Strategies for Enhancing Teachers' Beliefs in Their Effectiveness: Research on a School Improvement Hypothesis
by John A. Ross - 1995
Reviews research on teacher efficacy, concluding that teachers who believe they are effective set more challenging goals for themselves and their students, take responsibility for student outcomes, and persist when faced with obstacles to learning. The article suggests that efforts to improve schools should include attention to teacher efficacy. (Source: ERIC)
Teachers beliefs in their effectiveness consistently predict desired student outcomes. This article argues that the achievement impact of Teacher Efficacy (TE) arises from goal-setting and attributional processes. Teachers who anticipate that they will be successful set more challenging goals for themselves and their students, accept responsibility for the outcomes of instruction, and persist through obstacles. These findings suggest that student achievement of cognitive and affective goals can be enhanced by strengthening TE. The hypothesis that school improvement will flow from enhanced TE has been tested in a variety of skill-development projects with mixed results. It is proposed that skill-development approaches be augmented by attending to teacher beliefs (particularly about the mutability of intelligence) and to conditions of teacher work.
In 1989, Susan Rosenholtz described research on teacher efficacy (hereafter TE) as being in its infancy.1 Our understanding of the origins and outcomes of teachers beliefs about their effectiveness has grown substantially since then, but the use of these findings in teacher development programs has not. The purpose of this article is to identify strategies for enhancing TE as a mechanism for school improvement. The article begins by defining TE, outlining its roots in social learning theory, and distinguishing it from related notions. The second section reviews evidence of a consistent association of TE with student outcomes and presents an argument for interpreting these correlations within a causal chain. The third section describes teacher development strategies that have been or could be used to strengthen teachers beliefs in themselves as competent instructors and argues that these strategies must be augumented with attention to teacher beliefs and conditions of teacher work.
DEFINING TEACHER EFFICACY
TE is a type of self-efficacy that can be distinguished from related constructs such as outcome expectancy, locus of control, and self-concept.
TEACHER EFFICACY IS A TYPE OF SELF-EFFICACY
TE measures the extent to which teachers believe their efforts will have a positive effect on student achievement. Most researchers (but not all)2 treat TE as a form of self-efficacy, defined by Bandura as individuals judgments of their ability to complete future actions.3 These appraisals are based on personal interpretations of past actions rather than on external criteria. Over time these interpretations stabilize as persistent, but not static, performance expectations. Expectations can be modified by new information, especially judgments about the results of subsequent efforts of oneself or peers undertaking similar tasks. Verbal persuasion (e.g., attempts by peers or supervisors to convince subjects that they are competent to perform the target actions) and physiological responses (e.g., physical symptoms communicating an inability to perform effectively) also contribute to expectations about future performance.
In Banduras theory, self-efficacy is a regulatory mechanism that influences behavior in four ways.4 Through cognitive processes, high self-efficacy contributes to the adoption of loftier goals, increased goal commitment, and the expectation that goals will be achieved despite setbacks along the way. Through motivational processes, high self-efficacy subjects take responsibility for the outcomes of their actions, attributing success and failure to their own efforts rather than to factors beyond their control. Through affective processes, those with high self-efficacy develop coping strategies enabling them to turn off negative thoughts that lower performance. Through selection processes, self-efficacy shapes lives by influencing the selection of activities and environments.
TEACHER EFFICACY IS A MULTIDIMENSIONAL CONSTRUCT
Although some researchers5 treat TE as a unitary trait, most distinguish two type. Personal TE is a teachers expectation that he or she can bring about student learning. General TE is the belief that teachers are able to do so despite the impact of environmental factors beyond their control. A few studies have examined within-teacher differences in TE, finding that a given teachers feelings of effectiveness fluctuate in response to the characteristics of different classes or subjects that he or she teaches6 and the instructional tasks undertaken.7
TEACHER EFFICACY CAN BE DISTINGUISHED FROM RELATED CONCEPTS
Some researchers8 have treated general TE as equivalent to out come expectancy but more recent formulations9 see it as distinct. Outcome expectations refer to the predicted consequences of an actionwhether the action will lead to the desired outcomes. Efficacy expectations are about the ability to perform actions, either ones own ability (personal TE) or that of teachers as a whole (general TE). It is possible to believe that certain actions will lead to desired actions and not be able to perform them, just as one can feel capable of doing things that one believes will have limited impact. Another construct closely related to TE is locus of control, which refers to teachers willingness to attribute student outcomes to their own performance. Locus of control differs from TE in that taking responsibility for outcomes does not mean that one anticipates that the outcomes will be positive. A teacher might have an internal locus of control grounded in an inability to perform teaching actions effectively.10 Finally, TE differs from self-concept in focus. Self-concept is a composite view of oneself, a global self-image.11 In contrast, TE is much more context-specific, varying across different activities and settings.
TEACHER EFFICACY AND STUDENT OUTCOMES
TE scores correlate with students cognitive and affective achievement. TE contributes to student achievement, primarily through goal-setting processes.
TEACHER EFFICACY CORRELATES WITH COGNITIVE ACHIEVEMENT
TE consistently predicts students cognitive growth, even when student ability is controlled, as it must be, because teachers with low TE are more likely to be assigned lower-achieving classes.12 Personal TE contributes to student achievement in curriculum domains involving language, such as reading, language arts, and social studies.13 In contrast, general TE contributes to student achievement in mathematics-based subjects.14 Combining personal and general TE into a single construct has produced mixed results: Rosenholtz found that total TE predicted mathematics and reading in grade 4, but not in grade 2.15 Armor and colleagues found it predicted large and consistent gains in reading achievement beginning in grade 6.16
One possible explanation for the interaction of the TE dimension with subject achievement might be that many teachers view mathematics as a talent that is given and see language as a set of skills that can be acquired. The extent to which teachers believe that natural endowments can be overcome by education (general TE) would thus play a larger role in mathematics and the belief that individual teachers are able to develop student skills (personal TE) would come to the fore in language.
The association of TE with cognitive achievement is based exclusively on correlational data. Even though the findings are consistent, the empirical relationship could be the product of an unexamined third variable or coincidence. Even if there is a causal link, its direction is not obvious: The claims of researchers that TE is a cause of achievement must be substantiated with argument, which will be attempted below.
TEACHER EFFICACY CORRELATES WITH AFFECTIVE GROWTH
TE consistently correlates with student acquisition of school-approved values and attitudes. For example, high TE is associated with enhanced student motivation, increased self-esteem, improved self-direction, and more positive attitudes toward school.17
Since the research linking TE with student affective outcomes comes from correlational designs, the findings are threatened by the same problems of reverse correlation, intervening variables, and coincidence noted for cognitive outcomes. One study provides stronger but not sufficient evidence of a causal linkage. Midgley and colleagues tracked students (N = 1329) over a year as they made the transition from grade 6 to grade 7 and compared student attitudes with the TE scores of their pre- and post-transition teachers. The researchers found that grade 6 students who had high TE teachers believed that they would do well and these students continued in such beliefs if their grade 7 teachers also had high TE. Student attitudes declined if they moved to classrooms taught by grade 7 teachers with lower TE. The latter group of students reduced their expectations about future performance and their beliefs about how well they were currently performing. The findings were especially strong for lower-achieving pupils.18
TEACHER EFFICACY CONTRIBUTES TO STUDENT ACHIEVEMENT
The claim that TE contributes to student outcomes is based on two arguments. Each argument proposes an intervening mechanism that might account for the covariation of student achievement with TE. A third argument considers the possibility of reverse correlation (student success increases teachers perceptions of their effectiveness) and concludes that the relationship between TE and student outcomes is to some unknown degree reciprocal.
Teachers with High Teacher Efficacy Set High Standards for Themselves
TE influences teachers behavior through goal-setting processes. In Banduras theory, the expectation that one will be successful encourages the adoption of more challenging goals and increases persistence.19 In contrast to those with low self-efficacy, individuals who anticipate that they will be effective set higher standards of performance for themselves, accept responsibility if the standards are not met, and respond to failure with renewed effort. These individuals persist because they believe that their diligence will be rewarded with success. One would predict from Banduras theory that teachers with high self-efficacy would set higher goals in their teaching.
Empirical evidence indicates that teachers with higher TE do set more ambitious goals, although the direct evidence is limited to preservice teachers. Brookhart and Loadman found that teachers beginning their careers with high confidence in their ability to perform various teaching functions were more likely to report that their reason for teaching content was to foster student development.20 In contrast, low TE teachers believed that the purpose of teaching was to cover the curriculum. Similarly, Czerniak and Schriver-Waldon found that preservice candidates with high personal TE chose instructional strategies based on their power to increase student learning, while the low TE candidates selected methods in terms of their potential to reduce noise and confusion.21
Evidence that high TE teachers have loftier aspirations can be implied from teacher decisions about their practice. High TE teachers are more willing and likely to implement new instructional programs, leading to the acquisition of new teaching skills.22 The result is that teachers with high TE are more likely to use instructional strategies that are powerful but difficult to acquire, such as small-group techniques, cooperative learning, and activity-based methods.23 These teachers are less likely to rely on approaches such as whole-class teaching,24 which are weaker but easier to adopt. High TE teachers are also more likely to involve parents in school conferences, volunteering, and home monitoring.25 If TE contributes to teachers learning and using more powerful teaching strategies, then student achievement is likely to be higher.
Many of the findings linking teacher practices to TE are based on self-reports. There is a risk that some subjects may be overreporting the use of strategies perceived to be desirable or that they may be implementing these techniques in inappropriate ways. The risk is somewhat offset by evidence that preservice and experienced teachers with higher TE receive higher ratings from their supervisors.26
TEACHERS WITH HIGH TEACHER EFFICACY SET HIGH STANDARDS FOR PUPILS
In Banduras account, causal attributions are closely linked to self-efficacy.27 Individuals with high self-efficacy, in contrast to those with low self-efficacy, attribute the outcomes of their actions to themselves rather than to factors beyond their control. From this theory one could predict that teachers with high professional efficacy would set higher standards for their students, would make students accountable for their behavior, and would persist until the students had met these goals.
Available research supports these predictions. Personal TE has been consistently linked to pupil-control ideology, particularly in preservice training. Teachers with high personal TE tend to promote student autonomy, are more likely to confront student management problems than to respond permissively, and are more successful at keeping students on task. Those with high general TE have more confidence in their classroom-management techniques and rate management problems as less severe. They are more humanistic in their orientations and are less reliant on custodial methods to control the class.28 If TE contributes to achievement-oriented student-management strategies, then higher achievement is likely to result.
Higher TE teachers may be more successful in producing student achievement because they attend to the needs of lower-ability students more closely. Ashton and colleagues found that low TE teachers concentrated their efforts on the upper-ability group; they had less regard for lower-ability students, viewing them as potential sources of disruption.29 In contrast, high TE teachers had positive attitudes toward low achievers, built friendly relationships with them, and set higher academic standards for this group than did low TE teachers. Midgley and colleagues observed that TE had a bigger impact on low than on high achievers, suggesting that lower-ability students are less certain about their competence and are more likely to be influenced by teacher expectations.30
There is consistent evidence that high TE teachers are more willing to develop programs for special-needs pupils within their own classrooms, rather than referring these cases to special services.31 But these findings are based on teacher responses to hypothetical case studies in which teachers made referral decisions on the basis of very limited information. In some instances teachers were making judgments about age groups with which they were not familiar.
By increasing expectations for lower performers and providing greater instructional support, high TE teachers may create changes in students perceptions of their academic abilities. As student efficacy becomes stronger, students may become more enthusiastic about schoolwork and more willing to initiate contacts with the teacher, processes that impact directly on achievement.32 Evidence that TE has a delayed impact on student achievement (one study found that teacher efficacy correlated with achievement in the spring, but not the fall33) is congruent with this view.
Student Achievement Contributes to Teacher Efficacy
Bandura argued that the most important source of information about self-efficacy is personal experience.34 Individuals generate expectancies about future performance based on their interpretations of how well they have handled similar tasks in the past. From this principle we would anticipate that if students are successful, their teachers would feel competent and would anticipate being successful in the future.
There is not much evidence to confirm or disconfirm this view and the evidence is ambiguous. TE is higher in higher-achieving schools, suggesting that TE is a consequence of student achievement.35 In addition, orderly behavior of students is one of the strongest predictors of TE. Teachers who report that student misbehavior and class-cutting are having a disruptive effect on their classrooms obtain lower TE scores.36 Teachers in low-achieving schools with disadvantaged or disruptive student populations are more likely to become estranged from their craft than teachers in schools with college-bound students.37
The credibility of this limited evidence is threatened by three problems. First, the direction of these relationships cannot be determined. Second, the influence of student success is mediated by teacher attributions. Student achievement would have an impact on TE only if teachers attributed student success to teacher efforts. Since internal attributions are characteristic of high but not low TE teachers, this might be a case of the rich getting richer and the poor being left alone. But teacher attributions take different forms.38 Some teachers engage in ego-defensive attributions, accepting responsibility for student success and attributing student failures to characteristics of the child or the environment. For this group student achievement could have only a positive effect on TE (if students learned); it would have no effect if students did not learn. Other teachers engage in counterdefensive attributions, giving students the credit for success and accepting responsibility for student failure (because the teacher failed to motivate pupils or was unable to find the instructional strategy that worked for a specific child). For this group, student achievement could have only a negative impact on TE (if students did not learn); there would be no effect if students did learn. Third, not all teachers believe that teaching outcomes are predictable. For some teachers, teaching is an act of artistry rather than engineering.39 This group would be less influenced by student success because the same actions might not lead to the same outcomes in the future. In a study of within- and between-teacher influences on TE, Cousins, Gadalla, and I presented evidence that teachers beliefs about the predictability of teaching influenced the extent to which their feelings of success in the classroom affected TE scores.40
These caveats weaken but do not eliminate the claim that student achievement has an independent effect on TE. Although the size of the impact cannot be determined, it is reasonable to conclude that there is an effect. The relationship between TE and student outcomes is reciprocal, at least in part.
STRATEGIES TO ENHANCE TEACHER EFFICACY
If TE is consistently associated with desired student outcomes, and if a defensible argument can be made that TE has an independent effect on these outcomes, would it be possible to increase student achievement by strengthening TE? A unidirectional model might pose two questions: Have interventions been devised that increase TE? If so, does increasing TE contribute to higher student achievement? It can be argued that as teachers acquire more skills, their perception of themselves as competent professionals will increase, particularly if they experience success in the classroom as they apply their new skills.
SKILL DEVELOPMENT STRATEGIES FOR STRENGTHENING TE
Several investigators have designed interventions to increase TE.
Entry into the profession is characterized by relatively rapid changes in TE, particularly when preservice candidates have their first practicum experiences. Hoy and Woolfolk conducted an exemplary pre/post comparison among three groups: education students who were to have their practice teaching experience in the coming semester, education students who were scheduled to practice teach in a later semester, and noneducation majors enrolled in a psychology course. The study found that the general TE of subjects who experienced classroom teaching declined while that of the two control groups was unchanged. Hoy and Woolfolk suggested that the decline may have been the result of preservice teachers rationalizing their inability to control unruly pupils.41 Other studies have reported a decline in general TE of preservice candidates after initial practice teaching.42 Even as general TE declines, there is some evidence that personal TE increases during preservice.43 This increased confidence could be the result of teachers recognizing that they are becoming more skilled in their craft.
Three studies attempted to enhance TE by modifying the preservice program. Guyton, Fox, and Sisk found that an eight-week summer residency program followed by a one-year supervised internship had no greater impact on TE than did a traditional preservice program. The study may have failed to find an impact because it did not measure TE before and immediately after practice teaching, the period of greatest growth. Instead the first TE measurement was not made until five months into the program. In addition, the validity of the findings was threatened by uncontrolled differences between the treatment and control groups (the treatment group had a higher proportion of black teachers and there were lower levels of parental support in the schools in which the treatment teachers worked).44
Volkman, Scheffler, and Dana were more successful. They found that a small sample of preservice teachers had higher end-of-the-year TE, after adjusting for pretest scores, than a control group who experienced traditional preservice. The treatment consisted of being assigned to a school that had a graduate assistant who conferred with the teacher after each lesson and met biweekly with each preservice candidate to discuss problems and solutions. The feedback procedures might have affected TE by influencing teachers interpretations of the outcomes of their teaching. The feedback might also have increased the absolute level of teacher effectiveness (by strengthening their teaching skills), which could have influenced their feelings of success. This study obtained a promising outcome, but the account of the investigation was brief, omitting such details as the duration of the treatment and the procedure for assigning subjects to conditions.45
Vitale and Romance designed a semester-long course to increase teachers knowledge of earth science concepts. Preservice teachers viewed videodisc presentations, completed workbook activities to consolidate their knowledge, and designed and delivered model lessons based on the concepts. The control group experienced a similar treatment with regard to the curriculum development and practice teaching components, differing only in the intensity of the review of key concepts. (The description of the control group treatment was terse.) There was evidence that the videodisc program improved teacher understanding of the science concepts and led to more positive attitudes toward teaching elementary science. On the TE variable, confidence in ability to teach science to elementary students, there were no significant differences between treatment and control groups.46 In this study increasing teachers disciplinary knowledge was not a productive strategy for enhancing their TE, but findings from other studies suggest that disciplinary knowledge has an impact: Teachers feelings about how well prepared they are to teach a given course correlate with their TE for that course.47 More generally, university coursework relevant to future teaching (e.g., science courses with laboratory requirements) and graduate-level training are correlated with TE.48 Although relatively little research has been conducted on how disciplinary preparation of teachers affects practice,49 there is evidence that competent experienced teachers adopt less powerful teaching strategies when they are teaching in an unfamiliar area.50 A preservice program that attempts to strengthen TE by improving disciplinary knowledge is worth another attempt.
Given this limited body of research one cannot say that the hypothesis (that TE can be strengthened by modifying the preservice program) has been tested. The most promising finding, that feedback might influence TE, came from an inadequately reported study. None of the studies examined whether TE increases led to changes in student behavior.
Six studies assessed the impact of skill development in-service education on TE. Two of these investigated the impact of an in-service program based on the Madeline Hunter model.51 This model provides teachers with the products of effective-schools research so that their classroom decisions will be based on consideration of content, characteristics of the learner, and principles of learning. Bolinger found that personal TE increased during Hunter-based in-service training, arguing that the training program provided teachers with effective teaching skills that increased their performance, thereby contributing to teachers confidence in their personal teaching abilities. General TE was unaffected because the in-service training made no attempt to influence teachers beliefs about the effectiveness of their peers.52 But the credibility of Bolingers findings was weakened by the absence of a control group, the failure to report the duration of the inservice and testing periods, and the use of change scores in a one-way analysis of variance rather than a more powerful processing technique. Corbitt traced the impact of a six-week Hunter program on nine resource teachers. There was no overall impact of the program on personal or general TE. Some teachers changed their practice; others did not. The response depended on the fit of the model with teachers preferred teaching practices.53
Dutton examined the impact of an in-service program to promote the use of cooperative learning techniques on personal TE. She found that teachers who had gone through the professional development program had greater confidence in their teaching abilities than those who had not had the experience. Dutton also found that certain variables in the in-service training (the provision of group sharing and problem solving during training) and variations in school follow-up to it (opportunities for discussion with colleagues, principal observation and feedback) were each associated with increased TE. Sharing experiences with peers may contribute to TE by confirming teachers beliefs about their own success and/or by vicarious participation in the successes of others.54 But Dutton presented no data to demonstrate that the groups were equivalent prior to the in-service (it appeared that volunteers were compared with nonvolunteers), the TE measure was unique to the study, the response rate varied considerably from school to school, and the investigator inflated Type I error by using a series of one-way anovas.
Cousins, Maynes, and I measured TE on three occasions during a cooperative learning in-service program extending over six months. TE scores were highly stable: There was a slight upward trend over the three occasions, but the differences did not reach statistical significance. Teachers who participated more extensively in the in-service program (e.g., who were more persistent in their attempts to implement cooperative learning techniques in their classrooms) showed higher increases in general TE than those who participated to a lesser degree. Personal TE was not affected. The study suggested that TE was most influenced by activities in which teachers shared their classroom experiences in small-group meetings held at local sites. These sessions were dominated by teachers with prior experience in using cooperative-learning techniques. Their reports may have persuaded less-experienced teachers that the teacher population could be successful (increasing general TE scores) while establishing a new standard that novices felt they were not meeting, even if they felt their skills had increased (which could constrain growth in personal TE). Changes in student attitudes toward cooperating with one another correlated with the TE changes of their teachers, although the correlations were not in the expected direction.55
The most persuasive evidence for a causal connection among in-service training, TE, and student achievement comes from a study by Stein and Wang. They observed changes in the TE of fourteen teachers during the implementation of a specific innovation over three school terms. Changes in teacher practice occurred between terms one and two, preceding changes in TE that developed between terms two and three. Teachers who changed the most in terms of their use of the intended practices showed the greatest increases in TE. Stein and Wang argued that achievement in the new teaching task fostered positive perceptions of teachers professional competence.56
In the final study, Guskey examined the impact of a mastery-learning workshop on teaching self-concept, considered here to be a form of TE. For one term following the workshop, treatment teachers taught one class in their usual way and a second using frequent feedback and corrective activities. Guskey found that teachers who implemented mastery learning increased achievement in the treatment class, whereas the nonimplementors and teachers in a no-treatment control condition showed no differences in outcomes between their two classes. Implementors who increased student achievement also became more willing to take responsibility for student successes and failures. But these successful teachers developed a more negative teaching self-concept. The workshop provided teachers with a new teaching strategy to compare with their existing techniques. When they tested it in their own classrooms they found that the new method was superior and their confidence in their existing techniques declined.57 The negative effect of initial implementation has also been observed by qualitative researchers. For example, Marx and colleagues reported an expert teacher who lost confidence in her professional competence when implementing a technique that gave students the responsibility to design their own investigations. She found that her previous methods had left students with inert knowledge they could not use. Another teacher lost the smooth flow of lesson elements when she attempted to have students work in groups. In both cases teachers recovered their confidence as students became more proficient.58
The results of attempts to increase TE through skill-specific training have been mixed, with four of the six studies reporting a change in TE following in-service training, although in one of these TE declined. Three studies found that TE had an impact on student achievement, but none of these unambiguously supported the unidirectional model of in-service training that predicts that increasing TE will improve student performance.
COMBINING SKILL DEVELOPMENT WITH TEACHER BELIEF STRATEGIES
Skill-development approaches to strengthening TE might be enhanced by encouraging reflection on teacher beliefs and practices, thereby addressing the cognitive underpinnings of teacher expectancies. One strategy might be to approach TE through teacher attributions. There is substantial evidence that high TE is associated with internal attributions.59 If teachers reflected on their impact on student learning they might become more willing to take responsibility for the outcomes of instruction, strengthening their TE.
A different strategy is to stimulate reflection on teacher beliefs about the nature of students. Dweck and Leggatt distinguished incremental theories of intelligence (the belief that ability can be increased) from entity theories (the belief that ability is immutable).60 Fletchers re-analysis of data from the High School and Beyond study found that differences in TE within a school were predicted by teacher beliefs about students ability to learn. Teachers with high TE were more likely to subscribe to the belief that ability is an acquired attribute.61
Another strategy might be to encourage teachers to modify their definitions of classroom success. Some teachers define it exclusively in terms of cognitive outcomes whereas others are more concerned with students social development. The latter are more likely to maintain their TE if achievement standards become inappropriate for an increasingly disadvantaged student population.62
Only one study could be located that combined attention to teacher beliefs with a skill-development strategy for strengthening TE. Ohmart evaluated the impact of an in-service program designed to increase the quality of education for black students by strengthening the TE of their teachers. The five-day program (offered in two blocks separated by two weeks) consisted of activities intended to revise teachers theories of intelligence (toward an incremental view) and between-session activities in which teachers used Madeline Hunters strategies to improve the performance of one low-ability child. The program had an immediate positive impact on participants TE (both personal and general), but the effect disappeared on the delayed posttest.63 The study suffered from severe sample attrition in the treatment group, the procedures for assigning teachers to treatment and control conditions were not described, and there was evidence that the control group was contaminated.
The intervention evaluated by Ohmart relied on lecture and persuasion to influence teachers conceptions. It may have had only a temporary effect because verbal persuasion is the weakest source of information influencing self-efficacy: Illusory boosts in efficacy are readily disconfirmed by the results of ones own actions.64 Subsequent information about student outcomes may have swamped the initial glow of success with the target student. Changing teacher efficacy may require a radical restructuring in conceptions about students, teachers, and learning.65 The treatment assessed by Ohmart may not have been of sufficient strength and duration to enable teachers to make the transition.
TEACHER EFFICACY AND CONDITIONS OF TEACHER WORK
Skill-development approaches to strengthening TE might be more effective if they were part of a broader effort to redesign conditions of teacher work. These strategies might be of two types: attempts to build on the positive impact of restructuring and mechanisms to reduce the negative effects of externally imposed change.
Building Collaborative School Cultures
Rosenholtz, in a path analysis involving a very large sample of elementary teachers (N = 1213), found that four variables under school control had a direct effect on TE: receiving positive feedback on teacher performance, collaboration with other teachers, parent involvement in the school, and schoolwide coordination of student behavior.66 Other investigators have supported these findings, particularly the correlation of TE with teacher collaboration. I found that teachers who intereacted more frequently with peer coaches (from their own schools and with expert teachers from other schools) had higher general TE.67 Miskel reported a similar pattern: Teachers who engaged in joint work, sharing important instructional decisions (such as lesson preparation) with other teachers or with learning-disability specialists, had higher TE.68 Hoover-Dempsey found that TE was higher among teachers who were more aware of the expectations of teachers in grades above and below them.69 Curriculum coordination within the school and within the district is positively associated with TE.70
Teacher collaboration might affect TE in several ways. Since there is evidence that teacher effectiveness is higher in schools with heightened teacher collaboration,71 it may be that teachers feel more efficacious because mutual help-giving has increased their ability to bring about learning. It may also be that higher collaboration facilitates the development of technical norms that individual teachers can use to reassure themselves about the quality of their work and sharpen their expectations about future performance.72 Collaboration might influence teachers perceptions of how effective they are by developing and maintaining shared appraisals. Newmann and colleagues found that degree of consensus within a school about efficacy was the greatest predictor of individual TE.73 Collaboration might also influence TE by contributing to the development of a prevailing ethos, a shared set of norms and understandings about purposes, that influences teacher and student outcomes.74 Fuller and Izu found that teacher-belief convergence was associated with TE.75 Schools that reached consensus on schoolwide planning, school-improvement ideology, and the integration of personal and school philosophies had higher TE.
On the other hand, increased collaboration might, in some instances, reduce the confidence of some teachers if they receive negative feedback on their performance from their peers.76 This might explain why Ashton and colleagues found TE to be greater in a low- than in a high-collaboration school.77 Collaboration is complexly related to other teacher and school variables. The interaction of age, experience, collaboration, and TE was explored by Chester.78 He found that teachers who were older when they began their careers experienced greater increases in personal TE in the first few months of teaching than teachers who were younger. Chester also found that the disadvantage of age for younger teachers could be overcome if they had opportunities to collaborate with experienced peers. Complex interactions have also been reported among collaboration, TE, and program implementation.79 In addition, an argument can be made that TE is an effectnot a causeof collaboration. For example, Rosenholtz found a reciprocal relationship between collaboration and TE.80
These findings suggest that stimulating collaboration among teachers might be a productive strategy for enhancing TE. But no such attempts have been reported and there are significant challenges confronting those who would implement the strategy. The first is that the relationships between collaboration and TE are mediated by factors that are not easily controlled. Failure to attend to these interactions might frustrate the initiative. Second, collaboration must be deep and substantive if it is to influence teachers expectancies about their impact on students. But this depth of collaboration is difficult to bring about,81 and doing so might intrude on teachers professional autonomy, a condition of work that itself is positively correlated with TE. Hargreaves identified a series of negative effects that arise when teachers thoughtfully chosen individualism is overpowered by bureaucratic requirements to engage in joint planning, a process he labelled contrived collegiality.82
Redirecting Supervisory Practices
By coordinating, supervising, and rewarding teachers, principals might be able to influence teachers appraisals of their performance, heighten the exchange of vicarious experience, and engage in verbal persuasion. There is some evidence that this is the case.
Leadership actions positively correlated with TE include emphasizing accomplishment, increasing teachers certainty about the worth of their practice, being responsive to teacher concerns, promoting an academic emphasis in the school, and providing supervision perceived to be useful by teachers.83 The relationships between perceived TE and leadership behavior are especially strong when individual efficacy is aggregated to the school level.84 Because these findings are correlational, it is not clear whether the principals contribution to school TE occurs by changing existing staff or by attracting different teachers to the school. It is also possible that principals might adjust their behavior in response to the expectations of their teachers: High TE teachers might elicit supportive practices from school administrators.
There have been no reported attempts to increase TE by modifying the behaviors of school principals. The strategy has its greatest potential if the supportive behaviors of principals focus on teachers interactions with students, attempting through constructive feedback to influence teacher appraisals of their work. Principals also have the opportunity to influence teacher collegiality, which might lead to increased TE.
Increasing Participation in School and District Decision Making
Giving teachers a greater role in decision making is an affirmation of their competence. This might influence teachers perceptions of their past effectiveness and increase their expectations of future success.
Having a greater role in school decision making is consistently correlated with TE.85 Since increasing participation in school decision making influences teacher satisfaction, it is no surprise that TE is higher in schools with satisfied teachers, as measured by commitment to teaching, willingness to stay in the profession, satisfaction with current role, and willingness to re-choose teaching as a career.86 But participation in decision making is associated with increased productivity only when it focuses on instructional rather than managerial decisions.87
Reducing the Negative Effects of Externally Imposed Change
Several studies have found that state or national efforts to improve schools can have a negative effect on TE. Rosenholtz, in one of two studies investigating the effect of reform on TE, found that a statewide minimum competency testing program reduced teachers autonomy and feelings of success.88 They had to cut important topics and adopt a pace that was inappropriate for their students. The tests demanded high preparation time, reducing teacher-student interaction time. The testing program increased the tendency of some teachers to attribute student failures to external forces beyond their control. TE declined, except for a small group of teachers with classes similar to those of the developers and who shared the curriculum conceptions of state organizers.
Rosenholtz found that a career-ladder plan had negative effects on TE when teachers were excluded from setting evaluation standards and poor teachers were promoted.89 A sense of injustice prevailed that reduced effort. Further decreases in TE occurred when teachers were given little feedback by the evaluators and when teachers concluded that the portfolio method of assessment was unrelated to actual teacher performance. In contrast, Rosenholtz found positive effects, as did Ebmeier and Hart, when teachers were involved in designing the career-ladder scheme.90 Little s review of research on mentoring programs suggests that teacher leadership programs are likely to have a positive impact on the TE of those selected for leadership roles if they demonstrate their expertise to teachers.91 But leaders are judged on their consulting, not their teaching skills, and feelings of success as a leader may not transfer to perceived effectiveness in the classroom. The impact of leadership programs on the TE of unselected teachers is likely to be negative if teachers interpret their nonselection as a critique of their teaching.
McKeiver, Hogaboam-Gray, and I found that a government initiative to detrack grade nine had a negative effect on the TE of math teachers.92 These teachers felt capable of teaching different ability groups in separate classes, but found their skills could not be readily integrated to teach a mixed-ability group. This study also found that the negative effects of restructuring dissipated over time. There was a resurgence of teacher confidence as teachers developed new ways of working with heterogeneous classes and discovered that achievement, particularly of lower-ability performers, exceeded teacher expectations. The renewal of TE was associated with personal coping strategies (especially certainty about professional goals and control of emotional states) and social processes (particularly collaboration with same-subject peers).
A case has been presented for the hypothesis that changing teacher beliefs about their effectiveness will have a positive impact on students cognitive and affective achievement. But empirical tests of the hypothesis have produced mixed results, suggesting that TE may not be an inviting entry point for school improvement efforts.
The lack of consistent confirmation may be attributable to the weaknesses of the treatments that have been introduced. By focusing exclusively on the development of teachers skills, researchers have failed to attend to the cognitive underpinnings of teacher expectancies. Teachers may become more skillful, yet be no more confident about their future success than they were before, if teacher theories about the immutability of student intelligence are not challenged or if teachers maintain criteria for defining classroom success that are unattainable for the clients they serve.
Another factor that might account for the flaccid impact of skill-development strategies on TE is the failure to attend to the conditions in which the skills will be performed. Increased skill in teaching may not lead to increased confidence if the positive effects on students are not recognized by peers and supervisors or if teachers have so little control over their professional lives that they are unwilling to attribute school outcomes to their own efforts.
More powerful strategies for influencing TE can be built by extending skill-development approaches, for example, by confronting teacher beliefs that inhibit TE and by enhancing collaboration and control in the work place. But there are additional considerations that must be examined before renewing the vitality of teacher confidence becomes a credible strategy for educational reform.
The first of these concerns the multidimensionality of TE. Interventions to strengthen TE have tended to be diffuse, treating TE as if it were a trait, rather than targeting specific dimensions. And it has been assumed that change in one constituent element of TE will generalize to others. The problem is that the dimensions that constitute TE have not been well defined, nor have the within-construct correlations been measured. Studies examining within-teacher variance in personal TE suggest there might be a relatively stable core surrounded by dimensions that fluctuate in response to the characteristics of specific teaching assignments.93 Further research, involving a variety of teacher populations, may be required to refine conceptions of what TE is before practitioners can intensify their current efforts.
A second concern is the simple unidirectional model (change in skill leads to change in TE) that has guided much of the research. A more productive set of predictions might posit a process that is more interactive and open to influence of other variables. In addition, past findings suggest that a curvilinear representation of change might be more accurate. In this model TE might fluctuate through the change process: (1) High TE might contribute to the implementation of new teaching ideas by influencing teachers goal setting; (2) TE could decline as the new techniques disrupted the smoothness of existing practice; (3) TE might remain depressed even with early success if the perceived superiority of the new techniques persuaded teachers of the inadequacy of their routine practice; (4) TE might begin to increase as teachers integrated the new methods into their repertoire and began to enjoy increased student performance consistently; (5) enhanced TE might motivate the search for new skill-development opportunities. The multistage conception of the relationship between TE and change in teacher practice might supplant the debate between those who believe that alterations in teaching practice precede changes in teachers beliefs and those who believe the reverse.94
These considerations suggest prescriptions for future research exploring the hypothesis that school effectiveness can be enhanced by developing teacher confidence. First, researchers should build multiple treatments that combine skill-development efforts with attention to teacher cognitions and work-place constraints. Second, efforts to apply TE research to school improvement should be integrated with research efforts to identify the stable and variable elements of the construct and to clarify the interrelationships among its dimensions. Third, the research must be longitudinal. Planned change studies are likely to be the most productive but they could be accompanied by naturalistic observations during events that impact substantially on classroom practice (e.g., detracking, mandated testing programs) or professional relationships (e.g., career-ladder schemes).
Research on TE is no longer in its infancy. Enormous strides have been made in the last five years in our understanding of the construct and its influence on student outcomes. Although the results of attempts to use this newly acquired knowledge have been inconsistent, the next five years hold more promise. Harnessing what has been learned about TE to practical ends is within reach. But whose reach will it be? TE, a personal characteristic of teachers, could be manipulated as a tool for implementing reforms designed by nonteachers. Yet vigorous teacher confidence is inextricably linked to teacher control of instructional decision making. A limiting condition must be added to the hypothesis: Strengthening TE will contribute to enhanced student achievement only if the tactics used nurture teacher professionalism.
This research was funded by the Ontario Ministry of Education of Training through a grant to the Ontario Institute for Studies in Education and by the Social Science and Humanities Research Council of Canada. The views expressed in the article do not necessarily represent the views of the ministry or the council.