Abstract
More than 50 years after Brown v. Board, African American students continue to trail their White peers on a variety of important educational indicators. In this article, we investigate the political foundations of the racial “achievement gap” in American education. Using variation in high school graduation rates across the states, we first assess whether state policymakers are attentive to the educational needs of struggling African American students. We find evidence that state policymaking attention to teacher quality—an issue education research shows is essential to improving schooling outcomes for racial minority students—is highly responsive to low graduation rates among White students, but bears no relationship to low graduation rates among African American students. We then probe a possible mechanism behind this unequal responsiveness by examining the factors that motivate White public opinion about education reform and find racial influences there as well. Taken together, we uncover evidence that the persisting achievement gap between White and African American students has distinctively political foundations.
Public education occupies a unique place in the hierarchy of American social policymaking. Beyond comprising the nation’s largest social welfare program devoted to promoting equal opportunity through social mobility, surveys consistently show that public education is one of the few American social programs that commands broad public support (Hochschild, 2004). However, pitted against the democratic ideal of universal public schooling operating as American society’s “great equalizer” 1 is a large body of evidence documenting significant disparities in the high school graduation rates and standardized test scores of African American and White students (Coleman, 1966; Heckman & LaFontaine, 2010; Jencks & Phillips, 1998; Neal, 2006; Rivkin, 1995; Thernstrom & Thernstrom, 2003). 2 These persisting differences in educational outcomes between African American students and their White peers are important because they translate directly into social inequalities later in life including future earnings, employment status, and incarceration rates (Fryer, 2013; Heckman & Masterov, 2007; Lochner & Moretti, 2004). 3
The education achievement gap between Whites and African Americans likely has consequences for political equality as well. American public education has long been responsible for training future citizens (Campbell, 2006; Gutmann, 1999) and, whatever the precise mechanism(s) involved, level of educational attainment stands apart as the most robust predictor for whether an individual actively participates in politics (Nie, Junn, & Stehlik-Barry, 1996; Sondheimer & Green, 2010; Verba, Schlozman, & Brady, 1995; but see Kam & Palmer, 2008, 2011). In light of the special role public education plays in equipping citizens with the tools and motivation necessary to participate in American democracy, educational inequalities between White and African American students may serve to perpetuate and even exacerbate existing inequalities in rates of political participation for future generations (Verba, Burns, & Schlozman, 2003). 4
In response to the persisting education achievement gap, policy entrepreneurs have focused recent education reform efforts around improving the quality and effectiveness of America’s teaching workforce, with particular emphasis on identifying policies that increase the supply of effective teachers in underserved communities (B. Berry, 1988; Goldhaber & Hannaway, 2009). 5 This increased attention to reforming the policies that govern the American teaching profession is driven, in part, by the fact that in the 40 years since James Coleman (1966) penned his widely influential Equality of Educational Opportunity report, the most consistent finding in education research is that teacher quality/effectiveness stands as the most significant school level variable influencing student achievement (Chetty, Friedman, & Rockoff, 2011; Goldhaber, 2002; Rivkin, Hanushek, & Kain, 2005; Sanders & Rivers, 1996). 6
Although research shows that sustained access to effective teachers can help narrow the achievement gap (Hanushek & Rivkin, 2004), racial minority students are far less likely to be assigned to and taught by highly effective teachers (Clotfelter, Ladd, & Vigdor, 2005; Feng, 2010; Lankford, Loeb, & Wyckoff, 2002; Peske & Haycock, 2006; Scafidi, Sjoquist, & Stinebrickner, 2008). Equally noteworthy is the fact that disparities in who has access to effective teachers is partly a function of politics—namely, the policy choices states and school districts use to govern the teaching profession (Corcoran & Evans, 2008; Grissom, Loeb, & Nakashima, 2012; Hanushek, Kain, & Rivkin, 2004). 7 Since state policymakers maintain significant discretion over many teacher workforce policies (most notably educator pay and evaluation systems) that influence students’ access to highly effective teachers (Cohen & Walsh, 2010; Manna & Tydgat, 2008; Wong, 2008), the persisting achievement gap between African American and White students provokes important questions about political representation and racial inequality. Yet, apart from a few notable exceptions (Meier & England, 1984; Meier, Stewart, & England, 1989), existing research on education policymaking tends to overlook these fundamental questions about fairness and political equality.
To better understand the linkage between political inequality and educational opportunity in the United States, we begin by developing a set of theoretical expectations to explain why so few states adopt the policy reforms that education research suggests could help narrow the disparity in educational outcomes between White and African American students. Specifically, we hypothesize that racial inequality in state education policymaking can be explained, in part, by considering the “demand” for education reform among the public. To determine the viability of this explanation for the persistence of the education achievement gap between White and African American students, we first examine if state policymakers are equally responsive to the demonstrated educational needs of White and African American students when they enact education policy reforms aimed at improving teacher quality.
We find that teacher quality reform policymaking is responsive to poor education outcomes among White students but not African American students. We then investigate one possible mechanism behind this relationship by examining the factors that motivate White public opinion about (and demand for) education reform and find distinctively racial influences there as well. Specifically, we show that White opinions in support of education reforms are significantly tied to the performance of White but not African American students. White citizens are also less likely to identify the existence of a racial achievement gap in American education and less likely to prioritize it as a policy concern compared to African American citizens. In sum, we uncover evidence that educational inequalities between African American and White students are sustained, in part, by systematic political inequalities.
The Politics of the Education Achievement Gap
We investigate the existence of a racial undercurrent in the politics of education policymaking. Historically, elementary and secondary education policymaking, particularly at the subnational level, has been embedded within a larger struggle over racial and ethnic politics (Henig, Hula, Orr, & Pedescleaux, 1999; Hochschild & Shen, in press; Meier et al., 1989). From Southern governors who wed their political fortune to maintaining segregated schools to more contemporary controversies over busing and affirmative action, elementary and secondary education policymaking has been created and recreated within the contours of a racially polarized society.
Even in the aftermath of government mandated desegregation of public education, scholars uncovered systematic patterns of “second-generation discrimination” in which public schools suspend racial minorities and track them out of college preparatory classes at rates much higher than Whites (Meier et al., 1989). Political scientists have offered additional evidence of a link between these sorts of exclusionary school practices and more general biases in the American political system. For example, the use of “at-large” electoral systems by many school districts across the United States serves to minimize descriptive representation on school boards, which in turn is associated with racial minority students being subjected to heightened levels of second-generation discrimination (Leal, Martinez-Ebers, & Meier, 2004; Meier & England, 1984; Meier, Fraga, & England, 1986; Meier et al., 1989).
After a period of dramatic improvement in African American student outcomes relative to Whites (whether measured by student performance on the National Assessment of Education Progress or by attainment rates of high school and postsecondary degrees), since the early 1990s the size of the achievement gap between African American and White students has stagnated or even slightly increased (Neal, 2006; Thernstrom & Thernstorm, 2003). Although research in psychology demonstrates that there exist no quantifiable cognitive differences between White and African American infants (Fryer & Levitt, 2013), only 2 years into schooling African American children already significantly trail their White peers on age appropriate academic measures (Fryer, 2013; Fryer & Levitt, 2004). Moreover, on every subject at every grade level, there are large achievement gaps between African Americans and Whites that continue to grow as children progress through school—gaps that remain large and statistically significant even after controlling for socioeconomic and family background factors (Jencks & Phillips, 1998).
Partly on the basis of these stubborn achievement gaps, the 2002 No Child Left Behind Act took the unprecedented step of requiring states to document and publicly report student test score data by racial and ethnic subgroups. In fact, the law went so far as to list among its chief goals holding states accountable for “closing the achievement gap between high- and low-performing children, especially the achievement gap between minority and nonminority students.” 8 More recently, President Obama’s signature education policy initiative—the “Race to the Top” (RTTT) competitive grant program—made decreasing achievement gaps between racial subgroups a key component of its scoring rubric for awarding federal money to states willing to make significant reforms, particularly reforms directed toward increasing racial minority students’ access to highly effective teachers (U.S. Department of Education, 2009). 9
To date, the majority of scholarly research examining public policy and the education achievement gap has focused on identifying the effects, rather than causes, of state education policies (Harris & Herrington, 2006; Lee, 2002). We take a different approach by reversing the causal arrow and examining whether state education policymaking (i.e., the propensity of policymakers to undertake education reforms) can be explained by differential responsiveness to educational outcomes among White and African American students within a state and by investigating whether White public opinion about education reform is the mechanism linking educational outcomes to policymaking.
Theoretical Expectations
What aspects of student performance and school quality, if any, stimulate the demand necessary for state policymakers to enact politically controversial education policy reforms? Given widespread variation in student and school outcomes both within and across states as well as gaps between subgroups of states’ student populations, elected officials have multiple constituent concerns competing for their limited attention. Beyond the fact that these multiple competing policy problems limit the attention of politicians (Jones & Baumgartner, 2005), education reformers must also face off with the well-known status quo bias in the American political system—a bias that makes blocking newly proposed measures relatively easy for organized interests who favor the status quo (Moe, 2006). In the specific domain of state education policymaking, organized teacher union interest groups have been particularly effective in opposing policy proposals that would differentiate teacher pay and link tenure and evaluation decisions to student learning outcomes (Hartney & Flavin, 2011; Moe, 2011). In short, policy entrepreneurs who wish to convince elected officials to enact policy reforms that would link teacher pay and evaluations to student outcomes face a politically contentious environment in which organized interests can be expected to push back with significant zeal. Consequently, we hypothesize that these teacher reform proposals will only gain traction when state policymakers sense an acute demand among a significant majority of the public for a departure from the status quo.
We further theorize that demand for education reform among the public is largely shaped by the public’s perception of the quality and overall health of the public school system in their state and/or local community. 10 When citizens perceive that student performance is lagging and that the quality of their community schools are poor, only then are they likely to agitate enough to overcome policymakers’ hesitations to pursue controversial education reform policies. This raises an interesting question: Under what circumstances will the public’s unease over the quality of their schools reach a breaking point? More specifically, will academic deficiencies among African American students register in the public’s consciousness enough to catalyze citizen demand for education reform? Our expectation is that the vast majority of the public will rarely become alarmed over the state of K-12 education, and almost never alarmed over the education outcomes of African American students, for several reasons.
First, the organization of public education in the United States is designed in such a way that a broad constituency for school reform is unlikely to materialize (Warren, 2011). Specifically, the traditional financing system in American K-12 education is heavily reliant on the property tax, which ensures that tastes and preferences for public school spending and school performance remains commensurate with the majority of homeowners in a given school district (Fischel, 2001, 2009; Tiebout, 1956). The implications of this system for the demand side of education reform are obvious—middle and upper income Americans can avoid feeling the pain of poorly performing urban school systems by retreating to suburban enclaves where property values are high and the presence of a large contingent of poorly performing racial minority students is nonexistent. Therefore, the likelihood that a large majority of White citizens will clamor for education reforms to address the needs of struggling racial minority students is slim (Kozol, 1991; Roza, 2010; Warren, 2011).
Second, the performance of African American students is unlikely to generate demand for school reform among a majority of the White public because some teacher reform proposals are likely understood by White citizens as a zero-sum game in which their own school districts have nothing to gain and everything to lose. For example, proposals to pay teachers more to teach in low-performing urban schools are easily framed as redistributive in nature since they effectively use government subsidies to draw high-quality teachers out of suburban schools and provide incentives for them to instead work in urban schools. In short, the current status quo approach to teacher compensation (a “single salary” schedule that rewards experience and credentials) biases teacher labor markets in favor of advantaged schools and against low-performing urban ones where turnover is higher on account of more challenging working conditions (Hanushek et al., 2004). Consequently, proposals to differentiate teacher pay (e.g., additional compensation for working in “hard-to-staff” schools) are unlikely to solicit widespread support from White suburban parents who may view their children as the primary losers as a result of these policy changes.
Third, and perhaps most importantly, a growing literature reveals that African Americans generally exert less political influence than Whites in terms of having their opinions reflected in elected officials’ in-office behavior and policy decisions (Griffin & Flavin, 2007; Griffin & Newman, 2008; Hajnal, 2010). At election time, state legislators are faced with the often competing demands of White and African American constituents, especially on questions relating to education policy. To maximize voter support, we can expect state legislators to prioritize the needs and opinions of White constituents except in instances where African Americans make up a majority. While African Americans do make up a majority in many state legislative districts across the states, African Americans do not constitute a majority (or even a plurality) in any single state. As a result, we expect the needs and opinions of White constituents to win out over the needs and opinions of African American constituents when the two come into conflict in the state education policymaking process.
In sum, we have little theoretical reason to expect that either White parents or the state legislators who represent them will demand significant education reforms in the face of poor performance among racial minority students. However, such demand may materialize when White students are faring poorly. In other words, policymakers might be differentially responsive to the educational needs of White versus African American students. To empirically evaluate this claim, we first turn to state-level data on student educational outcomes and education reform policymaking.
Black–White Student Educational Outcomes and Teacher Quality Policy Reforms
We begin by examining whether state education reform policymaking responds differentially to the education outcomes of White and African American students. For both substantive and methodological reasons, we focus on education policy reforms aimed at increasing students’ access to highly effective teachers. Since this study is focused on state efforts to narrow gaps in student learning between White and African American students, focusing on teacher policy reform is especially appropriate given the educational research literature’s link between teacher quality and student outcomes. Moreover, unlike other salient state-level education reforms (e.g., school finance reform, vouchers, or charter schools), efforts to overhaul teacher pay and evaluation policies are among the most recent temporally, which is important from a methodological standpoint given our desire to understand how student outcomes influence policymaking (rather than how changes in policy influence students’ educational outcomes). In other words, because we assess teacher reform policies that were generally debated and enacted after our measures of student performance, we can have greater confidence that policymaking is responding to deficiencies in student outcomes rather than the other way around.
To measure educational outcomes for White and African American students separately by subgroup, we use each state’s high school graduation rate for the class of 2005 disaggregated by race. 11 Our graduation rate measure is taken from the U.S. Department of Education’s Averaged Freshman Graduation Rates for Public School Students (AFGR). 12 In addition to our main explanatory variable of interest—White and African American education outcomes as measured by high school graduation rates—we consider and control for alternative factors that may influence a state’s propensity to pursue teacher policy reforms. Specifically, we include a measure of the general ideological liberalism of state government (W. D. Berry, Ringquist, Fording, & Hanson, 1998) with the expectation that more liberal states will be less likely to enact reforms given the general propensity of political liberals to rely on teacher union interest groups for electoral support. We include a (logged) measure of a state’s public school student population and a measure of the percentage of education financing that comes from the state (relative to local or the federal government) with the expectation that larger states and states that provide a greater percentage of school dollars will be more empowered and thus more likely to enact state-level education reforms (Shen & Wong, 2006). We also include a measure of teacher union political campaign contributions and the strength of a state’s teacher collective bargaining law with the expectation that states with politically and organizationally strong teacher unions will be less likely to enact reforms (Hartney & Flavin, 2011). Finally, we include a measure from Education Week’s annual Quality Counts report grading states on their efforts to improve teacher quality for the year 2005 as a way to control for a state’s past policy climate on teacher reform issues (Editorial Projects in Education, 2005). By including this control variable, we are more confident that our dependent variable in the present is not simply a function of policy reforms made previously or a function of policymakers’ efforts to address past deficiencies in the overall qualifications of the state’s teaching workforce. 13 Because the dependent variable is a count (0-12) of policies enacted, we use a negative binominal regression estimator.
Table 1 reports the results of regressing the count of teacher reform policies on White and African American high school graduation rates in 2005 and the set of control variables explained above. 14 As expected, states with larger student enrollments are more likely to enact teacher quality reforms while states with more politically active teacher unions are less likely to enact these types of reforms. Moving to our primary independent variables of interest, the coefficient for White high school graduation rate is negative and statistically different from zero, indicating that across the states, as White graduation rates decline states are more likely to enact reforms aimed at improving the quality and effectiveness of their teaching workforce. Substantively, moving from a state at the 90th percentile for White student graduation rate to a state at the 10th percentile predicts an additional two policy reforms enacted on the 12 item policy scale. 15 In contrast, the coefficient for African American graduation rate is not statistically different from zero, providing evidence that state policymaking appears unrelated to the performance of African American students. 16 These results suggest that, across the states, policymakers are unequally responsive to the demonstrated educational needs of White and African American students. 17
Teacher Quality Policy Reforms Linked to White But Not African American Student Graduation Rates.
Note: Dependent variable: Count (0-12) of state teacher quality policy reforms in 2009.
Cell entries are negative binomial regression coefficients with standard errors in brackets.
p < .10. **p < .05. ***p < .01 using a two-tailed test.
What Motivates White Citizens’ Opinions on Education?
One possible mechanism that helps explain the unequal policy responsiveness documented above is that, consistent with the findings of several previous studies (Griffin & Flavin, 2007; Griffin & Newman, 2008; Hajnal, 2010), policymakers are more likely to respond to the political preferences of White citizens compared to African Americans. To investigate whether this mechanism is at play in education policymaking, we next examine whether White citizens’ opinions on education are sensitive to student outcomes using two sets of nationally representative public opinion data. First, we use the 2004 and 2008 waves of the National Annenberg Election Survey (NAES) to investigate whether student performance indicators predict if White respondents identify “education” when asked: “In your opinion, what is the most important problem facing our country today?” Specifically, we model responses to the most important problem item 18 (coded as “1” if a respondent identifies education as the most important problem and “0” if he/she identifies anything else as most important such as “crime,” “unemployment,” “the environment,” etc.) as a function of the high school graduation rate 19 for Whites and African Americans in their state of residence the year before the survey (2003 or 2007), their state’s unemployment rate the year before the survey, whether there are children under the age of 18 (i.e., “school-aged”) in a respondent’s household, and the respondent’s gender, age, income, level of education, partisanship (Democrat coded higher), political ideology (liberal coded higher), and an indicator for year of survey. If student outcomes have an effect on citizen opinion about the importance of education issues, the coefficient for both graduation rates will be negative such that, across the state-years, as graduation rates get lower, respondents will be more likely to identify education as the most important problem. We also expect that the coefficient for state unemployment rate will be negative such that as a state’s unemployment rate goes up, respondents will be more likely to focus on jobs and the economy and less likely to identify education as the most important problem. 20
We report the results of this probit estimation in Column 1 of Table 2 and find that White citizens’ propensity to identify education as the most important problem increases as White graduation rates decline but bears no statistical relationship with graduation rates for African American students. Combining the results from Table 1 and 2, the graduation rates (i.e., for Whites but not for African Americans) that predict policy change similarly predict White public opinion, making White opinion a likely link between educational outcomes and subsequent policies. Substantively, we find that moving from one standard deviation above the mean for White graduation rate to one standard deviation below the mean across the state-years leads to nearly a 1 percentage point increase in the probability of identifying education as the most important problem which, by comparison, is larger than the difference between a respondent living in a household without versus with school-aged children. Looking at the other coefficients in the model, we also find that women, Democrats, and ideological liberals are more likely to identify education as the nation’s most important problem.
White Respondents Identify Education as the Most Important Problem/Schools in Need of Major Change in Response to White But Not African American Graduation Rates.
Note: Dependent variable listed above each column.
Cell entries are probit/ordered probit coefficients with standard errors clustered by state-year/state in brackets. Dummy variable for survey wave included in Column 1 but not reported.
p < .10. **p < .05. ***p < .01 using a two-tailed test.
Second, we use data from a 1996 survey on education reform carried out by Moe (2001) 21 to investigate if student achievement indicators predict White responses to the question: “Which of the following best describes how you feel about the schools in your state as a whole?” The response categories are: (a) They are doing well, (b) They need minor changes, or (c) They need major changes. This question nicely taps citizen demand for school reforms. We model White citizens’ responses to this survey question as a function of White and African American graduation rates in the respondent’s state for 1996 and the same set of control variables used in the analysis of the NAES data above. 22 Similar to above, if student achievement has its expected effect on public opinion, the coefficient for both graduation rates will be negative such that, across the states, as graduation rates get lower, respondents will be more likely to identify education as needing major changes.
The results of this ordered probit estimation are reported in Column 2 of Table 2. We again find that White citizens are more concerned about education (i.e., they are more likely to report that changes are needed to schools) as White students’ graduation rates decline. Substantively, moving from one standard deviation above the mean for White graduation rate to one standard deviation below the mean across the states leads to an 11 percentage point increase in the probability of a respondent favoring “major changes” for their state’s public education system. In contrast, no such relationship exists for African American graduation rates. Looking at the other coefficients, females, more affluent respondents, and political conservatives are more likely to express demand for changes in schools. More importantly, however, is the fact the NAES and this analysis from Moe’s (2001) survey data nearly a decade earlier point to the same conclusion: White respondents sound the alarm about school quality when White students are performing poorly, but not when African American students are performing poorly. 23
One concern with examining the relationship between student achievement and public opinion using state-level measures of student educational outcomes is that citizens are better positioned to pay close attention to and be knowledgeable about the performance of students and the quality of schools in their local community (Henderson, Chingos, & West, 2010). Schools are, after all, traditionally a local matter in the United States (Berkman & Plutzer, 2005). Therefore, we next consider how White and African American student achievement at the local school district level influences how White citizens evaluate the quality of their local schools.
To better understand these microfoundations of White public opinion about education reform policymaking, we draw on a unique set of surveys commissioned by the Center for Evaluation and Education Policy (CEEP) at the University of Indiana-Bloomington. CEEP conducted a total of six “Public Opinion Surveys on Indiana Education” 24 between 2003 and 2008 querying Indiana residents’ perceptions about the quality of their local public schools. Fortunately, many of the items were asked identically throughout all six waves enabling us to pool the data and generate a reasonably large sample of respondents located in 277 of Indiana’s 295 school corporations (districts). 25
In order to link citizens to their actual school district of residence, we matched respondents’ home telephone numbers to their residential addresses and by extension their school district. 26 Drawing on administrative data from the Indiana Department of Education we then gathered information on the percentage of White and African American students who scored proficient in 3rd, 6th, 8th, and 10th grade on the Indiana Statewide Testing for Educational Progress-Plus (ISTEP) Assessment 27 in mathematics and reading for each school district and linked these outcome measures to each one of our respondents based on their school district of residence. 28 It is important to note that these data were matched to respondents by school district and year so that, for example, the percentage of students proficient on ISTEP math in the Indianapolis Public School District (IPS) during 2006 was linked only to respondents in the sample who lived within IPS and were contacted as part of the 2006 survey wave. Thus, unique data were obtained for each district and for each year for all variables to ensure use of the most accurate contextual measures of respondents’ districts.
We model respondents’ evaluations of their community schools as a function of White and African American student achievement (measured as the average math and reading ISTEP passage rates for 3rd, 6th, 8th, and 10th grades) in the respondent’s school district of residence, whether a respondent is retired or not, whether a respondent reported voting in the most recent election, whether a respondent has school-aged children living at home, a respondent’s age, level of education, income, and gender, the percentage of students in the respondent’s school corporation who are African American, and dummy variables for the year of the survey (leaving one as a reference category). Because we theorize that state policymakers responsible for enacting education reform policies tend to be responsive to White but not African American opinion (Griffin & Flavin, 2007, Griffin & Newman, 2008; Hajnal, 2010), we again limit our analysis to Whites only. Doing so allows us to examine what factors motivate the opinions of the racial group that state policymakers are most attentive and responsive to.
In Column 1 of Table 3 we report the results of regressing evaluations of local community schools on the set of covariates explained above. The dependent variable is coded such that a higher number (1-4) indicates greater approval. 29 We find that White citizens’ evaluations of their community school are strongly linked to White student performance in their school district of residence such that higher White student performance predicts more positive evaluations. In contrast, a White citizen’s evaluation of the local public schools is not linked to African American student performance. 30 Put another way, White Indianans rate the quality of their schools in relation to how well or poorly White students are performing, but pay comparatively little attention to the performance of African American students.
White Citizens’ Evaluations of Their Community Schools Linked to White But Not African American Student Achievement.
Note: Dependent variable: Evaluation of community schools (1-4, more favorable evaluation coded higher).
Cell entries are ordered probit coefficients with standard errors clustered by school corporation in brackets. Dummy variables for year of survey included in both models but not reported.
p < .10. **p < .05. ***p < .01 using a two-tailed test.
To check if White citizens incorporate African American student achievement into their evaluations of community schools as the percentage of African American students grows, we create an interaction term between African American test scores and the percentage of school corporation residents who are African American and include it as a covariate in the same model specification. The interaction term is not statistically different from zero (see Column 2 and Table 3), indicating that a greater percentage of African American students in a school corporation does not make White citizens more likely to link African American student achievement with evaluations of community schools. Consistent with the analysis of national survey data reported earlier in this section, the achievement of African American students seems to have little influence on the education opinions of White citizens which, in turn, influence educational policymaking decisions.
Do Whites and African Americans Have Different Opinions About the Achievement Gap?
Given the results uncovered in the previous section, what explains the fact that White but not African American student performance influences White citizens’ opinions about education? Are White citizens unaware of a racial achievement gap? Do they simply not view it as a major policy concern? We probe these questions using data from Phi Delta Kappa/Gallup’s Annual Survey of the Public’s Attitudes Toward the Public Schools. Controlling for an array of demographic characteristics (partisanship, level of education, income, gender, age, and whether a respondent has school-aged children or not), we examine whether Whites have different opinions on educational inequality (and its possible remedies) compared to African Americans. 31 In Table 4, we report the results of five estimations that investigate differences in responses between Whites and African Americans to the following survey items:
Is the academic achievement of White students, nationally, higher (3), lower (1) or about the same (2) as that of Black and other minority students?
In your opinion, do Black children and other minority children in your community have the same educational opportunities as White children? (1 = Yes, 0 = No)
In your opinion, how important do you think it is to close the academic achievement gap between White students and Black and Hispanic students—very important (4), somewhat important (3), not too important (2), or not important at all (1)?
In your opinion, is the achievement gap between White students and Black and Hispanic students mostly related to the quality of schooling received, or mostly related to other factors? (Schools = 1, Other factors = 0)
In your opinion, is it the responsibility of the government to close the achievement gap between White students and Black and Hispanic students, or not? (1 = Yes, 0 = No)
Whites and African Americans Have Different Opinions About the Existence and Importance of the Education Achievement Gap.
Note: Dependent variable listed above each column.
Cell entries are probit/ordered probit coefficients with standard errors in brackets.
p < .10. **p < .05. ***p < .01 using a two-tailed test.
The first two survey items (Columns 1 and 2) reveal that Whites are significantly less likely to identify the existence of a racial achievement gap or perceive inequality in educational opportunity within the public school system. Controlling for all other demographic variables, White respondents are 23 percentage points less likely than African Americans to think that, nationally, White student achievement exceeds African American student achievement. White respondents are also 40 percentage points more likely than African Americans to think racial minorities have the same educational opportunities as Whites.
The final three columns of Table 4 reveal that, compared to African Americans, Whites also view the education achievement gap as less of a policy concern. White respondents are 12 percentage points less likely than African Americans to think that the achievement gap is “very important” and 21 percentage points more likely than African Americans to report that the racial achievement gap in student learning is due to other factors outside of schools. Moreover, Whites are 14 percentage points less likely than African Americans to think that the government has a responsibility to close the education achievement gap. Taken together, these results indicate that, compared to African Americans, White citizens are less likely to identify that the education achievement gap exists, less likely to think it an urgent or fixable problem, and less likely to think a government response or solution is needed. Consistent with the analyses reported throughout the article, there seem to be distinctively racial considerations that drive White public opinion about educational opportunity in the United States.
Conclusion
As Hochschild (2004) argues, schools occupy a “unique place . . . in American political and social culture.” (p. 225). Education is one of the few policy areas where Americans have long maintained a consensus in favor of universal public access and provision. As a result, schools consume considerable portions of local, state, and even federal budgets, and are among the most significant sources of public employment. Perhaps most importantly, public schools provide not only crucial training in the skills and knowledge required for work and life, but also for effective future citizenship (Campbell, 2006; Gutmann, 1999). Because education plays such a crucial role in the cultivation of civic and political engagement (and in shrinking the gap in rates of political participation between Whites and African Americans), the results we uncover in this article are especially concerning from a political equality standpoint.
Building on previous research that has uncovered significant representational inequalities along racial lines (Griffin & Flavin, 2007; Griffin & Newman, 2008; Hajnal, 2010), we hypothesize that educational inequalities between African American and White students are sustained, in part, by systematic political inequalities. We first show that across the states, teacher reform policymaking is responsive to poor educational outcomes among White students but not African American students. We then probe a possible mechanism behind this relationship by examining the factors that lead White citizens to advocate for greater education reform and find distinctively racial influences there as well. That is, for White citizens, education issues become salient and school reform a major policy concern when White students perform poorly; however, such concerns are divorced from the success or failure of African American students. Moreover, White citizens are less likely to identify the existence of a racial achievement gap and less likely to think educational opportunities are unequal. This may explain why, even after controlling for an array of other factors, Whites are far less likely to prioritize the achievement gap as a policy concern or believe that government has a responsibility to help narrow the gap. Whether analyzed at the policymaking level or the level of individual citizens’ political attitudes, White students receive far more attention and subsequent response compared to African American students. Instead of promoting equality of opportunity, America’s system of K-12 education—with its heavy reliance on state and local control—may instead serve to further exacerbate existing political inequalities between Whites and African Americans.
Importantly, our analysis is confined to examining White citizens’ demand for school reform and state policymakers’ responsiveness without considering how the federal government may strive to attenuate some of these inequalities. Although the story we uncover focuses on the constraints of federalism—namely that the American tradition of ceding education policymaking authority to subnational governments tends to diminish the prospects of a broad constituency for school reform (Warren, 2011)—the federal government has, in recent years, played an increasingly important role in tackling issues of educational inequality (Manna, 2006). As Peterson’s (1981) work on federalism argues, redistributive policymaking will typically remain the domain of national, rather than subnational, governments. Therefore, it is not surprising that the most recent and widespread efforts to inject attention to educational inequality into K-12 school reform have come from policymakers in Washington, DC. Indeed, President George W. Bush’s No Child Left Behind Act and President Obama’s Race to the Top initiative have leveraged variants of fiscal federalism as a strategy to compel (in the former case) or entice (in the latter) states to enact a variety of reform policies aimed at reducing the education achievement gap—reforms which we have shown have little promise of receiving attention on the sole basis of a state’s internal political environment. Although we leave it to other scholars to evaluate whether these two federal education policies have worked as policymakers’ intended, we simply note that they further reflect the unique “tangled web” (Epstein, 2004) of American education governance that can limit policy attention to struggling students who need it most.
Footnotes
Appendix
Data Sources for Table 1
| Dependent variable | Measurement/descriptive statistics | Data source |
|---|---|---|
| Teacher quality reform policies | Additive scale of 12 state-level policies governing teacher pay and evaluation. NCTQ’s database tracked whether state policy provided for the following reforms to teacher pay and evaluation systems: Allow for additional pay for teaching certain subjects “high-needs” subjects? Allow for additional pay for high performance? Allow for additional pay for teaching in a “high-needs” (e.g., high poverty) school? Provide credit on the salary schedule for private school teaching experience? Provide credit on the salary schedule for postsecondary teaching experience? Provide credit on the salary schedule for private sector nonteaching experience? Is teacher performance defined/measured as improved student achievement outcomes? Can student test scores be used as a component of tenured teachers’ evaluations? Can student test scores be used as a component of nontenured teachers’ evaluations? Does the state required use of a statewide teacher evaluation instrument? Is student performance (however defined) used in the evaluation of a tenured teacher? Is student performance (however defined) used in the evaluation of a nontenured teacher? |
NCTQ’s Teacher Rules, Roles, & Rights (TR3) Database (2009) |
| Mean = 3.27, SD = 3.11, Range: 0 to 11 | ||
|
|
||
| Independent variables | Measurement/descriptive statistics | Data source |
| White/African American student graduation rate | Rates taken from NCE’ Averaged Freshman Graduation Rate (0-100%) White: Mean = 80.09, SD = 7.25, Range: 62.7 to 91.8 African American: Mean = 65.30, SD = 13.77, Range: 19.3 to 96.2 |
National Center for Education Statistics (NCES) 2005-2006 |
| State government liberalism | Measure of state government ideological liberalism for 2006 (more liberal coded higher) Mean = 48.94, SD = 26.85, Range: 10.00 to 94.77 |
W. D. Berry, Ringquist, Fording, & Hanson (1998) |
| Size of student population (logged) | Logged value of state’s public elementary and secondary student enrollment Mean = 13.31, SD = 1.07, Range: 11.46 to 15.67 |
National Center for Education Statistics (NCES) Common Core Data (2006) |
| % of funding from state | Percent of total per-pupil spending that comes from state government Mean = 49.55, SD = 14.10, Range: 25.9 to 89.9 |
National Center for Education Statistics (NCES) Common Core Data (2006) |
| Teacher union campaign Contributions | Total percentage of political contributions made to candidates for state political office that were made by state teacher union/association affiliates (1998-2006) Mean = 1.07, SD = 0.95, Range: 0.05 to 3.37 |
Hartney and Flavin (2011) |
| Teacher collective bargaining laws | State law mandates teacher collective bargaining (= 3); permits bargaining (= 2); or prohibits bargaining (= 1) Mean = 2.57, SD = 0.67, Range: 1 to 3 |
NCTQ’s State Bargaining Rules Database (2009) |
| Previous teacher quality reforms | Overall measure of state policymaking efforts to improve teacher quality in 2005. Education Week assigned a grade (0-100) to each state based on its overall commitment to improving teacher quality as demonstrated by teacher reform policies and the actual qualifications of a state’s teacher workforce. Mean = 76.37, SD = 8.54, Range: 61 to 93 |
Education Week’s Quality Counts 2005 |
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
