The Role of Expectancy and Self-Efficacy Beliefs

Pintrich, P. & Schunk, D. (1996). Motivation in Education: Theory, Research & Applications, Ch. 3.
Englewood Cliffs, NJ: Prentice-Hall

A. Historical Perspectives on the Expectancy Construct
B. Current Perspectives on the Expectancy Construct
     1. A recent model of the expectancy for success construct
     2. Research on self-perception of competence and ability
     3. A model of self-efficacy beliefs
C. Developmental and Group Differences in Expectancy Constructs
     1. Developmental differences
     2. Gender and ethnic differences
D. Implications for Teachers
E. Summary

Several junior high students have the following conversation during lunch.

Kevin: How are you doing in social studies? Did you think what we just did was too easy?

Rachel: Well, not really. It's more just boring. I mean, who cares about what happened 300 years ago? That has nothing to do with my life.

Kevin: Really? I like it. It's interesting. I think what happened before does matter today, plus I'm really good at it. I think I did well on the test today. I always think Iíll do well in social studies because I always get a good grade on the tests and the reports we have to do. I've been getting good grades in social studies ever since elementary school. I might even major in social studies when I go to college.

Sam: Yuck, are you kidding? What a nerd! I hate it! I never seem to do well on those tests. I'm sure I flunked this one. All those truelfalse questions, they just confuse me. I think the teachers just make those up to trick us. Plus, do you know how hard it is to write a report? I can never think of what to say. It's too hard. I just end up copying out of the book. Iíd rather play soccer I'm really good at that. I'm not good at social studies.

Rachel: I hate sports, too. I'm not good at social studies or sports.

Kevin: Well, I'm not too good at soccer I never do well at that, but I'm pretty good at baseball.

At the same time, several teachers are talking in the faculty lounge.

Mr. Dearborn: I always have a hard time getting the kids to understand photosynthesis. No matter what I do, it doesn't seem to make a difference. They just don't get it. I even expect them to have a hard time with it.

Ms. Morgan: Yes, that is a hard topic. I'm not sure I understand it myself, but at least you have all the top kids. I have the lower kids and you know how tough they are to handle. I just can't seem to get control of them, especially the sixth hour. It's the end of the day and we're all tired and I can't seem to get them under control. I just don't think I can manage those kids, let alone teach them complex things like photosynthesis.

Ms. Rivera: I have some of those same low kids and I don't have that many problems with them. I have my system set up and they know I mean business and they don't dare fool around with me. I can handle just about any kid, no matter who it is or what time of day. I know it's hard, but I make sure that they understand who is boss. I never have management problems.

Ms. Morgan: Well, it's easy for you. You've been doing this a lot longer than I. I'm still trying to figure out my system, get my units and lessons together, and actually teach. It's a lot harder than I thought it would be.

Mr. Dearborn: Well, don't worry, you'll figure it out. I remember my first few years of teaching, I thought I would never be able to do it. I had no confidence in my ability to manage a classroom. I used to wake up every morning all anxious and worried about the day and wonder if I would get through it without a major disaster I used to visualize all sorts of bad things happening in my rooms: fights, kids throwing things at me, everyone standing up and telling me to get lost. I look back now and realize how ridiculous some of those thoughts were. Eventually, I got more confidence in myself and figured out ways to manage my room. Now, I only worry about my instruction and the kids' understanding. The management stuff takes care of itself.

These students and teachers are concerned with their ability to succeed at a task, whether they have the skills or knowledge to do well, and what they expect will happen if they do attempt to do the task. All of these individuals are discussing, in one way or another, a central construct in many motivation theories-the role of expectancy beliefs. The concept of expectancy represents the key idea that most individuals will not choose to do a task or continue to engage in a task when they expect to fail. They may be interested in and value a task, but if they try a task and experience repeated failure, then eventually they will not engage in the task. Accordingly, whereas most lay views of motivation stress that interest or value are the most important aspects of motivation (e.g., see Rachel's statement about not being interested in social studies vs. Kevinís statement of interest), most formal motivational models stress that the expectancy construct is very important. In fact, much of the empirical research in achievement motivation situations has focused on the role of the expectancy construct, not the value or interest constructs (Parsons & Goff, 1978).

Go to top of page HISTORICAL PERSPECTIVES ON THE EXPECTANCY CONSTRUCT

The use of the expectancy construct springs from a general cognitive perspective on motivation and reflects the cognitive metaphor of the individual as an active and rational decision maker in contrast to earlier behavioral models of motivation. However, even early on in experimental psychology there were behavioral psychologists who stressed the cognitive nature of learning and utilized the expectancy construct in their models. For example, Tolman (1932) proposed that rewards did not just "stamp in" associations between stimuli and behavior, but that even animals learn expectancies about what will happen to them if they perform a certain behavior. They then come to expect the reward (or punishment) when they engage in the behavior in the situation. For Tolman, this cognitive notion of expectancy replaced the mechanistic concept of habit from Hullian drive models (Weiner, 1992).

As part of this general shift from mechanistic, behavioral models to more organismic, cognitive models, there was less of a conceptual need for postulating some instigator or energizer of behavior (i.e., an instinct, a drive, a need, a habit). When concerned with a rat in a maze, there is a place for thinking about what motivates the rat to run the maze. However, once a general cognitive model is proposed that assumes that humans are innately active learners, constantly seeking to learn and adapt to their environment (White, 1959), then the problem of what motivates behavior is not an issue, but the issue of directionality becomes paramount. Accordingly, cognitive motivational theories became concerned with how individuals make decisions about which goals or paths they will choose to pursue, about the direction in which they will focus their innate energy, curiosity, and activity.

Lewin (Lewin, Dembo, Festinger, & Sears, 1944) proposed that the construct of level of aspiration could capture this decision-making process. Level of aspiration is defined as the goal or standard that individuals set for themselves in a task, based on past experience and familiarity with the task. The traditional research paradigm that was used to assess level of aspiration was the ringtoss game in which individuals were asked to toss rings over a peg while standing at different distances from the peg. Distances further away from the peg were given more value because they obviously made the task more difficult. The subjects were given some experience with the game (usually 10 trials) and then asked to state their goal for the next 10 trials (How many are you going to try to get over the peg in the next 10 trials?). The combination of different distances and values allowed the experimenter and subject to estimate both expectancy (probability of success for a toss) and value (distance from peg).

A great deal of empirical research was done using this level-of-aspiration paradigm and several important findings emerged (Weiner, 1992). First, subjects were most likely to feel successful when they met the goals they set for themselves (subjective goal or level), not the actual objective level of attainment (e.g., five successful tosses). This type of situation occurs in the classroom when two different students get the same good grade of 85 out of 100 points on a test (same objective level of attainment), but one student is quite unhappy with the grade because of expectations of a higher grade, whereas the other student, who had a lower level of aspiration, is quite content with the grade. A second general finding was that level of aspiration was related to prior experience with the task and that prior success generally led to increases in level of aspiration, but failure usually decreased level of aspiration. Finally, the research found that there were individual and group differences in level of aspiration. Subjects high in ability tended to set higher aspirations than those low in ability In addition, subjects were influenced by group goals and performance and would adjust their level of aspiration to the group norms (Weiner, 1992).

Building on these general constructs of expectancy and level of aspiration from Tolman and Lewin, Atkinsonís (1957, 1964) model of achievement motivation attempted to combine the constructs of needs, expectancy, and value into a comprehensive theory. His model proposed that behavior was a multiplicative function of these three components, which he labeled motives, probability for success, and incentive value. Motives represented learned but stable and enduring individual differences or dispositions and included two basic achievement motives: to seek success (need for achievement or motive to approach success) and to fear failure (motive to avoid failure). These motives are affective in nature but do include an aspect of expectancy in terms of emotional anticipation (Covington, 1992). The motive for success was assumed to represent individuals' hope for or anticipation of success and reflect their " capacity to experience pride in accomplishment" (Atkinson, 1964, p. 214). If the motive for success was high, then individuals would likely approach and engage in achievement tasks. In contrast, the motive to avoid failure represents individuals' capacity to experience shame and humiliation when they fail, and when the motive is high, this would lead individuals to avoid engaging in achievement tasks.

In the voluminous research on this model, the motive for success was usually measured using the Thematic Apperception Test, TAT (Atkinson, 1958; McClelland, Atkinson, Clark, & Lowell, 1953), a projective test where individuals are asked to tell stories about a set of pictures they are shown. The pictures display people at everyday tasks (e.g., two men in a machine shop) and the stories are scored for achievement imagery (e.g., The two men are working on a new invention that will change the industry.). In contrast, the motive to avoid failure was usually measured using the Test Anxiety Questionnaire, TAQ (Mandler & Sarason, 1952), which is a more objective self-report measurement asks people to rate their anxiety and worry about testing situations (e "While taking an intelligence test, to what extent do you worry?"). The TAT been criticized for problems of reliability and validity, although a recent analysis of research on the TAT and questionnaire measures of need achievement suggests that both types of measures show approximately the same relation to actual behavior (Spangler, 1992). There continues to be development and research on the use of the TAT and other more objective measures of the need for achievement (Smith, 1992).

In Atkinson's original model, these two motives were assumed to be orthogonal to one another, although most subsequent research usually examined the motive for success and operated as if there was a single continuum anchored by high motive for success on one end and high motive to avoid failure on other end. Covington (1992) has suggested that the original model based the independence of these two motives provides a more comprehensive picture of different approaches to achievement. If the two motives are conceptualized as orthogonal, a two-by-two matrix is created with four cells generated to describe four different approaches to achievement (Covington, 1992; Covington & Omelich, 1991; Covington & Roberts, 1994). This quadripolar model suggests that there are four general types of students and ways of approaching achievement tasks: success oriented, failure avoiders, overstrivers, and failure accepters (see Figure 3.1).

Go to top of page

The success-oriented student who is high in motive for success and low in fear of failure would be highly engaged in achievement activities and not be anxious or worried by performance. In Chapter 1, Sharon represents this type of student because she achieves at a high level and does not worry or become anxious about doing well. A student who is a failure avoider, high in fear of failure and low in motive for success, would be very anxious and attempt to avoid failure by procrastinating and using other self-handicapping strategies (Covington, 1992; Garcia & Pintrich, 1994). This type of student would be very reluctant to even engage in academic achievement work. These two types are the most often researched and described by a simple continuum between the two motives.

The off-diagonal cells in the quadripolar model, however, suggest that there are two other types of students (Covington, 1992). Overstrivers are students who are high in both motives; they try to approach success but simultaneously fear failure greatly. These students work very hard at achievement tasks but also feel very anxious and stressed because of their fear of failure. In more social cognitive terms, these students are similar to defensive pessimists (Garcia & Pintrich, 1994; Norem & Cantor, 1986). Overstrivers are students who almost always do well in class, but are constantly asking the teacher about their grade and show signs of anxiety and worry about doing well. Finally, Covington labels the students who are low in both motives as failure accepters. These students are basically indifferent to achievement, although this indifference may be due either to lack of concern and caring or active anger and resistance to achievement values, as some have suggested is the case for minority students (Covington, 1992).

These two motives (to succeed and to avoid failure) represented the internal and personal contribution to motivation. Atkinson also included expectancy and value constructs in his model that represented the environmental side of the equation because they were assumed to be more closely tied to the situation or task. In Atkinson's model, his expectancy construct was very similar to Tolman's idea about the formation of expectancies regarding the associations between responses and rewards. In Atkinson's model and research, he used human subjects, not rats, so the construct was more cognitive in nature and reflected a person's subjective belief about the probability of success. Expectancy (or probability) for success was often measured using the same ringtoss method as developed by Lewin (e.g., How many rings will you be able to get on the peg?) or by giving subjects a set of puzzles and asking them to estimate how many they would be able to solve (Weiner, 1992). These subjective beliefs about expectancy for success, although certainly reflecting an individual's own beliefs, also were assumed to represent one of the environmental influences on motivation because they could reflect task difficulty (e.g., how far away from the peg the person stood, normative information about how many puzzles others solved). In this sense, the model tried to take into consideration external factors, although it is still basically an organismic model focused on the individual.

The third component of motivation in Atkinson's model was the incentive value of success. Incentive value of success was defined as an affect, specifically, pride in accomplishment. For example, the incentive value of receiving an A in a difficult course would be higher (more pride experienced) than the same grade in an easy course (Weiner, 1992). The incentive value of success was assumed to be inversely related to the probability of success (incentive value = 1.0 - probability of success). For example, as the expectation for success went up, as in an easy task, the incentive value would go down because it was assumed that the person would not value succeeding at an easy task. In the same way, for a difficult task where probability of success is low, the incentive value would be high in Atkinsonís model, incentive value was an affect, but in more general uses of the value construct, it has become a belief about the attainment value, importance, or interest in a task (Eccles, 1983). Nevertheless, the inverse relation between value and expectancy still may be applicable. For example, people may value certain careers or professions more highly and assign them more prestige and salary because of the perceived difficulty of attaining them (Weiner, 1992).

Given the inverse relation between incentive value and the probability success, Atkinson's mathematical model predicts that motivation will be highest when tasks are of an intermediate level of difficulty. When the probability of success is .5 (the person will succeed at the task about half the time), the incentive value of success also will be .5 (incentive value = 1.0 - .5, which is the value for probability of success). The product of multiplying the probability of success by the incentive value (as in all expectancy-value models) is greatest at this intermediate level (e.g., .5 X .5 = .25) in comparison to other conditions. For example, suppose that the probability of success is .1 (the student will succeed only 1 time out of 10), then incentive value will be high for such a difficult task (1.0 - A = .9) and the multiplicative product of these two numbers (.1 X .9 = .09) is lower than the number generated at the intermediate level of difficulty (.25). This is true for all values of probability of success and incentive value given the assumed inverse relation between probability of success and incentive value.

This generalization that motivation is highest at levels of intermediate task difficulty is one of the most often cited findings from achievement motivation research. Variations on this generalization are represented in behavioral theories such as Gagne's, which suggested that instruction should be designed to be within several steps of a student's prior level of competence (Gagne, 1985). Of course, in applications of Piagetian theory, it was suggested that students should only be taught concepts that were representative of the next immediate stage (+ 1 modeling/instruction), not concepts that were more than one stage above the students' current level. More recently, we have seen a great deal of interest in applying Vygotsky's concept of the zone of proximal development, which also represents the notion that instruction should be designed to be just beyond the students' current level of development, not too difficult to be far outside the students' capabilities or too easy to be repetitive for the students (Vygotsky, 1978). Although these theories are not conceptualized as motivational theories and tend to downplay or ignore the role of motivational constructs, the general principle that learning will be fostered when students are working at a task that is somewhat beyond their range of current capability fits with these cognitive theories as well as with most motivational theories. The real challenge is determining the students' current capabilities and the most propaedeutic difficulty level of the academic task.

Weiner (1992) summarizes the motivational research on this generalization regarding an intermediate level of difficulty and concludes that the laboratory research on task choice and persistence is supportive. However, he does note that in Atkinson's model, the third determinant of motivation, the motives for success and fear of failure, plays an important role beyond the expectancy and value components. Atkinson's model predicts that individuals high in the motive for success and low in fear of failure (what he called "high need for achievement") will be most likely to choose tasks of intermediate difficulty, whereas individuals who are high in fear failure and low in the hope for success (what he called "low need for achievement") will choose very easy or very difficult tasks. In the latter case, for individuals high in fear of failure, a choice of a very easy task ensures success, thereby minimizing fear of failure. The choice of very difficult tasks by individuals high in fear of failure does not maximize the fear of failure because there is the expectation that very few people can succeed at the most difficult tasks. Atkinson's model and reasoning here is based on the general hedonic principle of maximizing positive affect and minimizing negative affect (Weiner, 1992). However, the actual empirical findings do not support the differential predictions for individuals high in fear of failure. It appears that most people, regardless of their motives for success and failure, choose tasks of intermediate difficulty, although there is a tendency for individuals high in the motive for success to choose intermediate tasks more often than those high in fear of failure (Weiner, 1992).

Weiner and others (e.g., Trope, 1975) suggest that this choice of intermediate tasks can be explained in terms of an informational principle rather than the hedonic principle. They suggest that choosing tasks of intermediate difficulty provides the most information to individuals about their own effort and capabilities (too easy and too difficult tasks provide more information about the task than the individual). This information about personal capabilities is useful for self-evaluation and, in the long run, accurate self-information is functional and adaptive (Weiner, 1992). In any event, the finding of very little differential choice of tasks by motives and of a larger role for expectancy for success and incentive value suggests that these more cognitive and environmental constructs might be more important than the stable personality dispositions of motives.

In sum, the early research on expectancy constructs was important because of the focus on cognition and beliefs in contrast to overt behavior and the related constructs of drives, needs, and habits. These theories and models moved motivational psychology away from a dependence on a simplistic S-R psychology to a more rational and cognitive paradigm that is still dominant today. Moreover, these early cognitive models of motivation stressed the importance of the individual's perceptions and beliefs as mediators of behavior, thereby focusing motivational research on the subjective and phenomenological psychology of the individual. In particular, these early models developed the distinction between beliefs about being able to do the task (probability and expectancy for success) and beliefs about the importance, value, and desire to do the task (motives, incentive value) and posited that it is the combination of the two that results in motivated behavior. Accordingly, we may feel capable of doing a task, but if we do not value it, then we will be less likely to engage in it. In the same way, we may value a task, but if we do no feel able to do it and expect to fail, we will be less likely to engage in the task Current research on expectancy and value constructs continue in this tradition of focusing on these two general beliefs of the individual, although they do attempt to include contextual influences in their models. In the remainder of this chapter, we will discuss students' expectancy beliefs: their perceptions of their ability to perform a task. Chapter 8 covers students' value beliefs.Go to top of page

CURRENT PERSPECTIVES ON THE EXPECTANCY CONSTRUCT

Although there are many current motivational theories that include some type of expectancy construct, we will focus on three general approaches that reflect current research programs that are based on academic achievement in classroom settings. The first model comes from the work of Eccles and Wigfield and their colleagues (e.g., Eccles, 1983; Eccles et al., 1989; Wigfield, 1994; Wigfield & Eccles, 1992) and is derived from and bears the closest resemblance to the early expectancy-value models of Lewin and Atkinson (for other current expectancy-value models, see Heckhausen, 1977; Pekrun, 1993). This cognitive model focuses on the role of students' expectancies for academic success and their perceived value for academic tasks and springs from a general organismic perspective based in personality and social psychology. A second related model is the research on the development of children's perceptions of competence (e.g., Harter, 1982, 1985a; Phillips & Zimmerman, 1990; Stipek, 1981). This model also derives from a general organismic perspective that focuses on the role of the individual but has a more cognitive developmental flavor, reflecting its roots in the early writings of White (1959) on competence and motivation. Finally, the third model to be discussed concerns the application of self-efficacy theory to achievement situations. Bandura (1986) has proposed a general social cognitive model of cognition and behavior that has been applied to many different domains of human behavior (e.g., depression, phobias, sports, decision making in organizations) as well as achievement contexts. Schunk (1991b) has been the most prominent researcher on the application of the self-efficacy construct in achievement contexts. Although this model focuses on the individual as in the other two organismic models, it also represents a somewhat more mechanistic metatheoretical perspective, reflecting its development from social learning theory.Go to top of page

A Recent Model of the Expectancy-for-Success Construct

A model that is a direct descendent from Atkinson's expectancy-value model is the one developed by Eccles and Wigfield and their colleagues (Eccles, 1983; Wigfield & Eccles, 1992). A somewhat simplified version of their elaborate model is displayed in Figure 3.2. It is clear from the figure that the two most important predictors of achievement behavior are expectancy and task value. It is important to note that these are basically two of the three components specified in Atkinsonís model. In Atkinson's model, the expectancy construct was called probability for success and the task value construct was labeled incentive value. Reflecting a social cognitive and more situational approach and in line with the findings that motives don't differentially predict task choice, the current model does not include the stable and enduring personality constructs of motives for success and fear of failure. In the Eccles and Wigfield version, task value has a much more important and differentiated role to play than just being the inverse of probability for success as in Atkinson's original model. However, we will defer discussion of the task value component and the goals that lead into task value until Chapter 8. In this chapter, we will focus on the expectancy component and the constructs that lead into it.

As can be seen in Figure 3.2, achievement behavior is predicted by two general components: expectancy and value. These constructs are within the dashed lines in the figure to represent the fact that they are internal, cognitive beliefs of the individual in contrast to the achievement behaviors, which are overt and observable. In colloquial terms, the value construct refers to a student's response to the question, Why should I do this task? Responses would include those having to do with goals (I want to become a doctor.), values (I think biology is useful.), and interests (I'm interested in this topic.). Rachel's expressions of boredom and lack of care about social studies and "hate" for soccer are expressions of value beliefs. In contrast, the expectancy construct refers to the question, Am I able to do this task? (Eccles, 1983; Pintrich, 1988a, 1988b; Wigfield, 1994; Wigfield & Eccles, 1992). In Figure 3.2, expectancy refers to actual beliefs of students about their future expectancy for success; that is, whether they believe that they will do well on an upcoming test or some future event. In our example, Kevin expects to do well in social studies, whereas Sam and Rachel do not expect to do as well. This construct is usually measured by asking students to predict how well they will do in the future on some task or in some domain (see Table 3. 1). Accordingly, expectancy for success is more future oriented than simple self-perceptions of competence. In a great deal of research by Eccles and Wigfield (as well as many others), higher expectancies for success are positively related to all types of achievement behavior, including achievement, choice, and persistence (Eccles, 1983; Wigfield, 1994; Wigfield & Eccles, 1992).

Referring again to the model in Figure 3.2. the next two expectancy components include students' task-specific self-concept and their perceptions of task difficulty. The self-concept belief represents students' perceptions of their competence in different domains; it is their evaluation of their ability to do a task currently (in contrast to the future orientation of expectancy for success). Eccles and Wigfield are careful to note that these self-concept beliefs can be relatively domain specific and may vary by subject area in school contexts (see also Harter, 1983, 1985a; Marsh, 1989, 1990b).

Go to top of page

Task perceptions concern students' judgments of the difficulty of the task (not unlike Lewinís and Atkinsonís models). Again, Eccles and Wigfield assume that these task difficulty perceptions are relatively domain specific. Although they have not explicitly discussed the nature of the domain boundaries, in their own empirical work they have used general school subject areas (e.g., English, math, science) as their operationalization of the domain, not unlike other self-concept researchers such as Harter and Marsh. Of course, they would allow that there may be even more domain specificity in these beliefs that could influence achievement (e.g., differential self-concept and task difficulty beliefs for algebra vs. geometry vs. calculus), but more research is needed on the level of specificity for conceptualizing domains. Specificity of domains is one of the biggest issues that needs to be resolved for any cognitive or motivational theory that proposes domain specificity of constructs. What are the boundaries of a domain and at what level-the individual task level; the project, unit, or topic level; the course level; the discipline level-do we need to specify them?

Regardless of the domain issue, the general model proposes that self-concept beliefs and task difficulty beliefs are then considered together to produce an expectancy judgment. So, for example, Kevin has a belief that he is good in social studies and that it is not that difficult a task; hence, he has high expectations for social studies. Sam, on the other hand, has a lower self-perception of his competence and thinks that social studies tasks (writing reports) are hard and, therefore, has a lower expectancy for success for future social studies tasks. Table 3.1 provides some examples of how self-concept and task difficulty perceptions are measured in this expectancy-value model.

These beliefs are influenced by two other variables in the model (see Figure 3.2). One includes another cognitive and internal process concerning how students perceive and interpret different events that happen to them. In particular, this interpretative process is driven by the types of attributions a student makes for events and actual performance. Attributions are crucial to the formation of self-perceptions of competence and expectancies and are the focus of Chapter 4 so they are not discussed here. However, it is important to note that in keeping with a general cognitive and constructivist approach, this model assumes that students' perceptions of competence and task difficulty and subsequent expectancies for success are influenced by how students perceive their social environment and what happens to them as they move through it. The other variable that influences students' beliefs is the actual culture and environment, including the general cultural and societal milieu; the nature of the students' interactions with parents, peers, and other adults (e.g., teachers); and their past performance and achievements. These final influences are represented on the far left of Figure 3.2 outside the dashed lines because they are assumed to be external to the student. As noted in the figure, these influences can have a direct effect on children's self-concepts and task difficulty beliefs (a direct arrow), but their main effect is mediated by the students' perceptions and interpretations of these environmental influences.

In a series of large-scale correlational field studies, Eccles and Wigfield and their colleagues (Eccles, 1983; Eccles et al., 1989; Wigfield, 1994; Wigfield & Eccles, 1992) have investigated the role of expectancy constructs in achievement. These studies have used both cross-sectional and longitudinal designs where upper elementary and junior high students are given self-report measures of their self-perceptions of ability and expectancy for success in math and English at the beginning of one school year and at the end of that same year. In some of their studies, these same students are then followed into the next grade and given the same self-report measures, again at the beginning and at the end of the school year. At the same time, the researchers also have collected data on the students' actual achievement on standardized tests and course grades. They then used path analytic models that allowed them to examine the relative effects of expectancy perceptions versus grades on subsequent perceptions and grades. These studies have consistently shown that students' self-perceptions of ability and their expectancies for success are the strongest predictors of subsequent grades in math and English, even better predictors of later grades than are previous grades. This general finding that has emerged across a number of studies highlights the importance of students' expectancies and self-perceptions of competence as mediators between the environmental context and actual achievement behavior as proposed in Figure 3.21 again demonstrating the importance of student beliefs and the constructivist nature of motivation. In addition, these studies are conducted in actual classrooms and follow students over several years, resulting in high ecological validity and increasing the generalizability of the results.

Besides actual achievement on tests and teacher-assigned grades, other researchers have linked students' expectancies and perceptions of ability to students' cognitive engagement. For example, Pintrich and his colleagues (Pintrich, 1989; Pintrich & De Groot, 1990a; Pintrich & Garcia, 1991; Pintrich & Schrauben, 1992) have investigated the relations between expectancy beliefs and students' use of various cognitive strategies, such as elaboration (paraphrasing, summarizing), and metacognitive strategies (planning, checking, monitoring work) in a series of correlational field studies. These cognitive and metacognitive strategies have been shown to result in "deeper processing" of the material to be learned and higher levels of understanding and learning. Although they have used self-report measures of both motivation and strategy use, Pintrich and his colleagues have consistently found that higher levels of expectancy and perceptions of competence are correlated with more reported use of cognitive and metacognitive strategies in both junior high and college student samples. Other researchers have found similar relations (Paris & Oka, 1986; Shell, Murphy, & Bruning, 1989) between perceptions of competence and cognitive engagement.

In contrast to these relations between expectancy perceptions and cognitive engagement and actual achievement, the findings for choice are not as clear-cut. Although earlier research by Atkinson and others showed that expectancy beliefs did relate to students' choice of distances in the ringtoss experiments, more recent research on student academic choices does not show as strong a role for expectancy beliefs. Eccles and Wigfield and their colleagues find in their field studies that the value beliefs (see Chapter 8) concerning the students' perceptions of the importance, utility, and interest in the task are better predictors of their intentions to continue to take math and of their actual enrollment decision (Wigfield, 1994; Wigfield & Eccles, 1992; also see Feather, 1982, 1988). Accordingly, from an expectancy-value perspective, it appears that expectancy beliefs are more closely tied to actual achievement and cognitive engagement but that value beliefs are more closely tied to choice behaviors that would provide the student with the opportunity to achieve in the future.Go to top of page

Research on Self-Perceptions of Competence and Ability

The research on students' perceptions of their own competence (Harter, 1982, 1985a, 1990) is very similar in some ways to the research on expectancy-value models. Specifically, the definition of self-perceptions of competence is isomorphic with the definition of task-specific self-concept in Eccles and Wigfield's research. Self-perceptions of competence are students' self-evaluative judgments about their ability to accomplish certain tasks (Harter, 1985a). Self-perception of competence is the more cognitive evaluation of ability in domain (I can do math.), not a general measure of self-esteem or self-worth that concerns how individuals might feel about themselves (I am happy with the way I am.) (Wigfield & Karpathian, 1991). At the same time, this research on perceptions of competence springs from a more developmental perspective on the development of the self and personal identity in contrast to the focus on motivation in expectancy-value models. Given this different heritage, there are some differences in the research and theoretical underpinnings.

First, some of the research on students' perceptions of competence is closely related to the more general research on children's conceptions of their personal identity (e.g., Damon & Hart, 1988) and research on self-concept (Markus & Nurius, 1986; Marsh, 1984a, 1984b, 1990b). An important issue in the research on self-concept is the domain specificity of individuals' perceptions of themselves. Much of the early research on self-concept was not very theoretically based, good definitions of the construct were not developed, and many researchers assumed that self-concept was a rather global construct (Wylie, 1974, 1979, 1989). More recent research has focused on how children do distinguish their perceptions of competence by domain (Byrne, 1984) Although self-perceptions of competence become more differentiated with age, most researchers now accept the idea that even fairly young children (first and second graders) have self-perceptions of competence that are domain specific. In addition, whereas there is disagreement about the levels of specificity of the domains, most researchers at least distinguish among academic, social, and physical domains of competence (Harter, 1982; Wigfield & Karpathian, 1991).

The academic domains concern students' perceptions of competence at school tasks. The perceptions of competence in Table 3.1 are from Harter's perceived competence scale (1982) and reflect a general sense of competence for schoolwork. Another self-concept measure, Marsh's Self-Descriptive Questionnaire (SDQI for preadolescents, SDQII for adolescents, and SDQIII for late adolescents and adults), has domains for reading, math, and all school subjects. The social domain reflects individuals' perceptions of their competence in interactions with others. Harter's revised perceived competence scale, now called the Self-Perceptions Profile (Harter, 1985b), and Marsh's SDQ both have separate scales for relationships with parents and relationships with peers/close friends. In addition, both Harter and Marsh have scales regarding competence for romantic relations with others for their adolescent and adult measures. The physical domain on Harter's and Marsh's measures includes perceptions of competence at physical activities like sports as well as general perceptions of physical attractiveness/appearance. Students' scores on these different domain scales show moderately positive intercorrelations, but the research does suggest that these are empirically separate domains. Accordingly, students may have differential perceptions in different domains (high in math, low in reading, high in physical ability, low in peer relations, etc.). Researchers and teachers need to be sensitive to these domain differences in perceptions of competence and not assume that students have a single global self-concept that is related to their performance in all domains.

Although the simplistic idea of a global self-concept relating to performance across many domains is not accepted, the assumption that there are domain specific self-perceptions of competence still begs the question regarding the hierarchical nature of these self-perceptions. The issue revolves around the problem created by multiple domains and the specificity of those domains. For example, if a student has different self-perceptions of competence for different academic domains such as English, math, science, and social studies, the question becomes: Are these self-perceptions ever integrated into a general academic self-concept? This academic self-concept then might be at the same level in a hierarchy of selves as a physical self-concept (an integration of self-perceptions for different physical activities such as running, tennis, gymnastics, soccer, baseball, etc.) and a social self-concept (an integration of self-perceptions for relations with close friends, other classmates, parents, and other adults).

The research on this issue is somewhat mixed. Harter (1983, 1985a) suggests that the results of her studies with factor analysis of the domains from her student perceptions profile support a nonhierarchical model, reflecting support for the taxonomic position (Byrne, 1984). In contrast, the work of Marsh and Shavelson (Marsh, 1990b; Marsh & Shavelson, 1985; Shavelson & Bolus, 1982) suggests that the self-concept is hierarchically structured with separate domains of academic competence fitting into a global academic self-concept, although they find that the strength of the hierarchical structure differs with age (less hierarchy in older adolescents and adults). Wigfield and Karpathian (1991) suggest that these different findings may be due to developmental differences in the students and reflect a general developmental movement from nonhierarchical to hierarchical back to nonhierarchical self-concepts over time. Most of Harterís research has focused on younger students who may be less able or less motivated to integrate the different domains of their self-concept. In contrast, Marsh's early adolescents may show hierarchically organized structure because they are now able cognitively to integrate diverse information and given the adolescent "life task" of achieving identity, be more motivated to attempt to fit together the different self-domains. Finally, Marsh and Shavelsonís older adolescents and adults again may show less hierarchy because they are not actively involved in all domains (most adults are not involved in academic learning of math, science, social studies, etc.) and are less concerned with differences in domains and less upset with having multiple identities or self-concepts. Of course, as Wigfield and Karpathian (1991) point out, most of this research has been cross-sectional and there is a need for longitudinal research to examine the developmental trends and the suggested explanation of the findings.

A second important issue related to the domain specificity and hierarchical nature of perceptions of competence is the relation between perceptions of competence and global self-esteem. As already noted, after some confusion in the early research on self-concept, there is now fairly wide agreement that self-perceptions of competence and self-esteem are two theoretically and empirically distinct constructs (Wigfield & Karpathian, 1991). To reiterate, perceptions of competence are more cognitive judgments of personal skills and abilities, such as the belief that you are able to learn social studies, or that you can play soccer (recall Kevin, Rachel, and Sam), or that you can make friends. In contrast, self-esteem is a more global affective reaction or evaluation of yourself (e.g., you feel bad about yourself since you don't do well in social studies, or good about yourself because you play soccer well, or bad because you don't have many friends). Empirical research by both Harter (1985a, 1986) and Marsh (Marsh & O'Neill, 1984; Marsh & Shavelson, 1985) have shown that global self-esteem is an empirically separate dimension from self-perceptions of competence, although the two show positive moderate correlations. Obviously, perceptions of competence should be related to self-esteem, but the exact nature of this relation is unclear.

One of the early leading American psychologists, William James (1890), speculated that self-esteem was a ratio between one's "successes" and one's ďpretensions" and that self-esteem would be higher to the extent that individuals were able to succeed in meeting all their goals. More recently, Harter (1985a, 1986, 1990) has made a similar argument and suggested that global self-esteem is related to both self-perceptions of competence in different domains and the importance the individual assigns to those domains (paralleling expectancy-value theory predictions). Accordingly, if students have low perceptions of competence in several academic domains, they may not necessarily have low self-esteem if they do not think those domains are that important or central to them in comparison with other domains (e.g., social relations with peers, personal physical abilities). Harter (1986) found evidence for this type of "discounting" in a sample of fifth through seventh graders where students with high self-esteem had much less discrepancy between their importance and competence ratings in comparison with those with low self-esteem. Wigfield, Eccles, Mac Iver, Reuman, and Midgley (1991) also found similar results with junior high students. Steele (1988) has suggested that many minority students, particularly African Americans, use this type of discounting strategy by lowering their value for academic domains where they have lower perceptions of academic competence in order to affirm their self-worth and self-esteem. In contrast, in a study of high school and college students, Marsh (1986) did not find that self-esteem was related to perceptions of competence and importance, but rather self-esteem was best predicted by only perceptions of competence. Developmental differences may be responsible for the different results, but clearly more research is needed on the relations between self-perceptions of competence and global self-esteem (Wigfield & Karpathian, 1991).

A third issue that arises in the research on self-perceptions of competence concerns the accuracy of children's self-evaluations. Interestingly, this issue has not been a major concern of researchers working in the expectancy-value tradition. Expectancy-value theory is built on a motivational model that stresses the subjective nature of individuals' assessments of their expectancy for success. In this case, it is not whether individuals have accurate perceptions of their future expectations (e.g., that their expectations should "match" their past performance), but rather the importance of their personal and subjective expectations, regardless of the relation to previous performance. In contrast, the developmental research on children's self-perceptions of competence has been concerned with personal and self-concept development where issues regarding the accuracy of self-beliefs are important. The issue of accuracy assumes that there is some objective indicator of children's competence that can be used as a criterion against which to assess the congruence of their perceptions of their competence. In most studies, the objective indicators in the academic domain have been standardized tests, teacher grades, or teacher ratings. In general, the research suggests that in the middle elementary grades (third, fourth grade), there is less congruence between children's self-perceptions and more objective assessments, whereas the congruence becomes much better in the later elementary grades and later junior high school (eighth and ninth grades) (Harter, 1985a). Moreover, most of the incongruence in the early grades occurs because the younger students have fairly high self-perceptions of competence. Their perceptions then become more modest as they move into the later grades (Frey & Ruble, 1987; Harter, 1985a; Stipek, 1981, 1984; Weisz, 1983), thereby making them more veridical with other, more objective assessments (Phillips & Zimmerman, 1990).

Nevertheless, in a review of various studies on the validity of self-reports of perceived competence, Assor and Connell (1992) state,

Surveys of relevant research findings clearly demonstrate that, beginning in the third and fourth grade and extending through high school, in populations ranging from upper middle-class Caucasian American youth, to lower and middle class Israeli children, to primarily poor African American adolescents, there is no empirical justification for viewing self-report appraisals of academic competence and efficacy as invalid measures of performance affecting self-appraisals. (pp. 41-42) Assor and Connell (1992) also note that there may be some positive advantages for having "inaccurate" self-perceptions of competence when they are higher than should be expected given actual performance. They report that these "inflated" self-reports actually are related to positive performance outcomes 2 years later in a longitudinal study of high school students. They also note that students who have inaccurate self-perceptions that are lower than they should be (deflated self-assessments) perform at much lower levels, which coincides with other studies as well (Phillips & Zimmerman, 1990). Accordingly, in line with a motivational and constructivist perspective, individuals' personal and subjective self-perceptions are important for future achievement behavior, regardless of the "accuracy" of the perceptions in terms of their match to grades, standardized tests, or ratings by adults. Researchers must still be careful, however, to get accurate perceptions that really represent childrenís' beliefs about themselves, not modesty or social desirability effects.

Besides the general developmental differences in accuracy of self-perceptions, there also seem to be some individual differences in the inaccuracy of self-perceptions of competence. Inaccuracy in self-perceptions of competence can result from overestimation of competence (believing you are more competent than suggested by objective measures) and underestimation of competence (believing you are less competent than expected by objective measures), Phillips (1984, 1987; Phillips & Zimmerman, 1990) has examined three types of students in terms of their accuracy of competence perceptions. In her studies, she used samples of only high-achieving third-, fifth-, and ninth-grade students (defined in terms of a combination of standardized test scores and teacher ratings), but still found that there are overestimators (higher perceptions of competence than warranted by actual achievement), accurate perceivers (congruence between perceptions and achievement), and underestimators (lower perceptions than warranted by actual achievement). This latter group, all of whom have the "illusion of incompetence," was of particular concern to her because all of her sample students were actually achieving in the top 25% on nationally standardized tests. These underestimators were most likely to hold very low expectations for future success, to believe that their parents and teachers had low perceptions of them, to be more anxious, and to be less willing to try hard and persist on academic tasks (Phillips & Zimmerman, 1990). Accordingly, whereas most students do become more accurate in their self-perceptions of competence as they progress through school, there are still some very able children who have the illusion of incompetence and show negative motivational outcomes.

Finally, a fourth issue concerns the nature of the relation between perceptions of competence and self-concept and achievement. As noted in the section on expectancy-value theory, children's perceptions of their ability have been linked to most achievement behaviors, including effort, persistence, cognitive engagement, and actual achievement. However, within the tradition of research on self-concept, researchers have been concerned with the direction of the causal relation between self-concept and achievement. Some researchers have argued that self-concept is causally predominant over achievement, while others have taken the opposite position that achievement determines self-concept (see Byrne, 1984; Caslyn & Kenny, 1977; Hansford & Hattie, 1982; Scheirer & Kraut, 1979, for conflicting views). These different views obviously have implications for instruction. The self-concept enhancement version suggests that teachers should ensure that students believe they are capable and then their subsequent achievement will improve. The achievement-first model suggests that teachers should ensure that children have the academic skills to succeed and then their perceptions of competence will follow from their successes. The most recent view of this controversy is that "it is relatively fruitless to continue to pursue the general question of which causes which" (Wigfield & Karpathian, 1991, p. 256). Clearly, the relation is reciprocal; self-concept influences future achievement and actual achievement shapes and constrains self-perceptions of competence. Moreover, there may be individual, developmental, and contextual differences in the nature of the relations between self-concept and achievement (cf. Marsh, 1990a; Skaalvik & Hagtvet, 1990). In terms of future research and educational practice, it seems wise to realize that the relation is complex and that simple, linear models do not map onto the reality of classrooms. Future research should concentrate on understanding how self-perceptions and actual achievement work together to predict future behavior at different ages, for different students, and in different contexts. Teachers should focus on teaching their students the appropriate cognitive skills to master academic tasks and, at the same time, ensure that students have the motivational resources, including appropriate self-perceptions of competence, to engage in these tasks.]

In sum, the research on self-perceptions of competence has investigated the role of childrenís self-evaluative judgments, a construct isomorphic with task-specific self-concept from the expectancy-value models. The research has had a decidedly developmental perspective and has investigated these self-perceptions in young children, adolescents, and adults. Current research suggests that self-perceptions of competence are domain specific, not global, although there is not agreement about the specificity of the domain. Nevertheless, it seems important for future research as well as educational practice to recognize that in the academic domain, self-perceptions of competence need to be specified at least at the level of general subject areas (i.e., English, math, science, social studies, foreign language, art, music, etc.). It probably will not be that helpful, moreover it is potentially misleading, for teachers to label children as having low self-concept as a global trait that applies across all academic, social, and physical domains. Teachers need to differentiate more carefully and recognize that children can and do differ within those domains in terms of their self-perceptions of competence and that these intraindividual differences can result in differences in motivated behavior.Go to top of page

A Model of Self-Efficacy Beliefs

A third general model that has examined the role of students' perception competence is self-efficacy theory. Bandura (1982, 1986, 1989) has developed a social cognitive model of behavior that includes self-efficacy as a major construct. Self-efficacy theory grows out of Bandura's original social learning theory (Bandura, 1969, 1977) and, hence, has some behavioral and mechanistic aspects in contrast to the more organismic perspective on the individual represented in expectancy-value models and the research on self-perceptions of competence. Bandura and others have applied the model to a variety of domains, including mental health, such as coping with depression and phobias; health behavior, such as recovery from a heart attack and cessation of smoking; decision making and sales performance in business; athletic performance; career choices; and academic achievement. In the educational domain, Schunk (1989a, 1989c, 1991b) has been the leading theorist and researcher regarding the role of student self-efficacy in classroom settings.

In this model, self-efficacy is defined as "people's judgments of their capabilities to organize and execute courses of action required to attain designated types of performances" (Bandura, 1986, p. 391). In a general sense, this definition is fairly similar to the definitions of task-specific self-concept and self-perceptions of competence that have been discussed in the two previous sections. Self-efficacy does represent people's judgments of their abilities in the same way as Eccles and Wigfield represent task-specific self-concept and Harter represents self-perceptions of competence. There are, however, some important differences. First, the definition of self-efficacy includes "organize and execute courses of action," which represents the theory's more specific and situational view of perceived competence in terms of including the behavioral actions or cognitive skills that are necessary for competent performance. For example, self-efficacy in math would not be merely a self-recognition of being good in math but rather explicit judgments of having the skills for doing, say, two-digit subtraction problems (Schunk, 1991b). In fact, in assessing self-efficacy, researchers often show students the problems they will be asked to solve and then ask them to rate themselves on a scale from I to 100 as to how confident they are about their ability to solve the problems. This is usually done for all the problems in the set and students rate themselves for each problem, generating a series of ratings for self-efficacy for the problem set that are averaged to form a measure of self-efficacy for doing two-digit subtraction problems (see Table 3. 1).

A second aspect that distinguishes self-efficacy from self-concept and self-competence is that it is used in reference to some type of goal ("attain designated types of performance"). Again, this reflects the more situational perspective of efficacy theory in contrast to the personality and developmental heritage of expectancy-value and perceptions-of-competence research. The goal may be determined by the individual or by the task conditions and environment (or their interaction), but the important point is that judgments of efficacy are in reference to this goal. One implication of the inclusion of a specific goal is that self-efficacy judgments for very similar tasks may vary as a function of intraindividual or environmental differences. For example, skilled and experienced runners may lower their efficacy judgments for maintaining their usual time for a 10-kilometer run because of a nagging muscle pull. In an academic setting, a student's self-efficacy for learning a particular topic in mathematics may be lower because of the difficulty of the material to be learned in contrast to material covered earlier in the course. In colloquial terms, these individuals have lower than usual "self-confidence" in their capabilities to perform a specific task at a certain level of competence.

Efficacy theory also proposes that outcome expectations form a second construct related to motivational behavior and affect. Outcome expectations are judgments or beliefs regarding the contingency between a person's behavior and the anticipated outcome. This notion of contingency between response and outcome is similar to Rotter's (1966) construct of locus of control regarding the contingency between behavior and reinforcement (see Chapter 4). In addition, in terms of the anticipation of success, it is similar to expectancy for success from expectancy-value theories. As Bandura (1986) puts it, "The belief that one can high jump six feet is an efficacy judgment; the anticipated social recognition, applause, trophies, and self-satisfactions for such a performance constitute the outcome expectations" (p. 391). In the academic domain, students would have efficacy judgments of their capabilities, skills, and knowledge to master school-related tasks, but also have outcome expectations about what grades they might receive on the tasks. Although efficacy beliefs and outcome expectations are usually positively correlated, it is possible for a student to have a relatively high efficacy belief for a task, but low outcome expectations. For example, a college student in an organic chemistry class might have relatively high efficacy beliefs about personal capability to master the material, but low outcome expectations about grades on exams due to the very high competition among premed students and the grading curve instituted by faculty to weed out the weaker students from the medical school admission process.

Although both efficacy and outcome judgments are best thought of as continuums, Bandura (1982) has suggested that a simple high/low efficacy by high/low outcome expectation comparison provides some insight into behavior and affect. Figure 3.3 shows the expected behavioral and affective reactions for individuals who vary in efficacy and outcome expectation beliefs. The expected behaviors and affects listed in the figure are assumed to apply to all aspects of life, not just academic domains. Of course, it is best to be high in both efficacy and outcome expectations. Persons in this cell would be confident and assured in their performance, would show high levels of effort, would persist, and probably would have high levels of cognitive engagement in academic tasks. Students high in efficacy but low in outcome expectations, like our organic chemistry student, would be most likely to study hard and be engaged but, at the same time, may protest and lobby for changes in the grading system. In some cases, they would leave the environment by dropping out of the premed or science program (not the class), not because their efficacy is low, but rather because they don't perceive a contingency between their learning and the outcomes, in this case grades. Bandura (1982) also notes that many social activists who lobby for changes in social policies would be in this cell, in contrast to individuals who are low in both efficacy and outcome expectations who are often the "clients" of these activists. Bandura suggests that individuals who are low in both efficacy and outcome expectations would show resignation and apathy and an unwillingness or inability to exert effort. In the academic domain, this cell would represent students who have given up on learning and are very unwilling to exert effort. In most ways, they would be similar to learned helpless students (cf. Peterson, Maier, & Seligman, 1993, and our discussion in Chapter 4). The final cell in Figure 3.3 would include students who have low efficacy for their ability to master the task, but high outcome expectations. Students low in efficacy and high in outcome expectations believe they can't do the task but are very aware (by seeing others get rewarded) that if they were able to perform, the environment would be responsive and they would be appropriately rewarded. These students would evaluate themselves very negatively and focus on themselves as the cause of all failure, in contrast to students low in both efficacy and outcome expectations who also focus on the lack of environmental responsiveness.

Although Bandura proposes both of these motivational constructs, the theory and subsequent research focus on the role of self-efficacy beliefs. Bandura (1986) suggests that outcome expectations are heavily dependent on efficacy judgments: "If you control for how well people judge they can perform, you account for much of the variance in the kinds of outcomes they expect" p. 393). Bandura (1986) notes that outcomes are connected to actions; how one behaves largely determines the actual outcome and, in the same way, beliefs about outcome expectations are dependent on self-efficacy judgments. He gives the example that drivers who are not confident in their ability to negotiate a winding mountain road (low efficacy) will conjure up images of wreckage and injuries (one type of outcome expectation), whereas those confident in their ability will anticipate the grand views from the mountains. Recall the example of Mr. Dearborn at the beginning of this chapter. He remembers his anxiety when he began teaching and negative images about his classroom that were driven by his lack of efficacy for managing his classroom.Go to top of page

Similarly, in the academic domain, students' self-efficacy beliefs are very likely to be highly positively correlated with outcome expectations, represented by the top right and bottom left cells in Figure 3.3. Of course, as noted in the chemistry example, there can be occasions when students are high in efficacy but low in outcome expectations due to structural constraints in the environment such as grading curves. Another example of this high efficacy-low outcome expectation pattern would be the case of institutional discrimination on the basis of race, ethnicity, or gender. Students in the affected group (e.g., minorities in any school or classroom where they are discriminated against and women in math and science classes) might feel that they can master the material (high efficacy) but can't succeed due to the discriminatory practices in the setting. It should be noted that outcome expectations are beliefs, and in keeping with the general constructive perspective, students may perceive low outcome expectations due to discrimination when there may be very little actual discrimination in the setting.

This issue of accuracy in self-beliefs is even more important in terms of self-efficacy judgments. Bandura (1986), following all the other expectancy models, notes that people tend to avoid tasks and situations they believe exceed their capabilities, but they take on tasks and activities that they believe they can handle. This type of choice behavior can have a dramatic influence on personal development. His theory predicts that when self-efficacy perceptions are high, individuals will engage in tasks that foster the development of their skills and capabilities, but when self-efficacy is low, people will not engage in new tasks that might help them learn new skills. In addition, by avoiding these tasks, an individual will not receive any corrective feedback to counter the negative self-efficacy perceptions. In general, then, it is most adaptive to have self-perceptions of efficacy that slightly exceed actual skill level at any given time. Grossly optimistic efficacy beliefs that lead to tasks and situations that are far beyond the level of individual skill can result in quite aversive consequences. Consider an overly confident novice mountain climber who attempts a very difficult climb that results in a serious injury or even loss of life. In the academic domain, the consequences may not be as dire, but individuals who take on academic tasks far beyond their level of actual skill can suffer needless failure and subsequent debilitating efficacy beliefs. Individuals who grossly underestimate their efficacy, although the consequences might not be as aversive as in overestimation, will limit their potential for learning and development and, if they do undertake the task, will probably suffer from unnecessary anxiety and self-doubt that can increase the possibility of failure (Bandura, 1986).

Self-efficacy is related to choice behavior in terms of task choice, but it also has been related to career choices. For example, Betz and Hackett (1981, 1983; Hackett & Betz, 1981) have shown that although there are structural and social influences on career choices, self-efficacy is an important mediator of these external influences and has a direct bearing on career choice. In addition, they suggest that the gender differences that emerge in vocational choices are due to differences in self-efficacy; males are efficacious for all careers, whereas females are only efficacious for careers traditionally held by women and feel inefficacious for careers traditionally held by men.

Besides choice, self-efficacy has been related to the quantity of effort and the willingness to persist at tasks (Bandura & Cervone, 1983, 1986; Schunk, 1991b). Individuals with strong efficacy beliefs are more likely to exert effort in the face of difficulty and persist at a task when they have the requisite skills. Individuals who have weaker perceptions of efficacy are likely to be plagued by self-doubts and to give up easily when confronted with difficulties. However, there is some evidence that self-doubt (weak efficacy) may foster learning when students have not previously acquired the skills. As Bandura (1986) notes, "Self-doubt creates the impetus for learning but hinders adept use of previously established skills" (p. 394). Salomon (1984) found that students high in efficacy were more likely to be cognitively engaged in learning from media when the task was perceived as difficult, but they were likely to be less effortful and less cognitively engaged when the media were deemed easy.

Besides the quantity of effort, the quality of effort in terms of the use of deeper processing strategies and general cognitive engagement of learning has been strongly linked to self-efficacy perceptions (Pintrich & Schrauben, 1992). For example, Pintrich and De Groot (1990a) found that junior high students high in efficacy were more likely to report using various cognitive and self-regulatory learning strategies. In a series of experimental studies, Schunk (1982, 1983a, 1983b, 1983c, 1983d, 1984, 1987, 1991b) found that students who had stronger self-efficacy beliefs were able to master various math and reading tasks better than students with weaker efficacy beliefs. In addition, these studies showed that efficacy was a significant predictor of learning and achievement, even after prior achievement and cognitive skills were taken into consideration.

In sum, self-efficacy beliefs have been shown to be important mediators of all types of achievement behavior as well as many other types of behavior. Self-efficacy beliefs are very similar to the other expectancy constructs of task-specific self-concept and self-perceptions of competence because they all represent individuals' judgments of their capabilities. However, self-efficacy theory does differ from the other two traditions of research in assuming that self-efficacy perceptions are much more situation specific than the other expectancy beliefs. This assumption has led researchers to measure self-efficacy in a much more situationally sensitive fashion and at a much more microanalytic level (e.g., efficacy for very specific academic problems such as two-digit subtraction problems with borrowing). Related to this situational specificity, self-efficacy beliefs are assumed to be much more dynamic, fluctuating, and changeable beliefs than the somewhat more static and stable self-concept and self-competence beliefs. Self-efficacy beliefs, even for a very specific task such as subtraction problems in math, might fluctuate given the individuals' preparation on any given day, their physical condition (sickness, fatigue), and affective mood, as well as external conditions such as the nature of the task (length, difficulty, etc.) and the social milieu (general classroom conditions). In contrast, the other two traditions of research would assume a more global perception of competence (e.g., math competence) and, although they might recognize the existence of such daily shifts and fluctuations in self-perceptions of competence, would not be concerned theoretically or empirically with such microlevel instability of beliefs.Go to top of page

DEVELOPMENTAL AND GROUP DIFFERENCES IN EXPECTANCY CONSTRUCTS

In each of the three programs of research we discussed, there has been research on developmental and group differences. Given the similarity in the general construct of expectancy, however, we discuss developmental, gender, and ethnic differences together.

Developmental Differences

As noted in our discussion of the research on self-perceptions of competence there has been a great deal of research on the age-related changes in children's self-perceptions of competence. There is an important distinction in this developmental literature between the level and accuracy of children's self-perceptions of competence and their definitions of ability and effort (Blumenfeld, Pintrich, Meece, & Wessels, 1982). The latter have to do with the psychological meaning of ability and effort and will be discussed in Chapter 4 on attribution theory. In this chapter, we are concerned with the level and accuracy of children's self-perceptions of competence.

The research that has examined developmental differences self-perceptions of competence has consistently shown a decrease in the mean level of self-perceptions of ability as children move into adolescence (Eccles, Midgley, & Adler, 1984; Eccles & Midgley, 1989; Harter, 1990; Stipek & Mac Iver, 1989; Wigfield, Eccles, & Pintrich, forthcoming). In particular, the average level decrease seems to be the greatest when students move into junior high schools or middle schools in seventh or sixth grade, respectively (e.g., Eccles, 1993), although Marsh has found the lowest point to be in eighth or ninth grade (Marsh, 1989). Most of this research has been cross-sectional in nature, but the few longitudinal studies also show a decrease over time with age (Wigfield et al., 1991). There are a number of explanations for this age-related drop in mean level of self-perceptions.

First, there is a simple methodological explanation. Most measures of students' self-perceptions of competence use some version of a Likert-type scale whereby students rate themselves on one of several numerical scales from the lowest level of self-perceived ability (1) to the highest level of self-perceived ability (4, 5, 7, or 11, depending on the scale). It may be that young children (early to middle elementary grades) are more likely to use only the endpoints of the Likert scale, especially the higher end of the scale. Although this may be the case, it is not clear that this tendency to use the endpoints on a Likert-type scale reflects a methodological artifact of Likert scales per se. Assor and Connell (1992) argue that Likert scale items are acceptable and valid measures of children's self-perceptions of competence as long as some precautions are taken in their design and administration. They do note, however, that a four-point scale may provide all the needed precision with younger children and there may not be a need for scales with a larger range.

This brings us to the second general type of explanation for the developmental differences in perceptions of ability, which is more substantive and has to do with internal psychological mechanisms. There are two issues related to a substantive explanation of the overall drop; one concerns the accuracy of perceptions and the other concerns the overall optimistic perceptions of younger children. The accuracy of perceptions usually refers to whether the children's self-perceptions match or correspond to some external, more objective criterion. In general, it is probably best for individuals to have somewhat optimistic perceptions of their competence and efficacy (Assor & Connell, 1992), although not to be too overly optimistic and certainly not too pessimistic (the illusion of incompetence as pointed out by Phillips, 1984, 1987). Nevertheless, given the general social cognitive and constructivist perspective of current motivational theories, the accuracy of these perceptions in terms of their correspondence to objective measures of achievement is not as important as the fact that these perceptions do have motivational and achievement consequences. Assor and Connell (1992) review a number of studies of both the illusion of competence (children who overrate their competence, given their actual achievement) and the illusion of incompetence (children who underrate their competence, given their actual achievement) and conclude that in both cases these perceptions have important consequences. Students who overrated their competence were more likely to be engaged and achieve at higher levels, whereas those who underrated their competence were more likely to avoid tasks, report more anxiety, and not achieve that well. Assor and Connell conclude that these inaccurate perceptions are "real" because they have real consequences in motivational and behavioral terms. Accordingly, the accuracy of perceptions in terms of a match to external criteria like teacher ratings or achievement tests is not as important a validity consideration as whether measures of childrenís self-perceptions actually reflect the children's own perceptions of their abilities.

Given the argument that children's perceptions are real and younger children are more likely to have higher perceptions, we still need to account for this phenomenon, not just attribute it to inaccurate or invalid measures. A cognitive psychological explanation for this is based on younger children not having the information-processing skills available to integrate the information and make the types of social comparisons necessary (Blumenfeld et al., 1982). For example, Parsons and Ruble (1977) have shown that young children have difficulty utilizing all the cues presented to them to make accurate predictions about future performance. Other researchers have shown that children do not interpret feedback about task difficulty in the same manner as adults. Meyer and colleagues (Meyer et al., 1979) found that ninth graders and adults were more likely to believe that praise after success at an easy task implied low ability and criticism for failure at a hard task indicated high ability, but that third and fifth graders used praise as an indicator of ability regardless of the task difficulty level.

Besides these differences in information processing, there are a number of studies that suggest that younger children use less comparative standards or criteria to judge ability in contrast to the more relative standards of older children and adults. For example, Blumenfeld, Pintrich, and Hamilton (1987) have shown that younger children (second graders) are more likely to use an absolute standard to judge their competence and focus on completion of the task (I can read., I can do my math.). Older students (sixth graders) were more likely to invoke comparison information (I can do my math faster than other kids.) when asked to judge their competence. Ruble and her colleagues (Ruble, 1983; Ruble & Frey, 1991) have shown that younger children do not always seek out social comparison information, let alone use it to make a judgment of their ability. The lack of interest in and failure to use social comparison information by younger children suggest that they will maintain higher perceptions of ability. If younger children do not make comparisons of their performance to others, then they will be less likely to adjust their self-perceptions downward as they see some other children do better than they do on different academic tasks.

A third explanation for the drop in ability perceptions as children grow older is contextually based and focuses not on the changing cognitive and information-processing skills of children, but on the changing developmental environments through which they move. Obviously, one of the environment factors that changes is the nature of classrooms and schools. As a number-of researchers (Blumenfeld et al., 1982; Eccles et al., 1993; Marshall & Weinstein, 1984; Rosenholtz & Simpson, 1984) have pointed out, elementary and junior high schools differ in a number of organizational and structural ways that can influence students' ability perceptions. First, the nature of the class structure changes in junior high with children moving among several different classes and teachers for subject areas instead of having one teacher for most academic subject areas as in most elementary schools. Second, the nature of evaluation changes. Most elementary classroom teachers use a simple criterion-mastery system of grading. Children are graded on their performance on worksheets, projects, tests, and so on without reference to how others perform. In contrast, in junior and senior high classrooms, teachers often use some type of relative grading system whereby students' grades are partially determined in reference to how other children performed. In these cases, children are given information about their ability relative to other children. For those who do not do as well, it is not surprising that they begin to lower their self-perceptions of ability. in addition, tracking by ability level begins to emerge in the upper grades and this can have a detrimental effect on lower ability children's perceptions of their ability (Mac Iver, 1988; Rosenholtz & Simpson, 1984). In effect, many aspects of the classroom change as the children move to junior high and most of these changes serve to increase the information available to students upon which to make relative ability judgments. Not all students lower their ability judgments (Mac Iver, 1988; Reuman, 1989), but what both the cognitive developmental and classroom/school changes seem to do is to increase the variance in students' ratings, so that all children's ratings are not clustered around a higher mean. This situation results in more dispersion of ability self-perceptions and also a lower overall mean as children get older.Go to top of page

Gender and Ethnic Differences

Besides the developmental differences that are found in ability self-perceptions, there also have been studies that find gender and ethnic differences. However, there are a number of problems with making broad generalizations about gender and ethnic differences in motivation. First, by concentrating on overall group differences, many of the studies ignore the more significant within-group differences among females or among minority children. Second, it is often difficult to evaluate the relative contributions of gender and ethnicity to these group differences because the influence of socioeconomic status (SES) is not considered simultaneously (Graham, 1994, Pollard, 1993). Many of these studies confound SES with ethnicity by comparing middle or upper middle income Caucasian students with lower income African American students. Moreover, as Pollard (1993) points out, there may be important interactions among gender, ethnicity, and SES on childrenís motivation or achievement. Studies that do not take into account the multiple statuses and roles that students can play as a function of membership in different groups can be misleading and result in simplistic conclusions being drawn about one characteristic (e.g., ethnicity) when the reality is much more complex. Third, much of the research on ethnic differences has been atheoretical in terms of conceptualizing motivation and has often relied on deficit models and compared different ethnic groups on constructs that allow for simple dichotomies between those low and high in certain characteristics (Betancourt & Lopez, 1993; Graham, 1994; Pollard, 1993). Accordingly, any generalizations about gender or ethnic differences are problematic, albeit there are some trends that should be noted.

In most cases when a gender difference is found, the difference is that females have lower self-perceptions of ability than males (Wigfield, Eccles, & Pintrich, forthcoming). This is particularly surprising since many studies that have examined actual achievement or performance (see Linn & Hyde, 1989) show that there are few gender differences and that in many cases, females actually outperform males; yet other studies still show females as having lower self-perceptions. Although this discrepancy between actual achievement and self-perceptions of ability may be due to a response bias, with boys being more self-congratulatory and girls being more modest (Eccles, Adler, & Meece, 1984; Wigfield et al., forthcoming), the difference appears often enough to be taken seriously.

Eccles and Wigfield and their colleagues have consistently found gender differences in self-perceptions of ability. Males have higher self-perceptions in math and sports, whereas females have higher self-perceptions of their ability in English (Eccles, 1983; Eccles et al., 1989; Wigfield et al., 1991). Marsh (1989) also reports gender differences in his data on self-concept. Although he finds that the gender differences only account for 1% of the variance in self-concept, he finds that males have higher self-concept scores for their self-ratings of physical appearance, physical ability, and math, whereas females have higher self-ratings for verbal and reading tasks and general school self-concept (Wigfield et al., forthcoming). Phillips and Zimmerman (1990) also found that females had lower perceptions of their competence than males, although the gender difference did not emerge with third and fifth graders, only with ninth graders. Other studies, however, have found that gender differences in ability perceptions do emerge at earlier grades. For example, studies by Entwisle and Baker (1983) and Frey and Ruble (1987) both found that even in early elementary age children, females were more likely to have lower self-perceptions of ability than males. Clearly, there is a need for more research into the nature of these differences as well as more programs to change school and classroom practices that can give rise to these gender differences (Bailey, 1993; Meece & Eccles, 1993; Kahle, Parker, Rennie, & Riley, 1993).

Much less research has been done in terms of ethnic and racial differences, and the research often confounds race and ethnicity with social class differences by comparing middle-class white children with lower class minority children (Graham, 1989, 1994; Wigfield et al., forthcoming). There are two general issues that are often addressed in research that examines ethnic differences. First, are there ethnic group differences in the mean level of self-perceptions; of ability or other motivational constructs? Second, do the motivational constructs operate in the same fashion for minority group students as they do for other groups, or do we need different models of motivation for different ethnic groups?

In terms of the first question, Graham (1994) in a narrative review of published studies of African American students and their achievement motivation, finds little support for the general hypothesis that African American students should have lower expectancies for success or lower self-concepts of ability because of their poor school achievement or general economic disadvantage. In terms of expectancy for success, Graham (1994) reviews 14 experimental studies that used a common format of presenting a task to African American and Caucasian children and then asking them to predict their expectancy for success. In addition, some of the studies asked the students to make judgments of their expectancy for success after they had received either manipulated or actual success and failure. In 12 out of the 14 studies, African American students had higher expectations for success than Caucasian children. Graham (1994) also reviewed 18 studies that examined self-concepts of ability Again, she found very little evidence for the idea that African American students have lower self-concepts of ability. Only 2 of the 18 studies reported group differences in favor of Caucasian children, 7 favored African American children, and the remaining 9 had mixed or no significant differences between the two groups. Graham (1994) interprets these "counterintuitive" findings for the deficit hypothesis for African Americans in terms of the adaptive nature of maintaining optimistic expectancies and self-concept beliefs in the face of relative social and economic disadvantage.

In terms of the second issue, some studies find that self-perceptions of ability or efficacy are linked to academic achievement in the same fashion (e.g., moderate positive correlations) for minorities and other groups (Pollard, 1993; Taylor, Casten, Flickinger, & Roberts, 1994). However, Graham (1994) notes that in many of the studies she reviewed, the actual performance measures, such as grades or standardized achievement tests, showed that African Americans had lower levels of performance, yet they had higher self-perceptions of ability. This would suggest that the relation between self-perceptions of ability and actual achievement is not as strong in African American students as it is in Caucasian students. There have been many reasons proposed to explain this weaker relation. For example, some researchers (e.g., Fordham & Ogbu, 1986; Steele, 1988, 1992) have suggested that the motivational dynamics are different for black children and that they may devalue academic achievement (a task value belief) because of their repeated school failures. In this case, they may have relatively high self-perceptions of competence (or relatively low perceptions), but perceptions of competence are not linked as closely to actual achievement as they are in Caucasian children. Other reasons include the use of different social comparison groups; for example, African Americans compare themselves to other African Americans rather than the more advantaged Caucasian group, thereby maintaining high self-perceptions (Rosenberg & Simmons, 1971). They also attribute their lower performance to external factors such as prejudice, thereby maintaining high self-perceptions (Crocker & Major, 1989). Graham (1994), however, notes that these explanations are not supported in the general motivational literature and suggests that motivational principles that are used in current theoretical work on motivation should also apply to minority students. In this sense, it may be more important to examine within-group individual differences rather than between-group differences. For example, why do some minorities or females seem to maintain their achievement even in the face of many different obstacles? Graham suggests that there is less need for a differential psychology of minority student achievement, although more minority students must be included in future research on motivation, and samples must include a range of socioeconomic levels and different groups of minorities. In any event, there is a need for more research on these issues, both to evaluate the nature of the differences in perceptions of ability between minority and other children and to evaluate the claims that different theoretical models of motivation and development are necessary to explain the linkage between minority students' motivational beliefs about competence and their actual academic achievement.Go to top of page

IMPLICATIONS FOR TEACHERS

The research on expectancy, competence, and efficacy beliefs provides a number of implications for teachers. These general principles reflect the research, but given that psychology and educational psychology are probabilistic sciences, not deterministic ones, the principles may not apply in all contexts and situations. That is, psychology is not like chemistry where one can predict with great precision what will happen in a new situation based on previous research. Rather, findings from psychological research can be used to appraise the probability of different events occurring in the future. In this sense, research in psychology can be used as a guide for educational practice, not as the determiner of practice. In some cases, we can say that following certain principles will most likely result in positive motivational outcomes, but there are always exceptions and problematic situations where educators must use their best professional judgment. There is disagreement in the teacher education literature about whether teachers actually can use principled knowledge in their teaching or whether their pedagogical knowledge is all case based and develops from experience. Our position is that teachers can use and benefit from principled knowledge, albeit their use of it and their representation of it may be in the form of cases. Hence, we offer the follow suggestions as a guide for teacher practice. These principles will need to be adapted to the specific classroom context, but we trust that truly professional educators will be able to apply them in a thoughtful and pedagogically sound manner.

1. Help students maintain relatively accurate but high expectations and efficacy and help students avoid the illusion of incompetence. Mr. Dearborn tries to provide positive but accurate feedback to all students on their written work. He writes on their science reports about their demonstrated level of understanding of the content as well as their level of effort (at least as perceived by him). Some students do a very hasty report and he makes sure to note this in his written comments. He often speaks with the students individually to ask them why they did not try very hard or did not do very well. Given this information, he tries to help the students see how they can increase their effort or performance. Sometimes, he shows them other students' reports (with the names removed to protect confidentiality and limit social comparison) so they can see a model of a good report. He talks to the students about how they might go about doing a report at the level of the model. At the same time, Mr. Dearborn is very careful to make sure that all students know what they did incorrectly; he does not give insincere feedback to boost self-esteem.

As the research has shown, students are motivated to engage in tasks and achieve when they believe they can accomplish the task. Teachers need to provide accurate feedback to students to help them develop reasonable perceptions of their competence but, at the same time, communicate that their actual competence and skills will continue to develop.

2. Students' perceptions of competence develop not just from accurate feedback from the teacher, but through actual success on challenging academic tasks. Keep tasks and assignments at a relatively challenging but reasonable level of difficulty. Ms. Rivera has available many different levels of classroom assignments for her junior high school science students. Over the years, she has collected a number of different projects, experiments, labs, and workbook assignments that provide her with a diversity of tasks with differing levels of challenge. All the students in her class do a common set of seventh-grade science tasks, but she also has extra-credit assignments available for those who can go beyond the basic seventh-grade material. In addition, she has other extra-credit assignments that are below grade level that she uses with students who are struggling with the content. This mixture of common and individualized tasks allows most children the opportunity to be successful and at the same time be challenged.

Although practice on easy tasks is very helpful for building automaticity of skills, children also need to be challenged by tasks in order to be motivated and to actually learn new skills. Tasks should be set at a level of difficulty where most children in the classroom can master the assignment with some effort. They should not be too easy and especially not too difficult that most children fail at the task.

3. Foster the belief that competence or ability is a changeable, controllable aspect of development. At the beginning of this chapter, three students were talking about their self-perceptions of competence for social studies. Kevin had high efficacy beliefs, whereas both Rachel and Sam had lower beliefs about their capability to do social studies. In this case, if there are many students in the classroom who have the same level of beliefs as Rachel and Sam, the social studies teacher might want to have a class discussion about how different students study for social studies tests. During the discussion, different types of study strategies might be mentioned that some children don't know about. Most likely different levels of effort and amount of time studying would be mentioned as factors that influence test performance. The social studies teacher could then use the ideas generated in the discussion to help all the students in the class see that performing well on social studies tests is something that they all can learn to do. He could do this by reemphasizing the importance of effort, the amount of time spent studying, and the use of different study strategies. In addition, he could express his belief that all students can learn to do well in social studies. This type of teacher talk about social studies would communicate to students that the teacher has high expectations for all of them and that learning in social studies is not a stable ability or trait that some students have or do not have.

The vast majority of the knowledge and skills that are taught in K-12 schools can be learned by all children who do not have serious disabilities. Of course, some children may take longer to master the knowledge or skills than others, but there are very few inherent limitations that are stable traits of students. If students come to understand that they can master the material with some effort, they will be more likely to engage in the material. The teacher needs to communicate this type of positive high expectation for all students, high and low ability, females and males, and minority and others.

4. Decrease the amount of relative ability information that is publicly available to students. Ms. Morgan, a science teacher, has one bulletin board in her room that lists all the assignments for her class and all the students' grades for each assignment. She likes having the information posted so that everyone can see it. It helps her keep track of what students have done and it helps the students see what assignments they have finished and which ones they still need to complete. In particular, she finds that it is a very effective management tool because when students are done with their assignment for the day, they can check the board and see what else they have to do without bothering her. Since she started using this system, she is not constantly being asked by students what they should do next. This freedom allows her to work closely with small groups or individuals on their hands-on science projects without too many interruptions.

Ms. Morgan also has found as the semester goes along, that there are always a few students who fall behind in their work, get poor grades, and seem to give up. When she asks them what is going on, they often point to the board and say things like, "Look what I have to do. There's too much. I'm too far behind. What's the use?", or, "I'm too stupid to do science. Every day I come in here, some of the nerds make fun of me. They point to the low grades I got on the board and laugh. I really hate science and anyone who is good at it is just a nerd anyway." These comments are upsetting to Ms. Morgan because she wants her students to do well and also to like and enjoy science.

Some teachers facilitate social comparison by posting all students' scores and grades on wall posters in the room or by having students call out their test scores in class while writing them down in the grade book. These types of practices can increase the amount of social comparison information available to children and help to lower some childrenís (those doing less well) self-perceptions of competence. In Ms. Morganís case, her management practice with the bulletin board seems to be undermining some students' perceptions of their competence to do science. The students who are not doing so well are having negative interactions with some other students and making negative social comparisons about their ability to do science. The board presents to all students in the class, in a highly public manner, everyone's tests and assignment scores. Ms. Morgan should try to avoid these types of public practices that heighten the differences between students. She could keep the grading information private and use it in assigning final class grades, but it may not be that helpful for all students to have public access to it. If she still wanted to use the public bulletin board for management purposes, simple check marks for completion may serve just as well as the public posting of grades. On the other hand, there may be more individualized and more private ways (e.g., individual student folders or portfolios) to keep track of her students' completed work.

5. Students' perceptions of competence are somewhat domain specific and are not equivalent to global self-esteem. It is more productive for academic learning to help students develop their self-perceptions of competence rather than their global self-esteem. At the beginning of this chapter, Rachel, Kevin, and Sam all showed the domain specificity of their beliefs about school tasks and athletic tasks. Their teachers provide them with accurate feedback about their performance in their subject area. The teachers avoid global and nonspecific feedback ("You are a good person.", "You are all special in some way.", or "You should feel good about yourself.") in favor of specific feedback about their actual performance.

Although global self-esteem can be important for general mental health, in the academic domains, it is more important for students' learning that they have accurate feedback about their performance and begin to develop accurate and positive perceptions of their competence. General self-esteem improvement may not be that helpful, particularly when students can see that they can't do a certain type of math or science problem. In this case, older children from later elementary school onward will quickly surmise the insincerity of the praise and discount it in terms of their perception of ability to perform the specific task.Go to top of page

SUMMARY

The expectancy construct, in various guises, is one of the most important mediators of achievement behavior. Early research on expectancy constructs moved motivational psychology away from a dependence on a simplistic behavioral psychology to a more rational and cognitive paradigm that is still dominant today. Moreover, these early cognitive models of motivation stressed the importance of the individual's perceptions and beliefs as mediators of behavior, thereby focusing motivational research on the subjective and phenomenological psychology of the individual. These early theories focused on people's expectations for success and their value for the task. Current research on expectancy and value constructs continues in this tradition of focusing on these two general beliefs of the individual, although researchers do attempt to include contextual influences in their models.

There are three general research traditions on expectancy beliefs: expectancy-value theory, self-perceptions of competence research, and self-efficacy theory. Eccles and Wigfield and their colleagues have revised Atkinson's expectancy-value model by making it more social cognitive in nature to reflect the current cognitive paradigm of motivation. They have focused on students' expectancy for success and perceptions of ability for academic tasks in a number of large-scale correlational and longitudinal studies in schools. This research reflects some of the best ecologically valid studies of student motivation, given the focus on school and classroom tasks and the fact that they have followed the same students over time to actually investigate how motivational beliefs predict future behavior. They consistently find that students' expectancy beliefs about their capabilities to do a task and succeed at it are closely related to actual achievement on standardized tests as well as course grades.

The research on self-perceptions of ability has had a developmental focus and has investigated these self-perceptions in young children, adolescents, and adults. Current research suggests that self-perceptions of competence are domain specific, not global, although there is not agreement about the specificity of the domain. Self-perceptions of competence also are theoretically and empirically distinct from self-esteem. Perceptions of competence concern students' judgments of their capability in a domain and, hence, are more cognitive evaluations, whereas self-esteem is a more affective and global reaction to the self. Most students, even relatively young elementary school children, have relatively accurate self-perceptions of their own competence in specific domains. Moreover, when they are inaccurate, most children overestimate their competence, not underestimate it, although there are negative consequences for those who do underestimate their competence. Current research suggests that perceptions of competence are related in a reciprocal manner to actual achievement and performance.

Self-efficacy theory is a third model that stresses the importance of an expectancy construct for motivated behavior. Self-efficacy is based in social cognitive theory, which has its roots in the more behavioral learning theory, in contrast to the more personality-based expectancy-value theory and the developmental perspective of the perceptions-of-competence research. As such, self-efficacy theory assumes that self-efficacy is much more situation specific and defines efficacy in terms of judgments of capability to perform specific actions in light of specific goals. Research has consistently shown self-efficacy beliefs to be related to academic achievement and performance on standardized tests and actual school tasks in addition to self-report measures of cognitive engagement and self-regulated learning.

There are developmental, gender, and ethnic differences to consider in children's expectancy beliefs. In terms of developmental differences, research suggests that younger children are more likely to have relatively high perceptions of their competence and that the overall mean level of these perceptions declines with age. In particular, the research suggests that the decline is greatest when students make the transition to junior high school. There are both psychological and more sociological explanations for this drop. The psychological explanations focus on the changes in children's cognitive skills and beliefs as mediators of this age difference, whereas the more sociological explanations stress the changes in the nature of the school environment as the children move into secondary schools. The trends for gender and ethnic differences are much harder to summarize given the vagaries and confounding of gender, SES, and ethnicity in the samples or analyses in the research. Females do seem to have somewhat lower or inaccurate perceptions of competence given their performance, but there is little evidence for the idea that African Americans have lower self-perceptions of competence or lower expectations for success.Go to top of page