonsignificant difference. In light of the LD students' documented disability in reading, the authors commented that these students felt "rather efficacious" at reading; that is, the students with LD expressed about equal confidence for reading, even though their skill level was considerably lower.

The final study discussed in this section (Slemon & Shafrir, 1997) investigated efficacy beliefs for performance on the Reading, Spelling, and Arithmetic subtests of a standardized test (the WRAT-R) and for test performance on an intelligence test (the WAIS-R). Both the LD and NA groups showed overestimates, underestimates, and accurate estimates for achievement as well as intelligence test performance. Differences were most clearly seen on predictions of the Verbal Scale of the WAIS-R, where the LD students were generally accurate (three of five subtests) and the NA students generally four of five subtests) overestimated their level of performance. The authors interpreted the findings as showing that LD students were lacking in the optimistic beliefs shown by the NA group.

One study included comments on the relative accuracy of the LD students' self-evaluations. Although the majority of the writing studies showed that LD students overestimated self-efficacy beliefs, Page-Voth and Graham (1999) found pretest self-efficacy estimates "generally neutral," with mean pretest scores just above 3 on a 5-point Likert-type scale. Other studies appeared to show results of "neutral" efficacy beliefs, although the authors did not comment on the accuracy of calibration. The two studies by Wong et al. (1996, 1997) found both pretest and posttest efficacy measures at about 30 out of a possible 50; in other words, a mean efficacy response of about 3 on a 5-point Likert scale. The studies investigating goal setting, strategy training and attributional feedback in arithmetic (Schunk, 1985; Schunk & Cox, 1986) used efficacy measurement techniques that were similar to each other. Students first received training in using the efficacy scale by estimating their capabilities to jump progressively longer distances. After teaching the students the properties of the scale, the researchers showed them 25 pairs of arithmetic questions for about 2 seconds each. Students estimated their capability to solve each of the pairs of problems. Efficacy judgments in this case were moderate, with pretest self-efficacy means about 50 (on a 10-100 scale with 10-unit intervals) in each of the studies. When this method of efficacy measurement is used--estimating confidence to perform the actual task being shown--increased accuracy appears to be the result in some cases. Similar methodology was used in the Slemon and Shafrir study (1997), in which some predictive beliefs were found to be accurate as well.

DISCUSSION

As shown in the preceding section, for the most part, self-efficacy ratings were predictive of subsequent functioning and increase in conjunction with intervention and subsequent performance increases. Some of the areas explored--gender differences and self-efficacy differences between LD and non-LD students--offered few obvious trends, while others, such as self-efficacy in the writing domain, offered more promising findings. In the following section, some of the most clear findings are discussed, with attention paid to two central questions pertaining to calibration and problems with self-efficacy measurement.

Do Students with Learning Disabilities Miscalibrate Their Efficacy Beliefs?

In 8 of 22 studies, the authors commented on students' miscalibrations of efficacy beliefs with the criterial task. For the most part, the miscalibrations were errors of overestimation, with only one study (Slemon & Shafrir, 1997) finding LD students underestimating their self-efficacy for spelling and a short-term memory task, although they overestimated their efficacy for arithmetic skills and a long-term memory task. Authors of the remaining 14 studies either did not comment on the accuracy or calibration of students' self-efficacy beliefs or found them "neutral."

Investigations into the domain of writing showed the most consistent problems with calibration: Five of the eight studies of the writing of students with LD revealed that the students were overly optimistic about their writing abilities, even though the students had previously identified writing disabilities. Of the remaining three studies that did not find overestimates of efficacy, only one (Page-Voth & Graham, 1999) included students showing a previously identified specific weakness in written language. The other two studies, conducted by Wong and her colleagues (1996, 1997), consisted of a mixed group of low achievers and students identified as having learning disabilities because of low reading scores.

In other words, in the studies investigating the writing skills of students identified as having LD in writing, five of six studies showed these students to overestimate their writing capabilities. Similarly, in the lone study investigating the reading efficacy of students specifically identified as reading disabled (Pintrich et al., 1994), the ratings of self-efficacy overestimated actual level of performance. These findings support Graham and Harris' (1989b) statement that there is "a growing body of literature that indicates that learning disabled students have difficulty accurately assessing or predicting their performance capabilities" (p. 212).

What Is the Evidence That Students with LD Accurately Calibrate Their Efficacy Beliefs?

There is some evidence that in the area of mathematics functioning students are generally accurate in their calibrations of efficacy and performance. Of the five studies specifically targeting math and math-related self-efficacy, only one (Alvarez & Adelman, 1986) commented on evidence of overestimates of efficacy beliefs, but even in that study 68% of students were generally accurate in their efficacy estimates. It might be concluded, then, that students with learning difficulties are somewhat accurate in estimating their ability to complete specific math exercises.

One factor contributing to this accuracy might be the methods used to measure efficacy beliefs. In math, it is possible (e.g., Schunk, 1985; Schunk & Cox, 1986) to briefly show students the actual criterial task, have them rate their efficacy to complete that task, and then have them attempt the task. Thus, the task analysis component of self-efficacy beliefs is much clearer for students contemplating a specific arithmetic task in comparison to a writing task.

Why Might Some Students with Learning Disabilities Overestimate Their Efficacy?

The results from this review suggest that in some cases, students with learning disabilities overestimate their capabilities to perform certain tasks. The findings are somewhat tentative and may not generalize to all subject matters. But in some circumstances--particularly in the writing performance of students with previously identified writing difficulties--students with learning disabilities appear to be unrealistically optimistic about their capabilities. A number of possible factors underpin the finding of optimistic efficacy beliefs in certain students with LD. For example, estimations of self-efficacy can be construed as a form of metacognition, and students with learning disabilities have been found to display significant metacognitive deficiencies (e.g., Butler, 1999; Wong, 1985, 1986).

In this review, the study by Pintrich et al. (1994) found LD students to possess significantly lower metacognitive skills than students without LD. Bandura and Schunk (1981) proposed that discrepancies between perceived efficacy and performance may be the result of task misunderstanding and deficiencies in self-evaluation. Writing is a complex task, and self-assessment or reflection about challenging tasks like writing is a difficult process for many children with LD. Wong, Wong, and Blenkinsop (1989) found students with LD to focus on lower-order writing processes whereas typically achieving students focused on higher-order processes. In this review, Graham et al. (1993) suggested that students with LD possess a less mature understanding of the process of writing composition than students without learning disabilities. The question "How sure are you that you can write a good descriptive paragraph?" poses a challenging metacognitive task for the student with a specific learning disability. That is, in a short period of time, the student needs to process the question, define and assess task demands, evaluate his or her capabilities to carry out the task, perhaps consider Bandura's four sources of efficacy beliefs--past experience, vicarious experience, social persuasion, and physiological state--and finally come up with a numerical representation for the complete operation. All of this takes place while the researcher--probably a stranger--waits, tapping a pencil!

What Is the Impact of Miscalibration of Efficacy Beliefs?

In this review we have seen several examples of inflated efficacy beliefs of students with learning disabilities. Pajares (1996) posed the question: "But how much confidence is too much confidence [and] when can overconfidence be described as excessive and maladaptive in an academic enterprise?" (p. 565). Social cognitive theory holds that optimistic self-efficacy beliefs typically foster increased perseverance and effort; however, in this review, it is suggested that for students with learning problems, positive efficacy beliefs--especially in the face of specific academic weaknesses--might not operate in the same way as for typically achieving students. Bandura (1997) warned that "Deficient information leads to poor academic preparation" (p. 65). If self-evaluation can be viewed as a reflection of self-knowledge or as metacognition, then gross misjudgments about one's efficacy can be seen as misleading and potentially academically harmful. Butler (1998a) stressed the importance for LD students of constructing accurate metacognitive understandings and mindfully reflecting about learning processes. To a certain extent, cognitive strategy use depends on accurate self-awareness or self-knowledge. Deficient self-knowledge, by contrast, may result in appropriate strategies not being used, faulty task understanding, and difficulties with self-regulating and monitoring one's progress. Thus, the overly optimistic self-efficacy beliefs seen in LD students in some situations may, in fact, result in inferior academic functioning.

For those working with students with learning disabilities, miscalibration of efficacy beliefs may be seen as a naive overconfidence for writing tasks. This overconfidence might be considered a warning sign of faulty task analysis or poor self-knowledge. Pajares (1996) suggested that the teacher's response should focus on improving the student's calibration skills through improved task understanding, rather than focusing on lowering the student's efficacy beliefs. The findings from this review suggest that the domain of writing is particularly problematic for students with LD: Training that develops both cognitive and metacognitive knowledge (Butler, 1998a) about writing tasks may lead to better understanding of the writing process, and eventually, to better writing performance.

What Are Some of the Methodological Problems Found in These Studies?

Conceptual blurring. The problem of "conceptual blurring" was seen in several of the studies. Some of the measures described strayed considerably from the basic definition of self-efficacy provided by social cognitive theory: A number of individual items purporting to measure self-efficacy in fact were measures of other related self-beliefs. For example, one of the studies included items such as "I like to write" and "I do writing on my own outside of school" as part of, the measure of self-efficacy. Another efficacy measure included having the students rate their ability as "very below average" to "very above average" in a specific subject. Yet another self-efficacy measure asked students to use peer comparisons as the basis of their self-efficacy measure ("1 = much worse than people my age, 10 = as well as people my age, 19 = much better than people my age"). Several of the studies included items such as "When my class is asked to write a report (or essay, or story), mine is one of the best."

The above items are perhaps meaningful, potentially useful measures of students' beliefs about their capabilities, but they are not theory-grounded measures of self-efficacy. Thus, Pajares (1996) suggested researchers take "methodological precautions" when assessing unfamiliar motivation constructs, and to operationalize and use measures "in a manner consistent with the construct's theoretical home" (p. 570). In other words, measures claiming to evaluate perceptions of self-efficacy need to be carefully constructed, and based on a careful reading of the relevant theoretical background.

Self-efficacy beliefs are perceptions of the ability to carry out certain types of performance and to achieve certain results: "How sure (or confident) are you that you can read a Dr. Seuss story out loud?" Efficacy beliefs may be related to perceptions of competence ("I'm a good reader") and academic self-concept ("I feel that I'm pretty good at reading," or I'm a better reader than others in my class"), but they are conceptually distinct and are supported by different theoretical underpinnings. Suggestions have been made to ensure that the self-beliefs termed self-efficacy beliefs are, in fact, self-efficacy beliefs. For example, Bandura (1997) and Pajares (1997) suggested that items be "phrased in terms of can do rather than will do. Can is a judgment of capability; will is a statement of intention" (p. 43). Klassen (2002) proposed including the term "confidence" or another closely related term in the measure (i.e., "How confident are you that you can ...?" or "How sure are you ...?"). Conceptual blurring of self-beliefs results in unclear findings that do not further understanding of the mechanisms of efficacy beliefs in the context of social cognitive theory.

Global or generalized measures. It is important to strike the correct balance in deciding at what level of detail and in what way performance tasks should be represented on the efficacy measure. Bandura (1997) cautioned against using efficacy measurements of decontextualized minute subskills to represent broader performance functioning. Conversely, Bandura also warned that measurement of an overly broad domain, like efficacy for general academic functioning, results in poor prediction because such a domain consists of distinct subdomains that may vary markedly within an individual. Many of the studies in this review included efficacy measures that gauged important subskills of discrete academic domains like writing or arithmetic. For example, the writing measure used by Graham and Harris (1989a) showed reasonable specificity (level of precision or detail of the subtask measured) and appropriate correspondence (congruence between measure and performance task). In the Graham and Harris study, the authors asked 10 questions probing efficacy beliefs to write a story. The items were preceded with the stem "Can you write a story that," and included "(a) tells about the main character's feelings? (b) clearly tells about the setting? (c) has a good beginning?" and so on (p. 356).

However, other studies included measures that assessed self-efficacy in much broader terms, and that might be considered decontextualized, global measures of self-efficacy. Pajares (1996) warned "Generalized self-efficacy instruments assess people's general confidence that they can succeed at tasks and in situations without specifying what these tasks or situations are" (p. 547). One of the reasons for the problem of decontextualized self-efficacy measures is that some studies are designed to assess self-efficacy beliefs as if they were fixed personality traits, rather than context-based perceptions of abilities to perform certain tasks. For example, several of the studies reviewed (Baum & Owen, 1988; Gresham et al., 1988; Hampton, 1998; Saracoglu et al., 1989) measured general academic and social self-efficacy without reference to a specific domain of functioning. Although it could be argued that the domain concerned is "academic functioning," it remains unclear that people possess firmly rooted self-efficacy beliefs about domains that are made up of numerous distinct parts--individual subjects or courses in the case of academic functioning. The difficulty with measuring global domains of functioning is that respondents cannot assess their efficacy with a particular activity in mind. Students with specific learning disabilities, especially, may view their various school-related capabilities in quite contrasting ways: "Well, I'm not at all confident I can read very well, but I'm pretty certain I can do difficult math questions," and the measurement of overall academic self-beliefs may not uncover this range of perceptions. Certainly, the strong predictive power of self-efficacy, as compared to other more global self-beliefs, is based on the very fact of its being placed in a meaningful context.

Implications for Practice

For teachers, it is one thing to work at boosting the beliefs of students who are pessimistic about their self-efficacy (see Pajares & Johnson, 1996), but dealing with students' excessive overconfidence in certain domains may not be so easily managed. Findings from this review suggest that some students with learning disabilities overestimate their efficacy to complete writing tasks. For students with LD, optimistic efficacy beliefs may not increase effort and persistence, but mask strategy and skill deficits. If these miscalibrated beliefs are based on inadequate self-awareness (Butler, 1999) or deficient information (Bandura, 1997), students' approaches to some writing tasks may prove inadequate. Approaches that foster the self-awareness and self-regulation functions associated with metacognition (e.g., Butler, 1998; Meltzer et al., 1998) may improve the accuracy of students' calibration and subsequent performance. Attempting to lower overestimates of efficacy is discouraged (Pajares, 1996), but promoting academic self-awareness may result in a more realistic self-appraisal.

Self-efficacy scales may prove to be of use to teachers of students with learning disabilities. For example, in the classroom, the inclusion of a writing self-efficacy measure might form part of a prewriting exercise. An assessment of efficacy beliefs would provide the teacher with insight into the perceptions held by students about their abilities, and help identify miscalibrations. Self-efficacy measures for specific tasks can be brief and informative. A five-item measure using the stem "How confident (or sure) are you that you can ...?" (create an outline, use correct punctuation, write a story, etc.) with a 7-point response scale (as suggested by Bandura, 2001) could be completed by most students in a minute or two. Teachers would gain information that might help identify those students whose overconfidence is interfering with the development of appropriate skills and strategies.

Suggestions for Future Research

One important finding from this review is that in some domains, especially writing, students with learning disabilities appear to overestimate their capabilities. But the research on the self-efficacy beliefs of LD students in other domains is inconclusive and somewhat contradictory. The idea of self-evaluation as a form of metacognition has been discussed, and LD students' difficulties with metacognitive tasks have been documented. Accurate measurement of self-efficacy beliefs in students with LD is more complex than indicated by previous practice. One of the studies included in this review (Alvarez & Adelman, 1986) issued this warning: "Findings of self-evaluative overstatement have serious implications for both researchers and practitioners. Self-evaluative overstatements obviously can confound efforts to study a variety of phenomena" (p. 570).

Self-efficacy research and measurement with LD populations needs to be conducted in different ways than has been the case so far. A few researchers have proposed expanding measurement techniques used in investigating self-efficacy (e.g., Hampton, 1998; Pajares, 1996). Pintrich and DeGroot (1990), for example, suggested that results of self-report instruments "need to be replicated with other measures, such as think-aloud protocols, stimulated recall procedures, [or] structured interviews" (p. 38). Using qualitative methodology to validate quantitative self-efficacy measures seems called for in the measurement of self-efficacy beliefs. It may be the case that LD students' metacognitive and self-evaluative deficits can be circumvented through in-depth qualitative assessment of self-efficacy beliefs. Clearly, more research efforts are needed to explore and validate methods with which to uncover accurate self-beliefs of students with learning disabilities.

Only 1 of the 22 studies reviewed examined self-efficacy beliefs in reading. This finding is unexpected in light of the preponderance of reading disabilities among all learning disabilities (Shaywitz et al., 1992), and in light of the fact that reading disabilities and dyslexia are the subject of much research. The one study (Pintrich et al., 1994) that did examine the efficacy beliefs of students with LD included a specifically reading disabled sample, a measure clearly assessing efficacy beliefs and appropriate levels of specificity and correspondence between measure and criterial task. Findings of miscalibrated efficacy beliefs for reading bear replication and extension.

Gender differences in self-efficacy beliefs were seen in some studies, but most did not assess gender differences in efficacy beliefs. More research on gender differences in self-efficacy beliefs in LD populations would further our understanding of how these self-beliefs may operate in different contexts. Also, developmental differences or changes in efficacy beliefs with LD students were not addressed in the studies reviewed here. Longitudinal or cross-sectional studies investigating changes in efficacy beliefs would help further understanding of the development of self-efficacy in special populations.

None of the reviewed studies included control groups made up of low-achieving students. Instead, when a control group was used, studies relied on the measured perceptions of typically achieving students. It may be that low-achieving children not labeled as "learning disabled" display similar self-efficacy profiles, but no empirical evidence was found to support this supposition. As suggested by Grolnick and Ryan (1990), "many of the motivational and self-evaluative problems that children with LD have may be nonspecific (i.e., they may be apparent in other children who have difficulties in learning)" (p. 183). Future research should investigate efficacy beliefs of low-achieving as well as LD children to determine if students with learning disabilities form self-beliefs in ways that are significantly different from other academically struggling students.

Finally, to ensure the content validity of the self-efficacy measures used, researchers are directed to Bandura's most recent (2001) suggestions for constructing self-efficacy scales. In this thorough guide, the author addresses issues of domain specificity, content validity, phrasing of items, and most effective response scales.

Table 1
Summary of Reviewed Studies

Author         # of stds           Performance
(date)         age/grade           task or domain

Alvarez        19 stds (some       Arithmetic
& Adelman      LD) aged 9.6
(1986)         to 15.2

Baum           112 grade 4,        General academic
& Owen         5, and 6            functioning (no
(1988)                             performance task
                                   given)

Bryan          18 junior high      (Math) 50 addition
& Bryan        and high            and subtraction
(1991)         school              questions

Butler         6 college stu-      Student-chosen
(1995)         dents with LD       task--writing,
                                   reading, math--
                                   depending on need

Butler         30 (over three      Student-chosen
(1998b)        studies) college    task--writing,
               and university      reading, math--
               stds with LD        depending on
                                   need

Graham         22 LD and 11        Writing
& Harris       NA (a) (con-
(1989a)        trol); grades
               5-6

Graham         3 sixth-grade       Writing
& Harris       stds
(1989b)

Graham,        4 fifth-grade LD    Writing
MacArthur,     stds
Schwartz, &
Page-Voth
(1992)

Graham,        39 LD and           Writing--no
Schwartz, &    29 control in       performance task
MacArthur      grades 4, 5,
(1993)         7 & 8

Gresham,       336 stds incl.      Social and aca-
Evans,         mildly handi-       demic functioning
& Elliott      capped (incl.       (no performance
(1988)         LD), gifted, and    task)
               NA in grds 3-5

Hampton        109 high            Academic func-
(1998)         school and          tioning
               vocational
               rehab stds. with
               LD; 87 people
               without LD

Omizo,         60 6- to 8-year-    Arithmetic
Cubberly,      old LD stds--       achievement
& Cubberly     20 in each of
(1985)         three groups

Page-Voth      30 grade 7 and      Writing (essays)
& Graham       8 stds
(1999)

Panagos        96 high school      Career interest:
& DuBois       stds                A career interest
(1999)                             inventory
                                   (14 areas) was
                                   administered

Pintrich,      19 LD, 20 NA        Reading: Two
Anderman,      grade 5 stu-        reading compre-
& Klobucar     dents               hension tasks
(1994)                             were given

Saracoglu,     34 LD and 31        General and social
Minden,        NA university       self-efficacy
& Wilchesky    students
(1989)

Sawyer,        33 5th- and         10-item scale
Graham,        6th-grade LD        measuring SE for
& Harris       stds; 10 LD         writing a "made-
(1992)         control             up story"

Schunk         30 6th-grade        Stds were briefly
(1985)         LD stds             shown 25 pairs of
                                   subtraction q's
                                   and asked to rate
                                   on 10-point scale

Schunk         90 grade 6-8        Stds were briefly
& Cox          LD stds             shown 25 pairs of
(1986)                             subtraction q's
                                   and asked to rate
                                   on 10-point scale

Slemon         92 LD and           Students estimated
& Shafrir      40 NA college       their score (1-19)
(1997)         stds                on the WAIS-R
                                   (9 subtests) and
                                   the WRAT-R
                                   (3 subtests)

Wong,          38 LD and           Questionnaire on
Butler,        low-achiev.         attitudes towards
Ficzere,       gr. 8 & 9           writing and SE
& Kuperis
(1996)

Wong,          21 LD and           10-item, 5-point
Butler,        LA stds in          scale
Ficzere, &     gr. 9 & 10
Kuperis
(1997)

Author         Self-efficacy          Intervention or
(date)         measure                research question

Alvarez        20-item, 11-point      Why do students with
& Adelman      scale (0 to 10)        learning problems
(1986)                                (including LD) overstate
                                      their capabilities?

Baum           SEAT--34 items         What are self-efficacy
& Owen         measuring general      differences between
(1988)         academic self-         high- and avg.-ability
               efficacy               LD students?

Bryan          Estimated number       Positive mood induc-
& Bryan        (/50) of accurately    tion: thinking of
(1991)         completed arith-       happiest day of
               metic items in         their lives
               5 minutes

Butler         16-item SE ques-       SCL: Std-generated
(1995)         tionnaire; also        strategies guided by
               1 item asking stds     instructor
               to rate their
               ability on task

Butler         One item rating        SCL: Std-generated
(1998b)        task ability; mea-     strategies guided by
               sure judging task      instructor
               competence, task
               preference, and
               general SE; 8-item
               "across-tasks" SE

Graham         10 items measuring     Self-instructional strat-
& Harris       stds' confidence to    egy training; added self-
(1989a)        write stories          reg. training

Graham         Five items             Strategy instruction for
& Harris       assessing per-         planning and writing
(1989b)        ceived ability to      essays
               write a good essay

Graham,        10-item, 5-pt.         Planning and writing
MacArthur,     scale measuring        strategies
Schwartz, &    SE for writing
Page-Voth      tasks and
(1992)         cognitive strats

Graham,        10 items measuring     Study measured stds'
Schwartz, &    efficacy for           knowledge and attitudes
MacArthur      composing process      --no intervention
(1993)         and writing tasks

Gresham,       Group-administered     Purpose: to explore SE
Evans,         28-item, 5-point       beliefs in MH (incl. LD)
& Elliott      scale assessing        gifted, and NA students
(1988)         academic and social
               efficacy (ASSESS)

Hampton        Sources of Academic    To devise an instrument
(1998)         Self-Efficacy Scale    to determine the sources
               (SASES)--46 items      of SE beliefs and to ex-
                                      plore differences
                                      between LD and NA
                                      students

Omizo,         20-item scale mod-     Three groups: control,
Cubberly,      eled after Bandura     teacher- and participant-
& Cubberly     & Schunk (1981)        modeling
(1985)

Page-Voth      6-item scale           Goal-setting; goal-setting
& Graham       measuring efficacy     and strategy instruction;
(1999)         to write essays        control group

Panagos        14-item (1 for each    How are career SE beliefs
& DuBois       career area) career    linked with vocational
(1999)         self-efficacy          interests? Also, what is
               scale; four-item,      the role of the four
               5-point sources of     sources of SE beliefs?
               efficacy informa-
               tion scale

Pintrich,      10-item, 7-point       How does SE for reading
Anderman,      scale measuring        comp. differ between
& Klobucar     reading efficacy       groups?
(1994)         beliefs

Saracoglu,     23-item self-          Do SE beliefs differ for
Minden,        efficacy scale         LD and NA univ. stds?
& Wilchesky
(1989)

Sawyer,        Writing a story        Three types of self-regu-
Graham,                               lated strategy develop-
& Harris                              ment plus control group
(1992)

Schunk         Mathematics: sub-      Goal-setting: self-set
(1985)         traction               goals, assigned goals,
                                      and no goals

Schunk         Mathematics: sub-      Verbalization and effort
& Cox          traction               feedback
(1986)

Slemon         Verbal and nonver-     What are the SE beliefs
& Shafrir      bal cognitive          (predicted scores) for
(1997)         functioning on         LD and NA postsec-
               WAIS-R and 3           ondary students?
               achievement areas      (no intervention)
               on the WRAT-R

Wong,          Writing: opinion       Planning, drafting, and
Butler,        essays                 revising strategies
Ficzere,
& Kuperis
(1996)

Wong,          Writing: compare-      Writing strategies
Butler,        and-contrast essays
Ficzere, &
Kuperis
(1997)

Author
(date)         Outcomes

Alvarez        "Students' positive self-evaluations
& Adelman      represent a selective tendency and are
(1986)         not due to an inability to make accurate
               self-evaluative judgments."

Baum           High-ability LD students displayed
& Owen         lower efficacy beliefs than did high-
(1988)         ability non-LD, or avg.-ability LD

Bryan          Positive affect increased SE and
& Bryan        performance. However, SE was not
(1991)         changed with a control group of
               younger, non-LD ("at-risk") stds

Butler         The stated components of the SE measure
(1995)         --perceptions of competence, task
               preference, rating of on-task ability
               --all showed significant increase

Butler         Task-specific SE increased in all studies;
(1998b)        global SE not changed in posttest;
               SE for "non-instructed" tasks increased
               in one study, but not in others

Graham         SE increased in both treatment groups
& Harris       (strategy training with/without self-
(1989a)        reg. training). No difference between
               treatment groups

Graham         Two of three stds showed SE increases
& Harris       with intervention
(1989b)

Graham,        Confidence for writing dropped for
MacArthur,     three stds (all male) and rose slightly
Schwartz, &    for one std (female) after treatment
Page-Voth
(1992)

Graham,        No difference found in SE for either
Schwartz, &    composing process or writing tasks
MacArthur      between LD and NA groups or
(1993)         between older and younger students

Gresham,       MH stds (LD, MR, and BD) reported
Evans,         lower academic and social SE than NA
& Elliott      and gifted stds. Also, MH stds were
(1988)         reported by teachers as lower in
               academic and social SE

Hampton        LD stds rated each of four sources
(1998)         lower than NA stds. Also, Social
               Persuasion and Physical Arousal did
               not significantly contribute to the
               regression equation for LD students;
               that is, only Past Performance and
               Vicarious Learning were significant

Omizo,         Both conditions resulted in sig.
Cubberly,      increased SE beliefs; participant
& Cubberly     modeling sig. raised SE beliefs over teacher
(1985)         modeling

Page-Voth      No changes in self-efficacy beliefs in
& Graham       any of three groups
(1999)

Panagos        Ratings of SE beliefs were a significant
& DuBois       predictor of career interest. Also,
(1999)         Bandura's four sources of efficacy
               beliefs contributed to the develop-
               ment of career SE beliefs

Pintrich,      LD stds did not show sig. different SE
Anderman,      beliefs compared to the NA stds, in
& Klobucar     spite of lower performance levels
(1994)

Saracoglu,     LD and NA stds showed no diff. in
Minden,        social and general SE. SE correlated
& Wilchesky    positively with adjustment to
(1989)         university

Sawyer,        SE levels increased in all four groups:
Graham,        Posttest SE levels did not differ among
& Harris       the four intervention groups
(1992)

Schunk         Participation in goal-setting resulted
(1985)         in sig. higher SE judgments than other
               two groups

Schunk         Verbalization of the steps of a problem
& Cox          enhanced SE. Also, effort-
(1986)         attributional feedback enhanced SE

Slemon         The LD group "tend to lack the
& Shafrir      optimistic beliefs about ability of
(1997)         the NA stds"

Wong,          Posttest SE measure showed significant
Butler,        increase
Ficzere,
& Kuperis
(1996)

Wong,          Self-efficacy beliefs did not change
Butler,        from pretest to posttest
Ficzere, &
Kuperis
(1997)

(a) NA = normally achieving.

(b) Not all studies offered posttest self-efficacy ratings.
Table 2
Calibration: Studies Identifying Initial Self-Efficacy
Beliefs as Overestimates

Author                Domain          SE Scale

Alvarez & Adelman     Arithmetic      Two scales: (a) 20-item,
(1986)                                11-point scale (0-10) and
                                      (b) 2 items measuring
                                      global math efficacy

Graham & Harris       Writing         10 items (10-100 with 10
(1989a)                               intervals) measuring stds'
                                      confidence to write stories

Graham & Harris       Writing         5 items (10-pt. scale)
(1989b)                               assessing perceived ability
                                      to write a good essay

Graham,               Writing         10-item, 5-pt. scale meas-
MacArthur,                            uring SE for writing tasks
Schwartz, &                           and cognitive strategies
Page-Voth (1992)

Graham, Schwartz,     Writing         10-item, 5-point scale
& MacArthur (1993)                    measuring efficacy for
                                      composing process and
                                      writing tasks

Pintrich,             Reading         10-item, 7-point scale
Anderman,                             measuring reading effi-
& Klobucar (1994)                     cacy beliefs

Sawyer, Graham,       Writing         10-item, 10-point scale
& Harris (1992)                       measuring SE for writing
                                      a "made-up story"

Slemon & Shafrir      Reading,        "Self-Estimate of
(1998)                Spelling, and   Intellectual Functioning
                      Arithmetic      and Academic
                      tests of the    Achievement Test"
                      WRAT-R; and
                      9 subtests of
                      the WAIS-R

Author                Calibration

Alvarez & Adelman     30% of judgments (to do
(1986)                specific math problems)
                      were overestimates; 2%
                      were underestimates

Graham & Harris       "LD stds consistently
(1989a)               overestimated their compo-
                      sition abilities" (p. 360)

Graham & Harris       "The stds clearly overesti-
(1989b)               mated their composition
                      abilities" (p. 212)

Graham,               3 of 4 stds overestimated
MacArthur,            their SE; SE est. became
Schwartz, &           more realistic after treatment
Page-Voth (1992)

Graham, Schwartz,     "Students with LD overesti-
& MacArthur (1993)    mate their writing capabili-
                      ties" (p. 248)

Pintrich,             LD stds felt "rather effica-
Anderman,             cious" at reading
& Klobucar (1994)

Sawyer, Graham,       Overestimate: "Stds with LD
& Harris (1992)       began with relatively high
                      pretest self-efficacy scores,
                      despite their relatively poor
                      pretest writing performance"
                      (p. 350)

Slemon & Shafrir      The LD group underest.
(1998)                their SE for Spelling and
                      Digit Span; were accurate for
                      Reading, all Performance
                      subtests, and 3 of 5 Verbal
                      subtests; and overest, for
                      Information and WRAT
                      arithmetic

Author                Self-Efficacy Ratings

Alvarez & Adelman     Mean global math efficacy
(1986)                ratings were 8.47 and 8.73 on a
                      1-10 scale, with 10 indicating "very
                      sure" (that they would do well in
                      their current and future math classes).

Graham & Harris       Initial mean rating was 75.2 (out of
(1989a)               100). The 70 point was described as
                      "Pretty sure." Posttest rating, 88.8

Graham & Harris       (a) Initial mean rating 68 (out of 100).
(1989b)               The 70 point was described as
                      "pretty sure." Posttest rating, 79.

Graham,               (a) Initial ratings were 2.5 for
MacArthur,            writing tasks and 2.1 after treatment.
Schwartz, &           Perceived competence for executing
Page-Voth (1992)      cognitive strategies was 3.75 (on a
                      5-point scale) pretest and 3.3 posttest.

Graham, Schwartz,     The LD students' self-efficacy beliefs
& MacArthur (1993)    showed a mean score of 3.21 on a
                      5-point scale with 1 = strongly dis-
                      agree and 5 = strongly agree.

Pintrich,             The mean rating given by LD stu-
Anderman,             dents for self-efficacy for reading
& Klobucar (1994)     tasks was 5.37 on a 7-point scale.

Sawyer, Graham,       (a) Pretest mean score (across four
& Harris (1992)       groups) was 74.4. Posttest mean
                      score across four groups was 83.6.
                      On the 10-100 range scale, 10 was
                      described as "not sure," 40 as
                      "maybe," 70 as "pretty sure," and
                      100 as "real sure."

Slemon & Shafrir      For example, on the WRAT-R LD
(1998)                students predicted standard scores
                      of 104, 98, and 102 for Reading,
                      Spelling, and Arithmetic, respec-
                      tively; actual scores were 106, 102,
                      and 96.

(a) Mean ratings not given in study, but calculated from data provided.

REFERENCES

Alvarez, A., & Adelman, H. S. (1986). Overstatements of self-evaluations by students with psychoeducational problems. Journal of Learning Disabilities, 19, 567-571.

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. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175-1184.

Bandura, A. (1993). Perceived self-efficacy in cognitive development and functioning. Educational Psychologist, 28(2), 117-148.

Bandura, A. (1995). Exercise of personal and collective efficacy in changing societies. In A. Bandura (Ed.), Self-efficacy in changing societies (pp. 1-45). Cambridge, UK: Cambridge University Press.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman and Company.

Bandura, A. (2001). Guide for constructing self-efficacy scales. Available through Frank Pajares' Information on Self-Efficacy Web site: http://www.emory.edu/EDUCATION/mfp/effpage.html#effguide

Bandura, A., & Schunk, D. H. (1981). Cultivating competence, self-efficacy, and intrinsic interest through proximal self-motivation. Journal of Personality and Social Psychology, 41, 586-598.

Baum, S., & Owen, S. V. (1988). High ability/learning disabled students: How are they different? Gifted Child Quarterly, 32, 321-326.

Borkowski, J. G. (1992). Metacognitive theory: A framework for teaching literacy, writing, and math skills. Journal of Learning Disabilities, 25, 253-257.

Bryan, T., & Bryan, J. (1991). Positive mood and math performance. Journal of Learning Disabilities, 24, 490-494.

Butler, D. L. (1995). Promoting strategic learning by post-secondary students with learning disabilities. Journal of Learning Disabilities, 28, 170-190.

Butler, D. L. (1996). Promoting strategic content learning by adolescents with learning disabilities. Exceptionality Education Canada, 6, 131-157.

Butler, D. L. (1998a). Metacognition and learning disabilities. In B.Y.L. Wong (Ed.), Learning about learning disabilities (2nd ed., pp. 277-307). New York: Academic Press.

Butler, D. L. (1998b). The strategic content learning approach to promoting self-regulated learning: A report of three studies. Journal of Educational Psychology, 90, 682-697.

Butler, D. L. (1999, April). Identifying and remediating students' inefficient approaches to tasks. Paper presented at the annual meeting of the American Educational Research Association, Montreal, Quebec.

Butler, D. L., & Winne, P. H. (1995). Feedback and self-regulated learning: A theoretical synthesis. Review of Educational Research, 65, 245-281.

Carr, M., Borkowski, J. G., & Maxwell, S. E. (1991). Motivational components of underachievement. Developmental Psychology, 27, 108-118.

Chapman, J. W. (1988). Learning disabled children's self-concepts. Review of Educational Research, 58, 347-371.

Flavell, J. H. (1976). Metacognitive aspects of problem solving. In L. B. Resnick (Ed.), The nature of intelligence (pp. 231-235). Hillsdale, NJ: Erlbaum.

Graham, S., & Harris, K. R. (1989a). Components analysis of cognitive strategy instruction: Effects on learning disabled students' compositions and self-efficacy. Journal of Educational Psychology, 81, 353-361.

Graham, S., & Harris, K. R. (1989b). Improving learning disabled students' skills at composing essays: Self-instructional strategy training. Exceptional Children, 56, 201-214.

Graham, S., MacArthur, C., Schwartz, S., & Page-Voth, V. (1992). Improving the compositions of students with learning disabilities using a strategy involving product and process goal setting. Exceptional Children, 58, 322-334.

Graham, S., Schwartz, S., & MacArthur, C. (1993). Learning disabled and normally achieving students' knowledge of writing and the composing process, attitude toward writing, and self-efficacy for students with and without learning disabilities. Journal of Learning Disabilities, 26, 237-249.

Gresham, F. M., Evans, S., & Elliott, S. N. (1988). Self-efficacy differences among mildly handicapped, gifted, and nonhandicapped students. The Journal of Special Education, 22, 231-241.

Grolnick, W. S., & Ryan, R. M. (1990). Self-perceptions, motivation, and adjustment in children with learning disabilities: A multiple group comparison study. Journal of Learning Disabilities, 23, 177-184.

Hampton, N. Z. (1998). Sources of academic self-efficacy scale: An assessment tool for rehabilitation counselors. Rehabilitation Counseling Bulletin, 41, 260-277.

Jones, E. D., Wilson, R., & Bhojwani, S. (1997). Mathematics instruction for secondary students with learning disabilities. Journal of Learning Disabilities, 30, 151-163.

Klassen, R. (2002). Writing in early adolescence: A review of the role of self-efficacy beliefs. Educational Psychology Review, 14, 173-203.

Meltzer, L., Roditi, B., Houser, Jr., R. F., & Perlman, M. (1998). Perceptions of academic strategies and competence in students with learning disabilities. Journal of Learning Disabilities, 31, 437-451.

Omizo, M. M., Cubberly, W. E., & Cubberly, R. D. (1985). Modeling techniques, perceptions of self-efficacy, and arithmetic achievement among learning disabled children. The Exceptional Child, 32, 99-105.

Page-Voth, V., & Graham, S. (1999). Effects of goal-setting and strategy use on the writing performance and self-efficacy of students with writing and learning problems. Journal of Educational Psychology, 91, 230-240.

Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational Research, 66, 543-578.

Pajares, F. (1997). Current directions in self-efficacy research. In M. Maehr & P. R. Pintrich (Eds.), Advances in motivation and achievement. Volume 10 (pp. 1-49). Greenwich, CT: JAI Press.

Pajares, F., & Johnson, M. J. (1996). Self-efficacy beliefs in the writing of high school students: A path analysis. Psychology in the Schools, 33, 163-175.

Pajares, F., & Kranzler, J. (1995, April). Competence and confidence in mathematics: The role of self-efficacy, self-concept, anxiety, and ability. Paper presented at the meeting of the American Educational Research Association, San Francisco.

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.

Panagos, R.J., & DuBois, D. L. (1999). Career self-efficacy development and students with learning disabilities. Learning Disabilities Research and Practice, 14, 25-34.

Pintrich, P. R., Anderman, E. M., & Klobucar, C. (1994). Intraindividual differences in motivation and cognition in students with and without learning disabilities. Journal of Learning Disabilities, 27, 360-370.

Pintrich, P. R., & DeGroot, E. V. (1990). Motivational and self-regulated learning components of classroom academic performance. Journal of Educational Psychology, 82, 33-40.

Saracoglu, B., Minden, H., & Wilchesky, M. (1989). The adjustment of students with learning disabilities to university and its relationship to self-esteem and self-efficacy. Journal of Learning Disabilities, 22, 590-592.

Sawyer, R. J., Graham, S., & Harris, K. R. (1992). Direct teaching, strategy instruction, and strategy instruction with explicit self-regulation: Effects on the composition skills and self-efficacy of students with learning disabilities. Journal of Educational Psychology, 84, 340-352.

Schunk, D. H. (1985). Participation in goal setting: Effects on self-efficacy skills and skills of learning disabled children. Journal of Special Education, 19, 307-317.

Schunk, D. H., & Cox, P. D. (1986). Strategy training and attributional feedback with learning disabled students. Journal of Educational Psychology, 78, 201-209.

Shaywitz, B. A., Fletcher, J. M., Holahan, J. M., & Shaywitz, S. E. (1992). Discrepancy compared to low achievement definitions of reading disability: Results from the Connecticut longitudinal study. Journal of Learning Disabilities, 25, 639-648.

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.

Slemon, J. C., & Shafrir, U. (1997, March). Academic self-efficacy of post-secondary students with and without learning disabilities. Paper presented at the annual meeting of the American Educational Research Association, Chicago.

Swanson, H. L. (1989). Strategy instruction: Overview of principles and procedures for effective use. Learning Disability Quarterly, 12, 3-15.

Wigfield, A., Eccles, J. S., MacIver, D., Reuman, D. A., & Midgley, C. (1991). Transitions during early adolescence: Changes in children's domain-specific self-perceptions and general self-esteem across the transition to junior high school. Developmental Psychology, 27, 552-565.

Wong, B.Y.L. (1985). Metacognition and learning disabilities. In T. G. Waller, D. Forrest-Pressley, & E. MacKinnon (Eds.), Metacognition, cognition, and human performance (pp. 137-180). New York: Academic Press.

Wong, B.Y.L. (1986). Metacognition and special education: A review of a view. Journal of Special Education, 20, 9-29.

Wong, B.Y.L. (1987). How do the results of metacognitive research impact on the learning disabled individual? Learning Disability Quarterly, 10, 189-195.

Wong, B.Y.L. (1991). The relevance of metacognition to learning disabilities. In B.Y.L. Wong (Ed.), Learning about learning disabilities (pp. 231-256). New York: Academic Press.

Wong, B.Y.L., Butler, D. L., Ficzere, S. A., & Kuperis, S. (1996). Teaching low achievers and students with learning disabilities to plan, write, and revise opinion essays. Journal of Learning Disabilities, 29, 197-212.

Wong, B.Y.L., Butler, D. L., Ficzere, S. A., & Kuperis, S. (1997). Teaching adolescents with learning disabilities and low achievers to plan, write, and revise compare-and-contrast essays. Learning Disabilities Research and Practice, 12, 2-15.

Wong, B.Y.L., Wong, R., & Blenkinsop, J. (1989). Cognitive and metacognitive aspects of learning disabled adolescents' composing problems. Learning Disability Quarterly, 12, 300-322.

Zimmerman, B. J. (1995). Self-efficacy and educational development. In A. Bandura (Ed.), Self-efficacy in changing societies (pp. 202-231). Cambridge, UK: Cambridge University Press.

Requests for reprints should be addressed to: Rob Klassen, 13624 Blackburn Ave., White Rock, BC, Canada V4B 2Y8.

ROB KLASSEN, M.A., is a Ph.D. candidate, Simon Fraser University, Burnaby, BC, Canada.