+ All Categories
Home > Documents > Performance-based teacher education - Gage & Winne 1975 PBTE Chapter

Performance-based teacher education - Gage & Winne 1975 PBTE Chapter

Date post: 20-Oct-2015
Upload: joe-chee
View: 40 times
Download: 0 times
Share this document with a friend
Gage, N.L., & Winne, P.H. (1975). Performance-based teacher education. In K. Ryan (Ed.), Teacher Education (74th Yearbook of the National Society for the Study of Education. Part 2) (pp. 146–172). Chicago, IL: University of Chicago Press.
CHAPTER VI Performance-Based Teacher Education N. L. GAGE AND PHILIP H. WINNE The dominant new movement in teacher education during the last decade has unquestionably been performance-based teacher education (PBTE). Judging from the volume of literature, the number of organizations, and the breadth and depth of the contro- versies concerned with PBTE, every student of teacher education must consider this new approach. In this chapter we first define PBTE and then consider briefly its origins and short history. Next, we explore four basic issues: humanistic objections to PBTE, the relationship between teacher performance and student achievement, the foundations of train- ing, and the cost of PBTE. If PBTE is in some degree a func- tioning reality it is necessary to ask how well it achieves its ob- jectives. Thus, we also consider problems that arise in evaluating PBTE programs. The chapter ends with a statement of our own position. A Definition of Performance-Based Teacher Education PBTE cannot be adequately portrayed in a brief definition. Differing and slightly contradictory definitions have been offered by advocates of the movement/ and also by its opponents. 2 For working purposes a brief definition would be that PBTE is teacher training in which the prospective or inservice teacher acquires, to a prespecified degree, performance tendencies and capabilities that 1. \V. Robert Houston, "Competency-Based Education," in Exploring Competency-Based Education, ed. W. Robert Houston (Berkeley, Calif.: McCutchan Publishing Corporation, 1974), pp. 3-15. 2. Harry S. Broudy, A Critique of Performance-Based Teacher Education (Washington, D.C.: American Association of Colleges for Teacher Education, 1972).


Performance-Based Teacher Education


The dominant new movement in teacher education during the last decade has unquestionably been performance-based teacher education (PBTE). Judging from the volume of literature, the number of organizations, and the breadth and depth of the contro­versies concerned with PBTE, every student of teacher education must consider this new approach.

In this chapter we first define PBTE and then consider briefly its origins and short history. Next, we explore four basic issues: humanistic objections to PBTE, the relationship between teacher performance and student achievement, the foundations of train­ing, and the cost of PBTE. If PBTE is in some degree a func­tioning reality it is necessary to ask how well it achieves its ob­jectives. Thus, we also consider problems that arise in evaluating PBTE programs. The chapter ends with a statement of our own position.

A Definition of Performance-Based Teacher Education

PBTE cannot be adequately portrayed in a brief definition. Differing and slightly contradictory definitions have been offered by advocates of the movement/ and also by its opponents.2 For working purposes a brief definition would be that PBTE is teacher training in which the prospective or inservice teacher acquires, to a prespecified degree, performance tendencies and capabilities that

1. \V. Robert Houston, "Competency-Based Education," in Exploring Competency-Based Education, ed. W. Robert Houston (Berkeley, Calif.: McCutchan Publishing Corporation, 1974), pp. 3-15.

2. Harry S. Broudy, A Critique of Performance-Based Teacher Education (Washington, D.C.: American Association of Colleges for Teacher Education, 1972).

Text Box
Gage, N.L., & Bown, P.H. (1975). Performance-based teacher education. In K. Ryan (Ed.), Teacher Education (74th Yearbook of the National Society for the Study of Education. Part 2) (pp. 146 -172). Chicago, IL: University of Chicago Press.


promote student achievement of educational objectives. "Teacher performance" refers to observable behaviors, both verbal (oral and written) and nonverbal. "Tendencies" refers to what the teacher typically does in the average or normal teaching situation. "Capa­bilities" refers to what the teacher is able to do when trying his best. Both tendencies and capabilities are assessed in terms of an explicitly stated level of mastery so that if the teacher does not perform at or above this level, he is considered to be inadequately trained. Finally, the performance tendencies and capabilities are selected and defined with reference to their effects on student achievement.3 Student achievement is used here to refer to edu­cational objectives of all kinds-cognitive, social-emotional, and psychomotor.

Beyond these core elements of PBTE several ancillary features can be described. First, while the degree to which the teacher's performance has consistent causal relationships to student learning may be known or unknown, student learning remains the ideal criterion for deriving valid teaching actions. Short of this ideal, teacher performances are derived from explicit conceptions of teacher behaviors hypothesized to promote students' learning.4

These behaviors may be assessed in several ways: in the context of teaching an actual class, in microteaching or some other approxi­mations to actual teaching, or merely by paper-and-pencil mea­sures of the trainee's knowledge of "what to do when."

Second, PBTE is often guided by a systematic, cybernetic, and individualized model of instruction, a model adapted to the needs of each trainee. In this sense PBTE is seen as a training program in which teaching performances are derived, defined, trained, and evaluated in an integrated and systematic way. Corrective feed­back follows evaluation. If the predefined level of mastery is not achieved by the trainee, further instruction, perhaps in a dif­ferent form, may be provided, and the training may be reevalu-

3· This definition most closely corresponds to that proposed by Stanley Elamin Performance-Based Teacher Education: What Is the State of the Art? (Washington, D.C.: American Association of Colleges for Teacher Educa­tion, 1971).

4· See Marsha Weil, "Deriving Teaching Skills From Models of Teaching," in Houston, op. cit., pp. IIS-43·


ated. The training may be individualized by modifying the per­formances to be trained, the method of training, and the rate of training, although most existing PBTE programs individualize only the last of these. The individualization often results in the extensive use of instructional modules, which are discussed later in this paper.5

There has been widespread confusion about these ancillary fea­tures. Hence it should be emphasized that "the specific technique (for instructing trainees) is not unique to the concept of perform­ance-based instruction and therefore does not enter into the defi­nition." 6

The definition of PBTE has called fonh comparisons of the terms "performance-based" and "competency-based." For some writers, teaching competency refers to cognitive knowledge only. For others, teaching competencies eritail effects on students' learn­ing. 7 In contrast, performance falls between these conceptions, de­noting the ability to perform according to a model of teaching. In our judgment, all three elements-knowledge, effects, and per­formance-are essential to an adequate conception of PBTE, in that the teacher must know what to do, must be able to perform according to tbis knowledge, and must promote students' learn­ing. We use the term PBTE to denote all three competencies.

The Origins and History of PBTE

One source from which PBTE emerged was behavioral psy­chology and its application to training, particularly in industrial and military settings. In such training, the repertoire of skills to be achieved was analyzed systematically to specify less complex

5· See Robert L. Arends, et al., Handbook for the Development of In­structional Modules in Competency-Based Teacher Education Programs. (Syracuse, N.Y.: Center for the Study of Teaching, 1971).

6. American Association of Colleges For Teacher Education, Committee on Performance-Based Teacher Education, Achieving the Potential of Per­ftn'711mlce-Based Teacher Education: Recommendations. (Washington, D.C.: American Association of Colleges for Teacher Education, 1974), p. 8.

1· James M. Cooper and Wilford A. Weber, "A Competency-Based Sys­tems .Approach to Teacher Education," in Competency-Based Teacher Educa­titm: z. A Systems Approach to Program Design, eds. James M. Cooper, Wil­ford A. Weber, and Charles Johnson (Berkeley, Calif.: McCutchan Publishing Corpontion, 1973), pp. 7-18.


behavioral components and their interrelationships. Training usu­ally began with verbal instruction in the behavioral components, then proceeded with practice in performing the skills to be learned, and the practice was followed by corrective feedback. Additional cycles of instruction, practice, and feedback would be provided as needed. The individual components were integrated so as to approximate more and more closely the complete repertoire.

This general strategy was first transferred to education in the form of programed instruction. Although not successful in the form or to the extent originally proposed, programed instruction made apparent the advantages of (a) detailed analyses of educa­tional objectives in behavioral terms, (b) sequencing and organ­izing elements of the knowledge and skill to be acquired, and (c) individualizing the rate at which students progressed.

In the middle 196os the model was adapted to teacher educa­tion in the form of microteaching.8 Teaching strategies were ana­lyzed into relatively discrete teaching skills. The trainee then practiced these skills individually with a small number of students for a brief period. Following the microteaching, the trainee re­ceived corrective feedback and usually did additional microteach­ing.

Subsequently microteaching was formally incorporated into more comprehensive teacher training curricula in the form of "minicourses" consisting of self-contained packages of teacher training materials.9 Variations of these kinds of teacher training methods are generally found in contemporary PBTE programs.

Symptoms of inadequacy in teacher education also spurred the development of PBTE. Among these, the chronic discontent with teacher education must be ranked foremost. For many decades educators and laymen alike had expressed dissatisfaction with the status quo in teacher education. Such expressions became especially

8. Dwight Allen and Kevin Ryan, Microteaching (Reading, Mass.: Addi­son-Wesley Publishing Co., r¢9); David C. Berliner, "Microteaching and the Technical Skills Approach to Teacher Training," Technical Report No. ro8 (Stanford, Calif.: Stanford Center for Research and Development in Teaching, r¢9).

9· Walter R. Borg, et al., The Minicourse: A Microteaching Approach to Teacher Education (Beverly Hills, Calif.: Macmillan Educational Services, 1970).


strong in 1963 with the critique by Conant 10 and the condemna­tion by Koerner.11 Further evidence of malaise was obvious in widely read complaints by teachers concerning the usefulness of their training in actual classroom practice.12

The dissatisfaction with teacher education was accompanied by demands for teacher accountability-demands that teachers be held responsible for the achievement, or lack of achievement, of their students. But practicing teachers replied that their training had not provided them with the skills and strategies necessary for insuring student achievement to the degree demanded by proponents of ac­countability. Thus, reform in teacher education was urged by teachers, educational statesmen, journalists, parents, and taxpayers.

Recognizing the discontent and the demand for reform, the United States Office of Education invited proposals in 1967 for the development of model elementary teacher education programs. Approximately one hundred responses were received, ten of which were developed further with federal support. A common element in all ten models was an attempt to develop programs that would more effectively foster the skills needed in teaching.13 This focus on new methods for training teachers in essential teaching skills signaled a turning point in teacher education. The models called for modification of traditional curricula to incorporate in some form an analysis of complex teaching strategies into specific teach­ing skills, explicit skill practice, and corrective feedback.

The American Association of Colleges for Teacher Education and other organizations sponsored conferences and committees that built upon the major themes of the model elementary teacher edu­cation programs. The establishment of new and federally supported

10. James B. Conant, The Education of American Teachers (New York: McGraw-Hill, 1¢3).

u. James D. Koerner, The Miseducation of American Teachers (Balti­more, Md.: Penguin Books, 1¢3).

12. John Holt, HO'W Children Fail (New York: Pitman Publishing Cor­poration, 1¢4); Herben Kohl, 36 Children (New York: New American Library, 1¢7).

13. Joel L. Burdin and Kaliopee Lanzillotti, A Reader's Guide to the Com­prehensive Models for Preparing Elementary Teachers (Washington, D.C.: ERIC Clearinghouse on Teacher Education, 1969).


teacher training programs, such as the Teacher Corps, also accen­tuated the need for concrete, replicable, and effective methods for educating teachers.

Where and when the term "performance-based teacher educa­tion" first appeared is unknown to the authors. In any event, the many publications,14 conferences, research projects, and other in­dices reflect much serious consideration of the concept. Indeed, approximately twenty states have already adopted or are planning to adopt a performance-based approach to teacher education and certification.l5 Few movements in education have received such vigorous attention so quickly. The rapid growth has been accom­panied by internal dissension and external criticism on basic issues.

Basic Issues in PETE

Should PBTE be supported? Should existing teacher training programs become performance-based? These questions require us to deal with several substantial controversies concerning PBTE. Major issues have been further complicated by minor matters of vocabulary and perspective. Thus, many disputants speaking to the same issue talk past each other.

We consider here four central problems and judge most of the other issues identified in the literature to be merely variants or rephrasings of these basic concerns: (a) the humanistic criticisms of PBTE; (b) the relationship between teachers' performances and student achievement; (c) the trainability of desirable teacher behaviors; and (d) the costs incurred in developing and installing PBTE programs.


One strong protest against the PBTE movement was that made by humanistic educators. Advocates of this position claimed that

14. See Allen A. Schmieder, "The Last Relatively Complete Tentative Bibliography on Competency-Based Education," in Houston, op. cit., pp. 367-426.

15. John G. G. Merrow, "The Politics of Competence: A Review of Competency-Based Teacher Education" (Washington, D.C.: Office of Re­search and Exploratory Studies, National Institute of Education, 1974); Melvin G. Villeme, "Competency-Based Certification: A New Reality?" Educational Leadership 31 (January 1974): 348-49.


PBTE has philosophical and practical underpinnings deleterious to teachers and students alike. In a paper representing this perspective, Combs outlined the critical features of a humanistic teacher edu­cation program: (a) effective teacher education is highly personal and dependent on the prospective teacher's development of an ap­propriate system of beliefs; (b) educating effective teachers is a process of promoting the "becoming" of a teacher, rather than one of educating a person in how to teach; (c) "becoming" an effec­tive teacher has its origins in security and acceptance; (d) teacher education should emphasize meanings rather than behaviors; and (e) teacher education should focus on the teacher's subjective im­pressions, with less emphasis on objectively gathered information about the processes and effects of teaching.16

Is PBTE the antithesis of these and similar positions? In our view, it is not. Since PBTE does not stipulate training of one sort or another, it can respond to these implicit criticisins. The emphasis on individualization could insure that PBTE programs will attend to the needs of each prospective teacher. But, more important, each of Combs's assumptions can be met by core elements of the PBTE approach. Fostering the trainee's system of beliefs, security, and acceptance can readily result from the instruction and practice that are part of PBTE. The process of "becoming" a teacher can be encouraged by the practice of teaching skills. Also, teachers should not rely solely on objective data and exclude "meaning" from assessment and adaptations of their teaching. Indeed, PBTE should require that each teacher search for meaning in experience. One source of such meaning can be objective data from research, but this is not the only source. Finally, to the extent that human­istic teacher education defines goals for its trainees, specifies meth­ods for achieving these goals, and determines the achievement of the goals with measures of teacher performance and student achievement, it is itself performance-based. Thus, the humanistic and PBTE orientations are not necessarily antagonistic, and, in our judgment, benefits can be derived from their concurrent use.

16. Arthur W. Combs, ''Some Basic Concepts for Teacher Education," Journal of Teacher Education 22 (Fall 1972): 286-l)o.




As defined earlier, PBTE depends on the ability to specify teacher performances that promote student achievement. Ideally, evidence on such teacher behaviors would take the form of in­ferences of causal relationships drawn from experiments in which a teacher behavior was manipulated as the independent variable and student achievement was measured as the dependent variable. Furthermore, the ideal experiment would have students randomly assigned to teachers and teachers randomly assigned to treatments, with appropriate controls for extraneous variables which might af­fect the relationship between teacher behavior and student achieve­ment. Because of the way schools are currently organized, it is extremely difficult to carry out such true experiments on the effects of teacher behaviors. In fact, the authors know of no such research. Instead, most experimental studies of teaching are quasi experiments in which students are not randomly assigned to teachersP The findings of these few quasi-experimental studies thus far have pro­vided only an extremely limited empirical basis for PBTE pro­grams.

At a third level of utility for identifying teacher behaviors related to student achievement are nonexperimental (that is, corre­lational) studies. In such studies teacher behaviors are observed as they occur naturally and without manipulation. The relationships between measures of teacher behavior and student achievement are then examined with correlation coefficients or other statistics. More than one hundred nonexperimental studies relating naturally occur­ring teacher behaviors to student achievement have been identified.18

The relationships obtained in these studies are generally low. The typical correlation coefficients between a single kind of teacher

17. Barak Rosenshine, Teaching Behaviours and Student Achievement (Lon­don: National Foundation for Educational Research, 1971); Michael J. Dun­kin and Bruce Biddle, The Study of Teaching (New York: Holt, Rinehart & Winston, 1974); Donald T. Campbell and Julian C. Stanley, ''Experimental and Quasi-Experimental Designs for Research on Teaching," in Handbook of Research on Teaching, ed. N. L. Gage (Chicago: Rand McNally, Inc., 1¢3), pp. 171-246.

18. Rosenshine, op. cit., Dunkin and Biddle, op. cit.


behavior and a measure of student achievement, adjusted for stu­dent aptitude, is of the order of .25. Since the average nonexperi­mental study involves only about twenty-five teachers,19 and since a correlation coefficient must equal ·32 to be statistically signifi­cant at the .05 level when N = 25, the general conclusion drawn by readers of the reviews by Rosenshine and by Dunkin and Biddle is that teacher behaviors of the kind studied thus far are usually unrelated to student achievement.

Thus the correlational studies to date seem at first glance to provide no foundation for identifying significant types of teacher performance. Not only are the correlation coefficients low, but they are generally statistically nonsignificant. And in much re­search on the same type of teacher behavior, the correlations do not even consistently have the same sign, so that some studies indi­cate positive relationships while others indicate negative relation­ships. Several researchers have addressed the shortcomings of the work to date and have offered recommendations for improving research on teaching.20 These recommendations, however, speak to future possibilities for PBTE. The issue at hand is the value of the evidence obtained thus far.

A case can be made for the proposition that the existing corre­lational results have some positive value. One path to extract­ing positive value from existing research is to consider the import of clusters of studies of a given kind of teacher behavior rather than their significance when considered one at a time. The tech­nique for combining evidence advocated here is a variant of what Light and Smith termed "vote-counting." They described this method as follows:

19. Indeed, this is the median sample size for fifty nonexperimental stud­ies reviewed by Rosenshine, op. cit.

zo. Dunkin and Biddle, op. cit., pp. 419-47; Barak Rosenshine and Marilyn Martin, "Teacher Behavior and Education," paper presented at the annual meeting of the American Educational Research Association, 1974; Philip H. Winne, "Measuring and Evaluating Teacher Knowledge, Performance, and Outcome Competencies: Taming Mischievous Demons," paper presented at the annual meeting of the American Educational Research Association, 1974; N. L. Gage, "Evaluating \Vays to Help Teachers to Behave Desirably," paper presented at the National Conference on Competency Assessment, Research, and Evaluation, University of Houston, March 13-15, 1974-


All studies which have data on a dependent variable and a specific independent variable of interest are examined. Three possible outcomes are defined. The relationship between the independent variable and the dependent variable is either significantly positive, significantly negative, or there is no significant relationship in either direction. The number of studies falling into each of these three categories is simply tallied. If a plurality of studies falls into any one of these three categories, with fewer falling into the other two, the modal category is declared the win­ner. This modal categorization is then assumed to give the best estimate of the direction of the true relationship between the independent and the dependent variable.21

The modification of the Light-Smith vote-counting method that we recommend, in the light of the small sample sizes and low expected correlations mentioned previously, is that the statistical significance of the obtained relationship in single studies be disre­garded. Instead, votes should be counted only on the basis of the direction of the relationship obtained. At most, a lower limit of, say -+- • I 5 for correlation coefficients should be imposed, and co­efficients smaller than . r 5 in absolute value should be counted as "present but not voting."

When available research on teacher behavior is combined in this manner, several promising teacher performance variables can be identified. For example, the seventeen studies of teacher disapproval yielded twelve votes for a negative relationship with student achievement, three mixed or doubtful votes, and only two votes for a positive relationship. Of nine studies of teacher talk, as mea­sured by the temporal duration of teachers' utterances or by the number of lines of typescript or words spoken by the teacher within a given time period, four showed positive relation to stu­dent achievement, three were "nonvoting," and two cast mixed or doubtful votes. Of nine studies using Flanders' category of teacher acceptance of student ideas, five yielded positive correla­tions, one yielded a negative correlation, and three were "non­voting."

These compilations serve to strengthen the hypotheses that fre­quency of teacher disapproval is negatively related to student

21. Richard J. Light and Paul V. Smith, "Accumulating Evidence: Proce­dures for Resolving Contradictions Among Different Research Studies," Har­vard Educational Review 41 (November 1971): 433·


achievement, and that amount of teacher talk and acceptance of student ideas are positively related to student achievement. These and similar findings could provide at least the beginning of a basis for the design of PBTE programs. Yet, again, several caveats are in order. First, the validity of vote-counting as method for accumulating evidence from distinct studies depends on whether the studies themselves are valid. The previously discussed problems of appropriately defining and measuring teacher performance vari­ables are one source of invalidity in individual studies. Vote­counting also requires that the teacher performance variables in the different research studies be properly clustered. In some cases, different reviewers of the research have disagreed as to the ap­propriateness of including certain studies in a given cluster. The validity of the clusters depends on subjective judgments as to the similarity of teacher behaviors in different studies. Whether the findings of studies within a cluster are generalizable across con­textual variables, such as subject matter, grade level, and socio­economic level, is an empirical question.

Second, the results of these vote-countings are limited to in­ferences of co-occurrence, not causation. Beyond the accumulated results of nonexperimental studies, the empirical base for PBTE needs to be bolstered with the results of experiments. Experiments would require developing methods for training teachers in par­ticular teacher behaviors. Thus, if such experiments reveal causal relations between a given teacher performance variable and student achievement, they will make a double contribution to PBTE. They will (a) yield effective training methods that (b) result in effec­tive teacher behaviors.

Beyond accumulating results across studies, teacher behavior vari­ables should be combined. In almost all of the one hundred non­experimental studies of teacher effectiveness, only one dimension of teacher performance at a time was considered in the correlational statistics used. That is, multiple correlations and other methods of simultaneously relating two or more teacher behaviors to student achievement have almost never been used. For example, question­ing behaviors were considered alone, to the exclusion of other teacher behaviors performed to facilitate learning. Furthermore, measures of single kinds of teacher performance were almost al-


ways simple frequencies of occurrence of the teacher behavior of interest. Yet this behavior was part of a complex teaching strategy which combined several relatively distinct variables, some of which may promote and others of which may impede learning. These problems are illustrated by the studies which examined teacher praise of students. When the frequency of praising statements was counted in isolation, such statements were not differentiated accord­ing to other features, such as the presence or absence of informa­tion as to why praise was given. And no record was made of whether the praise was given to student behavior relevant to the measure of achievement. So it is not surprising that teacher praise exhibits inconsistent relationships with student achievement. In fact, of twenty studies cited by Rosenshine and by Dunkin and Biddle in which frequency of teacher praise was correlated with student achievement, ten yielded a positive relationship, with a median co­efficient of about ·37• and six showed a negative relationship, with a median coefficient of about -.23. (For two of the studies it was possible to determine only the sign of the relationship, and for two it was possible to determine neither the sign nor the magnitude of the relationship.)

Any single kind of teacher behavior should not be expected to have a large effect on student achievement. Teacher behaviors, even those which are complex, should reasonably be expected to be more influential when considered in combinations. Thus, when Wright and Nuthall combined six teacher behavior variables through multiple correlational methods, the overall relationship with student achievement was observed to be .891.22 Because the number of teachers in this study was only seventeen, this multiple correlation probably capitalized on chance covariance and should not be considered well-established.

The preceding argument suggests that definitions of teacher performance need to be framed in terms of distinct variables known to be related to learning. It also suggests that major efforts should be made to define and study the effects of patterns of interrelated teacher behavior variables as they influence student learning. Com-

zz. Clifford J. Wright and Graham Nuthall, "Relationships between Teacher Behaviors and Pupil Achievement in Three Elementary Science Lessons," American Educational Research Journal 7 (November 1970): 477-91.


puter techniques for the study of patterns are already available. 23

Further work is required to sharpen these techniques and apply them to the range of teaching patterns observed in the classroom.

The foregoing describes aspirations rather than accomplish­ments in providing an empirical basis for PBTE. Most of the needed systematic research and development remains to be done. Knowl­edge of the kind available from accumulated correlational studies needs to be increased. Such studies should be followed by experi­ments or quasi-experiments where independent variables of teacher performance are defined in the light of the accumulated findings of correlational studies. Until substantial work of this kind has been done, the empirical basis for PBTE will remain insufficient.


PBTE needs efficient methods of helping teacher trainees ac­quire teaching skills. The definition of PBTE, in focusing on the results of training, leaves open the method by which teachers are trained. But teacher behavior may to a large extent be determined by methods of training. Hence the problem of obtaining the best methods of training teachers has commanded much attention. Many techniques for training teachers have been developed and some have been evaluated. This section briefly examines the evidence on the training methods typically used in PBTE programs.

Methods of training for teaching skill vary widely. At one ex­treme these methods include student teaching and internships of a year or more in regular classrooms, with the intern engaging in the complete range of school activities. At the other extreme, the methods include short-term intensive textbook modules, or prod­ucts, lasting only several hours and dealing with a single, limited teaching skill, such as providing praise in language teaching in the elementary school. In between are various training methods, such as feedback of observations, rnicroteaching, and minicourses.

Because the field of teacher training products is young, the effectiveness of most products is as yet unsubstantiated by em­pirical data. Indeed, of 657 products which involved the trainee in practice or post-training activities that could provide data for

23. Ned A. Flanders, "The Changing Base of Performance-Based Teach­ing," Phi Delta Kappan 55 (January 1974): 312-15.


evaluating the product's effectiveness, only 13 percent reported evaluation data of any sort. 24 Of these, many used merely an in­dex of acceptance of the training products by trainees or anec­dotal reports on the effectiveness of training. Thus, little informa­tion is available concerning the effectiveness of training products in changing teaching behavior. But one study is especially note­worthy: it demonstrated the longevity over thirty-nine months of the effects of training in a specific set of teaching behaviors.25

Indirect evidence of a different type supports the possibility of meeting the training requirements of PBTE. This evidence consists of information implicit in the reviews by Rosenshine and by Dun­kin and Biddle.26 In these reviews, it is noted that teachers were sometimes trained to teach in some more or less specific way to some criterion of performance. The studies reviewed occasionally provided useful information on the methods of training, the mea­sures used to ascertain training effectiveness, and the degree to which trainees actually performed according to the objectives of training. Thus some of these studies show that teachers can be trained to exhibit teaching behaviors such as accepting student ideas, increasing pupil initiation, or using higher-level questions.

Studies directly aimed at the effectiveness of various techniques for changing teacher behavior have also been conducted. Some of these studies have evaluated the traditional approach to teacher training-a combination of university courses in teaching methods and university and cooperating teacher-supervised teaching in the classroom. These two training techniques, and particularly the latter (that is, the student teaching experience), have dominated conventional education for decades. One discussion of this form of training characterizes it as assuming that

. . . if the student teacher is exposed to what purports to be effective teaching, the osmosis process automatically will enable him to absorb

24. Stanford Center for Research and Development in Teaching, Program on Teaching Effectiveness, "Teacher Training Products: The State of the Field," Research and Development Memorandum No. u6 (Stanford, Calif.: Stanford Center for Research and Development in Teaching, 1974), pp. 63-64.

2 5. Walter R. Borg, "The Minicourse as a Vehicle for Changing Teacher Behavior: A Three-Year Followup," Journal of Educational Psychology 63 (December 1972): 572-9.

26. Rosenshine, op. cit., passim; Dunkin and Biddle, op. cit., passim.


from the supervising teacher an approach or style that is effective. At the same time, it is assumed that the process automatically filters out any approach or style that is not effective.27

If the supervising teacher is "excellent," as was reportedly the case in the study by Bradtmueller,28 the process may prove effec­tive. But typically the student teaching experience is unanalyzed, and what happens to the trainee is unstudied. It is understandable that Peck and Tucker, in their review of research on teacher edu­cation, concluded that " ... at least by the end of student teaching, there are some almost universally reported decrements in attitude and in teaching behavior, as compared with the starting position of students prior to their field experience." 29

Beyond student teaching as a whole, various adjuncts have been used. One has been the use of feedback about teaching perform­ance in helping trainees to acquire improved teaching behaviors. The subject matter of the feedback and the means of transmitting it to trainees have varied considerably. For example, Heinrich and McKeegan coded teacher behavior as desirable or undesirable and instantaneously conveyed this information to trainees by raising color coded cards while the trainee taught.30 More generally, how­ever, trainees have received feedback in verbal forms, sometimes in a discussion or seminar-like setting and sometimes coordinated with listening to an audiotape or viewing a videotape record of their teaching.

Evidence for the efficacy of feedback about teaching perform­ance is fairly consistent. When the information is explicit, clear, and keyed to specific aspects of teaching behavior, feedback re-

27. Anhur H. Oestreich, "The Professional Growth of the Student Teacher," Phi Delta Kappan 55 (January 1974), 335-7·

z8. W. G. Bradonueller, "An Investigation of the Development of Prob­lem Solving Skills Related to Teaching Reading During Student Teaching" (doctoral dissenation, Indiana University, 1¢4).

19. Roben F. Peck and James A. Tucker, "Research on Teacher Educa­tion," in SecO'IId Htmdbook of Research on Teaching, ed. R. M. W. Travers (Chicago: Rand-McNally, Inc., 1973), p. 967.

30. Darlene Heinrich and Hugh F. McKeegan, "Immediate and Delayed Feedback Procedures for Modifying Student Teacher Behavior According to a Model of Instruction," paper presented at the annual meeting of the American Educational Research Association, 11)69.


suits in improvement in the trainee's ability to perform according to a model of teaching. 31 Further, most of the research emphasizes the value of having another person, usually a supervisor, present during the feedback to discuss particular changes to be made in teaching behavior.

Several studies have been made of feedback to the trainee in the form of observational data, such as the kind obtained with Flanders' interaction analysis categories. One way to use such data is simply to train the teacher to use the categories of teach­ing behaviors in the observational system for analyzing his own teaching behavior. A second way, usually incorporating the first, is to provide teacher trainees with observational data for discussion following teaching. After reviewing eleven studies conducted in preservice settings and seven conducted in in-service programs, Flanders concluded that "attention to teaching behavior, practice in analyzing it, and performing it with feedback" tend to influence the behavior in the teacher's repertoire.32

Microteaching, another major approach to teacher training, often incorporates some of the feedback methods already men­tioned. In this well-known approach, trainees teach perhaps five students for a period of about ten minutes. The subject matter to be taught is usually restricted to a small range of concepts. Only one or two specifically defined teaching skills, such as asking high­order questions, inducing a set, or providing reinforcement, are to be practiced in a given microteaching session. Often the train­ee's teaching is recorded on videotape, which is analyzed by the trainee and his supervisor shortly after the microteaching episode. This technique, and its close relative, the minicourse, have consist­ently been found to be effective in changing trainees' teaching behaviors to parallel more closely those of a particular teaching strategy.38 As already noted, this training method is the only one which has been examined as to its long-term effect on teacher be­havior.114

31. Berliner, op. cit., passim.

32· Ned A. Flanders, Analyzing Teaching Behavior (Reading, Mass.: Addison-Wesley Publishing Co., 1970), p. 352·

33· Peck and Tucker, op. cit., pp. 951-54·

34· Borg et al., op. cit., passim.


The foregoing approaches have much in common with the gen­eral approaches to teacher training derived from behavior modi­fication. McDonald characterized this approach as consisting of the precise analysis of the desired teaching performance in behavioral terms and the matching of a treatment and feedback process to these specified behaviors.35 In addition to subsmning microteaching this approach includes trainee observation of a model that exhibits the teaching skill to be learned. The model may be printed, audio­taped, videotaped, filmed, or presented in some other manner. The research evidence directly relating this approach to teacher train­ing is sparse. But the power and generality of principles of be­havioral analysis in other settings suggest that these training tech­niques could successfully influence teaching performance in desired ways.36

In short, the evidence strongly supports the claim that trainees can be instructed in ways that bring about specified changes in their behavior. From the foregoing, it appears that such training should include a means by which the trainee can receive specific information about his own teaching as he performed it. If de­sirable teaching behaviors . are explicitly defined, trainees can be helped to use those behaviors in teaching.


It was concluded above that specific causal connections between teacher behaviors and student achievement were at present inade­quately supported with research evidence. The evidence reviewed did suggest, however, that more adequate support could be forth­coming from research on teaching of the kinds described. Thus, a major part of the cost of putting PBTE into effect will arise in conducting the research needed to identify and validate teacher competencies.

Estimates of this cost are hazardous. A small number of research

35· Frederick J. McDonald, "Behavior Modification and Teacher Educa­tion," in Behavior Modification in Education, Seventy-second Year book of the National Society for the Study of Education, ed. Carl E. Thoresen (Chicago: Distributed by University of Chicago Press, 1973) pp. 41-76.

36. Albert Bandura, Principles of Behavior Modification (New York: Holt, Rinehart & Winston, 1969).


and development programs are currently seeking to develop a sound basis for PBTE. Their efforts, however, are considerably smaller in scale than what is needed to make rapid progress. Rather than speculate about the funds required for an adequate effort, we note only that, short of unforeseeable scientific breakthroughs, a substantial increase over current research and development ex­penditures will be necessary to obtain the knowledge and products needed for substantially more valid PBTE than is now possible.

Assuming that effective teacher behaviors and training tech­niques have been identified, one may ask "How much will it cost to install PBTE?" Rosner and his co-workers provided detailed estimates of these costs as of 1972.37 A five-year test of the value of PBTE for improving the performance of 2o,ooo educational personnel per year was projected at a total cost of $I 14 mil­lion. This plan had five major components: a group to design and coordinate the work ($1.5 million); the development and evaluation of 100 teacher training laboratories ( $7 5 million); the development and field testing of instructional materials for use in training ($I 9 million); the design and testing of performance evalu­ation instruments ( $s. 5 million); and the establishment of career development boards to deal with issues of certification following training ($13 million).

In addition to estimating the cost for developing and testing PBTE, the same group estimated the operating expenditures needed for a single PBTE program. Again, the costs were spread over a five-year period for development and implementation, totaling $I9·9 million for approximately 450 trainees (presumably per year, although the source is unclear on this point). It was also pointed out that operating costs would increase to the extent that indi­vidual PBTE programs adopted differing models of teaching and methods of training teachers.

Another, seemingly independent, estimate of the expenditure needed for developing a single PBTE program at a single teacher training institution came to approximately $5 to $6 million. The. cost of implementing a national PBTE program was estimated at

37· Benjamin Rosner (chairman) et al., The Power of Competency-Based Teacher Education: A Report (Boston, Mass.: Allyn and Bacon, 1972) pp. 24-37·


around $wo million over about twenty years.38 Hite estimated that PBTE will cost 150 percent more than traditional programs.39

These figures are admittedly uncertain; at least, they are dis­crepant. It is also noteworthy that the Committee on Performance­Based Teacher Education of the American Association of Colleges for Teacher Education subsequently recognized a pressing need for more reliable cost analyses relating to PBTE.4() Despite the foreg@ing estimates, the Committee concluded that adequate cost analyses for PBTE are, as yet, unavailable. Without better means for projecting these costs, only experience will provide the answer as to how much investment PBTE will require.

Estimates of the cost of PBTE may be somewhat spurious if they are based on the assumption that the installation of PBTE will take the form of a direct and sweeping changeover. The PBTE program at Weber State College (Ogden, Utah) was still in a formative stage six years after PBTE was planned and imple­mented.41 The need for more valid research on teacher perform­ances effective in promoting student achievement, advances in computer and other technological support for teacher training, the changing base for teacher certification, and other factors make the several available cost analyses relatively ungeneralizable beyond their publication dates. Adaptations of features of various models of teaching and performance-based training by individual training institutions will also tend to make the available cost estimates in­appropriate for a given PBTE program.

Finally, the point should be made that the foregoing estimates, however speculative, say nothing about costs above and beyond those of present non-PBTE programs. Since PBTE would pre­sumably replace present programs, the Gosts of present programs

38. Theodore E. Andrews, "What We Know and What We Don't Know," in Houston, op. cit., pp. 31-36.

39· Herbert Hite, "The Cost of Performance-Based Teacher Education," Journal of Teacher Eiucation 24 (Fall1973): 224.

40· Committee on Performance-Based Teacher Education, op. cit., p. 23.

41. Caseel Burke, The Individualized, Competency-Based Sysetm of Teacher Education at Weber State College (Washington, D.C.: American Association of Teacher Education, 1972).


should be subtracted from those of PBTE programs as the latter go into effect.

Assessment in PBTE

To this point our discussion has centered on whether PBTE should be considered a useful approach and adopted. We now con­sider the two kinds of assessment problems that arise in PBTE: those of monitoring the performance of trainees as they move through the PBTE program, and those of assessing the quality of the PBTE program itself.

The problems of assessment in PBTE are similar to those of identifying causal relationships between teacher behaviors and stu­

dent learning. Elam's early statement on the state of the art of PBTE noted that, for the operating PBTE program, ". . . the overriding problem before which others pale to insignificance is that of the adequacy of measurement instruments and procedures."42

Many subsequent statements on PBTE have dealt with assessment in some sense. Yet Schalock and others have concluded that to date " ... there simply is no sign of an emerging technology of assessment that meets the demands of the competency-based teacher education movement." 43 This is not to say that no research has been done. The essence of work done to date will be reviewed below. But proponents and adversaries of PBTE have limited most of their work on assessment in PBTE to statements of models for assessment or integrations of models already put forth.44 Empirical work on assessment should now occupy a large part of the efforts of the PBTE movement. We turn now to a discussion of these efforts.


One enduring problem of research on teaching, and thus of

4z. Elam, op. cit., p. z1.

43· H. Del Schalock, "Notes on a Model of Assessment That Meets the Requirement of CBTE," in Houston, op. cit., pp. Z09-49·

44· For example, see the papers by Bruce Joyce, H. Del Schalock, Patricia M. Kay, and David A. Potter in Houston, op. cit. See also Jack C. Merwin, Performance-Based Teacher Education: Some Measurement and Decision­Making Considerations (Washington, D.C.: American Association of Colleges for Teacher Education, 1973). ·


research on PBTE, has been that of developing appropriate meas­ures of student achievement of cognitive and social-emotional ob­jectives. It is now generally recognized that standardized tests of cognitive achievement are inappropriate for assessing the effects of an individual teacher or a particular teaching strategy.45 Further­more, greater emphasis should be placed on developing and validat­ing measures of student learning in the domains of self-referent affect, attitudes, and social relations. Content validity appears to be a critical problem in developing improved measures of out­comes for validating teaching skills. The ideas developing out of the work of Glaser, of Anderson, of Bormuth, and of others may provide a path to improvement in this area.46

The brevity of this treatment of issues in measuring student achievement should not be regarded as implying that these prob­lems are simple or unworthy of concentrated effort. They equal in importance the problems of developing useful measures of teach­ing performance in the effort to identify causal relationships be­tween teaching behaviors and student learning. But research related to assessment in PBTE has focused primarily on measuring teacher behavior. Hence, this larger and better established area will be considered in most of the following discussion.

Empirical work on assessing teacher ability to use teaching skills which promote student achievement requires knowledge con­cerning the teacher behaviors that affect student learning. One response to the inadequacy of such knowledge was to ignore the specifics of teaching and directly measure the degree to which teachers promoted student learning. But a review of such studies by Glass revealed a major difficulty in this approach: the meas­ures of teaching effectiveness of a single teacher in terms of mean student achievement were quite unreliable; that is, the teaching

45· For example, see Richard C. Anderson, "How to Construct Achieve­ment Tests to Assess Comprehension," Review of Educational Research 42 (Spring 1972): 145-70.

46. Robert Glaser and Anthony J. Nitko, "Measurement in Learning and Instruction," in Educational Measurement, ed. Robert L. Thorndike (Wash­ington, D.C.: American Council on Education, 1971) pp. 625-70; Anderson, op. cit.; and John Bormuth, On the Theory of Achievement Test Items (Chicago: University of Chicago Press, 1970).


performance tests failed to differentiate reliably between teachers.47

The difficulty was due primarily to the inadequate sampling of teacher performances.

Using student achievement as a basis for evaluating teacher ef­fectiveness may prove fruitful for evaluating the overall ability of graduates of PBTE programs. It is limited, however, to the :final, or summative, assessment of trainees. More relevant to the training of teachers is assessment during training of specific teacher behaviors and their patterning.

In performing this more specific behavioral assessment, one must decide on an appropriate level of behavior for assessment. Quirk pointed out that dividing teacher performance into a large number of individually measured teaching skills lowers the reliability of each measure of teacher performance.48 Yet training and measur­ing complex skills is also hazardous in that the components of a complex skill may vary considerably and produce a large amount of error variance.

In judging the overall competency of a trainee, one faces a sig­nificant question as to whether a conjunctive or a compensatory model should be used for assessing whether a trainee has acquired each of the individual teaching skills as well as the complete set of skills. The compensatory model means that a deficiency in one skill can be compensated by superior performance in another. Thus, the overall assessment of training may not reflect accurately the trainee's skill in individual teacher behaviors. In contrast, a conjunc­tive model means that the trainee must perform at least at a speci­fied minimal level in each teaching skill. Here, a lack of one skill will result in failure on the total complex of skills.49 Quirk noted that the most commonly used conjunctive model, the multiple-cutoff

47· Gene V. Glass, "Statistical and Measurement Problems in Implement­ing the Stull Act," in Mandated Evaluation of Education: A Conference on California's Stull Act, ed. N. L. Gage (Washington, D.C.: Capitol Publica­tions, Inc., 1973) pp. 53-98.

48. Thomas J. Quirk, "Some Measurement Issues in Competency-Based Teacher Education," Phi Delta Kappan 55 (January 1974): 316-19.

49· A more complete discussion of these models and their relative value can be found in J. R. Hills, "Use of Measurement in Selection and Place­ment," in Thorndike, op. cit., pp. 680-732.


model, requires specifying a cutoff score. This score may be arbi­trary owing to insufficient evidence relating the teacher behavior to student achievement.110 When this teacher behavior has only limited validity, the problem is further compounded. The compen­satory model may be more useful in similar circumstances. The choice between these models should depend on the teacher vari­ables involved, that is, the nature of the teaching and learning situation.


The quality of a PBTE program can be evaluated in terms of (a) comparisons with non-PBTE programs, (b) comparisons with other PBTE programs, and (c) the value of its components, both singly and in combination. Some writers have argued that it is not possible to compare PBTE and non-PBTE programs because they differ too greatly in both means and ends. It may be replied that all teacher training programs are comparable in terms of the ability of their graduates, as a group, to promote student attainment of educational objectives. The comparability depends, in tum, on our ability to agree on a set of objectives and their operational defi­nitions. To claim that teacher training programs can determine educational objectives is to imply that students of a teacher from one program should have a different set of objectives than students of a teacher from another program. The opposite seems to be the more justifiable situation; that is, teacher training programs should be responsive to a set of educational objectives. Different methods of training, derived from different educational philosophies, may characterize different programs. But, for example, if all first grade teachers should help students learn to read, then teacher education programs for first grade teachers are comparable in rerms of the proficiency of their graduates in teaching children to read. In short, teacher training programs are comparable to the degree that the students of their trainees have the same educational goals.

We also hold that comparing one PBTE program with another is reasonable. The goal of all PBTE programs is to change the trainee's behavior in ways that promote student attainment of

so. Quirk, op. cit., 318.


educational objectives. PBTE does not dictate that one or another method of teacher training be used. But insofar as PBTE programs are aimed at the same objectives for the students of their trainees, their effectiveness can be compared.

The third, or "internal," type of program assessment, namely, determining the value of single and combined components of a teacher training program, is also meaningful. Because of the close, but not necessary, association of PBTE with an individualized, cy­bernetic model of training, we discuss internal program assessment only for this type of program organization.

Discussions of the internal evaluation of PBTE programs are abundant. But they reflect too little empirical work. This condi­tion can be seen in evaluations of one of the the most common tools of PBTE, the instructional module. As described by Cooper and Weber, such a module consists of a statement of rationale, a set of objectives, a listing of prerequisite entry behaviors, a pre­training assessment, learning procedures, a post-training assessment, and suggestions for remedial training to be used if the trainee fails to achieve the criterion level of achievement for the module.111

Lawrence surveyed more than two hundred such modules from several sources to evaluate their quality.52 Of the modules examined, two-thirds used procedures for measuring training and criteria for judging training that seemed substantially inconsistent with the teacher performance to be acquired. Nearly half of the modules did not provide specific criteria for evaluating the teaching skill to be learned. Half of the modules used measurements that did not cover the range of skill learning expressed in the training objectives. Per­haps most disturbing, one-third of the modules required or sug­gested training and assessment activities that were judged not likely to produce the kind of evidence necessary to assess the trainee's acquisition of the teaching skill. If these findings apply to most modularized PBTE programs, considerable improvement is needed in the validity and utility of such modules. This condition should not be taken to imply that the method of modularized instruction

51. Cooper and Weber, op. cit., p. 17.

sz. Gordon Lawrence, "Delineating and Measuring Professional Compe­tencies," Educational Leadership 31 (January 1974): 198-301.


has failed; rather, it means that modules currently used in PBTE are inadequate.

Parker reported an evaluation of the Individualized Perform­ance-Based Teacher Education program at Weber State College. The evaluation was based on the opinions of students, graduates, and faculty in the program. Parker concluded that "the students can apply the competencies taught in the program as first-year teachers" but he cited no supporting data. He also concluded that graduates of the program "feel that they are effective in applying the competencies learned in the program" and that "these com­petencies seem to be what they need as classroom teachers [in the opinion of] cooperating teachers, school principals, and school district supervisory personnel." 53 We need data to validate these subjective appraisals. The evidence on the utility of feedback in microteaching suggests that, unless specific training in self-evalu­ation is provided, teachers cannot accurately evaluate their own performance. Furthermore, unless school personnel receive specific training in teacher performances to be used, they probably cannot make valid judgments about PBTE training. In short, better in­formation is needed on the degree to which Weber State College graduates can perform the teaching skills at which their PBTE program was aimed.

Probably because the PBTE movement is new, most evaluations of PBTE emphasize program characteristics such as the efficiency of the instructional management system and the ways of rearrang­ing faculty duties, as against relationships between training pro­cesses and the behavioral outcomes of training. Much of the evalu­ation effort focuses on improving the capability of the PBTE program to do whatever it does rather than assessing the effec­tiveness of the program.

In summary, the evaluation of PBTE programs is still in its early stages. It faces problems of assessment at the level of indi­vidual program components in terms of matching skill instruction to skill learning. It also faces assessment problems at the level of evaluating the PBTE graduate's ability to teach as trained. Em­pirical work on methods for evaluating PBTE should be aimed at

53· Reese Parker, "Weber State College Evaluates PBTE after Three Years,'' Phi Delta Kappan 55 (January 1974): 320-24, emphasis added.


obtaining objective data about various kinds of effectiveness of PBTE programs. As the PBTE movement matures, this kind of assessment should begin to appear.

A Position on PBTE

Having reviewed the definitions, origins, and problems of PBTE, we now attempt to take a reasonable position on its present status and the alternatives that are available to teacher educators. We be­lieve it important to avoid the two extremes either of returning to the state of teacher education before PBTE or of advocating the legally enforced universal installation of PBTE.

The first of these alternatives would throw away the gains achieved by the PBTE movement. These gains take the form of framing the problems of teacher education in ways to which the tools of the behavioral sciences can be more effectively applied. These applications of behavioral psychology, systems analysis, and the like, make good sense, have already yielded some improvements in practice, and provide a feasible prospect for continued advance. Without the kinds of behavioral, pedagogical, and philosophical analysis that are the essence of PBTE, teacher education will of necessity flounder with overly global and complex variables (for example, unanalyzed student teaching, "humanistic gestalts," and large but vague and unattainable goals) that hindered teacher edu­cation before the advent of PBTE.

On the other hand, we see dangers in disseminating and instal­ling PBTE with legal and political force. This chapter has not neglected the serious and difficult problems that remain to be solved before PBTE can make good on its promise. Officials favor­ing PBTE in the federal and state agencies concerned with teacher education have not been unaware of those problems. But they have often not succeeded in sufficiently restraining the actions of gov­ernment agencies. State laws have been adopted to install PBTE with all the coercive power implicit in legislated improvements in education. In Texas, for example, PBTE has been legislated in such a way that university faculties and administrators protested that academic freedom was being infringed upon. 54 If PBTE had been

54· Ellis Sandoz, "PBTE: The Nays of Texas," Phi Delta Kappan 55 (January 1974): 304-6.


as well established as, say, Salk vaccine, it could and should have been enforced as desirable for the children of Texas. No one would now claim that PBTE has established itself that well-scientifically, educationally, or ethically. Accordingly, it was premature to make PBTE mandatory, and the Texas law was later interpreted in ways that eliminated its coercive impact.

Thus, we advocate that PBTE be encouraged, pursued, tried, studied, and improved, but not that it be required, legislated, man­dated, or enforced. To do the latter would, in our judgment, be similar to having established a national airline in 1913, only ten years after the first flight. Although the basic principles of aviation had been discovered and applied by then, an enormous amount of technical detail and sophistication was still needed before a na­tional airline could succeed. In 1975, it seems safe to say, certain basic approaches to PBTE have been well formulated and primitive models have made promising flights, but analogous problems of technique and values still remain. What is needed at this point is the kind of inventiveness, boldness, and perseverance, including "tinker­ing," that eventually made aviation succeed. Overly hasty installa­tions of the approach-installations that will inevitably disappoint as they go beyond the state of the art-will only produce setbacks.