In the Storm's Eye: How Race, Experience, and Exposure Shape Arizona Teachers' Attitudes Toward School Choice

by Frederick Hess, Robert Maranto, Scott Millman & Kathleen Grammatico Ferraiolo - 2002

Amidst the growing research on choice-based school reform, little attention has been paid to examining how teachers view school choice or what factors shape their attitudes. Teachers can impact the success of choice experiments through their willingness to launch and staff schools of choice, share information with teachers at those schools, and support change in their unions and communities. Examining Arizona, the state with the nation's most developed system of choice, we explore how personal traits, including race, tenure, partisanship, and familiarity with charter schooling, influence teachers' attitudes toward charter schools and school vouchers. In addition, we examine how other school- or district-level variables, including culture and charter penetration, inform teachers' views. Finally, to assess the effect of experience with school choice, we examine the factors that shape charter schoolteachers' views about choice. Using a 1998 survey, we find that White, experienced, unionized, Democratic educators and those working in "positive" school environments are less supportive of school choice. Charter-school teachers are significantly more positive about school choice than their public school counterparts, although they do not respond to the same variables. The results can help illuminate the likely political prospects and practical effects of choice-based reform.

Amidst the growing research on choice-based school reform, little attention has been paid to examining how teachers view school choice or what factors shape their attitudes. Teachers can impact the success of choice experiments through their willingness to launch and staff schools of choice, share information with teachers at those schools, and support change in their unions and communities. Examining Arizona, the state with the nation's most developed system of choice, we explore how personal traits, including race, tenure, partisanship, and familiarity with charter schooling, influence teachers' attitudes toward charter schools and school vouchers. In addition, we examine how other school- or district-level variables, including culture and charter penetration, inform teachers' views. Finally, to assess the effect of experience with school choice, we examine the factors that shape charter schoolteachers' views about choice. Using a 1998 survey, we find that White, experienced, unionized, Democratic educators and those working in "positive" school environments are less supportive of school choice. Charter-school teachers are significantly more positive about school choice than their public school counterparts, although they do not respond to the same variables. The results can help illuminate the likely political prospects and practical effects of choice-based reform.

The 1990s witnessed growing attention to the adoption of choice-based school reforms, particularly charter schooling and school vouchers. Although researchers have approached the school choice debate from theoretical, normative, and empirical directions, there has been relatively little attention to how public educators respond to market-based education reform. Such an oversight is significant because the nature of their profession allows teachers significant influence over how choice-based reforms develop and how traditional schools respond to competition from schools of choice. By agreeing or refusing to cooperate with or staff schools of choice, and by supporting or opposing policy change in their unions and communities, teachers can affect the success of the school choice movement in powerful ways.

There has been relatively little rigorous consideration of how teachers view school reform efforts. In the case of choice-based reforms, but also more generally, we have little systematic understanding of which teachers embrace or reject reform initiatives and why they do so.ⁱ The only published multivariate analysis of the topic was a study that used a national 1995 survey of high school teachers to investigate how teachers' background and school environment affected their attitudes toward choice (Hess, Maranto, & Milliman, 2000). These authors found that unionized, Democratic, experienced teachers who majored in education and those who reported working in positive school cultures were less supportive of school choice. Their work provided a valuable contribution to the literature on how teachers influence the fate of school reform, and in the current study we seek to build on their analysis in several ways.

First, instead of employing the ordinary least squares regression model that Hess et al. (2000) used, we use a more appropriate ordered probit model. Ordered probit analysis allows more accuracy and precision in predicting the effects of relevant variables on continuous, ordered dependent variables with a limited number of categories, such as survey responses on attitudes toward school choice. Second, by introducing teacher race as an explanatory variable, we consider whether the attitudes of non-White teachers differ from those of their White counterparts. School choice is often touted as a solution for the ills of largely minority urban school systems. It will be particularly interesting to examine whether non-White teachers embrace school choice as a means of improving poorly performing schools or reject it as a source of further decline of those schools. Third, unlike Hess et al. we focus exclusively on the attitudes of Arizona educators, who have considerable experience working in a marketlike educational environment. By evaluating the effects of exposure to competition on Arizona teachers' support for school choice, we are able to shed light on how teachers are likely to respond as charter schooling continues to expand at its current rapid rate. Further, as Table 1 demonstrates, on a number of demographic and educational variables (with the notable exception of charter school enrollment), Arizona is similar to other states with large numbers of charter school campuses. The results from this study can thus be viewed as akin to a "canary in a coal mine" approach and may provide insight into how teachers' opinions everywhere may evolve as charter competition spreads. The central question of how educators respond to the introduction of choice-based competition and to reform measures more generally (and what factors shape their attitudes) has a lasting importance for the whole of education policy. Finally, after considering the primary question of what factors influence traditional schoolteachers' views about school choice, we undertake an inquiry into the views of charter-school teachers. A comparison of the views of charter and traditional schoolteachers, and the factors that shape those views, will help us assess how exposure to charter schooling influences teachers' attitudes about market-based reform. However, because of the self-selection issues raised by our examination of die views of charter school educators, we are hesitant to offer strong conclusions about our findings in this section.

We test our hypotheses concerning the factors that influence teachers' attitudes about school choice in Arizona, a state with a unique charter-schooling environment.ⁱⁱ Republican party control and a weak teachers' union helped Arizona's legislature forge what is often regarded as the nation's most ambitious charter-schooling law. Arizona has the largest charter school sector and the most expansive public choice legislation in the nation (Center for Education Reform, 2000; Gresham, Hess, Maranto, & Milliman, 2000). In 1998, when the survey we analyzed was administered, 3.3% of Arizona public school students attended the state's 222 charter school campuses (Gresham et al., 2000). Currently there are more than 400 charter schools in Arizona, enrolling 7% of public school students.

School choice aims to improve public education by promoting responsiveness, accountability, and innovation in an experimental setting free from bureaucratic constraints. Although they may be uninvolved in the process of crafting education policy, teachers are responsible for carrying out reforms in the classroom and thus play an important role in determining their success (Cohn & Kottkamp, 1993; Hess, Maranto, Milliman, & Gresham, 1999; Lipman, 1997; Lipsky, 1980). Because of the independence and discretion they exercise, educators can choose how to respond to reform initiatives (Tyack Be Cuban, 1995), and administrators have few tools at their disposal to encourage them to comply with change efforts (Hess, 1999; Wilson, 1989). Although the teachers' role in the implementation of school choice is not directly analogous to their role in the implementation of other school reforms, we believe that there are important comparisons to be made. At least in the short term, the success of school choice will depend on teachers' willingness to adopt new practices developed in schools of choice, to embrace choice in the political arena through their unions (Morken & Formicola, 1999), and to leave district schools to launch and staff new schools of choice (Nathan, 1996). And once they are working in schools of choice, teachers will continue to exercise substantial discretion in implementing the new instructional programs those schools adopt. Indeed, some research suggests that charter schools offer unique instructional programs characterized by low student-faculty ratios, personalized learning, and interdisciplinary approaches to learning but that an ongoing, integrated curriculum of professional development that would assist teachers in implementing those programs is usually absent (Griffin & Wohlstetter, 2001). Through their behavior in their classrooms and communities Griffin and Wohlstetter suggest charter-school teachers will prove pivotal in shaping the prospects and effects of choice-based reform.

HYPOTHESES

Apart from the research by Hess et al. (2000) discussed earlier, only a few studies have used survey research to examine the factors that influence teachers' opinions about school reform, and even fewer have considered market-based reforms. A study of 12,000 Chicago teachers designed to assess their response to the 1988 Chicago School Reform Act found that teachers in schools with shared decision making, strong leadership, teacher collegiality, and community support were more likely to approve of change (Sebring & Camburn, 1992). Katz, Dalton, and Glacquinta (1994) surveyed 280 middle-school home economics teachers in New York, finding that in-service training, increased resources, and job security enhanced receptiveness to state reform. In a survey of the attitudes of 302 state teachers of the year concerning several school reforms, Page and Page (1989) found significant differences in opinion between secondary and elementary teachers, older and younger teachers, and teachers with different educational backgrounds.

In this study, we examine how several variables related to teachers' background, school environment, and familiarity with school choice influence their attitudes about market-based education reform. Professional investment, which includes union membership and teacher experience, is the first set of factors we suspect will influence teachers' attitudes. Teachers' unions view privatization as a threat to union autonomy and to member compensation and working conditions (Eberts & Stone, 1984), and the two major national unions now tolerate charter schooling but remain strongly opposed to school vouchers (Hassel, 1999; Lieberman, 1997; Witte, 2000). Unions typically exhibit more opposition to private choice plans than public choice efforts; thus, we predict that union members will be less supportive than their nonunion counterparts of school choice in general and school vouchers in particular.

The second professional investment variable, length of service, is, along with school culture, among the most frequently cited determinants of teacher responsiveness to reform. Having witnessed many reform efforts come and go, veteran teachers are often cynical about change (Chubb & Moe, 1990; Hess, 1998; Hill, Pierce, & Guthrie, 1997; Johnson, 1996; Tyack & Cuban, 1995; Wagner, 1994). Discouraging personal experiences with change often lead veteran employees to view new reforms as impositions (Kettl, 1998). Further, experienced teachers are more likely to oppose change because they have more time and energy invested in the system of public education; consequently, we expect longer serving teachers to be less supportive of school choice.

Third, we hypothesize that partisan affiliation will influence teachers' attitudes about school choice. Along with teachers' unions, the Democratic Party endorsed charter schooling in the 1990s but remains firmly opposed to school vouchers. At the state and national level, Republicans are generally more supportive of both public and private choice (Hassel, 1999; Witte, 2000). Republican party identifiers are likely to have more faith than Democrats in the ability of market competition to improve a product such as education.ⁱⁱⁱ Therefore, other things equal, we hypothesize that Republican teachers will be more supportive than Democratic teachers of school choice.^iv

Fourth, because race has been shown to influence individual-level opinions in a number of policy areas, we expect Black, Hispanic, and White teachers to differ in their opinions concerning school choice. Because Black and Hispanic students tend disproportionately to be educated in poorer quality schools, they may support school choice, viewing it either as a way of allowing students to exit from low-performing schools or of improving the quality of those schools through competition. Indeed, polling data have shown that non-Whites, and Blacks in particular, are more supportive of some forms of school choice (Moe, 2001; Rose & Gallup, 1999). We therefore expect non-White teachers to be more supportive than their White counterparts of school choice.

School culture is a fifth variable that has been found to influence teachers' attitudes toward education reform. Hess et al. (1999, 2000) found that teachers working in negative school cultures were more supportive of school choice. Market-based reform is more threatening to employees who are satisfied with their work setting, whereas those in negative, unsupportive environments are more likely to welcome change. Consequently, we hypothesize that teachers reporting negative school culture will be more likely to support school choice.

The impressions teachers acquire through experience with charter schooling may also inform their attitudes about school choice. Three variables measure teachers' familiarity with public choice: whether a teacher has considered a charter school job, whether he or she has spoken with a colleague who works at a school of choice, and the charter school market share. By considering a job at a charter school or by speaking with a colleague who works at a charter school, teachers obtain firsthand experience with school choice in the real world. We recognize that the first variable raises concerns about endogeneity because teachers who consider a charter school job may be more supportive of choice-based reform to begin with. Teachers may consider a job at a charter school for many different reasons: to escape from a negative school environment, to be able to experiment with new teaching practices, or to join colleagues who have already taken such jobs. It is possible that support for school choice leads teachers to consider a job at a charter school, rather than the consideration of such a job producing support for choice. Still, we believe there are important reasons to include this variable in the study.^v In particular, any relationship between exposure to charter schooling and attitudes toward school choice will have important public policy implications, as traditional schoolteachers contribute to the launch and spread of charter schools, or as policymakers seek to increase teachers' exposure to schools of choice to soften their attitudes. Some research suggests that exposure to innovation breaks down employee opposition to reform (Dirks, Cummings, & Pierce, 1996; Sarbaugh-Thompson, 1998). Therefore, we hypothesize that teachers who have experience with charter schooling will be more supportive of public choice. It is possible that familiarity with public choice programs may persuade teachers to view charter schooling as a less controversial way to reform education and voucher programs as a radical, unnecessary threat to public education. Alternatively, teachers who gain exposure to limited school choice efforts may become more supportive of all forms of choice, including voucher programs. Accordingly, we predict that teachers who have considered taking a job at a charter school and those who know a colleague who has taken such a job will be more supportive of public and private school choice.

A third variable used to assess teachers' familiarity with school choice is charter school market share, which measures charter school enrollment by district. Because the survey we use to conduct our analysis sampled teachers in districts with both low and high populations of charter schools, we include the charter school market share variable to control for teachers' knowledge and familiarity with charter schooling. This variable is a robust measure of the presence of charter schooling in each teacher's district because it measures the number of charter schools per district independent of their size and the actual enrollment of students within those schools. Although we have predicted that familiarity with charter schooling will lead to greater support for school choice, it is possible that the experience that comes from living with competition is qualitatively different from the experience that comes from personal contact with a charter school. The variables that measure teachers' familiarity with competition may exert opposite effects, with intimate knowledge of charter schooling increasing support for choice, and mere awareness of the presence of a high student enrollment in schools of choice (and perhaps an enhanced sense of threat from those schools) increasing opposition. Still, we suspect it is more likely that teachers who work in a marketlike environment grow accustomed to their situation and come to accept and support choice in education. Consistent with our earlier predictions, we expect that a larger in-district charter presence will be associated with greater support for school choice generally.

As discussed earlier, we are also interested in exploring charter-school teachers' views about school choice and the factors that shape their attitudes. Therefore, we test identical models for charter-school teachers (with the exception of the variables that measure whether a teacher has considered a charter schooling job or know someone who has taken such a job) to determine whether their views about market-based education reform are subject to similar influences as those of traditional schoolteachers. We expect that the factors shaping charter-school teachers' attitudes may differ substantially from those of traditional schoolteachers, in part because of the self-selected nature of the former group. For example, charter-school teachers are more likely to be novice educators and may display less attachment to or membership in teachers' unions. In spite of the limitations of this part of the analysis, this first cut into the data will provide some insight into the factors that influence charter-school teachers' views about public and private school choice and the nature of those views themselves.

The data for this study were collected through a 1998 survey of public and charter-school teachers in Arizona. The Southwestern Teachers Survey was distributed by mail with stamped return envelopes to random samples of district elementary schoolteachers in 24 Arizona districts (54 schools) where 30% or more of public school campuses were charter schools (hereafter referred to as high-penetration districts), and 21 demographically similar Arizona districts (44 schools) with few or no charter schools (hereafter referred to as low-penetration districts); 19 of these districts had no charter schools. Up to 18 teachers were randomly selected in each school (the sample excluded librarians, administrators, and guidance counselors). We were particularly interested in sampling veteran teachers (defined as those who had been at their school for at least 3 years) because the survey asked teachers to compare how their schools had changed from 1994-1995 to 1997-1998. Respondents were paid $5 for participation, and two follow-up mailings were sent to nonrespondents. This yielded a 79.1% response rate and an N of 1065.

The average enrollment for all 45 districts sampled was 6,621, and total enrollment ranged from a high of 65,889 in the Mesa Unified District to a low of 576 in the Mayer Unified District. Average enrollment was slightly higher in high-penetration districts (7,527) than in low-penetration districts (6,191). Mean school size for both districts was between 614 and 617 students. All districts had a majority of White students; in high-penetration districts, 64.8% of students were White, and 57.8% of students in low-penetration districts were White. A slight majority of all districts sampled were located in suburban areas. Among the 21 low-penetration districts, 32.6% were located in urban areas, 37.9% in suburban areas, and 29.5% in rural areas. We find a similar pattern in the 24 high-penetration districts, 33.1% of which were located in urban areas, 40.1% in suburban areas, and 26.8% in rural areas. Finally, although both the high- (135.42) and low-(128.35) penetration districts scored slightly below the state average (145 combined score) on the Iowa Test of Basic Skills in 1994, the test scores were very similar for the two district samples. Overall, the low- and high-penetration districts were quite similar, with high-penetration districts displaying only slightly higher enrollment figures, percentage of White students, presence in suburban areas, and test scores.

In addition to the 1,065 traditional public schoolteachers who responded to the Southwestern Teachers Survey, 129 charter-school teachers also participated. Schools that opened in the fall of 1996 or before, and that had children enrolled in at least three of the six elementary school grades (defined as kindergarten through sixth grade), were eligible for selection. Thirty-five charter schools were selected, and we received responses from teachers in 29 of those schools, representing 20 traditional public school districts. The average enrollment in all charter schools sampled, 141 students, was significantly lower than that in the traditional schools selected. Out of 209 teachers sampled, 129 returned completed surveys, a slightly lower response rate of 63.2%, when compared with the traditional school sample. As Table 2 demonstrates, charter-school teachers are on average less unionized and less experienced than traditional schoolteachers, and they report more positive school cultures. The partisan affiliation and racial composition of charter-school teachers are similar to that of their traditional school counterparts. Finally, charter school educators are more likely to teach in more affluent schools and, not surprisingly, in districts that have a stronger charter-schooling presence.

Two dependent variables are used to measure teachers' attitudes toward school choice. The survey's public choice question defined charter schools as institutions that "can be started by teachers or parents and are largely autonomous public schools with open admission." The private choice question asked teachers to assess "vouchers for all non-religious schools, public or private, for all parents." School vouchers often allow low-income parents to use public funds to pay for half or more of the tuition at public or private schools of their choice (Peterson, 1998; Witte, 2000). We expect teachers' opinions about the two forms of choice to differ in magnitude but not in direction. Although a majority of teachers will likely oppose both public and private school choice, we predict that their attitudes toward the latter will be considerably more negative. Charter schooling is a more moderate form of choice that even traditional school choice opponents such as teachers' unions and Democrats now tolerate. School vouchers, meanwhile, remain extremely controversial, in part because critics argue that they increase segregation by race and class and drain resources .away from public schools.

Respondents reported their feelings about school choice using a 5-point scale, where the categories ranged from strongly oppose (variable value = 1) to strongly support (variable value = 5), which was recoded to eliminate the middle, no opinion category. The dependent variables were receded into four-category scales because there was no theoretical reason to expect that teachers interpreted the middle category as a neutral position, and results from models run using the original five-category scale were nearly identical. All percentages reported do not include a no opinion category. The exclusion of this category reduced the sample size from 1,065 to 879 respondents for the voucher question and from 1,065 to 863 respondents for the charter question. Finally, the N reported in Table 4 shrunk slightly when we performed the ordered probit analysis because not all teachers offered a response for every variable included in the analysis.

In the survey, teachers were asked to indicate their opinions concerning seven commonly discussed education reforms. A majority of teachers supported every non-market based reform the survey mentioned. But as Tables 3A and 3B show, 69.9% of traditional schoolteachers either strongly or moderately opposed charter schooling, and more than three quarters, 77.3%, opposed school voucher programs. Meanwhile, an overwhelming majority of charter-school teachers, 95.9%, supported charter schooling, and an almost equally large majority of 75.0% supported voucher programs.^vi

We hypothesize that attitudes toward school choice will be influenced by the explanatory variables previously discussed. We operationalize those variables in the following manner.

Union membership is measured by a dichotomous variable, with 1 as union membership and 0 as nonmembership.

Length of service is measured by subtracting the year the respondent joined the teaching profession from 1998, the year the survey used for this study was conducted.

Political party is measured by a self-reported variable that ranges from 1 (strong Republican) to 7 (strong Democrat).

Race is measured by a self-reported variable that was recoded into a dichotomous variable labeled non-White and coded 0 for Whites and 1 for teachers of all other races^vii

The school culture variable asks teachers to rate, on a 1 to 6 scale, whether colleagues share their beliefs and values about what the central mission of their school should be. Higher values indicate a more positive school culture.^viii

Finally, two measures of experience with charter schooling are dichotomous variables: Respondents who said they had considered leaving their current position for a job in a charter school were coded 1, otherwise, 0. Teachers who knew a colleague who had left for a teaching job at a charter school were coded as 1, else as 0. Charter school market share is a continuous variable measured as a percentage, calculated by dividing the number of children in both charter and traditional elementary schools in each district, with the number of children who attend charter elementary schools located in that district.

RESULTS

The four-category dependent variables made the use of ordered probit analysis appropriate. This technique has the advantage of allowing for the possibility of a nonlinear relationship between each explanatory variable and the probability that a respondent is in a given category of the relevant dependent variable. Results for the analysis of public and private school choice are presented in Table 4. Findings from both models were very much consistent with theoretical expectations. As hypothesized, the probability of offering an antivoucher response increases with union membership, teacher experience, and Democratic Party affiliation. Union teachers are also slightly less supportive of charter schooling. Non-White teachers are significantly more supportive of vouchers than Whites, and teachers working in negative school cultures are more supportive of charter schooling. Finally, although the presence of charter schooling has no significant effect on teachers' attitudes about school choice, personal experience with public choice is strongly associated with increased support for both public and private forms of school choice.^ix

To simplify interpretation of the ordered probit results, we established a baseline teacher respondent by setting all interval-level variables at their means and all dichotomous variables at their modes.^x Tables 5 and 6 display the baseline teacher's predicted probability of supporting school choice, as well as predicted probabilities for teachers with various combinations of attributes, based on the significance levels of relevant variables. The baseline teacher has a predicted probability of .59 of expressing strong opposition to private school choice and .26 probability of expressing moderate opposition. Professional investment and personal characteristics significantly influence receptivity to vouchers: The effects of race, union membership, length of service, and partisan affiliation are all consistent with expectations. Perhaps because minority students tend disproportionately to be educated in poorer quality schools, we find that a baseline respondent who is non-White is nearly 8% more supportive of vouchers than his White counterpart (p < .05). A baseline nonunion member is similarly 8% more likely to support private choice than a union teacher (p < .01), and an educator who has been teaching just 5 years is approximately 3% more supportive than a teacher of 20 years (p < .05).

Consistent with expectations, teachers who are more invested in public education 'through union membership or extensive service are opposed to vouchers, which they perceive might threaten traditional schools.^xi In addition, the 17% difference in support for vouchers between a strong Republican teacher and a strong Democrat (p < .01) suggests that rank and file party members reflect the views of the national party and its leaders. Overall, as Table 5 documents, movement from nonunion membership to union membership or strong Democratic affiliation to strong Republican affiliation, for instance, generally decreases the probability that teachers will strongly oppose school vouchers and increases the probability that they will moderately support vouchers. Even teachers whom we might expect to be the strongest advocates for school choice, such as Republicans, have only an 8% chance of strongly supporting voucher programs.

Results for the analysis of public school choice indicate that the charter-schooling model closely resembles the school voucher model; in fact, the models share two significant explanatory variables. The baseline teacher has a predicted probability of .46 of strongly opposing charter schooling and .33 probability of moderately opposing public school choice. Nonunion teachers are 5% more supportive of charter schooling than their union counterparts (p < .10). Teachers who know a colleague who has taken a charter school job are more than 8% more supportive of public school choice (p < .01). Finally, teachers working in a moderately negative school culture are 7% more supportive of public school choice than are teachers working in a very positive school culture (p < .10).^xii

As expected, union membership is a significant predictor of teachers' attitudes toward charter schooling. Although union leaders now tolerate some forms of public choice, rank-and-file members harbor more anti-choice views than nonmembers.^xiii Union membership remains an important attitude cue even for teachers who work in a strong charter-schooling state or who have seen charter schooling up close. Again in the case of charter schooling, the variables that enhance support for public choice tend more often to produce moderate rather than strong support for charter schooling. As we will discuss later, even among teachers who have considered a job at a charter school, whom we would expect to be the most favorably disposed to charter schooling generally, only about 18% are likely to be strong supporters of public choice, with another 37% supporting public choice only moderately.

Consistent with expectations and some previous research, teachers who report working in negative school cultures are more supportive than teachers in positive cultures of charter schooling. For teachers working in less positive settings, public schools of choice could represent either a source of competition that might encourage improvements in the culture of their own schools or an exit from unfavorable environments in the form of job opportunities. As the first three tables document, charter-school teachers' more positive rating of their school culture compared with that of traditional schoolteachers reinforces the possibility that teachers working in negative school cultures view schools of choice as more positive environments in which to work.

Although working in a marketlike environment alone does not foster support for school choice, teachers who have considered working in a charter school are two times more supportive of school vouchers and nearly three times more supportive of charter schooling (p < .01).^xiv However, it is unclear whether such teachers investigate employment in charter schools because of their already positive feelings about choice-based reform or whether seeking out such jobs leads them to view schools of choice more favorably. Both stories are plausible, and though the available data set does not allow us to tease out the precise nature of the relationship between the variables, the correlation between considering a charter school job and support for choice-based reform is significant from a public policy standpoint.^xv If positive feelings about school choice encourage teachers to seek out employment at charter schools, the prospects for choice are likely to improve as public school educators abandon traditional schools and contribute to the spread of schools of choice. If experience with charter schooling produces positive attitudes toward choice, policy makers may seek to encourage skeptical teachers to support choice-based reform by increasing their exposure to limited reform efforts, such as charter schools. Future efforts to understand teachers' attitudes toward school reform should devote particular attention to clarifying the nature of the causal relationship between exposure to public school choice and positive attitudes toward choice-based reform.^xvi

Finally, contrary to expectations, we find that charter-school market share has no significant effect on teachers' attitudes toward school choice. The presence of a flourishing or languishing charter school sector does not appear to influence teachers' attitudes. Combined with the findings discussed previously, these results suggest that a strong charter-schooling presence within teachers' district is a less important factor in explaining their attitudes than the personal experience with charter schooling that comes from considering a job at a charter school.

The results we present in Table 7 from the nearly identical model we tested for charter-school teachers are starkly different from the results for traditional schoolteachers. None of the variables discussed previously union membership, partisan affiliation, years of service, race, school culture, or charter market shareis significant in predicting charter-school teachers' attitudes about school choice. Although these findings suggest that the views of these educators are shaped by very different considerations than those of their traditional school counterparts, we are hesitant to draw firm conclusions from the data due to methodological concerns regarding self-selection and the small sample size of charter-school teachers. We discuss the implications of these findings in the, conclusion.

CONCLUSIONS

Teachers' opinions about choice-based school reform are relevant to scholars of the choice movement and to policy makers as well because the consequences of choice experiments will depend in large part on how the nation's three million teachers respond to reform. Even if school choice enjoys unanimous approval from lawmakers and the public, the willingness of teachers to support choice in their classrooms, unions, and communities will ultimately help determine how choice initiatives influence the practices and performance of traditional public schools. This study suggests that the future and impact of school choice may be powerfully affected by the makeup of the teacher workforce and the environments in which teachers work.

Our findings suggest that school choice is more likely to prosper in states with nonunion, novice, non-White teachers who have some familiarity with charter schooling. This result is particularly significant at a time when the rapid rate of teacher retirement and turnover may alter the makeup of-the nation's teacher workforce in the coming decade. The scant research done thus far on this topic has concentrated on how the introduction of school choice may change the population of teaching (Hoxby, 2001), but no attention has been paid to how the changing population of teachers may influence the future prospects of choice-based reform. In particular, efforts to recruit nontraditional teachers (who have backgrounds in fields other than education and lack certification) may help build a future teaching force that is significantly more receptive to school choice than the current one.

Teachers' professional investment, including union membership and length of service, along with variables such as partisanship and race, were found to significantly influence attitudes about market-based reform. Non-union teachers, novice teachers, Republicans, and non-White teachers are more supportive of school vouchers than their unionized, long-serving, Democratic, and White counterparts, respectively. It is likely that supporters of choice will continue to confront stiff opposition from the current teaching force, which is heavily unionized and has an average length of service of nearly 20 years. But the prospects for choice may improve in the coming decade as veteran teachers begin to retire and are replaced by younger, novice educators who are more receptive to choice experiments and practices.

School environment was also found to influence teachers' attitudes: Educators working in negative school cultures were slightly more receptive to public school choice. To the extent that public schools become less appealing places to work, teachers may begin to abandon these institutions and seek out more appealing and rewarding opportunities in schools of choice. Not surprisingly, a decline in the quality of the public school environment may produce, even among teachers, support for private sector alternatives.

We found that personal familiarity with charter schooling makes teachers more likely to moderately support choice-based reform. Simply working in a district with a strong charter-schooling presence does not soften opposition to voucher programs or charter schooling. On the other hand, teachers who learn about charter schooling on a personal level are significantly more supportive of private and public school choice. For traditional public schoolteachers, choice-based arrangements may be perceived as threatening because teachers may fear that they will force unpopular changes in traditional public schooling. The findings suggest that such concerns may be reduced as public educators acquire direct personal experience with choice-based alternatives. Despite the endogeneity issues the variable raises, the finding that personal experience with charter schooling is associated with support for school choice generally carries potentially significant public policy implications. If familiarity with public schools of choice leads to positive evaluations of choice-based reform generally, as teachers view reform up close they may come to support choice experiments and instructional methods in their union activities, classrooms, and communities. On the other hand, already pro-school choice teachers who pursue job opportunities in charter schools may contribute to the growth of choice-based reform by lending their experience and expertise to those schools.

Finally, in our analysis of the attitudes of charter-school teachers, we discovered that none of the variables that help us predict the views of traditional schoolteachers prove to be significant. Despite the methodological concerns regarding this analysis discussed earlier, these findings suggest that charter-school teachers' extremely positive views about both public and private school choice may be a result of their own personal exposure to choice experiments and practices. If this interpretation is correct, it indicates, along with the other measures of exposure that were significant predictors of traditional teachers' views about choice, that up-close, personal experience with school choice can powerfully increase teacher receptiveness to moderate reform in the form of charter schools, and perhaps even to more ambitious voucher programs.

As mentioned previously, we are cautious in interpreting and seeking to generalize from the results because we are unable to determine whether experience with charter schooling generates support for choice, or whether teachers who seek out charter-schooling jobs are more receptive to choice to begin with, and we strongly encourage other researchers to explore the nature of the causal relationship. Further, in using a survey of Arizona teachers to explore the personal and contextual traits that help influence teachers' attitudes about choice-based reform, we recognize that factors shaping teachers' attitudes are in part a function of local culture and organizational environment. Such considerations urge caution in generalizing these results, but its large public choice sector makes Arizona an important place to look for clues about how teachers everywhere will respond to the spread of school choice experiments.

The independence and discretion that teachers enjoy ensures that they will play a critical role in the evolution of market-based change in education. If public schoolteachers are willing to launch charter or voucher schools, are receptive to working in them, include other teachers in their professional networks, and do not oppose political leaders who advocate choice-based reforms, then the prospects of such reforms will be significantly brighter. On the other hand, to the extent that teachers refuse employment in choice-based schools, ostracize charter or voucher educators or freeze them out of professional networks, or publicly oppose choice-based reform, the political prospects of such proposals will be dimmer, and the rate at which choice schools might expand will be much slower. Moreover, any benefits from innovation will depend heavily on public school teacher receptiveness to techniques that are developed in schools of choice.

We do not mean to imply that teachers' attitudes alone determine the fate or the effects of school choice. The role of teachers, though important, is not deterministic. More broadly, the dearth of systematic research into teachers' attitudes about school reform is troubling for scholars interested in understanding how reform evolves and influences the practices and performance of public schools. Our hope is that this effort may inspire increased attention to understanding the nature and causes of teacher receptiveness or hostility to an array of education reforms. Such efforts can only help scholars and policy makers as they seek to promote school improvement.

REFERENCES

Center for Education Reform. (2001, October). Charter Schools in Arizona. Retrieved October, 2001, from http://www.edreform.com/charterschools/states/arizona.htm

Chubb, J., & Moe, T. (1990). Politics, markets, and America's schools. Washington, DC: Brookings Institution.

Cohn, M. M., & Kottkamp, R. B. (1993). Teachers: The missing voice in education. Albany: State University of New York Press.

Dirks, K. T., Cummings, L. L., &: Pierce, J. L. (1996). Psychological ownership in organizations: Conditions under which individuals promote and resist change. In R. W. Woodman &: W. A. Pasmore (Eds.), Research in organizational change and development (Vol. 9, pp. 1-23). Greenwich, CT: JAI Press.

Eberts, R. W., & Stone, J. A. (1994). Unions and public schools. Lexington, MA: Lexington Books.

Elmore, R. F. (1996). Getting to scale with good educational practice. Harvard Educational Review, 66(1), 1-26.

Finn, C. E., Manno, B., & Vanourek, G. (2000). Charter schools in action. Princeton, NJ: Princeton University Press.

Gresham, A., Hess, F. M., Maranto, R., & Milliman, S. (2000). Desert bloom: Arizona's free market in education. Phi Delta Kappan, 81(10), 751-757.

Griffin, N. C., & Wohlstetter, P. (2001). Building a plane while flying it: Early lessons from developing charter schools. Teachers College Record, 103(2), 336-365.

Hassel, B. C. (1999). The charter school challenge. Washington, DC: Brookings Institution.

Hess, F. M. (1998). Reconsidering the promise of systemwide innovation for urban districts. Educational Forum, 62(3), 249-257.

Hess, F. M. (1999). Spinning wheels: The politics of urban school reform. Washington, DC: Brookings Institution.

Hess, F. M., Maranto, R., Milliman, S., & Gresham, A. (1999). How Arizona teachers view school reform. In A. Gresham, F. M.

Hess, R. Maranto, & S. Milliman (Eds.), School choice in the real world (pp. 173-186). Boulder, CO: Westview Press.

Hess, F. M., Maranto, R., & Milliman, S. (2000). Resistance in the trenches: What shapes teachers' attitudes towards school choice. Educational Policy, 14(%), 195-213.

Hill, P. T., Pierce, L. C., & Guthrie, J.W. (1997). Reinventing public education: How contracting can transform America's schools. Chicago: University of Chicago Press.

Hoxby, C. M. (2001). Changing the profession: How choice would affect teachers. Education Matters, 1(1), 57-63.

Johnson, S. M. (1996). Leading to change: The challenge of the new superintendent. San Francisco: Jossey-Bass.

Katz, E. H., Dalton, S., & Glacquinta, J. B. (1994). Status risk taking and receptivity of home economics teachers to a statewide curriculum innovation. Home Economics Research Journal, 22(4), 401-421.

Kettl, D. F. (1998). Reinventing government: A fifth-year report card. Washington, DC: Brookings Institution.

King, G., Keohane, R., & Verba, S. (1994). Designing social inquiry: Scientific inference in qualitative research. Princeton, NJ: Princeton University Press.

Lieberman, M. (1997). The teacher unions. New York: Free Press.

Lipman, P. (1997). Restructuring in context: A case study of teacher participation and the dynamics of ideology, race, and power. American Educational Research Journal, 34(1), 3-37.

Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual in public services. New York: Russell Sage Foundation.

Long, J. S., &: Freese, J. (2001). Regression models for categorical dependent variables using Stata. College Station, TX: Stata Corporation.

Moe, T. (2001). Schools, vouchers, and the American public. Washington DC: Brookings Institution. Morken, H., & Formicola, J. R. (1999). The politics of school choice. Lanham, MD: Rowman & Littlefield.

Nathan, J. (1996). Charter schools: Creating hope and opportunity for American education. San Francisco: Jossey-Bass.

Page, F. M., Jr., & Page, J. A. (1989, November). The teaching profession and educational reform: Views of social science teachers. Paper presented at the meeting of the National Social Science Association Conference, New Orleans, LA.

Peterson, P. E. (Ed.). (1998). Learning from school choice. Washington, DC: Brookings Institution.

Rose, L. C., & Gallup, A. M. (1999). The 31st annual Phi Delta Kappa/Gallup poll of the public's attitudes toward the public schools. Phi Delta Kappan, 81(1), 41-56.

Sarbaugh-Thompson, M. (1998). Change from below: Integrating bottom-up entrepreneurship into a program development framework. American Review of Public Administration, 28(1), 3-23.

Sebring, P. A., & Camburn, E. M. (1992, April). How teachers are engaging reform in Chicago: Differences among schools. Paper presented at the meeting of the American Educational Research Association, San Francisco, CA.

Tyack, D. B., 8c Cuban, L. (1995). Tinkering toward Utopia: A century of public school reform. Cambridge, MA: Harvard University Press.

Wagner, T. (1994). How schools change: Lessons from three communities. Boston: Beacon Press.

Wilson, J. Q. (1989). Bureaucracy: What government agencies do and why they do it. New York: Basic Books.

Witte, J. F. (2000). The market approach to education. Princeton, NJ: Princeton University Press.

FREDERICK HESS is a resident scholar at the American Enterprise Institute. His books include Revolution at the Margins: The Impact of Competition on Urban School Systems (Brookings, 2002) and Spinning Wheels: The Politics of Urban School Reform (Brookings, 1999). His research interests include urban education,' school choice, education politics, and educational governance.

ROBERT MARANTO teachers political science at Villanova University and with others has produced numerous works including School Choice in the Real World: Lessons from Arizona Charter Schools (Westview, 2001) and Radical Reform of the Civil Service (Lexington, 2001). He can be reached at robert.maranto@villanova.edu.

SCOTT MILLIMAN is an associate professor of economics at James Madison University and coeditor of School Choice in the Real World: Lessons from Arizona Charter Schools. His education research has appeared in journals such as Political Science Quarterly, Policy Studies Journal, and Phi Delta Kappan.

KATHLEEN GRAMMATICO FERRAIOLO is a PhD candidate in politics at the University of Virginia. Her research interests include public opinion, political institutions, and public policy.

Endnotes

ⁱ Still, numerous anecdotal accounts exist concerning teacher responsiveness to reform (see Elmore, 1996, for a good overview of this literature).

ⁱⁱ School choice programs generally embrace the notion that parents and students should be given the freedom to select the schools that best satisfy their educational priorities. Charter schools, a form of public school choice, are characterized by an autonomy-accountability tradeoff: In exchange for exemption from most state education regulations, they must demonstrate satisfactory student achievement. Charter schools are funded on a per pupil basis and are free from many of the laws and regulations governing public schools. As of the start of the 2000-2001 school year, more than 1,700 charter schools operated in more than 35 states. See Finn, Manno, and Vanourek (2000), Hassel (1999), or Nathan (1996) for excellent overviews of charter schooling.

ⁱⁱⁱ Just as the views of union members may not perfectly mirror the preferences of union leaders, Republican and Democratic Party identifiers may not espouse all of the issue positions of their party platforms and leaders. The important point is that union members, who are more susceptible to influence by union leaders' views through paying dues, attending meetings, and receiving literature, are more likely than nonmembers to adopt those views and come to oppose school choice. And Republicans, who tend to support free markets .and competition, are more likely to support school choice than Democrats, who tend to prefer government involvement in and regulation of markets.

^iv We tested the relationship between union membership and partisan affiliation because of the strong connection between the national teachers' unions and the Democratic Party. The correlation was significant (.11), but substantively not very important, and a regression of the two variables yielded a very low adjusted R-square value of .01.

^v Although the operationalization of the variable is imperfect, we are still interested in testing the relationship between familiarity with charter schooling and support for school choice generally. However, we also test identical models that do not include the variable and evaluate any differences that exist between the two sets of results.

^vi Some might argue that teachers' responses to the questions about school choice might be influenced not only by their profession but also by their political beliefs or parental status. We have controlled for partisan affiliation by including it as one of the explanatory variables, and there is no reason to suspect that parental status is causally prior to parental status or that it makes respondents any more or less likely to support school choice. Although we acknowledge that including parental status as a variable might strengthen the explanatory power of our models, unless such a variable is causally prior, it will not introduce bias into the results (King, Keohane, & Verba, 1994).

^vii We briefly discuss the different choice opinions of African American and Hispanic teachers in footnote 12, but because of the extremely low sample size of each group (8 Black teachers and 66 Hispanics) we do not specify separate models for the two groups.

^viii In previous work with the data we use for this study, researchers created a school culture index, which included this variable and two others: whether a respondent reported feeling treated as a valuable employee and whether he or she usually looked forward to each working day (Hess et al., 1999). We included only the first variable after determining that the other two were measures less of school culture than of job satisfaction; the correlation between them, .38, was more than 50% greater than the correlation between either one and the first, more direct measure of school culture. When combined into a composite index, the job satisfaction variables had no effect on teachers' school choice attitudes, and when tested separately, only the variable that measured whether a teacher looked forward to each working day was significant. Because we were not confident that this variable alone was a valid measure of job satisfaction, we included only the first measure of school culture.

^ix We also tested models predicting teachers' support for the other five education reforms mentioned in the Southwestern Teachers' Survey using the same explanatory variables as in the school choice models (but not excluding the no opinion category, for the sake of simplicity and because in the choice models its exclusion did not change the substantive results). The results showed that union membership was significant in predicting attitudes toward two of the five reforms, length of service was significant in three cases, and school culture and partisan affiliation were each significant in two cases. Although certain personal characteristics may affect teachers' views about education reform generally, other variables, such as familiarity with charter schooling and teacher race, were only significant predictors of school choice attitudes.

^x The baseline teacher is White, a political independent, has 19.93 years of experience, and rates his or her school culture as moderately positive (4.9 out of 6). The teacher is a union member, does not know someone who teaches at a charter school, has not considered teaching at a charter school, and works in a district where the majority of students are from low-income families. By setting up this baseline respondent, and then varying the levels of variables of interest, we can determine the marginal effect of any explanatory variable on the probability of a teacher being in each category of school choice opinion.

^xi In addition, the variable that measures school culture was only slightly above the .10 cutoff for significance; the results suggest that teachers working in a negative school environment are slightly more supportive of voucher programs.

^xii The non-White variable achieves statistical significance in the school voucher model but not in the charter-schooling model. Testing Black and Hispanic respondents separately revealed that Hispanic teachers are significantly more supportive of vouchers than non-Hispanic Whites, but that African Americans do not differ significantly from Whites in their support for either form of school choice. Because of the very small number of Hispanic and, in particular, African American teachers sampled, however, we are not confident about the generalizability of these findings.

^xiii Still, union members are significantly more opposed to school vouchers than charter schools: The predicted probability of being in the antivoucher versus the anticharter categories is 4% higher for union teachers.

^xiv Concerns about possible endogeneity were noted previously, in footnote 5. The results from models that excluded this variable suggest that the effects of other variables are not being confounded; we discuss these models further in a later note. However, including this variable does not confound the results of other variables and cannot tell us the direction of the relationship between considering a charter school job and attitudes toward school choice.

^xv In models run with other commonly discussed education reforms as dependent variables, considering a job at a charter school was significant in only one caseteachers' standing on parental choice with regard to magnet schools. This finding suggests that familiarity with charter schooling is a particularly important predictor of teachers' attitudes toward many forms of school choicecharter schools, voucher programs, and magnet schoolsrather than of school reform generally.

^xvi Given the endogeneity concerns raised by the variable that measures teachers' knowledge of charter schooling, we tested identical models that excluded it. Overall, the results from these models differed in some slight but significant ways. In particular, in predicting charter attitudes, partisan affiliation was significant in the reduced model, while in predicting voucher attitudes, teachers' knowledge of someone who works at a charter school was significant. Other than these differences, the coefficients for the other explanatory variables and their significance levels changed only slightly. The major difference in the models that excluded the familiarity variable was the increase in the log likelihood function and corresponding decrease in the value of the model chi square and McFadden's R-square. Additional tests (obtained from modules for Stata created by Long & Freese, 2001) were run to determine the overall importance of each explanatory variable, and these results revealed that the two dichotomous variables that measure familiarity with charter schooling are the most important variables in the charter schooling model. In the voucher model, whether a teacher has considered a charter schooling job, along with union and partisan affiliation, have the most significant effect on teachers' attitudes toward private school choice. Still, the chi-square value remains significant in the reduced models, which indicates that although the variable that measures teachers' knowledge of charter schooling significantly increases the model chi square and McFadden's R-square values, the remaining coefficients offer a statistically significant (and substantively important) improvement over an intercept-only model.

Cite This Article as: Teachers College Record Volume 104 Number 8, 2002, p. 1568-1590 https://www.tcrecord.org ID Number: 11332, Date Accessed: 1/19/2022 6:54:14 AM Purchase Reprint Rights for this article or review