Teacher Well-Being and the Implementation of School-Wide Positive Behavior Interventions and Supports

Abstract

Teacher well-being has become a major issue in the United States with increasing diversity and demands across classrooms and schools. With this in mind, the current study analyzed the relationship between outcomes of teacher well-being, including burnout and efficacy, and the implementation of School-Wide Positive Behavioral Interventions and Supports (SWPBIS). Using a multilevel regression approach, the surveys of 184 teachers across 40 elementary schools were analyzed at individual and school levels. Results indicated that teachers in schools implementing SWPBIS with fidelity had significantly lower levels of burnout and significantly higher levels of efficacy. In addition, an interaction effect implied that teachers benefited most from SWPBIS implementation in schools of low socioeconomic status. Limitations of the study are discussed and directions for future interventions and research are recommended.

Keywords

well-being stress efficacy PBIS SWPBIS

Today’s teachers experience higher levels of accountability within school contexts that include increasingly diverse students, challenging school climates, lower socioeconomic resources, and an array of new initiatives. Teachers report experiencing a multitude of stressors ranging from student discipline problems, to poor working conditions, to a lack of emotional support—all of which have been linked to teacher burnout and, in many cases, teacher turnover (Ingersoll & Smith, 2003; Mitchell & Arnold, 2004; U.S. Department of Education National Center for Education Statistics [NCES], 2007). Conversely, positive school climates have been shown to support teachers’ emotional well-being and sense of competence and in turn improve student outcomes (Grayson & Alvarez, 2008; Jennings & Greenberg, 2009). To this end, school-wide behavioral support has been recommended as a means for supporting teachers (Oliver & Reschly, 2007) and is the focus of the present investigation.

School-Wide positive behavior interventions and supports (SWPBIS) improves school climate, student learning, and social behavior through the implementation of three tiers of support (e.g., Horner et al., 2009; Nelson, Martella, & Marchand-Martella, 2002; Tobin, Dickey, Horner, & Sugai, 2008). But to this point, the majority of research and emphasis has been placed on student outcomes, systems processes, and structures to support teacher implementation of SWPBIS (Horner, Sugai & Anderson, in press; Sugai & Horner, 2006; Sugai, Horner, Fixen, & Blase, 2010), with a paucity of research investigating the relationship between SWPBIS and teacher well-being, including efficacy beliefs and burnout. The purpose of this study was to investigate these relationships by comparing two groups of schools; those that implemented SWPBIS with high fidelity and those that did not. If SWPBIS is effective at improving the overall social culture of a school, it was hypothesized that schools implementing SWPBIS with high fidelity would be more likely to have teachers with higher self-efficacy and lower burnout.

Teacher Well-Being

Teaching requires emotional competence on the part of individual teachers and varies depending on the structures and expectations of the organization within which they are working (Hargreaves, 2000; Morris & Feldman, 1996). Experiencing and appropriately expressing emotions takes effort, especially in situations where emotions are incongruent with the expectations of the organization (Morris & Feldman, 1996). For example, a teacher who is coping with a stressor (e.g., a noncompliant student) may experience some emotional dissonance as he or she acts calm while feeling frustrated and fatigued. Conversely, the emotional labor of teaching may be most rewarding when it is aligned with teachers’ goals and involves circumstances where teachers can successfully reach their goals. Teachers commonly report feeling positive emotions when their students enjoy learning or show affection towards them, particularly in cases of difficult or demanding students (Hargreaves, 2000; Sutton & Wheatley, 2003). Behaviorally speaking, when teachers are reinforced for their efforts through improved academic or behavioral outcomes, their confidence and the likelihood that they will exert that effort in the future increases. But if that effort goes unrewarded, teachers learn over time that the reinforcement they need is not worth the emotional effort required. This can result in drastic changes to well-being, specifically in the areas of efficacy and burnout.

Teacher Efficacy

Teacher self-efficacy involves teachers’ self-judgments about their ability to affect student outcomes, especially for those students who appear unmotivated or difficult to teach. Teachers with high self-efficacy feel that they can be effective with students even in the face of challenging contexts (Guskey & Passaro, 1994; Tschannen-Moran, Woolfolk Hoy, & Hoy, 1998). Teacher efficacy involves not only personal skills and competencies, but how contextual factors, such as resources, affect effective teaching and student support (Tschannen-Moran et al., 1998). For example, when the school context does not support teachers’ efforts to effectively manage the learning environment of their students, this can result in lower student academic engagement and more problem behaviors. On the other hand, teachers’ sense of self-efficacy has been found to positively relate to teachers’ instructional behavior and student outcomes (Armor et al., 1976; Goddard, Hoy, & Hoy, 2000; Ross, 1992). Teachers with higher levels of perceived self-efficacy set more challenging goals, collaborate with colleagues and parents, take personal responsibility for student outcomes, such as individualizing instruction, and are more likely to successfully implement new programs (Allinder, 1995; Han & Weiss, 2005; Skaalvik & Skaalvik, 2007). Teacher efficacy also appears to have a protective function that plays an important role in mediating the relationship between perceived student misbehavior and exhaustion (Tsouloupas, Carson, Matthews, Grawitch, & Barber, 2010), a major component of teacher burnout.

Teacher Burnout

Burnout is defined as the result of repeated exposure to job-related stressors (Maslach, Schaufeli, & Leiter, 2001; Pines & Keinan, 2005). Negative teacher judgments about student behavior and other teaching tasks may contribute to teachers’ repeated experiences of unpleasant emotions, eventually leading to burnout (Chang, 2009). The implications of teacher burnout are far reaching and associated with frequent absences, increased health care costs, poor job performance, and mental health claims (Burke, Greenglass, & Schwarzer, 1996; Leithwood, Menzies, Jantzi, & Leithwood, 1999). When teachers experience burnout, they often feel less effective and struggle to actively engage in teaching practices, thereby reducing the quality and fidelity with which they implement programs (Han & Weiss, 2005). In addition, teachers experiencing burnout are less tolerant of classroom behavior problems and tend to have more negative relationships with their students (Kokkinos, Panayiotou, & Davazoglou, 2004; Yoon, 2002), which can further exacerbate their experiences of stress (Grayson & Alvarez, 2008).

Teacher Burnout has been defined by three linked components: emotional exhaustion, depersonalization, and diminished feelings of personal accomplishment (Kokkinos, 2006; Maslach et al., 2001; Maslach, Jackson, & Leiter, 1996). Emotional exhaustion is considered the tired and fatigued feeling that develops as energies are drained, resulting from repeated incidents where teachers’ emotional efforts go unrewarded. Emotional exhaustion leads teachers to believe that they can no longer give to students as they once could. As teachers experience this extinction of their effort, they often lose their positive feelings about their students and experience the second component of teacher burnout, depersonalization. When teachers experience depersonalization, they separate themselves emotionally from stressful experiences and often have indifferent or negative attitudes toward students. Numerous experiences of unrewarded effort may also result in reduced feelings of personal accomplishment, which may involve teachers experiencing profound disappointment and thoughts that they are no longer contributing to students’ development (Maslach et al., 1996).

Although teaching is a profession associated with fatigue, burnout, and high rates of teacher turnover (Boe, Cook, & Sunderland, 2008; Chang, 2009; NCES, 2007), many of the working conditions identified by teachers as reasons for leaving the profession (i.e., student discipline problems and poor administrative support) can be reduced by targeted and coordinated approaches to providing teacher support (Ingersoll & Smith, 2003). Interventions aimed at supporting teachers should focus not only on teaching individuals skills but on changing the job environment (Maslach et al., 2001). SWPBIS is a tiered intervention model that addresses both of these areas through the implementation of (a) effective practices (teaching individual skills) and (b) the creation of a more positive school environment (changing the job environment).

School-Wide Positive Behavior Interventions and Supports

SWPBIS is based on a three-tiered intervention model designed to match the intensity of the intervention with the severity of the problem behavior. The implementation of a continuum of evidence-based interventions provides the level of support needed for students to demonstrate expected pro-social behavior across various settings (Lewis & Sugai, 1999; Sugai & Horner, 2006). In a prototypic school, 80% of students will respond to primary interventions, 15% to 20% will benefit from secondary interventions, and approximately 5% from tertiary interventions. Across all three tiers, data are used to determine if students are responding to supports, and if not, the need for additional support is identified early, thereby maximizing the use of school resources (Sugai & Horner, 2002).

SWPBIS is embedded with opportunities for teachers to have positive interactions with their students. For example, an emphasis on recognizing students for meeting school-wide expectations increases positive interactions between students and teachers, which is reinforcing for both (McIntosh, Filter, Bennett, Ryan, & Sugai, 2010). Also, by reducing the number of problem behaviors exhibited by students, the implementation of SWPBIS decreases the potential stressors experienced by teachers and may increase teachers’ sense of self-efficacy.

In addition to the provision of simple, efficient, and effective skills, staff in SWPBIS schools learn to work together in the creation of a positive, supportive culture. A major impact of SWPBIS on teacher well-being occurs through the development of team skills, collaboration, and positive relationships, as well as the use of effective practices. The better a school staff learns to teach and reinforce appropriate behavior, discourage inappropriate behavior, monitor students and use data for decisions, the more efficacious they will feel and the more supportive they will perceive the rest of the school.

Successful implementation of SWPBIS allows for the delivery of more effective and efficient behavior support to students, reducing the amount of emotional resources required of individual teachers while increasing the amount of reinforcement (e.g., increased student engagement and positive peer interactions) for teachers’ emotional effort. In turn, teachers with more positive perceptions about their own capacity to affect student outcomes and manage student behavior may implement SWPBIS more effectively, thereby changing the learning environment of their students.

A pilot study examining these relationships was initially completed in 2006–2007 across four middle schools in the Northwest (Ross & Horner, 2007). In the pilot study, the Schoolwide Evaluation Tool (SET; Todd et al., 2003) was used to establish the extent of which schools were implementing SWPBIS with fidelity. Specifically, the high-implementation group consisted of two schools with scores higher than 80% on the SET, and the low-implementation group was composed of two schools with scores lower than 80%. Five teachers were randomly selected from each of the four participating schools (10 teachers per high- and low-implementation group, with a total of 20 participating teachers). Teachers were asked to complete measures of teacher stress (Index of Teaching Stress; Abidin, Greene, & Konold, 2004) and teacher efficacy (Teacher Efficacy Scale; Gibson & Dembo, 1984). Results of the investigation indicated a significant relationship between level of SWPBIS implementation and teacher efficacy, but no significant relationship between implementation and teaching stress.

Although these results were promising, two major limitations existed. First, the small sample size likely did not yield the adequate statistical power to detect a significant relationship between level of SWPBIS implementation and teacher stress. Second, the pilot study only implemented simple analyses of variance between teachers and failed to consider interschool variation. This is important to acknowledge because the results of a multilevel framework evaluating both teacher and school levels are likely to differ from those obtained by studying either individually.

Brownell and Smith (1993) propose a multilevel model for understanding educator outcomes. Their model, an application of Bronfenbrenner’s ecological model (1976), consists of a microsystem, a mesosystem, an exosystem, and a macrosystem. According to Brownell and Smith (1993), the microsystem involves variables in a teacher’s most immediate setting like their own personal characteristics, the characteristics of their own students, and other characteristics related to their own classroom. The mesosystem describes the next level of ecology for a teacher, which includes issues such as collegial and supervisory support, opportunities for professional growth, and power or influence in school-level decisions. The exosystem includes school district variables such as policies, salary and benefits, and the location of a school. Finally, the macrosystem includes the most removed setting from the teacher’s classroom and encompasses such variables as state or federal policies and economic conditions.

The present study was designed to extend investigation of the relationship between variables associated with SWPBIS and teacher perceptions of Burnout and Efficacy using a larger sample, more robust measures, and multilevel analyses that focused on the first two levels of Bronfenbrenner’s framework: the microsystem and the mesosystem. Thus, the primary research questions for this study were (a) Is there a relationship between the implementation of SWPBIS and perceptions of teacher efficacy? and (b) Is there a relationship between the implementation of SWPBIS and perceptions of teacher burnout?

Method

Sample

Elementary schools in Oregon were recruited for this study from a database at the University of Oregon (PBIS Evaluation Website). Based on the school’s preexisting scores on the SET (Todd et al., 2003), 26 low-scoring schools (less than 80% on the SET) and 25 high-scoring schools (80% or higher on the SET) were randomly selected from the 86 schools across the state that had completed the SET during the 2008–2009 school year. Of the 51 schools recruited, 20 high-scoring schools (M = 91.72%, SD = 0.05) and 20 low-scoring schools (M = 61.53%, SD = 0.10) agreed to participate and were selected to do the study. To encourage a high proportion of teachers returning the survey, principals of these schools were asked to randomly select five teachers to complete the anonymous online survey. Selected teachers were sent an email with a link to the survey. Of the 200 randomly selected teachers, we obtained 184 completed surveys (92%) with a mean of 4.65 teachers completing the survey from each school (M = 4.65, range = 4–5, SD = 0.49). The survey was administered in May 2009 and completed surveys were received through July.

Microsystem Measures

The 54-item survey was completed by each randomly selected teacher through an online survey program (www.surveymonkey.com), took approximately 15 minutes to complete, and each teacher received a 10-dollar gift card in appreciation of her or his participation. The survey was composed of three embedded measures. The first embedded measure was composed of demographic items and questions about the teachers’ experience. This included teacher gender, ethnic/racial identification, highest degree awarded, years of student teaching, years teaching, the number of positive school tokens (small rewards) the teacher had given out to students over the past month, the number of times the teacher had reviewed the school-wide expectations with her or his classes over the past month, and the number of office discipline referrals the teacher had completed for students over the past month.

The second embedded measure of the survey asked teachers about their perceptions of teacher stress and burnout as measured by the Maslach Burnout Inventory–Educators Survey (MBI-ES; Maslach & Jackson, 1981). The MBI-ES has been used extensively to identify burnout levels of individuals who work in school settings across the three factors as described in the introduction: Emotional Exhaustion, Depersonalization, and Personal Accomplishment. The MBI-ES, a 22-item survey, asks how often teachers experience certain emotions on a 7-point scale, ranging from never to every day. For example, one item on the scale asks, “How often I feel emotionally drained from my work?”

Two studies have substantiated the validity and reliability of the MBI-ES. Factor-analytic studies by Gold (1984), with 462 California students, and Iwanicki and Schwab (1981), with 469 Massachusetts teachers, support the three-factor structure of the MBI-ES. Gold reports Cronbach’s alpha estimates of .88 for Emotional Exhaustion, .74 for Depersonalization, and .72 for Personal Accomplishment, whereas Iwanicki and Schwab report estimates of .90, .76, and .76, respectively. The third embedded measure of the survey focused on the teachers’ perceptions of personal efficacy as measured by the Teachers’ Sense of Efficacy Scale (TSES; Tschannen-Moran & Woolfolk Hoy, 2001). The TSES is a 24-item survey based on the model of teacher efficacy by Tschannen-Moran et al. (1998), which suggests that “a valid measurement of teacher efficacy must assess both personal competence and an analysis of the task in terms of the resources and constraints in particular teaching contexts” (p. 795). The TSES asks how much teachers can do on a 9-point scale, with responses ranging from nothing to a great deal. It includes items such as “How much can you do to get through to the most difficult students?” and “How much can you do to get students to believe they can do well in school work?” Principal-axis factoring conducted with in-service and preservice teachers of Ohio (N = 366) on the measure indicated three teacher efficacy subscales (instruction, management, and engagement), but when additional analyses were conducted specifying one factor, all items loaded on this factor, with loadings ranging from 0.49 to 0.76 and a reliability of 0.94. Thus the authors contend that either the total score or the subscale scores can be used to assess the underlying construct of teacher efficacy.

Mesosystem Measures

Once information at the microsystem or teacher level was considered, the study analyzed predictors at the school or mesosystem level, which included environmental factors and school-level practices. From the survey of teachers, aggregate factor scores were computed for teacher efficacy and the three scales of teacher burnout. Also, averages for years teaching, years of student teaching, positive school tokens, school-wide expectation reviews, and office discipline referrals were computed for each school. We then analyzed existing information regarding socioeconomic status (SES) as measured by the percentage of students receiving free and reduced-price lunch. The mean SES for high-scoring schools was 49.73% (M = 49.73%, range = 10%–88%, SD = 0.24) and the mean SES for low-scoring schools was 51.31% (M = 51.31%, range = 12%–85%, SD = 0.22). Finally, we considered average number of students per teacher full-time equivalent (FTE) and level of implementation of SWPBIS using the SET (Todd et al., 2003). The SET evaluates 28 research questions across seven feature areas (Horner et al., 2004). The feature areas include (a) expectations defined, (b) behavioral expectations taught, (c) acknowledgement procedures, (d) correction procedures, (e) monitoring and evaluation, (f) management, and (g) district-level support. Scoring for the SET involves assigning 0, 1, or 2 (0 = not implemented, 1 = partially implemented, 2 = fully implemented) for each of the 28 items. Multiple sources are used to gather information necessary for the SET such as interviews with administrators, teachers, staff members, and students; reviewing permanent products including school policies, training curricula, and meeting minutes; and examining current data systems. The internal consistency reliability of the SET has documented an overall alpha of .96 and exceeds standard psychometric criteria for discriminability, internal consistency, and test–retest reliability. All SET scores were attained within spring before the completion of all other measures.

A multilevel regression approach was used to analyze the individual teacher-level (microsystem) and school-level (mesosystem) variables. Specifically, we used the HLM6 (Raudenbush, Bryk, Cheong, & Congdon, 2000) software. This approach allowed the simultaneous analysis of teacher- and school-level variables and the interactions between the two. We applied models using the teacher- and school-level variables thought to affect teacher burnout and teacher efficacy. We then refined the models to retain only those predictors that were statistically significant. Full maximum likelihood was used to simultaneously estimate the fixed effects and the variance components. Predictor variables were grand mean centered throughout the analysis, and surveys with missing data were eliminated from the analysis. To clarify, the survey entry format did not allow for missing responses and only 16 participants (8%) did not complete the survey in its entirety.

Results

Correlations, means, and standard deviations of selected teacher (microsystem) and school (mesosystem) variables are presented in Table 1 (additional information on other variables not included here is available on request). Three models were applied to the efficacy and burnout data. The first model was an unconditional model that examined within- and between-school variation. This model is equivalent to a random effects analysis of variance testing whether there were significant differences between the schools. It was also used as a baseline model for comparison with other models. Application of the unconditional model showed that there were significant differences between schools for each of the outcome measures: teacher efficacy, t(39) = 127.96, SE = 1.31, p < .001; Emotional Exhaustion, t(39) = 43.76, SE = 0.46, p < .001; Personal Accomplishment, t(39) = 86.80, SE = 0.40, p < .001; and Depersonalization, t(39) = 33.73, SE = 0.29, p < .001.

Table 1.

Correlations, Means, and Standard Deviations for Selected Teacher (Microsystem) and School (Mesosystem) Variables

Variable	1	2	3	4	5	6	7	8	9	M	SD
Level 1 (n = 184)
1. Years of student teaching	1.000									2.28	0.81
2. Years teaching	−.303	1.000								4.78	1.94
3. Positives given last month	.105	.062	1.000							3.00	1.09
4. Expectations reviewed	−.028	.105	.242	1.000						2.22	1.08
5. Office discipline referrals	−.146	−.083	−.451	−.110	1.000					2.09	1.04
6. Teacher efficacy	−.016	.101	.198	.226	−.150	1.000				167.47	14.84
7. Emotional exhaustion	−.042	−.110	−.229	−.189	.406	−.264	1.000			20.18	4.05
8. Depersonalization	.006	−.123	−.177	−.144	.203	−.186	.545	1.000		9.66	3.32
9. Personal accomplishment	.102	.116	.183	.175	−.190	.286	−.376	−.376	1.000	35.00	3.81
Level 2 (n = 40)
1. Socioeconomic status	1.000									0.51	0.23
2. SET score	−.151	1.000								0.77	0.17
3. Average students per FTE	−.008	−.150	1.000							19.80	2.17

Note. SET = Schoolwide Evaluation Tool; FTE = full-time equivalent.

Next we attempted to find the best set of predictor (microsystem) variables among individual teachers out of the following: years of teaching, years of student teaching, office discipline referrals, positives rewards, and reviews of school-wide expectations. For teacher efficacy, only the number of reviews of school-wide expectations was significant and therefore retained for further analysis. This reduced model resulted in a significant improvement in model fit, χ²(3) = 8.16, p = .042. For Depersonalization, only positive rewards given was significant, again resulting in improved model fit, χ²(3) = 11.47, p = .01. For the remaining two scales of teacher burnout, Emotional Exhaustion and Personal Accomplishment, no microsystem predictors were found to be significant and therefore retained for further analysis.

The third model used school-level (mesosystem) variables to predict variation in the microsystem parameters. Several initial analyses were conducted using exploratory analysis to identify the best predictors from the eight context variables: SET score, school socioeconomic status, average number of students per teacher FTE, average years teaching, average years student teaching, average positive school tokens, average school-wide expectation reviews, and average office discipline referrals. Among these predictors, school SET scores and school socioeconomic status were significantly related to all of the burnout and efficacy outcome measures. In addition, for Depersonalization and Personal Accomplishment, interaction effects between socioeconomic status and SET score were discovered. The final, refined version of each model containing only the significant predictors appears in Table 2.

Table 2.

Summary Statistics for the Final Multilevel Regression Models

Effect	Coefficient	t	df	p
Teacher efficacy
Fixed effects, Level 1, teachers
Intercept (B0)	167.44 (0.91)	184.86	37	<.001
Expectation reviews (B1)	1.96 (0.98)	2.00	39	.052
Fixed effects, Level 2, schools
SET score	19.90 (5.50)	3.62	37	.001
SES	−15.63 (4.08)	−3.83	37	<.001
Random effects
Intercept (U0)	0.13 (0.36)		37	>.500
Level 1 (R)	180.19 (13.42)
Emotional exhaustion
Level 1, teachers
Intercept (B0)	20.21 (0.35)	15.08	37	<.001
Level 2, schools
SET score	−8.57 (1.98)	−4.32	37	<.001
SES	4.62 (1.46)	3.15	37	.004
Random effects
Intercept (U0)	2.54 (1.59)		36	<.001
Level 1 (R)	9.91 (3.15)
Depersonalization
Level 1, teachers
Intercept (B0)	9.66 (0.17)	44.83	36	<.001
Positives (B1)	−.59 (0.23)	−2.55	39	.015
Level 2, schools
SET score	2.00 (2.74)	.73	36	.470
SES	11.39 (3.38)	3.37	36	.002
SET × SES interaction	−9.68 (3.93)	−2.46	36	.019
Random effects
Intercept (U0)	0.00 (0.05)		36	>.500
Level 1 (R)	10.34 (3.22)
Personal accomplishment
Level 1, teachers
Intercept (B0)	35.00 (0.25)	142.59	36	<.001
Level 2, schools
SET score	−5.01 (2.88)	−1.74	36	.090
SES	−22.93 (4.35)	−5.27	36	<.001
SET × SES interaction	23.02 (5.33)	4.32	36	<.001
Random effects
Intercept (U0)	0.37 (0.62)		36	.097
Level 1 (R)	11.03 (3.32)

Note. Standard errors are in parentheses. SET = Schoolwide Evaluation Tool; SES = socioeconomic status.

Teacher Efficacy

The average composite efficacy score across all teachers was 167.44. Although individual teacher expectation reviews was not significant in the final model, the two school-level variables assessed were far more predictive. The average level of teacher efficacy was predicted best by SET score and school socioeconomic status. This indicates that schools with higher PBIS implementation and higher socioeconomic status had significantly higher composite teacher efficacy scores. Application of the final, reduced Level 2 model accounted for 42.51% of the variance in school average differences in teacher efficacy and resulted in a significant reduction in unexplained variance over the unconditional model, χ²(5) = 30.91, p < .001.

Emotional Exhaustion

In the case of the first teacher burnout scale, only the two school-level variables retained a significant relationship with Emotional Exhaustion, indicating that schools with higher PBIS implementation and higher socioeconomic status had significantly lower average scores on Emotional Exhaustion, regardless of individual teacher responses. Application of the final, reduced Level 2 model accounted for 44.05% of the variance in school average differences in Emotional Exhaustion and again resulted in a significant reduction in unexplained variance over the unconditional model, χ²(2) = 22.90, p < .001.

Depersonalization

At the individual teacher level, “positives rewards given” remained significantly, negatively related to teacher scores on Depersonalization, but at the school level, the final Level 2 model resulted in some stark changes. When the Level 2 model was run without considering interaction effects, both SET score and socioeconomic status continued to have a significant relationship with average teacher score on Depersonalization, indicating that schools with higher SWPBIS implementation and socioeconomic status had lower scores on average teacher Depersonalization. But when the interaction between SET score and SES was included in the analysis, not only was the interaction effect significant but the amount of variance accounted for by school SET score alone was greatly reduced, resulting in a nonsignificant main effect. These results indicate that the relationship between SET score and Depersonalization was moderated by the school’s socioeconomic status. The negative coefficient of −9.68 indicates that in schools with lower socioeconomic status, the SET score–Depersonalization relationship was stronger, and in schools with higher socioeconomic status, that relationship was weaker. In other words, when a school had lower socioeconomic status, school implementation of PBIS had a larger impact on teacher Depersonalization. Application of this final, reduced Level 2 model accounted for 40.84% of the variance in school average differences in teacher Depersonalization and again resulted in a significant reduction in unexplained variance over the unconditional model, χ²(6) = 34.14, p < .001.

Personal Accomplishment

A similar situation was discovered with the final measure of teacher burnout. First, as in the case of Emotional Exhaustion, none of the Level 1 variables were significantly related to Personal Accomplishment when analyzed using the final model. And like Depersonalization, before the interaction between SES and SET was considered, the two school-level variables retained strong significant relationships with Personal Accomplishment, indicating that schools with higher PBIS implementation and higher socioeconomic status had significantly higher average scores on Personal Accomplishment. But when the interaction effect was included in the final model, a significant interaction effect was again discovered between school socioeconomic status and SET score, where the relationship between SET score and Personal Accomplishment was moderated by the school’s socioeconomic status. Figure 1 shows schools in the top, middle, and lower third on socioeconomic status, indicating the moderating relationship of socioeconomic status on the SET–Personal Accomplishment relationship.

Figure 1.

The interaction between socioeconomic status and SET score on Personal Accomplishment

In schools with high SES, the slope of the SET–Personal Accomplishment relationship was nearly flat and insignificant. In schools with middle or low SES, the slope of the SET–Personal Accomplishment relationship was steeper. Thus, in schools with higher average socioeconomic status, SET score was not as strongly related to Personal Accomplishment, but when a school had a low socioeconomic status, teacher Personal Accomplishment was associated more strongly with the school’s implementation of SWPBIS.

Application of this final, reduced Level 2 model accounted for 63.46% of the variance in school average differences in teacher Personal Accomplishment and resulted in a significant reduction in unexplained variance over the unconditional model, χ²(3) = 38.22, p < .001.

Comparison With National Norms

National norms provided by the Maslach Burnout Inventory Manual (Maslach & Jackson, 1981) were used to better understand how the teachers in our two samples compared to teachers across the country. As indicated in Table 3, teachers tend to have slightly higher mean scores on Emotional Exhaustion (teachers M = 21.25, overall sample M = 20.99); substantially higher scores on Depersonalization (teachers M = 11.00, overall sample M = 8.73); and lower scores on Personal Accomplishment (teachers M = 33.54, overall sample M = 34.58). Results indicated that teachers in our low-SET sample were similar to teacher national norms, but that teachers in high-SET schools scored considerably better.

Table 3.

Means and Standard Deviations for the MBI Subscales

	MBI subscales
	Emotional exhaustion		Depersonalization		Personal accomplishment
	M	SD	M	SD	M	SD
Overall sample (n = 11,067)	20.99	10.75	8.73	5.89	34.58	7.11
Teaching subgroup (n = 4,163)	21.25	11.01	11.00	6.19	33.54	6.89
Low SET score sample (n = 90)	22.06	3.88	10.83	3.86	33.32	3.94
High SET score sample (n = 94)	18.38	3.41	8.53	2.84	36.61	2.89

Note. MBI = Maslach Burnout Inventory; SET = Schoolwide Evaluation Tool.

Discussion

Findings indicate strong relationships between SWPBIS implementation and teacher perceptions of efficacy and burnout. From the multilevel, ecological perspective, SWPBIS affects both school-level (mesosystem) and teacher level (microsystem) variables. At the mesosystem level, by changing the culture of a school through systems and data usage, SWPBIS improves teaming structures, opportunities for collaboration, and positive interactions with adults and students. At the microsystem level, SWPBIS increases evidence-based practices, such as the teaching of expectations and the delivery of positive reinforcement. It is likely that both levels of intervention affect teacher well-being, and results highly encourage their adoption.

In addition, this study provides a couple of particularly interesting findings regarding school practices and teacher well-being. One finding is that none of the teacher or school-level variables involving gender, ethnicity, education, experience, or specific SWPBIS practices were effective predictors of teacher efficacy or teacher burnout in the final models. This is worth noting because in other studies, these variables often have been related to teacher well-being (Abel & Sewell, 1999; Brouwers & Tomic, 2000; Brown & Ralph, 2002; Caprara, Barbranelli, Steca, & Malone, 2006; Dewberry & Briner, 2007). A number of other variables suggested in the literature may have been related significantly to the outcome measures individually but were not significant when other predictors were in the equation. This was particularly true for the teacher variables involving SWPBIS practices: the delivery of positive school tokens and the teaching/reviewing of school-wide expectations.

A second, even more critical, finding in this study involved the relationship between SWPBIS implementation and school socioeconomic status. Research has clearly demonstrated that student outcomes benefit from the implementation of SWPBIS regardless of community status, but the results here indicate that the well-being of teachers in schools of lower socioeconomic status may benefit the most from SWPBIS implementation, where the practices of SWPBIS greatly increase their feelings of effectiveness and the emotional resources they have at their disposal. These results will be especially valuable to schools in financially struggling areas across the country, where teachers are often reluctant to implement new initiatives like SWPBIS because of feelings that their plates are already overflowing. In fact, these findings suggest just the opposite: through effective and efficient implementation, SWPBIS may reduce the amount (of problem behavior) on teachers’ plates and may even increase the size of those plates.

Limitations

The results of the present, exploratory analysis are limited in a number of ways. First, the analyses stretched the limits of the sample—we conducted numerous analyses with 184 teachers in 40 schools, which were selected from a relatively small list of 86 schools in Oregon that had completed the SET. Although we believe the results presented here paint a consistent picture and imply the relationship between SWPBIS implementation and teacher well-being, replication is paramount and with different students and teachers, and in different settings. Such replication will happen with ongoing studies, but we nevertheless caution against overgeneralizing from the present results. Second, although we attempted to analyze a substantial number of teacher- and school-level variables, other theoretically important variables (e.g., leadership skills, teacher age, IQ, professional development opportunities, percentage of minority students per school, school resources, and school climate) were not collected. With a sample of only 184 teachers, the problem of overfitting was a concern, and with that in mind, the analysis still may have led to some overfitting and spurious results (Babyak, 2004; Burnham & Anderson, 2002 ), which may be further exaggerated in the multilevel model (Snijders & Bosker, 1993). Finally, the inability to collect survey information prior to SWPBIS implementation in each of the study schools makes it impossible to determine the directionality of the relationship between teacher well-being and SWPBIS implementation fidelity. The results of this study indicate that schools implementing SWPBIS with high degrees of fidelity have teachers with increased well-being. But does SWPBIS implementation improve feelings of well-being in teachers, or are highly efficacious and stress-free teachers more likely to “buy in” and implement SWPBIS with fidelity? Both are likely, and longitudinal research assessing teachers throughout their implementation of SWPBIS will be necessary to effectively determine how much of a role each plays.

Future Directions

The use of SWPBIS is just a first step in supporting teacher well-being. Although SWPBIS involves multiple tiers of support for students, the system can only be considered a first tier of support for teachers. It is highly probable that a certain percentage of teachers will need support beyond what SWPBIS provides, so it is imperative that future efforts develop and evaluate secondary- and tertiary-level interventions that can work in an RtI-like system of support for teacher well-being. For example, the measures of burnout and efficacy used in the present study can be used as a “universal screener,” and teacher scores falling below a certain range can initiate a secondary-level response, where simple, generic, group-level interventions are provided and may include additional training in classroom/behavior management as well as strategies for responding to emotional challenges. Teachers not responsive to this level of intervention or those who express the need could receive tertiary support. One such intervention of promise in this area is Acceptance and Commitment Therapy or Training (ACT), which directly combats what is often considered the trap of the dedicated teacher (Hayes, Luoma, Bond, Masuda, & Lillis, 2006). From an ACT perspective, a teacher’s reason for entering the field often involves a value of caring for children and seeing them develop, but the difficult emotions associated with challenging school environments are incompatible with their self-images as competent and caring people. Teachers reluctant to have such feelings may experience “experiential avoidance,” but unfortunately, efforts to control or eliminate difficult feelings may actually exacerbate them (Wegner & Erber, 1992). This can lead to attempts to stifle problematic feelings through alcohol or drugs (Jarvis, 2002; Watts & Short, 1990), and ultimately may lead to attrition from the teaching field (Jackson, Schwab, & Schuler, 1986; Miller, Brownell, & Smith, 1999). Instead, ACT encourages people to accept—but not to take literally—their thoughts and feelings and to commit valued actions even amid those unpleasant thoughts and feelings. Findings of more than 20 randomized control trials have demonstrated that ACT can improve a wide range of psychological problems, including depression, anxiety, pain, job burnout, prejudice, and epilepsy (Hayes et al., 2006).

The present results suggest that school-wide positive behavior support may have positive effects not only on student outcomes but on teacher outcomes as well, a consideration that deserves more attention than it currently receives. Improved student outcomes remain the educator’s ultimate goal, but teacher well-being will be a critical step along the way. By providing teachers with systemic, RtI-like support starting with effective school-wide behavior support practices, we can give teachers the tools necessary to maintain effective educational efforts and mental wellness despite the increasing challenges that continue to be placed at their feet.

Footnotes

Action Editor: Lori Newcomer

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Office of Special Education Programs, U.S. Department of Education (H326S980003). Opinions expressed herein are those of the authors and do not necessarily reflect the position of the U.S. Department of Education, and such endorsements should not be inferred.

References

Abel

M. H.

Sewell

(1999). Stress and burnout in rural and urban secondary school teachers. Journal of Educational Research, 92, 287–293.

Abidin

R. R.

Greene

R. W.

Konold

T. R.

(2004). Index of teaching stress: Professional manual. Lutz, FL: Psychological Assessment Resources.

Allinder

R. M.

(1995). An examination of the relationship between teacher efficacy and curriculum-based measurement and student achievement. Remedial and Special Education, 16, 247–254.

Armor

Conroy-Oseguera

Cox

King

McDonnell

Pascal

Pauly

Zellman

(1976). Analysis of the school preferred reading programs in selected Los Angeles minority schools, REPORT NO. R-2007-LAUSD. Santa Monica, CA: Rand Corporation.

Babyak

M. A.

(2004). What you see may not be what you get: A brief, nontechnical introduction to overfitting in regression-type models. Psychosomatic Medicine, 66, 411–421.

Boe

E. E.

Cook

L. H.

Sunderland

R. J.

(2008). Teacher turnover: Examining exit attrition, teaching area transfer, and school migration. Exceptional Children, 75, 7–31.

Bronfenbrenner

(1976). The experimental ecology of education. Teachers College Record, 78, 157–204.

Brouwers

Tomic

(2000). A longitudinal study of teacher burnout and perceived self-efficacy in classroom management. Teaching and Teacher Education, 16, 239–253.

Brown

Ralph

(2002). Teacher stress and school improvement. Improving Schools, 5(2), 55–65.

10.

Brownell

M. T.

Smith

S. W.

(1993). Understanding special education teacher attrition: A conceptual model and implications for teacher educators. Teacher Education and Special Education, 15, 270–282.

11.

Burke

R. J.

Greenglass

E. R.

Schwarzer

(1996). Predicting teacher burnout over time: Effects of work stress, social support, and self-doubts on burnout and its consequences. Anxiety, Stress, and Coping: An International Journal, 9, 261–275.

12.

Burnham

K. P.

Anderson

D. R.

2002. Model selection and multimodel inference: A practical information-theoretic approach (2nd ed.). New York, NY: Springer.

13.

Caprara

Barbranelli

Steca

Malone

(2006). Teachers’ self-efficacy beliefs as determinants of job satisfaction and students’ academic achievement: A study at the school level. Journal of School Psychology, 44, 473–490.

14.

Chang

(2009). An appraisal perspective of teacher burnout: Examining the emotional work of teachers. Educational Psychology Review, 21, 193–218.

15.

Dewberry

Briner

(2007). Report for Worklife Support on the relation between well-being and climate in schools and pupil performance. London, UK: Worklife Support.

16.

Gibson

Dembo

(1984). Teacher efficacy: A construct validation. Journal of Educational Psychology, 76, 569-582.

17.

Goddard

R. D.

Hoy

W. K.

Woolfolk Hoy

(2000). Collective teacher efficacy: Its meaning, measure, and impact on student achievement. American Educational Research Journal, 37, 479–507.

18.

Gold

(1984). The factorial validity of the Maslach Burnout Inventory in a sample of California elementary and junior high school classroom teachers. Educational and Psychological Measurement, 44, 1009–1016.

19.

Grayson

J. L.

Alvarez

H. K.

(2008). School climate factors relating to teacher burnout: A mediator model. Teaching and Teacher Education, 24, 1349–1363.

20.

Guskey

T. R.

Passaro

P. D.

(1994). Teacher efficacy: A study of construct dimensions. American Educational Journal, 31, 627–643.

21.

Han

S. S.

Weiss

(2005). Sustainability of teacher implementation of school-based mental health programs. Journal of Abnormal Child Psychology, 33, 665–679.

22.

Hargreaves

(2000). Mixed emotions: Teachers’ perceptions of their interactions with students. Social Psychology of Education, 16, 811–826.

23.

Hayes

S. C.

Luoma

J. B.

Bond

F. W.

Masuda

Lillis

(2006). Acceptance and commitment therapy: Model, processes and outcomes. Behavior Research and Therapy, 44, 1–25.

24.

Horner

R. H.

Sugai

Anderson

C. M.

(2010). Examining the evidence base for school-wide positive behavior support. Focus on Exceptional Children, 42(8), 1-14.

25.

Horner

R. H.

Sugai

Smolkowski

Eber

Nakasato

Todd

A. W.

Esperanza

(2009). A randomized, wait-list controlled effectiveness trail assessing school-wide positive behavior support in elementary schools. Journal of Positive Behavior Interventions, 11, 133–144.

26.

Horner

R. H.

Todd

A. W.

Lewis-Palmer

Irvin

L. K.

Sugai

Boland

J. B.

(2004). The School-wide Evaluation Tool: A research instrument for assessing school-wide positive behavior support. Journal of Positive Behavior Interventions, 6, 3–12.

27.

Ingersoll

R. M.

Smith

T. M.

(2003). The wrong solution to the teacher shortage. Educational Leadership, 60, 30–33.

28.

Iwanicki

E. F.

Schwab

R. L.

(1981). A cross-validation study of the Maslach Burnout Inventory. Educational and Psychological Measurement, 41, 1167–1174.

29.

Jackson

S. E.

Schwab

R. L.

Schuler

R. S.

(1986). Toward an understanding of the burnout phenomenon. Journal of Applied Psychology, 71, 630–640.

30.

Jarvis

(2002). Teacher stress: A critical review of recent finding and suggestions for future research direction. In Stress News 2002. South Petherton, UK: International Stress Management Association.

31.

Jennings

P. A.

Greenberg

M. T.

(2009). The prosocial classroom: Teacher and social and emotional competence in relation to student and classroom outcomes. Review of Educational Research, 79, 491–525.

32.

Kokkinos

C. M.

(2006). Factor structure and psychometric properties of the Maslach Burnout Inventory Survey among elementary and secondary school teachers in Cyprus. Stress and Health, 22, 25–33.

33.

Kokkinos

C. M.

Panayiotou

Davazoglou

A. M.

(2004). Correlated of teacher appraisals of student behaviors. Psychology in the Schools, 42, 79–89.

34.

Leithwood

K. A.

Menzies

Jantzi

Leithwood

(1999). Teacher burnout: A critical challenge for leaders of restructuring schools. In Vandenberghe

Huberman

A. M.

(Eds.), Understanding and preventing teacher burnout: A sourcebook of international research and practice (pp. 85–115). New York, NY: Cambridge University Press.

35.

Lewis

Sugai

(1999). Effective behavior support: A systemic approach to proactive schoolwide management. Focus on Exceptional Children, 31, 1–24.

36.

Maslach

Jackson

S. E.

(1981). Maslach Burnout Inventory. Palo Alto, CA: Consulting Psychologists Press.

37.

Maslach

Jackson

S. E.

Leiter

M. P.

(1996). Maslach Burnout Inventory manual (3rd ed.). Mountain View, CA: Consulting Psychologists Press.

38.

Maslach

Schaufeli

W. B.

Leiter

P. M.

(2001). Job burnout. Annual Review of Psychology, 52, 397–422.

39.

McIntosh

Filter

K. J.

Bennett

J. L.

Ryan

Sugai

(2010). Principles of sustainable prevention: Designing scale-up of school-wide positive behavior support to promote durable systems. Psychology in the Schools, 47, 5–21.

40.

Miller

M. D.

Brownell

M. T.

Smith

S. W.

(1999). Factors that predict teachers staying in, leaving, or transferring from the special education classroom. Exceptional Children, 65, 201–218.

41.

Mitchell

Arnold

(2004). Behavior management skills as predictors of retention among south Texas special educators. Journal of Instructional Psychology, 31, 214–219.

42.

Morris

J. A.

Feldman

D. C.

(1996). The dimensions, antecedents, and consequences of emotional labor. Academy of Management Review, 21, 986–1010.

43.

Nelson

J. R.

Martella

Marchand-Martella

(2002). Maximizing student learning: The effects of a comprehensive school-based program for preventing problem behaviors. Journal of Emotional and Behavioral Disorders, 10, 136–148.

44.

Oliver

S. M.

Reschly

D. J.

(2007). Improving student outcomes in general and special education: Effective classroom management. Washington DC: Learning Point Associates, National Comprehensive Center on Teacher Quality.

45.

Pines

A. M.

Keinan

(2005). Stress and burnout: The significant difference. Personality and Individual Differences, 39, 625–635.

46.

Raudenbush

S. W.

Bryk

A. S.

Cheong

Congdon

(2000). HLM 5 for Windows [Computer software]. Lincolnwood, IL: Scientific Software International.

47.

Ross

J. A.

(1992). Teacher efficacy and the effect of coaching on student achievement. Canadian Journal of Education, 17, 51–65.

48.

Ross

S. W.

Horner

R. H.

(2007). Teacher outcomes of school-wide positive behavior support. Teaching Exceptional Children Plus, 3. Retrieved from http://journals.cec.sped.org/tecplus

49.

Skaalvik

E. M.

Skaalvik

(2007). Dimensions of teacher self-efficacy and relations with strain factors, perceived collective teacher efficacy, and teacher burnout. Journal of Educational Psychology, 99, 611–625.

50.

Snijders

T. A. B.

Bosker

R. J.

(1993). Standard errors and sample sizes for two-level research. Journal of Educational Statistics, 18, 237–259.

51.

Sugai

Horner

R. H.

(2002). The evolution of discipline practices: School-wide positive behavior supports. Child and Family Behavior Therapy, 24, 23-50.

52.

Sugai

Horner

R. H.

(2006). A promising approach for expanding and sustaining school-wide positive behavior support. School Psychology Review, 35(2), 245-259.

53.

Sugai

Horner

R. H.

Fixen

Blase

(2010). Developing systems-level capacity for RTI implementation: Current efforts and future directions. In Glover

T. A.

Vaughn

(Eds.), The promise of response to intervention: Evaluating current science and practice (pp. 286–309). New York, NY: Guilford.

54.

Sutton

R. E.

Wheatley

K. F.

(2003). Teachers’ emotions and teaching: A review of the literature and directions for future research. Educational Psychology Review, 15, 327–358.

55.

Tobin

T. J.

Dickey

C. R.

Horner

R. H.

Sugai

(2008). Comprehensive implementation of the three-tiered prevention approach to behavioral support. In Greenwood

C. R.

Kratochwill

T. R.

Clements

(Eds.), Understanding and preventing teacher burnout: A sourcebook of international research and practice (pp. 87–114). New York, NY: Guilford.

56.

Todd

A. W.

Lewis-Palmer

Horner

R. H.

Sugai

Sampson

N. K.

Phillips

(2003). The School-wide Evaluation Tool-: SET implementation manual. Eugene, OR: University of Oregon.

57.

Tschannen-Moran

Woolfolk Hoy

Hoy

W. K.

(1998). Teacher efficacy: Its meaning and measure. Review of Educational Research, 68, 202–248.

58.

Tschannen-Moran

Woolfolk Hoy

(2001). Teacher efficacy: Capturing an elusive construct. Teaching and Teacher Education, 17, 783–805.

59.

Tsouloupas

C. N.

Carson

R. L.

Matthews

Grawitch

M. J.

Barber

L. K.

(2010). Exploring the association between teachers’ perceived student misbehavior and emotional exhaustion: The importance of teacher efficacy beliefs and emotion regulation. Educational Psychology, 30, 173–189.

60.

U.S. Department of Education National Center for Education Statistics. (2007). Teacher follow-up survey (“Questionnaire for Current Teachers” and “Questionnaire for Former Teachers”), 2004–05. Washington, DC: Government Printing Office.

61.

Watts

W. D.

Short

A. P.

(1990). Teacher drug use: A response to occupational stress. Journal of Drug Education, 20, 47–65.

62.

Wegner

D. M.

Erber

(1992). The hyperaccessibility of suppressed thoughts. Journal of Personality and Social Psychology, 63, 903–912.

63.

Yoon

J. S.

(2002). Teacher characteristics as predictors of student-teacher relationships: Stress, negative affect, and self-efficacy. Social Behavior and Personality, 30, 485–494.