Printer Friendly

Observed supervisory behavior and teacher burnout in special education.

Observed Supervisory Behavior and Teacher Burnout in Special Education

Stress and burnout have been identified as significant problems in special education. "Burnout" is a particular reaction to stress, a coping mechanism involving emotional exhaustion, depersonalization of those whom one serves, and psychological disengagement from the job (Cherniss, 1980; Maslach & Jackson, 1984).

As recognition of the burnout problem has spread, teachers, administrators, and researchers have begun to search for ways of reducing the strain that seems to be endemic in this kind of work. Empirical research suggests that the supervisor may be a key factor in the amelioration of stress and burnout (Cherniss, 1980). For instance, Lawrenson and McKinnon (1982) found that special education teachers believed that "hassles" with administrators were a leading cause of resignations. Several other studies found that lack of support on the part of supervisors and administrators was one of the strongest and most frequently cited sources of stress, low morale, and attrition in special educators (Johnson, Gold, & Vickers, 1982; McKnab & Mehring, 1984; Thomas, 1984; Zabel & Zabel, 1982). Fimian (1986) found that special education teachers who reported not receiving supervisory support also reported stronger and more frequent work-related stress.

This previous research has been limited in two important ways. First, supervisory "support" has been treated in rather global terms; types of support provided by the supervisor have not been differentiated and there has been no attempt to gather information about the quantity, quality, and type of support. Second, supervisory behavior has not been assessed directly; instead, it has been measured through questionnaires completed by teachers. Focusing on specific supervisory behaviors that might be related to teacher burnout not only would provide more valid results; it also could be more potentially useful to administrators by showing realistic, concrete ways in which supervisors can provide support to teachers.

Research on supervision in the workplace provides some clues about the types of actual behavior that might be associated with lower levels of burnout in special eduation programs. Two general dimensions typically have emerged: consideration and initiating structure (Schriescheim & Kerr, 1974). Supervisors who score high in consideration are sensitive, sympathetic, tolerant, and supportive. Supervisors high in structure are task oriented, maintain high standards, and help subordinates to perform better by providing clear goals and frequent feedback. Cook (1983), in a study of regular education teachers, used a paper-and-pencil measure of these two dimensions of supervision and found that supervisor consideration, but not initiating structure, was significantly, negatively correlated with burnout.

The study reported in this article was an exploratory effort to develop a methodology for examining the realtionship between burnout and supervisory behavior, using an objective, structured, behavior observation measure of supervision rather than a paper-and-pencil measure filled out by teachers. Based on previous research and theory, it was expected that behaviors associated with emotional support and consideration, such as amount of time spent discussing their work-related problems with staff, would be strongly associated with lower burnout. It was also expected that more time devoted to certain structuring behaviors such as goal clarification, guidance, coaching, and feedback would be linked with lower levels of burnout. It was hoped that the observational method of data collection would lead to a more refined description of supervisory behavior and its relationship to burnout.


Subjects and Settings

The principals and staff in two schools for severely retarded children served as subjects. The first school was chosen because it was independently identified by the head of the agency and by a principal at a third school not involved in the study as a setting in which staff burnout was surprisingly low, and the principal was given much of the credit for the low level of burnout. These judgments were based on lower staff turnover, fewer grievances, and more positive parent reactions than found in other schools operated by the agency. This assumption about the burnout level was supported by objective data when the staff subsequently filled out the Maslach Burnout Inventory (MBI) (Maslach & Jackson, 1981) and scored in the low range (based on the MBI's norms) on both the Emotional Exhaustion and Depersonalization scales of the MBI. For this particular study the Frequency scores from the MBI were used instead of the Intensity scores because the internal consistency reliabilities were higher for the former on all three scales.

Staff at the "low burnout" school were 12 certified special education teachers, 5 teacher aides, and 5 ancillary staff (social worker, physical and occupational therapists, music teacher, and nurse). All were supervised by the principal, a 35-year-old woman who had been a special education teacher before becoming principal at this school 10 years earlier. The school was part of a large, urban association for mentally retarded citizens. The association's division of children's services operated a total of nine schools. The children at this particular school ranged in age from 5 to 17 years and came predominantly from poor Black or Hispanic families.

The second school was operated by the same association for retarded citizens, was similar in size, and served a similar population. Located in a different but equally impoverished section of the city, this school consisted of 11 teachers and 8 ancillary staff. Staff at the second school did not differ from those at the first in either age or sex, according to a t test and Chi square test respectively. The principal at this school was 28 years old, female, also a former teacher, and she had served as principal for 2 years. Staff scores on the MBI revealed that burnout levels were significantly higher at this school than at the low burnout school for two of the three scales. Staff scored in the moderate range on the MBI on two of the three scales, while the score on the third (Personal Accomplishment) was in the high burnout range, based on national norms.

Observational Instrument

Development of the Supervisor Behavior Observation Scale (SBOS) has been described elsewhere (Cherniss, 1986). The SBOS consists of six major dimensions, each of which is divided into several mutually exclusive categories. The six dimensions are mode of communication, function (or purpose), content, target, tone, and location. The specific categories for each dimension are listed in Table 1. The dimensions and categories were developed on the basis of extensive, unstructured, ethnographic field study in various settings and a review of the literature on supervisory leadership.

Reliability for the SBOS was high. It was calculated via 30 simultaneous observations conducted during three phases of the study (beginning, middle, and end). Only interactions were considered for reliability estimation; noninteraction, which could artificially inflate the reliability estimates, was not included.

Reliability for each category was assessed through use of Cohen's Kappa (Cohen, 1960). Kappas for all categories are presented in Table 1. In general, they are quite good, ranging from .71 to 1.00. However, it should be noted that several categories occurred infrequently (e.g., feedback); the Kappas for these cannot be considered conclusive.

Stability of the SBOS was assessed by comparing observations made in the morning with those made in the afternoon and by comparing observations between weeks. Using Chi square tests, there were no statistically significant differences between morning and afternoon observations or between weeks for any categories of the SBOS, suggesting a rather high degree of stability.

Observational Procedure

The observers were two graduate students from doctoral programs in psychology. The author trained them in use of the SBOS. Approximately 30 hours of training were required in order to reach the a priori reliability criterion of at least 80% agreement for every category.

The SBOS uses a time-sampling procedure in which an interaction is recorded only if initiated during the first 10 seconds of a 60-second interval, the latter 50 seconds being used to record the interaction. In order to minimize reactivity, the observer visited each setting several times prior to data collection. Also, the data collected were not considered valid until the subjects appeared to have habituated to the observers, as evidenced by the diminution of comments made to the observer as well as various nonverbal indications.

Initially, the subjects were told that the goal of the study was to develop a behavior observation instrument for use in future research on supervision. They were assured of confidentiality and thoroughly debriefed at the study's conclusion. The subjects were also told that they could ask the observer to leave them alone temporarily if the observer's presence might cause problems. An observer was excluded only once, for 3% of the total possible observation time.

Observations of each principal were done on at least 2 different days in 3 different weeks. For the moderate burnout school, there were a total of 953 observations (32 separate half-hour observation periods spread over 4 weeks). For the low burnout school, there were a total of 1,011 observations (34 separate half-hour observation periods spread over 5 weeks). Because the data did not significantly differ across weeks, this amount of observation seemed to be adequate. Observation times were distributed throughout the day in a random fashion. Observers were instructed to "shadow" the principal, following her wherever she went from the beginning of the designated observation period until the end.

At the end of the scheduled observations, observers checked the representativeness of the period chosen for observation by asking the subject if the pattern of activities observed was typical. The principal in the moderate burnout school said that the behavior was not typical because her secretaty had been out sick part of the time. Thus, the observer conducted one more week of observation in that school. A comparison of the data from the two periods revealed no significant differences in the pattern.


Table 1 shows how the two schools compared in frequency of supervisory behavior for all of the behavioral categories. Several differences are statistically significant.

Looking first at the total amount of interaction, the principal in the moderate burnout school interacted with others significantly more than did the low burnout school principal: [X.sup.2] (1, N = 2142) = 194.8, p [is less than] .001. The data on Target of Interaction indicate that in both schools, much of the interaction involved subordinates (teachers and ancillary staff). However, the differences between the two principals reached statistical significance on the Target dimensions: [X.sup.2] (6, N = 739) = 57.04, p [is less than] .001. A post hoc analysis indicated that the two principals differed significantly on two of the seven Target categories. First, the low burnout principal interacted significantly more frequently with other administrators: [X.sup.2] (6, N = 739) = 18.26, p [is less than] .001 (primarily the principal's own supervisor in the central office). Second, the low burnout principal also interacted significantly more frequently with clerical staff [X.sup.2] (6, N = 739) = 24.47, p [is less than] .001. None of the other post hoc comparisons for the Target dimension was statistically significant.

The differences in supervisory behavior on the Mode dimension also were statistically significant: [X.sup.2] (2, N = 708) = 39.28, p [is less than] .001. Post hoc comparisons suggested that the low burnout principal made statements significantly more often than did the moderate burnout principal: [X.sup.2] (2, N = 708) = 30.90, p [is less than] .001, and listened to others significantly less often: [X.sup.2] (2, N = 708) = 40.02, p. [is less than] .001. The difference in frequency of asking questions did not reach statistical significance.

The two principals also differed on the Function dimension: [X.sup.2] (5, N = 709) = 73.94, p [is less than] .001. The most striking difference was in the frequency of interactions where the purpose seemed to be to provide support. A post hoc comparison showed that the low burnout principal engaged in significantly more support than did the moderate burnout principal: [X.sup.2] (5, N = 709) = 41.03, p [is less than] .001. On the other hand, the low burnout principal observed others performing their jobs significantly less: [X.sup.2] (5, N = 709) = 27.23, p [is less than] .001. None of the other comparisons in the Function dimension was statistically significant.

The two principals also differed significantly in the Content of their interactions: [X.sup.2] (8, N = 709) = 128.48, p [is less than] .001. Post hoc comparisons indicated that the two principals differed significantly on four of the nine Content categories. The low burnout principal scored significantly lower in administrative content: [X.sup.2] (8, N = 709) = 22.64, p [is less than] .005, and in the "other" category: [X.sup.2] (8, N = 709) = 16.82, p [is less than] .05. Informal observation suggested that most of the communication scored "other" involved nonwork-related issues that were not personal problems, such as discussions of the weather or the latest movies playing in town. The low burnout principal scored significantly higher in personal/work-related content: [X.sup.2] (8, N = 709) = 74.74, p [is less than] .001, and in personnel content: [X.sup.2] (8, N = 709) = 20.17, p [is less than] .01.

There was no significant difference between the two principals in the observed Tone of their interactions, but the two principals did differ significantly in the Location of their observed interactions: [X.sup.2] (4, N = 760) = 104.73, p [is less than] .001. Post hoc comparisons for all dimensions, except "other," reached statistical significance. As Table 1 shows, the low burnout principal spent significantly more time in the office: [X.sup.2] (4, N = 760) = 104.50, p [is less than] .001, and significantly less time in the staff lounge: [X.sup.2] (4, N = 760) = 16.88, p [is less than] .005, in classrooms, [X.sup.2] (4, N = 760) = 15.70, p [is less than] .005, and in halls, [X.sup.2] (4, N = 760) = 14.21, p [is less than] .01.


The principals of the low and moderate burnout schools presented two distinctly different behavioral profiles. However, in considering these results one should keep in mind that this was an exploratory study of only two supervisors. Thus, the findings can only be suggestive. The study should be replicated in other settings with more subjects. Also one must be cautious about drawing cause-effect inferences; the observed differences in supervisory behavior may or may not be the causes of the differences in staff burnout levels. The two schools are closely matched in many respects, and thus it is plausible to infer that some of these supervisory differences did in fact account for the differences in burnout. But other interpretations are possible. In fact there has been growing recognition in recent years that followers can have a significant impact on leaders, and in some cases this may have happened in this study. Subtle differences in the two staffs may have contributed to the differences in supervisory behavior. Finally, supervisory style is but one of many factors that might reduce or exacerbate teacher burnout in special education; excessive paperwork, lack of discipline support, and unrealistic IEPs also may play a major role.

Thus, a number of limitations in the research method make the conclusions of this study limited when considered by themselves. But they become somewhat more compelling when combined with previous studies using different methodologies and yielding similar results (e.g., Fimian, 1986). Furthermore, the behavior observation procedure does provide some concrete and specific suggestions for supervisory practice. For instance, the findings, along with anecdotal information provided by the observers, suggest that supervisory might help reduce staff burnout by spending more time in planning, organizing, and advocating for staff. Furthermore, a supervisor who listens more and talks less, and who tries to engage in much "small talk" with staff, will not necessarily be more supportive. On the other hand, the low burnout principal in this study did spend more time discussing the significant work-related problems and concerns of her staff and made more of an effort to keep staff informed about critically important personnel matters, such as impending layoffs, salary raises, fringe benefits, promotions, etc.

A concrete example of how the low burnout principal supported staff in this manner occurred during one of the observation periods and was described by the observer in the context notes: A staff member was experiencing a crisis relating to some aspect of the job and came to the principal to discuss it. The principal immediately put aside her other work and talked to the staff member about the problem for some time. The principal and staff member discussed the problem again later in the day. Thus it was not through high levels of informal interaction that this principal became a meaningful source of support for staff, but rather through direct involvement in their work-related problems as needed.

In conclusion, the present study suggests a new way to measure supervisory behavior and its relationship to burnout in special education. Research and practice have suffered from a lack of objective behavioral measures of supervision. This study has suggested a viable means of overcoming this deficiency.


Chermiss, C. (1980). Professional burnout in human service organizations. New York: Praeger.

Chermiss, C. (1986). Instrument for observing supervisor behavior in educational programs for mentally retarded children. American Journal of Mental Deficiency, 91, 18-21.

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37-46.

Cook, K. E. (1983). The relationship between teacher perceptions of supervisory behavioral style and perceived teacher burnout. Unpublished doctoral dissertation, University of Connecticut.

Fimian, M. J. (1986). Social support and occupational stress in special education. Exceptional Children, 52, 436-442.

Johnson, A. B., Gold V., & Vickers, L. L. (1982). Stress and teachers of the learning disabled, behavior disordered, and educate mentally retarded. Psychology in the Schools, 19, 552-557.

Lawrenson, G. M., & McKinnon, A. J. (1982). A survey of classroom teachers of the emotionally disturbed: Attrition burnout factors. Behavioral Disordered, 8, 41-49.

Maslach, C., & Jackson, S. E. 1981). A scale measure to assess experienced burnout: The Maslach Burnout Inventory. Journal of Occupational Behavior, 2, 99-113.

Maslach, C., & Jackson, S. E. (1984). Burnout in organizational settings. In S. Oskamp (Ed.), Applied Social Psychology Annual, 5 (pp. 133-154). Beverly Hills, CA: Sage.

McKnab, P. A., & Mehring, T. A. (1984), April). Attrition in special education: Rates and reasons. Paper presented at the annual convention of The Council for Exceptional Children, Washington, DC.

Schriescheim, C. A., & Kerr, S. (1974). Psychometric properties of the Ohio State leadership scales. Psychological Bulletin, 81, 756-765.

Thomas, W. R. (1984, April). Occupational stress among exceptional education teachers. Paper presented at the annual convention of The Council for Exceptional Children, Washington, DC.

Zabel, R. H., & Zabel, M. K. (1982). Factors in burnout among teachers of exceptional children. Exceptional Children, 49, 261-263.

CARY CHERNISS is Associate Professor, Graduate School of Applied and Professional Psychology, Rutgers University, New Brunswick, NJ.
COPYRIGHT 1988 Council for Exceptional Children
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1988 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:includes bibliography
Author:Cherniss, Cary
Publication:Exceptional Children
Date:Feb 1, 1988
Previous Article:Comparison of academic and cognitive programs for young handicapped children.
Next Article:Relationship of temperament to preschoolers' interactions with peers and teachers.

Related Articles
Burnout in the lab: symptoms, stages, strategies.
Relationship between teachers' effectiveness and their tolerance for handicapped students.
Teachers' perceptions of problem behavior in general and special education.
Testing a model of special educators' intent to stay in teaching.
Supporting professionals-at-risk: evaluating interventions to reduce burnout and improve retention of special educators.
Strategies for inclusion: an annotated bibliography.
An analysis of relations among locus of control, burnout and job satisfaction in Turkish high school teachers.
Behavior management skills as predictors of retention among South Texas Special Educators.
Teacher stress and burnout in deaf education.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters