Workload Manageability Among Novice Special and General Educators: Relationships With Emotional Exhaustion and Career Intentions

Abstract

Novice special educators (those in their first 3 years) consistently report their workloads are unmanageable. Yet, it is not clear whether their perceptions of workload manageability contribute to outcomes of concern such as emotional exhaustion (a component of burnout) or intentions to continue teaching in their schools and districts. This pilot investigation used structural equation modeling to analyze data collected for the Michigan Indiana Early Career Teacher Study. We found (a) novice elementary and middle school special educators rated their workloads less manageable than novice elementary and middle school general educators; (b) novice special and general educators’ ratings of workload manageability predicted emotional exhaustion, which mediated a relationship between workload manageability and career intentions; and (c) the magnitude of the relationships was stronger for novice general educators. Results have implications for supporting and retaining novice special and general education teachers.

Keywords

special education leadership personnel preparation preparation teacher(s)workload beginning teachers

Novice special educators (those in their first 3 years teaching) often report feeling overwhelmed by trying to fulfill their complex responsibilities, while simultaneously learning how to teach and negotiating roles in schools (Billingsley, Griffin, Smith, Kamman, & Israel, 2009; Youngs, Jones, & Low, 2011). For instance, in Youngs and colleagues’ investigation, novice special education teachers (SETs) explained that their roles were poorly defined and they had to define their roles while learning; this led them to feel overwhelmed. In Billingsley, Carlson, and Klein’s (2004) investigation, more than 75% of novice SETs reported routine duties interfered with their efforts to teach, and more than a quarter rated their workloads “not at all” manageable. Similarly, novice SETs who participated in Griffin and colleagues’ (2009) investigation reported that time to do their work was their most pressing concern.

Novice SETs’ concerns about workloads are well-documented (e.g., Griffin et al., 2009), but no studies have investigated how SETs’ workloads compare with other novices. Prior studies also have not examined whether novice SETs’ perceptions of workload manageability have consequences of concern. Thus, it is unclear whether perceptions of workload manageability are important to address; should leaders improve novice SETs’ workload manageability, or are overwhelming workloads a normal part of beginning a new career? Furthermore, do all novices need support for managing workloads, or do novice SETs require particular attention?

Special and General Educators’ Workloads

SETs and general education teachers (GETs) fulfill different roles in schools (McCray, Butler, & Bettini, 2014). GETs are responsible for all students, and typically provide instruction in standards-based, grade-level content. They are also responsible for differentiating instruction and providing Tier 2 interventions to students who struggle, but their instruction focuses on general education curricula (Brownell, Sindelar, Kiely, & Danielson, 2010; McCray et al., 2014).

SETs’ roles are less clearly defined; there continues to be debate about the extent to which they should be supporting students in learning general education curricula versus implementing interventions to address individual skill deficits (McLaughlin, 2010). SETs are often responsible for developing accommodations and modifications to support students in learning general education curricula in all content areas, in addition to implementing interventions across some combination of reading, math, writing, adaptive skills, and social-emotional skills (McCray et al., 2014). They also collaborate with parents, paraprofessionals, related service providers, and GETs (Brownell et al., 2010). Because students with disabilities have unique learning needs, SETs also may have to make independent decisions about what, when, and how to teach—issues that, for GETs, are dictated by class schedules, curriculum calendars, and instructional materials (Youngs et al., 2011). Furthermore, SETs also have extensive administrative and supervisory responsibilities (Vannest & Hagan-Burke, 2010).

Recent studies have raised concerns about whether SETs’ workloads are reasonable. SETs have limited time for their most important responsibilities—planning, providing, and collaborating on instruction (Bettini, Kimerling, Park, & Murphy, 2015; Mitchell, Deshler, & Lenz, 2012). For instance, Vannest and Hagan-Burke (2010) found that SETs in one district provided instruction (including academic instruction, nonacademic instruction, and instructional support) during only 35% of the school day, engaged in collaboration during 8% of the day, developed instructional plans during 5% of the day, and assessed students during 4% of the day. Almost half of the day was dedicated to administrative and supervisory tasks (Vannest & Hagan-Burke, 2010). Other studies obtained similar results (Bettini et al., 2015; Mitchell et al., 2012). Novice SETs are still developing instructional skill, and thus, they are likely to be especially challenged by limited time for planning and providing instruction (Billingsley et al., 2009).

Conservation of Resources (COR): Consequences of Unmanageable Workloads

COR theory suggests that overwhelming workloads may reduce the energy novice SETs have to engage in their responsibilities, leaving them burnt out. COR theory posits that individuals have limited resources (e.g., time, energy) which they deploy strategically; employees who experience prolonged periods of high demands and low resources—for instance, extensive responsibilities with limited time and materials—respond by reducing the energy they invest in their work, their identification with their work, and their efficacy for completing their work well. In other words, they experience burnout, “a psychological phenomenon of prolonged exhaustion and disinterest . . . in the work” (Alarcon, 2011, p. 549). In a meta-analysis of research on COR theory, Alarcon found that employees who rated workloads less manageable were also more likely to plan to quit and to experience emotional exhaustion (a component of burnout, when “emotional resources are depleted” and employees “feel they are no longer able to give of themselves” emotionally; Maslach & Jackson, 1981, p. 99).

No studies to date have investigated whether novice SETs’ workload manageability is related to these outcomes. However, studies of GETs and of experienced SETs have obtained findings consistent with COR theory, identifying relationships between workload manageability, and (a) emotional exhaustion (a component of burnout) and (b) career intentions.

Workload Manageability and Emotional Exhaustion (a Component of Burnout)

Knowing whether workload manageability predicts novice SETs’ burnout is important because teachers experiencing burnout are more likely to plan to leave their schools (Brunsting, Sreckovic, & Lane, 2014), and they may invest fewer resources in providing effective instruction (Irvin, Hume, Boyd, McBee, & Odom, 2013; Ruble & McGrew, 2013). Studies have identified significant negative relationships between SETs’ burnout and (a) the rate at which adult words were used (Irvin et al., 2013), (b) the quality of Individualized Education Programs (IEPs), (c) SETs’ adherence to interventions, and (d) students’ attainment of IEP goals (Ruble & McGrew, 2013). In Ruble and McGrew’s study, 9.3% of variance in students’ IEP goal attainment was explained by teachers’ emotional exhaustion, a component of burnout. Thus, if workloads contribute to emotional exhaustion, ameliorating workloads would be an important priority.

Embich (2001) examined relationships between SETs’ workload manageability and burnout in a survey of 300 secondary SETs serving students with learning disabilities. Among SETs who co-taught, perceptions of whether workloads were a problem accounted for 7% of variance in emotional exhaustion. Among SETs in self-contained settings, perceptions of whether workloads were a problem accounted for 16% of variance in emotional exhaustion. These findings indicate workload manageability may predict SETs’ emotional exhaustion. However, only one study has examined this relationship, and the sample was drawn from a single district. Furthermore, it is possible this relationship may operate differently for novices.

Workload Manageability and Career Intentions

Knowing whether workload manageability contributes to novices’ plans to continue teaching is important because teacher attrition is consistently associated with lower student achievement. For instance, Ronfeldt, Loeb, and Wyckoff (2013) found that students in grades with more turnover had significantly lower achievement than students from the same school whose grades had lower turnover, and lower achievement than students in the same grade and school from a different year with lower turnover. Hiring and training new teachers is also financially costly; Milanowski and Odden (2007) found a first-year teacher’s attrition cost one urban district between US$9,000 and US$23,000. Teachers’ career intentions—their plans to stay in or leave their current school—are an important indicator of future attrition; Gersten, Keating, Yovanoff, and Harniss (2001) found 69% of teachers who planned to leave moved to new positions within 15 months. Thus, if workload manageability predicts career intentions, ameliorating workloads would be an important priority.

Consistent with COR theory, most studies investigating workload manageability and career intentions found that teachers with more manageable workloads were more likely to plan to continue teaching. Albrecht, Johns, Mountstevens, and Olorunda (2009) surveyed 776 SETs and related service providers; participants who planned to stay were significantly more likely to report having adequate time to complete paperwork. Westling and Whitten (1996) surveyed 158 SETs; SETs who intended to continue teaching for the next 5 years were 3 times more likely to state they had adequate time to complete paperwork, plan instruction, and prepare materials. They were also significantly more likely to report they were never asked to complete responsibilities not part of their job. These studies provide evidence of a relationship between SETs’ workload manageability and career intentions, though they did not focus on novices.

The only study to obtain nonsignificant results was conducted with novice GETs (Pogodzinski, Youngs, & Frank, 2013). Pogodzinski and colleagues surveyed 184 novice first- through eighth-grade GETs in 11 districts. Spring career intentions were regressed onto novices’ spring ratings of workload manageability and onto their prior (fall) career intentions. Workload manageability did not significantly predict changes in novices’ plans to continue teaching.

In spite of Pogodzinski and colleagues’ (2013) nonsignificant results, the collected studies provide relatively consistent evidence for a relationship between teachers’ perceptions of workload manageability and career intentions. However, no studies investigated this relationship among novice SETs. It is possible that novice SETs would experience workloads in the same ways as more experienced SETs. However, it is also plausible that novices may view overwhelming workloads as a temporary part of beginning a new profession, rather than an inherent part of the job, in which case the relationship may be weaker or nonexistent. No research has examined the relationship between novice SETs’ perceptions of workload manageability and career intentions.

Purpose

The purpose of this study is to explore novice SETs’ perceptions of workloads. We aim to determine whether novice SETs perceive workloads as less manageable than novice GETs, and whether perceptions of workload manageability predict career intentions and emotional exhaustion (a component of burnout). Results will help establish whether novices’ perceptions of workload manageability are important for school leaders to address. Research questions are as follows:

Research Question 1 (RQ1): Do novice SETs perceive their workloads significantly differently than novice GETs?

Research Question 2 (RQ2): Do novice SETs’ and GETs’ perceptions of workload manageability in the fall predict their emotional exhaustion and career intentions in the spring?

We hypothesized the following:

Hypothesis 1: Novice SETs would rate workloads significantly less manageable than novice GETs.

Hypothesis 2: There would be a significant negative relationship between workload manageability and emotional exhaustion.

Hypothesis 3: There would be a significant positive relationship between workload manageability and career intentions.

Hypothesis 4: Emotional exhaustion would partially mediate the relationship between workload manageability and career intentions.

Method

Data Source and Participants

This study is a secondary analysis of an existing data set, the Michigan Indiana Early Career Teacher (MIECT) study, an investigation funded by the Carnegie Corporation, led by Dr. Peter Youngs from 2006 to 2009, and approved by the Institutional Review Board at Michigan State University. Survey data collected for MIECT were used for this study because the survey included rich information about all constructs of interest. Few existing studies have focused simultaneously on the experiences of novice SETs and GETs, and no other extant data sets could be located that included measures of novice SETs’ and GETs’ perceptions of workload manageability, career intentions, and emotional exhaustion.

MIECT focused on large urban districts because novices in these settings often experience more significant challenges and are at higher risk for attrition. Thus, large (>9,000 students) urban districts in Michigan and Indiana were eligible for participation in MIECT. Eleven districts were selected and agreed to participate (see Table 1). Teachers in these districts were eligible for MIECT if they (a) taught first through eighth grade, (b) had 3 or fewer years’ experience, (c) taught core content (i.e., math, English language arts, science, social studies, special education), and (d) had completed traditional preparation program. Teacher labor markets in Michigan and Indiana at the time were tight, as declining student enrollments, a recession, and tightening educational budgets combined to reduce the number of new teachers being hired.

Table 1.

District Demographic Information.

District	Student population	Proportion of students from minority cultural and ethnic backgrounds (%)	Proportion of students eligible for free or reduced price lunch (%)
A	18,386	12	51
B	21,448	80	65
C	9,139	50	42
D	29,261	11	11
E	7,994	46	36
F	11,645	19	29
G	13,666	48	62
H	16,138	57	44
I	10,662	84	50
J	21,769	60	62
K	12,483	59	58

Survey

Researchers developed the survey in 2005–2006, primarily from previous scales (e.g., Penuel, Riel, Krause, & Frank, 2009). Researchers piloted the initial survey in 2006–2007, conducting cognitive interviews with a small sample of teachers and teacher educators. During cognitive interviews, teachers completed the survey in the presence of a researcher, describing what they were thinking while completing each question, why they answered in a particular way, and any questions. Their questions and comments revealed when items were unclear or could be interpreted differently than intended; researchers revised items accordingly.

Surveys were administered twice per year, in fall and spring of 2007–2008 and 2008–2009 school years, using Dillman’s (2007) five-contact approach. Teachers received links to the survey by email; if they did not respond to the survey invitation after several reminders, they received a paper copy. Only participants who completed the survey in both fall and spring were included in this analysis. Of 78 SETs eligible to participate in 2007–2008, 67% responded in fall, and 75% of them responded in spring, for a total of 39 SETs in 2007–2008. Of 50 SETs eligible to participate in 2008–2009, 90% responded in fall, and 93% of them responded in spring (i.e., 42 of 50 eligible participants responded both times); however, 20 had previously participated in 2007–2008 and were thus excluded from the 2008–2009 sample, yielding a total of 22 SETs in 2008–2009. Of 384 eligible novice GETs, 63% responded in fall and 76% of them responded in spring. The final sample included 61 SETs and 184 GETs. Descriptive data indicated no significant observable differences between responders and nonresponders, though there may be unobservable differences that we are unable to test. There were an average of two novice GETs per school and an average of 1.5 novice SETs per school; as such, school-level nesting could not be addressed. It would have been possible to address district-level nesting, but prior analyses of these data found district-level nesting accounted for a negligible proportion of variance (e.g., Jones, Youngs, & Frank, 2013). Therefore, multilevel methods were not necessary.

Measures

Data on workload manageability were obtained from fall surveys, while data on career intentions and emotional exhaustion were obtained from spring surveys, allowing us to examine whether workload manageability at one time point predicted emotional exhaustion and career intentions at a subsequent time point. Items were selected to represent workload manageability, emotional exhaustion, and career intentions (see Table 2). We analyzed scales’ internal structure and reliability, as described in the “Data Analysis” and “Results” sections.

Table 2.

Hypothesized Factors and Items from the Survey.

Hypothesized construct	Items
Workload manageability^a	I am teaching with adequate resources and materials to do my job properly.
	My workload is manageable.
	I feel I’m working too hard on my job. (reverse scored)
	Administrative duties/paperwork do not interfere with my teaching.
Career intentions^a	I would prefer to continue teaching in this school next year.
	I could see myself teaching at this school in 5 years.
	I would prefer to continue teaching in this district next year.
	I would prefer to continue teaching in this district in 5 years.
Emotional exhaustion^a	I feel emotionally drained from my work.
	I feel fatigued when I have to get up in morning and face another day on the job.
	I feel burned out from my work.
	I feel fed up with my work.

All items are rated on a 5-point Likert-type scale. The wording of the response categories varied depending on the wording of the question.

Workload manageability

Workload manageability is defined as a teacher’s perception of the degree to which his or her responsibilities can be completed within time allotted. We selected items consistent with this definition from the survey.

Emotional exhaustion

Emotional exhaustion is a component of burnout, consisting of feeling emotionally drained (Maslach & Jackson, 1981). Researchers developed items measuring emotional exhaustion from the Emotional Exhaustion subscale of Maslach’s Burnout Inventory.

Career intentions

Career intentions are a teacher’s plans to continue teaching in the current school and district over the following 1 to 5 years. This definition focuses on teachers’ plans to leave their current school or district, not plans to leave special education. We focused on plans to leave their schools and districts because effective teachers who change schools or districts are most likely to move away from low performing schools and districts, where students with disabilities are in greatest need of their newly acquired expertise, to schools and districts that are already highly performing (e.g., Boyd, Grossman, Lankford, Loeb, & Wyckoff, 2008); in contrast, a novice SET who exits special education, but remains in his or her school or district, is still using his or her expertise in the service of students with disabilities in that school who receive Tier 1 instruction. Thus, movement out of a school or district may be a greater concern than movement out of special education, in terms of reducing the impact of attrition on those students who are most likely experience a “revolving door” of inexperienced teachers (Ingersoll, 2001, p. 499). Also note, these items ask about leaving a school or district (not moving to another school or district); thus, teachers who answered negatively include both teachers who plan to leave teaching altogether and teachers who plan to move to a different school or district.

We selected items consistent with this definition. These items are also consistent with those used in prior studies (e.g., Albrecht et al., 2009; Billingsley et al., 2004; Jones et al., 2013). However, whereas most prior investigations (e.g., Albrecht et al., 2009; Billingsley et al., 2004) have treated career intentions as dichotomous (intend to stay vs. intend to leave), we treated it as a latent construct. This reflects the fact that teachers’ career intentions are an unobserved intention to engage in a particular kind of work within a particular location during subsequent year(s); this is not dichotomous, but rather includes multiple (highly correlated) components, such as intentions with respect to the school and with respect to the district.

Data Analysis

Comparing Novice SETs’ and GETs’ Perceptions of Workload Manageability

We conducted two-way repeated measures ANOVA (Cohen, 2008) on workload manageability items to determine whether differences existed between SETs’ and GETs’ perceptions of workload manageability (RQ1).

Relationships Among Workload Manageability, Emotional Exhaustion, and Career Intentions

We used structural equation modeling (SEM) to answer the second research question. A full hybrid structural regression model (i.e., simultaneously modeling both measurement components and path components; Kline, 2011) was not feasible given the small sample. Instead, we tested measurement models first, and the path model second, without the measurement components. The advantage of this approach is that, even with a relatively small sample, it permits modeling of complex relationships (Kline, 2011). The limitation is that path models assume zero error; therefore, to account for error in the path model, we fixed error variance for each scale to variance of the factor score times one minus reliability of the construct (Kline, 2011). For both measurement models and path models, at least five observations are required per parameter estimated (Jackson, 2001, 2003); all analyses meets these criteria. We addressed missing data using full information maximum likelihood estimation.

Measurement models

For each construct, we conducted confirmatory factor analyses (CFA) using robust maximum likelihood estimation in MPlus (Muthén & Muthén, 2010) to test hypotheses that items work together to measure unidimensional latent constructs. We tested measurement models separately for SETs and GETs, as constructs may operate differently for these populations. Testing measurement models separately may lead to a choice between (a) keeping a model that fits poorly for one population or (b) using different measurement models for two populations. When faced with this choice, we retained best fitting models. This supports validity of conclusions about each population, but complicates comparisons between them.

In each measurement model, we set variance of the latent factor to 1, to address scale indeterminacy (Kline, 2011). We examined model fit indices (chi-square, comparative fit index [CFI], Tucker–Lewis Index [TLI], and root mean square error of approximation [RMSEA]) to determine model fit. We also examined factor loadings to determine if the latent factor explained a significant proportion of variance in each item. We calculated Raykov’s (1997) composite reliability. With latent variables, composite reliability is more accurate than Cronbach’s alpha because it weights items based their contribution to measuring the latent factor; items that load more strongly count more highly in determining reliability (Raykov, 1997).

Structural model

To test the path model, we calculated factor scores (i.e., composites that weight items based on their contribution to latent construct) for each construct. We specified novices’ perceptions of workload manageability in fall as an observed exogenous variable, and specified emotional exhaustion and career intentions in spring as observed endogenous variables (see Figure 1). We fixed error variance for each scale to variance of the factor score times one minus reliability. We used nonparametric bootstrapping (Kline, 2011) to test indirect effects.

Figure 1.

Hypothesized path model testing the relationships among workload manageability, emotional exhaustion, and career intentions.

Results

Do Novice SETs Perceive Their Workloads Significantly Differently Than Novice GETs?

For the first item (I am teaching with adequate resources and materials to do my job properly), the interaction (time by teacher role; F[1, 236] = .389, p = .533) and the effect of time (F[1, 236] = .389, p = .533) were not significant (see Table 3). However, SETs were significantly less likely to agree (M = 2.482, SD = 1.232 in fall; M = 2.467, SD = 1.241 in spring) than GETs (M = 2.847, SD = 1.667 in fall; M = 2.934, SD = 1.047 in spring; F[1, 236] = 7.266, p = .008).

Table 3.

Workload Manageability ANOVA.

Item	Interaction effect (Time × Teacher Group)	Time effect (fall vs. spring)	Teacher group effect (special vs. general educators)
I am teaching with adequate resources and materials to do my job properly.	F(1, 236) = 0.389p = .533	F(1, 236) = 0.389p = .533	F(1, 236) = 7.266** p = .008
Administrative duties/paperwork do not interfere with my teaching.	F(1, 234) = 3.746p = .054	F(1, 234) = 3.034p = .083	F(1, 234) = 15.613*** p =.000
My workload is manageable.	F(1, 229) = 7.646** p = .006	F(1, 229) = 1.677p = .066	F(1, 229) = .204p = .652
I feel I’m working too hard on my job. (reverse scored)	F(1, 234) = 2.573p = .110	F(1, 234) = 3.409p = .066	F(1, 234) = .557p = .456

*p < .05. **p < .01. ***p < .001, two-tailed test of significance.

For the second item (Administrative duties/paperwork do not interfere with my teaching), the interaction effect, F(1, 234) = 3.746, p = .054, and the effect of time, F(1, 234) = 3.034, p = .083, were not significant. However, SETs were significantly less likely to agree (M = 1.746, SD = 1.347 in fall; M = 1.729, SD = 1.271 in spring) than novice GETs (M = 2.198, SD = 1.374 in fall; M = 2.520, SD = 0.893 in spring; F[1, 234] = 15.613, p = .000).

For the third item (My workload is manageable), the interaction was significant: GETs’ were more likely to agree in spring (M = 2.457, SD = 1.208 in fall; M = 2.803, SD = 0.783 in spring) whereas novice SETs’ were slightly more likely to disagree in spring (M = 2.603, SD = 1.059 in fall; M = 2.534, SD = 0.995 in spring; F[1, 229] = 7.646, p = .006). This could indicate changes in their capacity to manage workloads over the year, or changes in the nature of their workloads (i.e., SETs may have more caseload management responsibilities in the spring).

For the fourth item (I feel I’m working too hard on my job), no effects were significant (Interaction: F[1, 234] = 2.573, p = .110; Time: F[1, 234] = 3.409, p = .066; Group: F[1, 234] = 0.577, p = .456). Both groups showed a slight decrease, but novice SETs’ ratings (M = 2.438, SD = 1.112 in fall; M = 2.167, SD = 1.196 in spring) decreased more (not significantly more) than novice GETs’ (M = 2.461, SD = 1.184 in fall; M = 2.412, SD = 1.161 in spring).

Do Novice SETs’ Perceptions of Workload Manageability Predict Their Career Intentions and Emotional Exhaustion?

Measurement models

Workload manageability

For SETs, the chi-square was not significant (χ² = 0.023, p = .989), indicating exact fit, but the first item loaded poorly onto the construct (standardized coefficient = .349). We removed this item and tested the model again. With only three items, the model would have been just identified, so we set the first factor loading to 1. Chi-square for this model was not significant (χ² =.124, p = .725), indicating the model fit exactly (see Table 4). Composite reliability was adequate, .636.

Table 4.

CFAs for Workload Manageability, Career Intentions, and Emotional Exhaustion.

Hypothesized factor	Model fit	Item content	Factor loadings
Workload manageability, special educators	χ² = .124, p = .725CFI: 1.000TLI: 1.146RMSEA: 0.000, CI = [0.000, 0.245]	My workload is manageable.	0.701***
		I feel I’m working too hard on my job. (reverse scored)	0.406*
		Administrative duties/paperwork do not interfere with my teaching.	0.744***
Workload manageability, general educators	χ² = 5.493, p = .064CFI: 0.983TLI: 0.949RMSEA: 0.098, CI = [0.000, 0.200]	I am teaching with adequate resources and materials to do my job properly.	0.324***
		My workload is manageable.	0.708***
		I feel I’m working too hard on my job. (reverse scored)	0.964***
		Administrative duties/paperwork do not interfere with my teaching.	0.611***
Career intentions, special educators	χ² = 3.731, p = .053CFI: 0.988TLI: 0.929RMSEA: 0.212, CI = [0.000, 0.458]	I would prefer to continue teaching in this school next year.	0.749***
		I could see myself teaching at this school in 5 years.	0.930***
		I would prefer to continue teaching in this district next year.	0.854***
		I would prefer to continue teaching in this district in 5 years.	0.951***
Career intentions, general educators	χ² = 133.859, p = .000CFI: 0.746TLI: 0.521RMSEA: 0.864, CI = [0.744, 0.990]	I would prefer to continue teaching in this school next year.	0.654***
		I could see myself teaching at this school in 5 years.	0.690***
		I would prefer to continue teaching in this district next year.	0.947***
		I would prefer to continue teaching in this district in 5 years.	0.891***
Emotional exhaustion, special educators	χ² = .687, p = .709CFI: 1.000TLI: 1.041RMSEA: 0.000, CI = [0.000, 0.214]	I feel emotionally drained from my work.	0.799***
		I feel fatigued when I have to get up in morning and face another day on the job.	0.894***
		I feel burned out from my work.	0.864***
		I feel fed up with my work.	0.665***
Emotional exhaustion, general educators	χ² = 6.006CFI: 0.990TLI: 0.969RMSEA: 0.105, CI = [0.004, 0.206]	I feel emotionally drained from my work.	0.733***
		I feel fatigued when I have to get up in morning and face another day on the job.	0.769***
		I feel burned out from my work.	0.931***
		I feel fed up with my work.	0.774***

Note. CFA = confirmatory factor analyses; CFI = comparative fit index; TLI = Tucker–Lewis Index; RMSEA = root mean square error of approximation; CI = confidence interval.

For GETs, chi-square was not significant (χ² = 5.493, p = .064), indicating the model fit exactly. The first item loaded poorly (standardized coefficient = .324), but when it was removed, model fit indices were no longer adequate (χ² = 5.493, p = .064; CFI = .917; TLI = .876). Therefore, we retained it (see Table 4). Composite reliability was adequate, .796.

Emotional exhaustion

For SETs, chi-square was not significant (χ² = 1.080, p = .583), indicating exact fit (see Table 4). Composite reliability was .895. For GETs, chi-square was significant (χ² = 6.006, p = .050), indicating inexact fit; however, CFI (.990) and TLI (.969) were both high, above .90 (Kline, 2011); we retained this model. Composite reliability was .880.

Career intentions

For SETs, chi-square was significant (χ² = 25.983, p = .000), and CFI (.896) and TLI (.687) were also below .90. Items 2 and 4 correlated more strongly, which makes sense, as SETs planning to stay in a school implicitly plan to stay in a district. We tested CFA again, allowing these items to correlate. Chi-square was not significant (χ² = 3.731, p = .053), indicating exact fit (see Table 4). Composite reliability was .934.

For GETs, chi-square was significant (χ² = 134.232, p = .000) and CFI (.748) and TLI (.243) were also low. Allowing Items 2 and 4 to correlate did not improve model fit (see Table 4). We examined modification indices, but no changes were sufficient to obtain better fit. To enable comparison of SETs to GETs, we retained this model. Composite reliability was .871.

Structural model

We tested a path model (see Figure 1) using composite factor scores. The model was just identified, so we could not obtain model fit. For SETs, fall workload manageability significantly negatively predicted spring emotional exhaustion (p = .010), such that a 1 SD increase in workload manageability was associated with a .561 SD decrease in emotional exhaustion (see Table 5). In turn, emotional exhaustion significantly predicted career intentions (p = .021); a 1 SD decrease in emotional exhaustion was associated with a .430 SD increase in career intentions. The direct relationship between workload manageability and career intentions was not significant (p = .176), but the indirect relationship, via emotional exhaustion, was significant (p = .050), such that a 1 SD increase in fall workload manageability was associated with a .242 SD increase in spring career intentions.

Table 5.

Workload Manageability as a Predictor of Emotional Exhaustion and Career Intentions.

Educator	Effect		Regressed onto	Unstandardized coefficient	p value (one-tailed)
Special	Direct	Career intentions	Workload manageability	.261	.176
		Career intentions	Emotional exhaustion	−.430*	.021
		Emotional exhaustion	Workload manageability	−.561**	.010
	Indirect	Career intentions	Workload manageability via emotional exhaustion	.242*	.050
General	Direct	Career intentions	Workload manageability	.061	.358
		Career intentions	Emotional exhaustion	−.683***	.000
		Emotional exhaustion	Workload manageability	−.864***	.000
	Indirect	Career intentions	Workload manageability via emotional exhaustion	.591***	.000

p < .05. **p < .01. ***p < .001.

For GETs, fall workload manageability negatively predicted spring emotional exhaustion (p = .000; see Table 5); a 1 SD increase in workload manageability was associated with a .864 SD decrease in emotional exhaustion. In turn, emotional exhaustion negatively predicted career intentions (p = .000); a 1 SD decrease in emotional exhaustion was associated with a .683 SD increase in career intentions. The direct relationship between workload manageability and career intentions was not significant (p = .358), but the indirect relationship, via emotional exhaustion, was significant (p = .000), such that a 1 SD increase in workload manageability was associated with a .563 SD increase in career intentions.

Discussion

We examined whether novices’ perceptions of workload manageability predicted emotional exhaustion and career intentions. Results are consistent with prior studies (e.g., Embich, 2001); workload manageability predicted emotional exhaustion, which mediated a relationship between workload manageability and career intentions. The same relationships were significant for both SETs and GETs, but the magnitude of the relationships was larger for GETs.

Limitations

The sample included only novice elementary and middle school teachers in Michigan and Indiana who were certified through traditional preparation. This limits generalizability of results. Conclusions cannot be generalized to other populations of teachers (e.g., high school teachers, uncertified teachers, teachers in rural districts). Results cannot be generalized to alternatively certified teachers, who are often employed in urban districts (Mason-Williams, 2015), and who may experience more substantial problems with workload manageability. In addition, changes in schools (e.g., teacher evaluation policies, increasing student diversity) have occurred since data were collected; it is possible that some findings may not generalize to current school contexts.

The workload manageability factor was not as reliable for SETs as would be ideal, though we did account for reliability. With regard to the two-way ANOVA, the assumption of homogeneity of variance was violated for the first three items; variance in the SET sample was significantly larger than variance in the GET sample in spring. This may result in inflated Type 1 error rates; results of comparisons should be confirmed in future studies.

We did not have sufficiently large sample to conduct multigroup models (Kline, 2011, recommends at least 100 participants per group for multigroup models), which would have allowed us to formally test for measurement invariance. For two constructs (i.e., workload manageability, career intentions), measurement models that worked well for SETs did not work as well for GETs. Because we were primarily interested in SETs, we retained models that worked well for SETs; for GETs, we used a slightly different workload manageability scale and we used a poorly fitting career intentions scale. As such, differences between populations should be interpreted with caution and confirmed in future studies, as should results for GETs.

Implications for Future Research

More research is needed to confirm our findings and examine whether these relationships occur with other populations of SETs, including secondary SETs, SETs in rural and suburban districts, and SETs who are alternatively certified. National data sets (i.e., Schools and Staffing Survey) do not include enough items about workload manageability to permit a nationally generalizable examination of relationships among workload manageability, emotional exhaustion, and career intentions. However, surveys of SETs who are at high risk for attrition would be useful for understanding how workloads may be contributing to attrition.

Future research should also examine potential relationships between workloads and instruction. Consistent with COR theory, some studies indicate SETs who are overwhelmed may devote less time to instruction (Bettini et al., 2015; Vannest, Soares, Harrison, Brown, & Parker, 2010). More research examining whether workload manageability contributes to instruction could help make a stronger case for investing financial resources in reducing workloads.

Two scales had strong psychometric properties for SETs but not GETs. Future research should confirm findings using stronger scales for GETs. This also suggests scholars should validate scales for both SETs and GETs before aggregating data for these populations. A two-group methodology (e.g., multigroup SEM) with larger, balanced sample would be a promising approach for examining differences in the structural parameters across SETs and GETs by testing for cross-group invariance (Byrne, 1998; Schumacker & Marcoulides, 1998).

This study provides insights into how workload manageability may contribute to negative outcomes among novices, but it does not provide insights into what leads novices to feel overwhelmed. Future studies should explore why novices feel overwhelmed. Assigned workloads are one obvious factor worth exploring. In addition, qualitative research indicates curricular resources (Grossman & Thompson, 2004), collegial interactions (Kardos, Johnson, Peske, Kauffman, & Liu, 2001), mentorship (Grossman & Thompson, 2004), and school cultures of collective responsibility (Kardos et al., 2001) may play a role in novices’ workloads. Scholars should test whether these factors explain variance in novices’ perceptions of their workloads.

Future research would also benefit from a stronger workload manageability scale. Scholars should consider drawing upon organizational research investigating role overload, a related construct for which several scales have been developed (e.g., Bolino & Turnley, 2005).

Implications for Practice

The shortage of SETs appears poised to, again, become a major problem (Brownell & Sindelar, 2016). In this environment, retaining SETs becomes especially urgent (Dewey et al., in press). Our results indicate school leaders may be able to retain novices by providing them more manageable workloads. One obvious strategy would be to reduce new SETs’ workloads by, for instance, assigning administrative and supervisory tasks to classified staff, or by hiring more SETs and distribute their responsibilities among more personnel.

In addition, COR theory suggests social support is a resource that can reduce the effects of high demands on employees’ burnout and attrition (Alarcon, 2011). Consistent with COR theory, educational research indicates that school culture plays an important role in novice SETs’ experiences (Billingsley et al., 2004; Jones et al., 2013). Thus, school leaders may be able to help novice SETs manage their workloads by cultivating school cultures supportive of SETs’ work. A school culture of collective responsibility for students with disabilities’ learning may be especially important for novice SETs, because they rely on collaboration with colleagues to do their jobs (Jones et al., 2013; Otis-Wilborn, Winn, Griffin, & Kilgore, 2005). Thus, leaders may be able to help novice SETs manage their workloads by cultivating a school culture in which all staff share a sense of responsibility for students with disabilities.

Conclusion

Novice SETs have long reported their workloads are unmanageable (Billingsley et al., 2004; Griffin et al., 2009). Our results demonstrate that workload manageability may be associated with emotional exhaustion and career intentions. School leaders should carefully consider how to systematically relieve novice SETs’ workloads, while scholars should continue investigating how workload manageability and emotional exhaustion can be ameliorated.

Footnotes

Associate Editor: Wendy Oakes

Authors’ Note

The second and third author contributed equally to this article.

Declaration of Conflicting Interests

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

Funding

Data for this study were obtained from the Michigan Indiana Early Career Teacher study; that original study was made possible by a grant awarded to Dr. Peter Youngs from the Carnegie Corporation, New York. The author(s) conducted independent, secondary analysis of this data set and received no financial support for the present analysis, authorship, and/or publication of this article. The opinions expressed are the authors’, and do not represent the views of the original study’s funding agency.

References

Alarcon

G. M.

(2011). A meta-analysis of burnout with job demands, resources, and attitudes. Journal of Vocational Behavior, 79, 549–562. doi:10.1016/j.jvb.2011.03.007

Albrecht

S. F.

Johns

B. H.

Mountstevens

Olorunda

(2009). Working conditions as risk or resiliency factors for teachers of students with emotional and behavioral disabilities. Psychology in the Schools, 10, 1006–1022. doi:10.1002/pits.20440

Bettini

Kimerling

Park

Murphy

(2015). Responsibilities and instructional time: Relationships identified by special educators in self-contained classes for students with emotional and behavioral disabilities. Preventing School Failure, 39, 121–128. doi:10.1080/1045988X.2013.859561

Billingsley

B. S.

Carlson

Klein

(2004). The working conditions and induction support of early career special educators. Exceptional Children, 70, 333–347. doi:10.1177/001440290407000305

Billingsley

B. S.

Griffin

C. C.

Smith

S. J.

Kamman

Israel

(2009). A review of teacher induction in special education: Research, practice, and technology solutions (National Center to Inform Policy and Practice in Special Education Professional Development Document No. RS-1). Available from http://www.ncipp.org

Bolino

M. C.

Turnley

W. H.

(2005). Personal costs of citizenship behavior: The relationship between individual initiative and role overload, job stress, and work-family conflict. Journal of Applied Psychology, 90, 740–748. doi:10.1037/0021-9010.90.4.740

Boyd

Grossman

Lankford

Loeb

Wyckoff

(2008, May). Who leaves? Teacher attrition and student achievement (Working Paper No. 14022). Cambridge, MA: National Bureau of Economic Research. doi:10.3386/w14022

Brownell

M. T.

Sindelar

P. T.

(2016). Preparing and retaining effective special education teachers: Systemic solutions for addressing teacher shortages (Policy Brief for the CEEDAR Center). Gainesville, FL: Collaboration for Effective Educator Development, Accountability and Reform. Retrieved from ceedar.education.ufl.edu

Brownell

M. T.

Sindelar

P. T.

Kiely

M. T.

Danielson

L. C.

(2010). Special education teacher quality and preparation: Exposing foundations and constructing a new model. Exceptional Children, 76, 357–377. doi:10.1177/001440291007600307

10.

Brunsting

Sreckovic

Lane

K. L.

(2014). Special education teacher burnout: A synthesis of research from 1979 to 2013. Education and Treatment of Children, 37, 681–711. doi:10.1353/etc.2014.0032

11.

Byrne

B. M.

(1998). Structural equation modeling: Basic concepts, application, and programming. Mahwah, NJ: Lawrence Erlbaum.

12.

Cohen

B. H.

(2008). Explaining psychological statistics (3rd ed.). Hoboken, NJ: John Wiley.

13.

Dewey

Sindelar

Bettini

Boe

Rosenberg

Leko

(in press). Declining special education teacher employment, 2005–2012. Exceptional Children.

14.

Dillman

D. A.

(2007). Mail and internet surveys: The tailored design method (2nd ed.). New York, NY: Wiley.

15.

Embich

J. L.

(2001). The relationship of secondary special education teachers’ roles and factors that lead to professional burnout. Teacher Education and Special Education, 24, 58–69. doi:10.1177/088840640102400109

16.

Gersten

Keating

Yovanoff

Harniss

M. K.

(2001). Working in special education: Factors that enhance special educators’ intent to stay. Exceptional Children, 67, 549–567. doi:10.1177/001440290106700408

17.

Griffin

C. C.

Kilgore

K. L.

Winn

J. A.

Otis-Wilborn

Hou

Garvan

C. W.

(2009). First-year special educators: The influence of school and classroom context factors on their accomplishments and problems. Teacher Education and Special Education, 32, 45–63. doi:10.1177/0888406408330870

18.

Grossman

Thompson

(2004). District policy and beginning teachers: A lens on teacher learning. Educational Evaluation and Policy Analysis, 26, 281–301. doi:10.3102/01623737026004281

19.

Ingersoll

R. M.

(2001). Teacher turnover and teacher shortages: An organizational analysis. American Educational Research Journal, 38, 499–534. doi:10.3102/00028312038003499

20.

Irvin

D. W.

Hume

Boyd

B. A.

McBee

M. T.

Odom

S. L.

(2013). Child and classroom characteristics associated with the adult language provided to preschoolers with autism spectrum disorder. Research in Autism Spectrum Disorders, 7, 947–955. doi:10.1016/j.rasd.2013.04.004

21.

Jackson

D. L.

(2001). Sample size and number of parameter estimates in maximum likelihood confirmatory factor analysis: A Monte Carlo investigation. Structural Equation Modeling, 8, 205–223. doi:10.1207/S15328007SEM0802_3

22.

Jackson

D. L.

(2003). Revisiting sample size and number of parameter estimates: Some support for the N: q hypothesis. Structural Equation Modeling, 10, 128–141. doi:10.1207/S15328007SEM1001_6

23.

Jones

N. D.

Youngs

Frank

K. A.

(2013). The role of school-based colleagues in shaping the commitment of novice special and general education teachers. Exceptional Children, 79, 365–383. doi:10.1177/001440291307900303

24.

Kardos

S. M.

Johnson

S. M.

Peske

H. G.

Kauffman

Liu

(2001). Counting on colleagues: New teachers encounter the professional cultures of their schools. Educational Administration Quarterly, 37, 250–290. doi:10.1177/00131610121969316

25.

Kline

R. B.

(2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: The Guilford Press.

26.

Maslach

Jackson

S. E.

(1981). The measurement of experienced burnout. Journal of Occupational Behavior, 2, 99–113. doi:10.1002/job.4030020205

27.

Mason-Williams

(2015). Unequal opportunities: A profile of the distribution of special education teachers. Exceptional Children, 81(2), 247-262. doi:10.1177/0014402914551737

28.

McCray

E. D.

Butler

T. W.

Bettini

(2014). Sharing expertise and shifting expectations: The roles of general and special educators in effective, inclusive schools. In McLeskey

Waldron

N. L.

Spooner

Algozzine

(Eds.), Handbook of research for effective, inclusive schools (pp. 80–93). New York, NY: Routledge.

29.

McLaughlin

M. J.

(2010). Evolving interpretations of educational equity and students with disabilities. Exceptional Children, 76, 265–278. doi:10.1177/001440291007600302

30.

Milanowski

A. T.

Odden

A. R.

(2007). A new approach to the cost of teacher turnover (Working Paper No. 13). The School Finance Redesign Project. Available from http://www.schoolfinanceredesign.org

31.

Mitchell

B. B.

Deshler

D. D.

Lenz

B. K. B. H

. (2012). Examining the role of the special educator in a response to intervention model. Learning Disabilities, 10, 53–74.

32.

Muthén

L. K.

Muthén

B. O.

(2010). Mplus (Version 7) [Computer software]. Los Angeles, CA: Author.

33.

Otis-Wilborn

Winn

Griffin

Kilgore

(2005). Beginning special educators’ forays into general education. Teacher Education and Special Education, 28, 143–152. doi:10.1177/088840640502800401

34.

Penuel

W. R.

Riel

Krause

A. E.

Frank

(2009). Analyzing teachers’ professional interactions in a school as social capital: A social network approach. Teachers College Record, 111, 124–163.

35.

Pogodzinski

Youngs

Frank

K. A.

(2013). Collegial climate and novice teachers’ intent to remain teaching. American Journal of Education, 120, 27–54. doi:10.1086/673123

36.

Raykov

(1997). Estimation of composite reliability for congeneric measures. Applied Psychological Measurement, 21, 173–184. doi:10.1177/01466216970212006

37.

Ronfeldt

Loeb

Wyckoff

(2013). How teacher turnover harms student achievement. American Educational Research Journal, 50, 4–36. doi:10.3102/0002831212463813

38.

Ruble

McGrew

J. H.

(2013). Teacher and child predictors of achieving IEP goals of children with autism. Journal of Autism and Developmental Disorders, 43, 2748–2763. doi:10.1007/s10803-013-1884-x

39.

Schumacker

R. E.

Marcoulides

G. A.

(1998). Interaction and nonlinear effects in structural equation modeling. Mahwah, NJ: Lawrence Erlbaum.

40.

Vannest

K. J.

Hagan-Burke

(2010). Teacher time use in special education. Remedial and Special Education, 31, 125–142. doi:10.1177/0741932508327459

41.

Vannest

K. J.

Soares

D. A.

Harrison

J. R.

Brown

Parker

R. I.

(2010). Changing teacher time. Preventing School Failure, 54, 86–98. doi:10.1080/10459880903217739

42.

Westling

D. L.

Whitten

T. M.

(1996). Rural special education teachers’ plans to continue or leave their teaching positions. Exceptional Children, 62, 319–335.

43.

Youngs

Jones

Low

(2011). How beginning special and general education elementary teachers negotiate role expectations and access to professional resources. Teachers College Record, 113, 1506–1540.