Teacher Working Conditions in Charter Schools and Traditional Public Schools: A Comparative Study
by Yongmei Ni - 2012
Background/Context: Teachers affect student performance through their interaction with students in the context of the classrooms and schools where teaching and learning take place. Although it is widely assumed that supportive working conditions improve the quality of instruction and teachers’ willingness to remain in a school, little is known about whether or how the organizational structure of charter schools influences teacher working conditions.
Purpose/Research Question: This article compares teacher working conditions in charter and traditional public schools and among various types of charter schools. In doing so, it seeks to understand whether the different working conditions are influenced by the intrinsic institutional features of charter schools such as autonomy and competition, or by the extraneous factors such as measureable school and teacher characteristics.
Research Design: This study utilized data from the 2003–2004 Schools and Staffing Survey (SASS), the nation’s most extensive survey of K–12 schools and teachers, both for charter schools and traditional public schools (TPSs). This article is a quantitative analysis that involves three main steps. First, based on the responses to the SASS teacher questionnaire, confirmatory factor analysis was performed to generate multiple factors corresponding to key dimensions of teacher working conditions. Second, propensity score matching was used to pair charter schools with TPSs that are similar in terms of school location, educational level, school type, and student demographics. This matching process mitigates the confounding effects of these extraneous factors on teachers’ perceptions of working conditions. Finally, a series of weighted Hierarchical Linear Models were utilized to compare teachers’ perceptions of working conditions between charter and traditional public schools, controlling for teacher and school characteristics.
Conclusions/Recommendations: The results show that charter and traditional public school teachers perceive their working conditions to be similar in many regards, including principal leadership, sense of community and collegiality, classroom autonomy, opportunities for professional development, and adequacy of instructional supplies. However, charter school teachers perceive that they have significantly more influence over school policies, but a heavier workload than traditional school teachers. Among charter schools, district-granted charter schools show consistently more supportive working environments than charters granted by other organizations. This implies that state policy can have some indirect influence over charter school working conditions by providing substantial administrative support and oversight to charter schools authorized by independent organizations other than the established structure of school districts.
As one of the most prominent reforms in U.S. elementary and secondary education since the 1990s, charter schools have been widely advocated to utilize market incentives to improve student achievement. Evidence on the performance of charter schools, however, is quite mixed. While some studies find that students in charter schools perform better than their counterparts in traditional public schools (TPSs), other studies have found no difference or lower student performance in charter schools (Carnoy, Jacobsen, Mishel, & Rothstein, 2005; Gill, Timpane, Ross, & Brewer, 2001; Miron, Evergreen, & Urschel, 2008). Rather than focusing solely on student outcomes, scholars are increasingly recognizing that further progress will require research that probes inside the black box of school organizations to find out what conditions influence student achievement in charter schools (Berends, Watral, Teasley, & Nicotera, 2007; Zimmer & Buddin, 2007).
Teacher working conditions have an important influence on the quality of instruction and teachers willingness to remain in a school (Johnson, 2006). Individual teachers, regardless of their qualifications and experiences, affect student performance through their interaction with students in the context of the classrooms and schools where teaching and learning take place. Although it is widely assumed that supportive working conditions improve teacher efficacy and student achievement, little is known about whether or how the organizational structure of charter schools influences teacher working conditions (Berends et al., 2007; Betts & Loveless, 2005; Cannata, 2007; Hill, Angel, & Christensen, 2006; Zimmer & Buddin, 2007). Research on charter school working conditions has relied heavily on case studies and small-scale quantitative studies, which provide only tentative evidence on this issue. The great diversity of charter schools naturally raises concerns about the external validity of such research.
This paper undertakes a comprehensive comparison of working conditions in charter schools and TPSs, incorporating several methodological improvements over previous studies. First, the data in the analysis are from the 20032004 Schools and Staffing Survey (SASS), which specifies a nationally representative sample of charter schools and their teachers. Second, propensity score matching is used to select a sample of comparable TPSs from the SASS based on schools student demographics, curricular focus, and location. This matching process diminishes the influence of these extraneous factors on working conditions. Third, weighted hierarchical linear models (HLMs) are utilized to control for the interdependent nature of teacher and school characteristics. Finally, teachers working conditions in various types of charter schools are analyzed.
DEFINING TEACHER WORKING CONDITIONS
According to Johnson (2006), working conditions are influenced by the organizational structure and the social, political, psychological, and physical features of the work environment. However, there is no consensus among researchers about how to measure working conditions, mainly because there are so many dimensions of working conditions and many are difficult to quantify. Johnson (2006) identified 11 dimensions of teacher working conditions and defined what would constitute supportive conditions on each dimension. These categories included teaching assignments, working relationships among teachers, support for new teachers, support for students, curricular support, resources and materials, facilities, assessment, professional development, professional influence and career growth, and the principals leadership.
Other scholars argue that a comprehensive measure of working conditions can be unwieldy and complicated to interpret. According to this view, focusing on those elements that are most likely to be relevant to important policy outcomes is more promising (Ladd, 2009). Based on a survey of working conditions in North Carolina, Ladd (2009) identified six key dimensions of working conditions: leadership, facilities and resources, teacher empowerment, professional development, mentoring, and time. A review of studies on working conditions reveals that researchers tend to construct their own measures of working conditions. Although these measures are very diverse, they all contain the main elements of Johnson and Ladds framework (Johnson & Birkeland, 2003; Loeb, Darling-Hammond, & Luczak, 2005; Weiss, 1999). This paper draws on these previous efforts to conceptualize the relevant dimensions of teacher working conditions and identifies data in the SASS that reflect these dimensions.
Research has shown that supportive working conditions not only enable teachers to teach more effectively but also improve retention of high-quality teachers (Johnson, 2006; Loeb, Darling-Hammond, & Luczak, 2005; McLaughlin & Talbert, 2001; Weiss, 1999). Less supportive teaching conditions lead to high rates of teacher turnover and attrition, which undermines efforts to provide high-quality instruction to all students (Ingersoll, 2001, 2004). Some scholars have also started to examine which dimensions of working conditions matter most for teacher efficacy and retention (Buckley, Schneider, & Shang, 2005; Ladd, 2009), but such differential impacts of various aspects of working conditions are not the focus of this paper.
CHARTER SCHOOLS AND TEACHER WORKING CONDITIONS
Charter schools are publicly funded schools that operate independently of the conventional governance structure for local public school districts under charters granted by authorizing agencies. Students enroll in the charter schools through the choices of their families rather than residential assignment. Charter schools are intended to be more autonomous than TPSs and enjoy substantial authority over working conditions for teachers (Malloy & Wohlstetter, 2003). This section draws on research to distill several hypotheses regarding why charter schools may provide more or less supportive working conditions than TPSs.
As conceived by early proponents, charter schools would enjoy considerable autonomy from many of the regulations that govern traditional public schools, and in turn decentralize substantial decision-making authority to the school level (Wohlstetter, Wenning, & Briggs, 1995). This autonomy would permit more flexibility in hiring teachers who are committed to a schools instructional mission and help establish school environments that support collaboration among educators. Teachers meanwhile would self-select into charter schools whose mission matched the teachers own vision and with colleagues who shared views of instructional goals and practices. Together, teachers commit themselves to a common mission (Malloy & Wohlstetter, 2003).
Charter schools autonomy could also foster a change in employment relations such that teachers would be treated in a more professional fashion and afforded more discretion not only to innovate educational programs to suit the particular needs of their students but also to collectively participate in fundamental decisions regarding school design and organization (Hassel, 1999; Kolderie, 1990; Malloy & Wohlstetter, 2003; Nathan, 1996). Charter schools would encourage professional development opportunities for teachers to develop a professional culture as a means of creating more effective schools (Wohlstetter et al., 1995). For all these reasons, charter schools could be expected to encourage improvements in teacher working conditions relative to traditional public schools.
However, charter schools are under market pressure to compete for students, since the schools funding is directly linked to the number of students enrolled. Thus, charter schools must be responsive to the interests of the local community and organizations that establish or manage the schools (Crawford & Forsyth, 2004; Gawlik, 2007; Sykes, 1999). In other words, although charter schools are freed from many state and district regulations, these schools do not necessarily extend the autonomy to their teachers. In addition, many charter schools are not bound by collective bargaining agreements, and the majority of charter school teachers are not members of a teacher union. Without the unions protection in areas such as workload, salaries and benefits, and due process rules, working conditions for charter school teachers may or may not improve relative to those of teachers in unionized TPSs. In addition, even if a charter school provides teachers with a high level of autonomy, extensive involvement in school-wide policies, especially in issues unrelated to curriculum and instruction, is time-consuming. Such involvement has the potential to overwork teachers and not necessarily benefit teachers as professionals (Malloy & Wohlstetter, 2003). Yet charter schools that are converted from public schools may continue to operate much the way they did as TPSs, such that teacher working conditions are essentially unchanged (Buddin & Zimmer, 2005).
In general, it is clearly possible based on a priori reasoning to adduce predictions of either more or less supportive workplace conditions for teachers in charter schools compared to TPSs. Teacher working conditions are multidimensional, and charter schools might influence different dimensions of working conditions in different ways. To date, empirical research fails to provide systematic evidence on this issue. One main reason is that many studies tend to focus exclusively on describing teacher working conditions in charter schools but not making comparisons between charter schools and TPSs. Case studies have reported that charter school teachers tend to perceive strong principal leadership, express a sense of faculty collaboration, and value involvement in decisions regarding governance, curriculum, instructions, personnel, and professional development (Finnigan, 2007; Johnson & Landman, 2000; Malloy & Wohlstetter, 2003). At the same time, many charter school teachers are dissatisfied with salaries and benefits, facilities, workload, and a sense of job insecurity (Johnson & Landman, 2000; Malloy & Wohlstetter, 2003; Miron & Applegate, 2007; Wells, 1998).
Although case studies provide very rich descriptions of working conditions in charter schools, the case studies fail to answer the question of whether charter schools or TPSs provide more supportive working environments for teachers. So far, only a few quantitative studies have provided some tentative evidence. Based on surveys sent to teachers in 16 charter schools and 7 TPSs in Colorado, Bomotti, Ginsberg, and Cobb (1999) found that although TPS teachers rarely express a strong sense of a common mission and shared goals, the comparison of shared missions and collegiality between charter and TPS teachers shows no statistically significant difference. Charter school teachers exert more influence over classroom-related issues but have the same or less influence on school governance and policies than their peers in TPSs. Bomotti et al. also found that charter school teachers are more likely to be dissatisfied with their physical facilities and access to computers than teachers in TPSs. Cannata (2008) used 20032004 SASS data to compare the weekly hours worked by charter and TPS teachers and found little difference. However, none of these studies controlled for teacher or school characteristics. Charter schools and TPSs differ significantly in many ways, such as their location, curricular programs, student populations, and enrollment size. All these differences may result in systematic difference in teachers perceptions of working conditions. Even if we find differences in teachers perceptions of working conditions when comparing all charter schools with all TPSs, the differences cannot be solely attributed to the institutional and organizational difference between the two types of schools.
Moreover, in view of the great diversity among charter school, little research exploring how key features of charter policy design and implementation influence charter school workplace conditions has been conducted. These key policy features include the type of organizations authorized to grant charters, the use of management companies, and a schools status as start-up or conversion school. The following section details the studys research methods to overcome these limitations.
DATA AND METHODS
This study utilized data from the 20032004 SASS, the nations most extensive survey of K12 schools, teachers, and administrators. The 20032004 SASS surveyed about 43,000 teachers in public schools, including charter schools. Teachers in charter schools and TPSs were asked the same set of questions, including questions pertaining to the teachers perceptions, training, resources, and teaching assignment. The survey also requested detailed demographic information about the teachers. School and district questionnaires provide information about student composition and various organizational characteristics. Data from the teacher, school, and district surveys can be merged with unique identifiers.
To obtain a nationally representative sample, the SASS used a complex stratified sampling design. Although about 40 states had charter schools in operation in 20032004, charter schools were sampled in only 28 states based on the sampling framework. In total, 233 charter schools were sampled, representing more than 2,000 charter schools nationally. Since this paper compares teachers in charter schools and TPSs, the analysis is limited to the 28 states that had charter schools and teachers surveyed. To get unbiased estimates, the sampling weights used by the SASS are incorporated in this studys factor analysis and HLM analysis.
MEASURING TEACHER WORKING CONDITIONS
The measures of teacher working conditions were obtained from responses to the SASS teacher questionnaire. Factor analyses were performed since many closely related questions in the questionnaire pertain to given dimensions of school working conditions. The items negatively worded and scored in the questionnaire were rescored in a positive direction to match the other items. Factor analyses generated six distinct factors corresponding to key dimensions of school working conditions: (a) principal leadership and administrative support, (b) sense of community and collegiality, (c) teacher influence in school policy, (d) teacher control in classroom, (e) professional development, and (f) workload.
All the factors have eigenvalues greater than one. About 35 items on the surveys that have a factor loading of 0.40 or above loaded on the factors. Appendix A shows a detailed description of the survey items used and their associated factor loadings for each variable. The factor scores of all measures generated from the factor analysis were used for the subsequent comparison analysis. In addition, a seventh measure of working conditionsthe adequacy of instructional supplieswas captured directly by a teacher survey item: Necessary materials such as textbooks, supplies, and copy machines are available as needed by the staff. Since this was the only question related to the availability of instructional supplies, no factor analysis was involved in generating the measure.
MATCHING CHARTER SCHOOLS WITH TRADITIONAL PUBLIC SCHOOLS
In addition to specific institutional and organizational structures, charter schools tend to differ from TPSs in many other ways. For example, charter schools tend to locate in large metropolitan areas, where higher population density enhances the schools prospects of attracting enough students. Many charter schools provide only K8 education since it is generally less costly than high school education (Arsen, Plank, & Sykes, 1999). To some extent, charter schools implicitly select or shape their student body. Some charter schools attract students disproportionately from certain social, ethnic, or racial backgrounds or academically talented students with marketing strategies, admission criteria, and curricular focus (Scott & Villavicencio, 2009; Wells, Holme, Lopez, & Cooper, 2000). Many other charter schools are established specifically to draw students who are struggling and poorly served in their TPSs (Carnoy et al., 2005). All these differences may result in systematic difference in teachers perceptions of working conditions, regardless of the type of schools. For instance, working in urban schools with high concentrations of disadvantaged students might negatively influence teachers perceptions of working conditions if educating disadvantaged students entails more intensive and demanding instructional strategies, or more time for class preparation or individual student problems.
To mitigate the confounding effects of these factors on teachers perceptions of working conditions, propensity score matching (PSM) was used to pair charter schools with TPSs that are similar in terms of school location, educational level, school type, and student composition. The comparison between the matched charter schools and TPSs generates more reliable estimates because charter schools are compared to a TPS that has the same (or very similar) propensity to be a charter school. Indeed, PSM is regarded as one of the most robust methods for reducing bias due to the differences in observed factors (Rosenbaum & Rubin, 1983).
Several steps are involved in pairing charter schools and TPSs through PSM. First, all schools were stratified by urbanicity (whether the school is located in a central city, urban fringe or large town, or rural or small town), the level of the school (elementary, secondary, or combined), and school type (regular or nonregular schools, including special education schools, vocational or technical, or alternative schools). This created 3 x 3 x 2 = 18 strata so that each school was assigned to one of the strata and only schools within the same stratum could be matched. For example, a regular urban elementary charter school would be matched only with a regular urban elementary TPS.1 This stratification procedure greatly enhanced internal validity of matching (see Berends, Goldring, Stein, & Cravens, 2010; Zimmer & Buddin, 2007; Zimmer et al., 2003).
Second, within each stratum, charter schools and TPSs were matched using the MatchIt procedure in the R statistical package. A propensity score was generated for each school by estimating a logistic regression model conditional on multiple school characteristics, such as percentage of students eligible for the free or reduced lunch program, percentage of minority students, percentage of students with an individual education plan (IEP), and percentage of students who were of limited English proficiency (LEP). Other variables, including school size, class size, teacher characteristics, and teacher salary, were not included as predictors in the logistic regressions because charter schools tend to have smaller school sizes and hire inexperienced teachers. In addition, charter schools tend to pay significantly lower salaries than TPSs for teachers with similar credentials and experiences (Cannata, 2008). These policy variables are endogenous to the charter school philosophy (Zimmer et al., 2003) and systematically influence teachers perceptions of working conditions (Darling-Hammond, Ross, & Milliken, 2007). Including these policy variables in the PSM would reduce the organizational differences between charter schools and TPSs and would bias teachers perception of working conditions between the two types of schools.
Finally, based on the estimated propensity scores, the nearest neighbor matching method was used to match each charter school with a TPS that was in the same stratum and had the smallest average absolute propensity score distance from the charter school. Through the PSM procedure, each charter school in the sample was successfully matched with one TPS, resulting in an analysis sample of 233 charter schools and 233 TPSs.2 Table 1 describes the sample sizes and student composition of matched charter schools and TPSs in urban, suburban, and rural areas. The student compositions in charter schools and TPSs are similar except that the percentages of Hispanic and Asian students are not closely matched in rural areas.
Table 1. Student Composition of Charter Schools and Matched Traditional Public Schools, Unweighted
Table 2 shows some summary information for the charter schools. Among all sampled charter schools, more than 40% are granted by school districts, about one third are granted by state boards of education (SBE), and about 20% are granted by other organizations, including postsecondary institutions, state charter-granting agencies, and others. The average years in operation for all charter schools are about 5 years. About 72% of all charter schools are newly created, 22% are converted from a preexisting school, and 6% are otherwise created. The conversion charter schools were not further distinguished as formerly private or public in the analysis because of the small samples for each category. The next section describes the methods comparing teachers perceptions of working conditions in the matched samples of charter and traditional public schools.
Table 2. Charter School Statistics
Note. All the statistics are weighted by the SASS school final weight.
ESTIMATING TEACHER WORKING CONDITIONS IN CHARTER SCHOOLS AND TRADITIONAL PUBLIC SCHOOLS
HLMs were used in the analysis because of the nested structure of teacher and school data. For example, the perceptions of working conditions of individual teachers within the same school are interdependent, mainly because of shared experiences within the school and because of the ways in which individual teachers were initially drawn into the school. Traditional regression models analyzed at the teacher level fail to consider the dependence of individual teachers within schools. Alternatively, if data are analyzed only at the school level, variation in teachers perceptions of working conditions associated with individual teacher characteristics cannot be captured. Either way, a traditional regression model can introduce inefficiency and bias in estimations. HLMs resolve this problem by incorporating data for teachers and schools in a multilevel nested structure that allows the perceptions of teachers on working conditions to differ by individual teacher characteristics and to vary from one school to another simultaneously (Raudenbush & Bryk, 2002). The HLM software also takes into account the design effect arising from the multistage complex sampling of the SASS by incorporating weights for teachers and schools in the analysis.
The HLM analysis was sequentially processed starting with a set of basic two-level HLMs that can be expressed as:
Level 1: Yij = β0j + rij
Level 2: β0j = γ00 + γ01(CHARTER)j + Σγ0p(CHARTER * CHARTER TYPEp) j + u0j
Teachers are the unit of analysis in level 1 and schools in level 2. At level 1, the outcome variable, Yij, is the perception of the working conditions for teacher i in school j. In total, there are seven outcome variables, representing seven dimensions of working conditions. Except for the measure of instructional supply availability that was derived from an original survey item, all six other measures were continuous variables generated from the factor analysis described above. For easier interpretation in the regression analyses, all seven working condition measures were standardized to have a mean of 0 and a standard deviation of 1. No teacher characteristics variables were added in the level 1 equation at this stage. At level 2, the regression intercept, β0j, in the level 1 model for each school was conceived as the outcome variable that is hypothesized to depend on the dummy variable, CHARTER, indicating whether a school is a charter school or not. The coefficient γ01 captures the influence of charter schools on β0j, the mean difference in teacher perceptions of working conditions between charter schools and TPSs. To distinguish the working conditions in different types of charter schools, additional models were estimated where CHARTER TYPE is included to further describe the authorizer of the charter, the years in operation, and whether the charter school is newly created or converted from a preexisting school. Final weights for teachers and schools were incorporated at both levels and in all HLM estimations.
After the basic models, a set of fully conditional models were estimated by adding teacher and school characteristics to the level 1 and level 2 equations.
Level 1: Yij = β0j + Σβkj (teacher characteristics) kj + rij
Level 2: β0j = γ00 + γ01 (CHARTER)j + Σγ0p(CHARTER * CHARTER TYPEp)j + Σγ0q (school characteristicsq)j + u0j
βkj = γk0 for k = 1, …, K
The outcome variable Yij at level 1 remained the same. The explanatory variables in the level 1 models were a set of teacher personal and professional characteristics, including gender, race, educational degree, whether they hold a teaching certificate,3 whether they belong to a teachers union, whether they have less than 3 years of teaching experience, and teacher salary. At level 2, a set of school characteristics were controlled, including school size, pupil-teacher ratio, elementary or secondary level, urbanicity, and student composition. Teachers perceptions of working conditions in a school may be highly correlated with these school characteristics, regardless of whether the school is a charter or a traditional school.
In Table 3, the matched charter schools and TPSs on variables included in the HLM analysis are compared. After the SASS teacher and school final weights were applied, the matched sample represented about 38,000 charter school teachers in 2,179 charter schools and 119,000 teachers in 3,173 TPSs. Compared to TPS teachers, charter school teachers are younger and have fewer years of teaching experience. Charter school teachers are less likely to be licensed and earn significantly less than their TPS counterparts. While the majority of TPS teachers are members of a teacher union, only a quarter of charter schools teachers are union members.
Table 3. Descriptive Statistics: Weighted Mean Differences between Matched Charter Schools and Traditional Public Schools
Notes. aweighted by the SASS teacher final weight. bweighted by the SASS school final weight.
* p < .05. ** p < .01.
Turning to school characteristics, charter schools are much smaller but have similar pupil-teacher ratios to TPSs. Charter schools tend to have lower proportions of students of color and students from low-income families. The student composition comparisons between charter schools and TPSs are different from those shown in Table 1 because the statistics shown in this table have taken into account the SASS weights. The differences indicate that a larger share of charter schools are educating White and high-income students than TPSs. This highlights the necessity to control for student composition in comparing the working conditions even in the matched sample of charter schools and TPSs. The next section compares teachers perceptions of working conditions between charter schools and traditional public schools in a multilevel framework of HLM analysis.
This section reports the results of the two sets of HLM models assessing the influence of charter schools on teachers perceptions of working conditions. The first set of models are basic models including only variables describing the charter status of schools, while the second set controls for the full set of teacher and school characteristics.
For the first set of regressions, reported in Table 4, four basic HLM models were estimated for each domain of the working conditions. Model 1 includes only one explanatory variable CHARTER at level 2 indicating whether a school is a charter school. Models 2 to 4 further differentiate charter schools by years of operation, prior existing status, and authorizer, respectively. Since the dependent variables of all models are standardized to have a mean of 0 and a standard deviation of 1, the coefficients of the explanatory variables can be interpreted as effect sizes or as proportions of a standard deviation.
Table 4. Estimated Coefficients in the Basic Hierarchical Linear Models
The Model 1 results show that the estimated coefficient of the CHARTER variable in regressions assessing five of the seven working conditions domains is insignificant. This indicates that charter schools and TPSs have similar working conditions for teachers in many domains, except that charter school teachers reported 40% standard deviations greater influence in school-wide policies and about one-fifth standard deviations less in professional development opportunities than TPS teachers.
The results of Model 2 and Model 3 show that the variables measuring the operation years of a charter school and its prior existing status reveal no significant impacts on teachers perceptions of working conditions, while the coefficients on CHARTER remain similar to those in Model 1. This indicates teachers perceptions of working conditions do not vary depending on whether a charter school is newly created or converted from a preexisting school. Nor does the experience charter schools gain through additional years of operation significantly change teachers perceptions of working conditions.
Model 4 allows the perception of working conditions to vary among charter schools with different authorizers. Two dummy variables indicating the grantors of charter schoolsby SBEs and by other organizationsare added to the level 2 equations, and the charter schools granted by local school districts serve as the reference group. The coefficients on CHARTER, interpreted as the difference of teacher perceptions of working conditions between district-granted charters and TPSs, become larger and significant in several regressions, suggesting that teachers in district-granted charter schools perceive better working conditions than TPS teachers in principal leadership, sense of community and collegiality, and influence in school-wide policies and classrooms. The magnitudes of the differences are sizable, ranging from 20% to 60% of a standard deviation. In addition, the coefficient in the professional development regression becomes insignificant, suggesting that district-granted charter schools provide at least the same professional development opportunities for their teachers as TPSs. In contrast, the coefficients of SBE-granted charters in all regressions are negative and significant, implying that teachers in SBE-granted charters perceive significantly less supportive working conditions than teachers in district-granted charters in every domain. Further, indicated by the insignificance of the coefficient of charters granted by other organizations in most regressions, the teachers in these charter schools perceive working conditions less supportive than teachers in district-granted charters.
The basic models examine the working conditions in charter schools and TPSs without controlling for teacher and school characteristics. As shown earlier, charter schools differ systematically from TPSs in terms of the teachers the schools hire and students the schools serve. Even after matching, many of these differences between the two types of schools are still statistically significant. It is interesting to see how much of the variation of the teacher working conditions between the two types of schools can be explained by teacher and student characteristics and by the specific institutional features of charter schools, such as autonomy, freedom of regulation, and the pressure of market competition. Next, we turn to estimate the conditioned models by controlling the full set of teacher and school characteristics variables, so that we can separate the effect of these extraneous factors from that of the institutional differences between the two types of schools.
As shown in Table 5, for each working condition domain, two different models are estimated. In Model 1, charters are represented once again by the CHARTER dummy variable, while Model 2 differentiates the authorizers of charter schools. Although two additional models indicating charter schools years of operation and prior status were estimated, the results are not reported here because, similar to the basic models, the regressions produced no significant results for these variables.
Table 5. HLM Analysis of the Teacher Working Conditions in Charter Schools and Traditional Public Schools
Table 5. HLM Analysis of the Teacher Working Conditions in Charter Schools and Traditional Public Schools (Continued)
Note. The robust standard error was used. The SASS final weights for schools and teachers were incorporated. * p < .05. ** p < .01
Model 1 in Table 5 shows that charter school teachers still perceived significantly greater influence in school-wide policies than TPS teachers, although the magnitude of the difference decreased slightly from 40% of a standard deviation in the basic model to 36% of a standard deviation after controlling for teacher and school characteristics. However, after controlling for these characteristics, charter school teachers now perceive 20% standard deviations heavier workloads than teachers in TPSs, while the charter school teachers perceptions of professional development have become similar to TPS teachers.
Model 2 allows for the perception of working conditions to vary among charters with different authorizers. The results reinforce the pattern found in the basic models that teachers in district-granted charter schools perceive better working conditions than teachers in TPSs. For example, teachers in district-granted charters perceived about 59% standard deviations more influence on school-wide policies than TPS teachers. In addition, the heavy workload of charter school teachers presented in Model 1 is not reflected in district-granted charter schools, but is attributable specifically to the negative perceptions of workload in charter schools granted by SBEs and other organizations.
Moreover, district-granted charter schools seem to provide better working conditions than charters granted by SBE and other organizations. Consistent with the basic model results in Table 4, SBE-granted charter schools had significantly lower ratings on six of the seven working condition indexes than district-granted charter schools. For example, teachers in SBE-granted charters reported about 23% standard deviation less support from principals and about half of a standard deviation less influence on school-wide policies than teachers in district-granted charter schools. The only exception is teachers perception of professional development. Although teachers in SBE-granted charter schools perceived less professional development than teachers in district-granted charters, the coefficient is not statistically significant. In addition, teachers in charter schools granted by organizations other than SBEs and school districts also reported significantly less professional development and heavier workloads than district-granted charter schools, a result consistent with the basic models.
Table 5 also suggests that teachers with different personal and professional characteristics tend to have different perceptions of working conditions. For example, male teachers tended to report fewer professional development opportunities than female teachers. Novice teachers with less than 3 years of teaching experience tended to perceive fewer professional development opportunities and heavier workloads than more experienced teachers. White teachers expressed less principal leadership and less influence in school policy than minority teachers. Teachers with masters degrees tended to report weaker principal leadership but lighter workloads than teachers with bachelors degrees. Interestingly, salary did not influence teachers perceptions of many dimensions of working conditions. The only differences are that high-salary teachers reported more professional development opportunities but at the same time heavier workloads than teachers earning lower salaries. Surprisingly, whether teachers belong to a teachers union seemed to have little influence on their perceptions of working conditions, except that teachers who are union members reported significantly smaller workloads than nonunion teachers.
In addition to teacher characteristics, the characteristics of the schools the teachers work in also systematically influence teachers perceptions of working conditions. For example, teachers in schools serving higher proportions of minority students tended to report significantly more adverse working environments, including less support from the principal, less sense of community and collegiality, less influence on school policy, and less adequate instructional supplies. Teachers in large schools reported slightly less attractive working conditions, while large class size does not seem to be strongly associated with negative perceptions of working conditions. Compared to teachers working in suburban schools, teachers in rural schools perceived relatively adverse working conditions in two domains, including a weak sense of community and collegiality and inadequacy of instructional supplies. In addition, secondary school teachers tended to report a stronger sense of control in their classrooms but less professional development than elementary school teachers.
DISCUSSION AND IMPLICATIONS
This paper compared teacher working conditions in charter and traditional public schools and among various types of charter schools. The paper also distinguishes whether the different working conditions are influenced by the intrinsic institutional features of charter schools such as autonomy and competition or by extraneous factors such as measureable school and teacher characteristics. As the conditional weighted HLM results show, autonomy and competition influence teacher working conditions in charter schools. After controlling for teacher and school characteristics, teachers in charter schools reported about 40% standard deviations more influence in school-wide policies than TPS teachers. This finding implies that charter schools extend the autonomy to their teachers to some extent so that the teachers have more opportunities to participate in school-level decision making. The HLM analysis also shows that the workloads in charter schools are about one fifth of a standard deviation higher than in TPSs, indicating that charter schools are under market pressure to be more efficient and their teachers have to work extra hours.
In addition to these two aspects of working conditions, however, teachers perceptions of working conditions are not substantially different between charter schools and TPSs. After controlling for teacher and school characteristics, teachers in the two types of schools perceived similar levels of principal leadership, sense of community and collegiality, classroom autonomy, opportunities for professional development, and adequacy of instructional supplies. This finding suggests that autonomy in charter schools does not necessarily foster changes regarding classroom instruction, professional development, and teacher collaboration that are important in improving student achievement. This finding supports previous research that although charter school teachers have greater influence in school policies, the decision-making processes seem to be unrelated to curriculum and instruction and often demand extra time, which significantly adds to their workload (Johnson & Landman, 2000; Malloy & Wohlstetter, 2003).
The analysis of working conditions among different types of charter schools also generated some interesting findings. As the results of the weighted conditional HLM models show, teachers in district-granted charter schools perceived more power in school-wide decision making but have workloads similar to teachers in TPSs. This finding is intriguing because it suggests that autonomy and workload are not necessarily in conflict; it is possible to empower teachers in decision making and avoid overworking them at the same time. In addition, among charter schools granted by different organizations, district-granted charter schools generally provide more supportive working environments than charter schools granted by SBE, postsecondary institutions, and other organizations. This implies that district-granted charter schools not only provide more opportunities for teachers to influence school policies but also receive necessary support from their districts. The school districts may take on certain administrative and governance responsibilities that can demand considerable time and expertise, so that teachers are freed to focus on participating in decision making regarding classroom instruction. Since the designation of which organizations are authorized to grant charters is defined by each states laws, these results also imply that state policy can have some indirect influence over charter school working conditions. For states heavily relying on authorizers that were independent from the established structure of school districts, substantial administrative support and oversight need to be provided to improve working conditions in charter schools.
In addition to the institutional difference between charter schools and TPSs, systematic differences of measurable characteristics of schools and teachers between charter schools and TPSs also contribute to the different perceptions of working conditions. School size is usually negatively related to teachers perceptions of working conditions. As the HLM analysis shows, teachers working in large schools tended to report more adverse working conditions, such as less principal leadership and support, less sense of community and collegiality, less influence in school policies, and less classroom autonomy. As a result, teachers in charter schools tended to perceive more supportive working conditions because of small school sizes. However, the low funding level in charter schools, reflected in higher proportions of inexperienced teachers and lower salaries, is associated with fewer professional development opportunities than TPSs. This might suggest that charter schools do not have enough funding to support release time from teaching or tuition reimbursement. Since professional development is crucial for beginning teachers to grow as professionals, lack of professional development can be detrimental to teacher retention.
In addition, charter schools tend to serve lower percentages of minorities, which is associated, on average, with more favorable working conditions, such as more principal leadership, sense of community, and more influence on school policy. In this paper, the comparisons of working conditions are based on a matched sample of charter schools and TPSs with similar student characteristics, locations, and school level. Therefore, although these extraneous factors exert some independent explanatory power in predicting teachers perception of working conditions, these factors do not seem to make a significant difference in teachers perception of working conditions between the two types of schools.
Several limitations in this study need to be noted. The first is the potential for the bias that arises from the use of self-reported survey data in quantifying subjective characteristics of the schooling environment. Second, because of data limitations, the analysis missed several domains of teacher working conditions such as physical facilities and appropriateness of teacher assignment, which are expected to be different between TPSs and charter schools. In addition, certain measures of working conditions in this analysis were based only on responses to a limited number of survey items. For example, the measure of professional development is based on responses to two items, and the adequacy of teaching materials is based on one item. We need to be cautious about their interpretations.
Although research has shown that charter schools tend to have high attrition rates and often do not outperform their TPS contemporaries (Miron & Applegate, 2007; Miron et al., 2008), evidence on how teacher working conditions influence student achievement is scarce. The current study fails to address this issue directly due to the limitation of the data. However, the findings imply that if teachers in charter schools enjoy more autonomy only unrelated to classroom instruction and collaboration, then it may not improve student achievement in charter schools. This is consistent with the finding of a recent RAND study that, although charter schools have more control over decision making, that control fails to translate into higher test scores in charter schools (Zimmer & Buddin, 2007).
Given the existing literature that supportive working conditions increase teacher satisfaction and reduce teacher turnover rates, an important next step is to examine what aspects of working conditions matter most to charter school teachers as they make decisions about their careers. It is also important to consider what aspects of working conditions have an impact on building teacher capacity and whether they will translate into greater gains in student achievement. In addition, due to the finding that district-granted charter schools tend to provide more supportive teaching working conditions than other types of charter schools, further research is needed to explore how different authorizers affect working conditions for teachers and whether the more supportive working conditions in some types of charter schools translate into low attrition rates and high student achievement.
1. Some strata are further combined because of inadequate number of charter schools in each stratum. For example, rural charter schools are lumped into one stratum because they do not have much variation in educational level and school type.
2. A few combined (elementary and secondary) charter schools that failed to be matched with combined TPSs in the same stratum were matched to either an elementary or secondary TPS, based on the criterion of whichever generated the smallest distance between the propensity scores.
3. The SASS asks whether teachers have five different kinds of certificates. In this analysis, teachers are considered certificated if they have a regular, probationary, or provisional certificate. Teachers with temporary or emergency certificates are considered uncertified.
Arsen, D., Plank, D., & Sykes, G. (1999). School choice policies in Michigan: The rules matter. East Lansing: Michigan State University.
Berends, M., Golding, E., Stein, M., & Cravens, X. (2010). Instructional conditions in charter schools and students mathematics achievement gains. American Journal of Education, 116(3), 303335.
Berends, M., Watral, C., Teasley, B., & Nicotera, A. (2007). Charter school effects on achievement: Where we are and where were going. In M. Berends, M. G. Springer, & H. J. Walberg (Eds.), Charter school outcomes (pp. 243267). Mahwah, NJ: Erlbaum.
Betts, J. R., & Loveless, T. (2005). Getting choice right: Ensuring equity and efficiency in education policy. Washington, DC: Brookings Institution Press.
Bomotti, S., Ginsberg, R., & Cobb, B. (1999). Teachers in charter schools and traditional schools: A comparative study. Education Policy Analysis Archives, 7(22), 228.
Buckley, J., Schneider, M., & Shang, Y. (2005). Fix it and they might stay: School facility quality and teacher retention in Washington, DC. Teachers College Record, 107(5), 11071123.
Buddin, R., & Zimmer, R. (2005). Student achievement in charter schools: A complex picture. Policy Analysis and Management, 24(2), 351371.
Cannata, M. (2007). Teacher community and elementary charter schools. Education Policy Analysis Archives, 15(11).
Cannata, M. (2008). Teacher qualifications and work environments across school types. Tempe, AZ: Education Policy Research Unit, Education Public Interest Center.
Carnoy, M., Jacobsen, R., Mishel, L., & Rothstein, R. (2005). The charter school dust-up: Examining the evidence on enrollment and achievement. New York, NY: Teachers College Press.
Crawford, J. R., & Forsyth, P. B. (2004). Teacher empowerment and charter schools. Journal of School Leadership, 14(1), 6281.
Darling-Hammond, L., Ross, P., & Milliken, M. (2007). High school size, organization, and content: What matters for student success? Brookings Papers on Education Policy - 2006/2007.
Finnigan, K. S. (2007). Charter school autonomy: The mismatch between theory and practice. Educational Policy, 21(3), 503526.
Gawlik, M. A. (2007). Beyond the charter schoolhouse door: Teacher-perceived autonomy. Education and Urban Society, 39(4), 524553.
Gill, B. P., Timpane, P. M., Ross, K. E., & Brewer, D. J. (2001). Rhetoric versus reality: What we know and what we need to know about vouchers and charter schools. Santa Monica, CA: RAND.
Hassel, B. C. (1999). The charter school challenge: Avoiding the pitfalls, fulfilling the promise. Washington, DC: Brookings Institution.
Hill, P., Angel, L., & Christensen, J. (2006). Charter school achievement studies. Education Finance and Policy, 1(1), 139150.
Ingersoll, R. M. (2001). A different approach to solving the teacher shortage problem (No. 3). Seattle: Center for the Study of Teaching and Policy, University of Washington, Seattle.
Ingersoll, R. M. (2004). Why do high-poverty schools have difficulty staffing their classrooms with qualified teachers? Washington, DC: Center for American Progress.
Johnson, S. M. (2006). The workplace matters: Teacher quality, retention and effectiveness. NEA Research Best Practices Working Paper.
Johnson, S. M., & Birkeland, S. (2003). Pursuing a sense of success: New teachers explain their career decisions. American Educational Research Journal, 40(3), 581617.
Johnson, S. M., & Landman, J. (2000). Sometimes bureaucracy has its charms: The working conditions of teachers in deregulated schools. Teachers College Record, 102(1), 85124.
Kolderie, T. (1990). Beyond choice to new public schools: Withdrawing the exclusive franchise in public education. Washington, DC: Progressive Policy Institute.
Ladd, H. (2009). Teachers perceptions of their working conditions: How predictive of policy-relevant outcomes? Santa Cruz, CA: The New Teacher Center.
Loeb, S., Darling-Hammond, L., & Luczak, J. (2005). How teaching conditions predict teacher turnover in California schools. Peabody Journal of Education, 80(3), 4470.
Malloy, C. L., & Wohlstetter, P. (2003). Working conditions in charter schools: What's the appeal for teachers? Education and Urban Society, 35(2), 219241.
McLaughlin, M. W., & Talbert, J. E. (2001). Professional communities and the work of high school teaching. Chicago, IL: University of Chicago Press.
Miron, G., & Applegate, B. (2007). Teacher attrition in charter schools. Tempe, AZ: Education Policy Research Unit, Education Public Interest Center.
Miron, G., Evergreen, S., & Urschel, J. (2008). The impact of school choice reforms on student achievement. Tempe, AZ: Education Policy Research Unit, Education Public Interest Center.
Nathan, J. (1996). Charter schools: Creating hope and opportunity for Americans education. San Francisco, CA: Jossey Bass.
Raudenbush, S., & Bryk, A. (2002). Hierarchical linear models for social and behavioral research: Applications and data analysis methods (2nd ed.). Newbury Park, CA: Sage.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 4155.
Scott, J., & Villavicencio, A. (2009). School context and charter school achievement: A framework for understanding the performance black box. Peabody Journal of Education, 84(2), 227243.
Sykes, G. (1999). The new professionalism in education: An appraisal. In J. Murphy & K. S. Loius (Eds.), Handbook of research on educational administration (pp. 227276). San Francisco, CA: Jossey-Bass.
Weiss, E. M. (1999). Perceived workplace conditions and first-year teachers' morale, career choice commitment, and planned retention: A secondary analysis. Teaching and Teacher Education, 15, 861879.
Wells, A. S. (1998). Beyond the rhetoric of charter school reform: A study of ten California school districts. Los Angeles: UCLA Charter School Study.
Wells, A. S., Holme, J. J., Lopez, A., & Cooper, C. W. (2000). Charter schools and racial and social class segregation: Yet another sorting machine. In R. D. Kahlenberg (Ed.), A notion at risk: Preserving education as an engine for social mobility (pp. 169222). New York, NY: Century Foundation Press.
Wohlstetter, P., Wenning, R., & Briggs, K. L. (1995). Charter schools in the United States: The question of autonomy. Educational Policy, 9(4), 331358.
Zimmer, R., & Buddin, R. (2007). Getting inside the black box: Examining how the operation of charter schools affect performance. Peabody Journal of Education. 82(2-3), 231273.
Zimmer, R., Buddin, R., Chau, D., Daley, G., Gill, B., Guarino, C., . . . Brewer, D. J. (2003). Charter school operations and performance: Evidence from California. Santa Monica, CA: RAND.
MEASURES FOR TEACHER WORKING CONDITIONS: SURVEY ITEMS AND FACTOR LOADINGS
Note. The SASS final teacher weights were used. aOriginal Likert scale responses coded as 1= strongly agree, 2 = somewhat agree, 3 = somewhat disagree, and 4 = strongly disagree. For easier interpretation, the codes were reversed so that higher values represent higher levels of agreement. bLikert scale responses coded as 1 = no influence, 2 = minor influence, 3 = moderate influence, and 4 = a great deal of influence.