Politics, Markets, and America's Schools.

• Author: Liebman, James S.

You can actively flee, then, and you can actively stay put.(1)

In John Chubb and Terry Moe's book,(2) choice is hot; voice is not.(3) As influential as their book has become in current policy debates,(4) however, its data and reasoning may support policies the reverse of those that the authors and their "New Paradigm" disciples propose.(5) In this review, voice is hot; choice is not.

Chubb and Moe conclude that dropping SAT scores and American students' poor performance on internationally standardized math and science exams demonstrate that America's public schools are in crisis.(6) Based on an empirical study of a large sample of students attending American public and private high schools in the early 1980's,(7) the authors give the following account of the causes of educational failure and success: The "inputs" on which educational policy debates typically have focused--per capita spending, class size, teacher experience, and the like--have almost no effect on educational quality (measured by achievement test scores) at the high school level.(8) Putting aside student ability and family resources, which exert the largest impact on performance but are more or less immutable,(9) the only policy-sensitive input that is associated with strong academic performance is a school's "effective organization"--its "clear goals, ... ambitious academic program, strong educational leadership, and high levels of teacher professionalism."(10) Schools permitting a lot of professional "autonomy," that is, ones with minimal "bureaucratic influence," are more likely to be effectively organized than schools that do not permit such autonomy.(11) Private and parochial schools exhibit a lot of autonomy; public schools do not.(12)

Translating correlation into cause, the authors restate their thesis in reverse: Private schools give teachers and principals a lot of autonomy, which in turn facilitates good school organization, which in turn causes high achievement. Conversely, bureaucratically administered public school discourage autonomy, which leads to poor organization, and low achievement.(13)

The authors conclude that politicians can solve the crisis in education by causing all schools to emulate private schools--that is, by substituting "market" control for "democratic" (by which the authors mean "interest group" and "bureaucratic") control of the schools.(14) Because market-driven excellence would come at too high a cost if it entirely sacrificed a public system's welfare and equity features, the authors would locate a set of new or overhauled schools in the public sector, subject to loose public regulation.(15) Within the public sector, the authors would replace existing systems of funding and student assignment with a system in which: * The state (alone) provides funds for participating schools by giving

them "scholarships," in amounts fixed by law, for each child who

enrolls in each school. * Parental and school choice determines the assignment of children to

schools: Parents may apply to any school in the state, and schools may

select applicants and retain or expel matriculated students on the basis

of any criteria the schools choose, exclusive of race but not, for

example, of intelligence, achievement, motivation, disciplinary record, and,

possibly, religion.(16)

In Part I, I argue that Chubb and Moe exaggerate the impact of school structure on student achievement. They do so by underestimating the impact of parent and student educational "tastes," mislabeling evidence of personal tastes as elements of school structure, and treating mundane structural effects of good or bad achievement as important policy-driven causes of that achievement. As a result, the authors fail to rule a plausible alternative explanation of their data: that the educationally oriented parents and children (call them "educational connoisseurs"(17)) who congregate in private schools demand and receive higher quality educational services than do consumers with less exacting educational tastes.

Part II concludes that enhancement of choice along the lines the authors propose would put even more children at educational risk than the existing system and thus would make an educationally (and equitably) bad situation worse.

Drawing on theories that Albert O. Hirschman developed twenty years ago,(18) Part III explains how an excess of the market mechanism of "choice" caused the crisis in public education and how enhancement of the democratic mechanism of "voice"--the conceptual opposite of choice--might make public schools better. Rejecting the authors' curious elision of "bureaucracy" and "democracy," and their juxtaposition of both to markets, Part III argues that bureaucracy actually stands between markets and democracy as a third form of organizational control, and that bureaucracy only came to dominate urban public schools in response to increasing parental exit from, and declining parental voice in the administration of, those schools. Accepting the authors' critique of bureaucratic control, and doubting the wisdom of market control, Part III sketches a proposal for enhanced parental voice in, and democratic control of, the public schools.

1. WHY CAN'T JANE AND JOHNNY READ?

   1. The Authors' Statistical Story Stated

      Chubb and Moe ask a question that has occupied researchers for years: Why do some children learn more in school than others? The authors' answer is new and sophisticated: Market forces favor school autonomy; school autonomy fosters effective organization; effective organization promotes achievement.(19) Or, even more simply, market forces promote achievement.(20) This section describes each step in the authors' argument in order to begin examining whether the argument's sophistication illuminates or obscures.(21)

      Step 1: Researchers have long known that parental resources and student ability strongly affect successful learning.(22) The authors' significant discovery is a third factor associated with learning--effective school organization, which accounts for a quarter of the variance in achievement that the authors' regression model explains.(23) The discovery is significant because organization may be more sensitive to public policy than are wealth and ability.

      The authors' of "effective school organization" is not an established metric in educational research, however, but rather their own invention.(24) Actually, the authors use two measures of effective organization. The first is a composite of: (1) students' satisfaction with the fairness and effectiveness of their school's disciplinary mechanisms; (2) teachers' feelings of "efficaciousness" vis-a-vis their students, "collegiality" and "cooperativeness" vis-a-vis each other, "clarity" vis-a-vis their principals' "vision" and goals, and "influence" vis-a-vis school policy; and (3) principals' confidence in the excellence of their teachers, commitment to academic excellence as opposed to assuring literacy, and dedication to running their schools as opposed to advancing their carrers.(25)

      The authors' second measure of effective organization adds one additional factor--"academic program," i.e., the proportion of students in a school who are enrolled in an academic as opposed to a "general" or a vocational course of study.(26) Although it is not surprising that students in academic programs score higher on academic tests than students in nonacademic programs, the authors are interested in why students enroll in one or the other program. In their first regression analysis--of the factors associated with student achievement--the authors find that academic program has explanatory power even after parental resources and student ability are controlled. Based on this finding, the authors hypothesize that school traits also influence academic program, and they sometimes include academic program in their measure of "school organization."(27)

      Step 2: The authors conduct a second regression analysis to determine whether school policy or other factors explain why schools are well organized according to the author's first definition of organization--to determine, that is, why students say they respect the educational enterprise, why teachers say they are happy with it, and why principals say they are dedicated to it. The authors again find that certain parent traits (resources and frequency of contact with schools) and student traits (ability and comportment) account for close to three-fourths of the variance in the condition under examination (effective organization).(28) Likewise, the authors again discover a policy-sensitive factor--"absence of bureaucratic influence" or "autonomy"--that accounts for a quarter of the explained variance in the condition being studied.(29) By "autonomy" the authors mean the propensity of principals to credit or blame themselves, rather than unions and central office staff, for their schools' program and personnel.(30)

      Interestingly, at this second stage of their analysis, the authors introduce two new parent-student factors--parent contact with schools and student comportment--that were not studied at the first stage of their analysis and that seem to reflect the importance that parents and students attach to schooling. The belated introduction of this new causal factor--"taste" for education--in the second regression raises the question whether educational taste also might have proven causal had it been introduced, in some form, in the initial regression of achievement.

      Step 3: The authors conduct a third regression analysis to see whether manipulable policy, rather than some other factor, determines how much responsibility principals attribute to themselves as opposed to outsiders. In this study, parent traits (contact with schools) and student traits (ability, comportment, and achievement gain) explain only 20-30% of the fluctuation in the dependent variable (autonomy).(31) Notably, the authors again introduce a new student-related causal factor at this...