usually justified in terms of the higher levels of expenditure incurred in advanced courses in high school. In this study no account is taken (or "credit" given) for such differences; the disparities as stated here may therefore overestimate disparities in some States (while possibly 11 underestimating disparities in others).

Using these measures and presentation methods, changes in resource 'distribution between 1970 and 1975 are presented in three sections:

The first analysis is of variations in resources within each
State, without reference to characteristics of the districts

The second is of the characteristics of the districts that
have gained or lost from 1970 to 1975

The third section presents estimates of the cost of further
equalization.

Changes in Overall Disparity

Table I presents, for each State the disparity in 1975 and the percentage between 1970 and 1975. Column 1 lists the disparity ratio in 1975 as calculated by the 95:5 measure, and column 2 shows the ratio of the 1975 value to the 1970 value; a ratio greater than 1.0 shows that the disparity in 1975 was greater than in 1970. Columns 3 and 4 present

similar calculations for the coefficient of variation.

Column 5 summarizes

the results of columns 2 and 4, indicating whether or not the change from 1970 to 1975 appears to exhibit an equalizing trend. The use of question 12. marks is intended to reflect uncertainties in a result.

Several characteristics of the behavior of the two measures should

first be noted. The disparities according to the two measures (columns 1

and 3) are broadly in agreement, although not in every instance. When

When one

either measure shows a clear indication of disparity, the other also does. The percentages of change according to the two measures (columns 2 and 4) are also in close agreement, except for Alaska (which is an outlier in both the geographic and the statistical sense) and Delaware. measure is close to 100 percent, however, the other sometimes indicates a change in the opposite direction. Finally, the percent change in either direction usually appears greater under the coefficient of variation: If the 95:5 measure shows 102, the coefficient of variation may show 133 (as it does for Idaho), while if the 95:5 measure shows 89 the coefficient of variation may show 63 (as it does for Kansas).

They

The States with the greatest disparities (columns 1 and 3) show no regional patterns, including States from every part of the country. do include, however, several of the nation's largest States. The 12 States with greatest disparities enroll approximately 38 percent of the nation's public school students. At the opposite end of the ranking, States with the least disparity tend to be those with the smallest populations, the only exception being Florida. It is noteworthy that

the six States with the least disparity also operate relatively few school 13

systems.

Columns 2 and 4 show the change from 1970 to 1975, as a percent. Only about one-third of the States appear to have made any progress in reducing disparities in the intervening years, even including some that improved by a percentage point or less each year in the five-year period. Among the 12 States with the greatest disparities, only Vermont exhibits a clear change toward equalization

but it must be pointed out that Vermont's

starting point in 1970 was extremely unequalized.

*

Thus far, the study has described the disparities in the States with

out addressing the question, How much disparity is too much? In considering this question it is helpful to characterize two views that represent the ends of a spectrum.

Some argue that local control of the schools is essential to effective delivery of educational services and that localities must be free to choose the level and types of resources appropriate to the community's needs and desires. They would argue, on principle, that any constraint on a locality's freedom to use its own wealth to support education is counter to the spirit of the free enterprise system and, on practical grounds, that imposition

of such a constraint would lead to a "wealth flight" from public schools like the "white flight" that is sometimes claimed to follow school desegregation.

-

-

there

In opposition, others would hold that aside from justifiable differences arising from the costs of education or differing pupil needs is no reason to permit any variation. They would point to decisions by State courts in school finance cases for support of the view that a State is, in effect, a large school district that establishes local school districts for purposes of educational effectiveness and administrative convenience but not in order to permit school resources to vary from place to place.

These views do not exhaust the issues that arise in considering the

*

In Vermont in 1970, the 95:5 measure showed a disparity ratio of 3.36,. and the coefficient of variation was 0.38, higher figures than for any other State in that year.

extent of permissible variations within a State. It has been argued, for instance, that the costs of equalizing constitute a burden that the already hard-pressed educational system is not equipped to bear; the last part of this section derives estimates of these costs for varying levels of equalization. Moreover, it has been pointed out that there is little evidence that the level of school resources influences the achievement of students.

It is not the purpose of this study to evaluate these views. It suffices that disparities have been and are a concern not only of the State courts but also of the Federal government. How much disparity, then, is too much? One source of guidance is a criterion set by the Office of Education in determining how a State can qualify to be able

to count Federal impact aid as State aid.14 The State must be operating

an effective school finance system, where effectiveness is determined by whether or not its 95:5 measure shows a disparity no greater than 25 percent. According to the results in Table I, only two States, Hawaii and Nevada, could meet this requirement in 1975. The Office of Education test does not count toward disparity the spending made for special needs, while the present study has not separately identified such funds. Moreover, these results are based on approximate measures and on a sample of districts. Nonetheless, it is difficult to believe that more precise measurements would result in more than a handful of additional States qualifying in 1975.

Disparities throughout the distribution of a State's districts are not the only concern in school finance. The courts and legislatures have also been concerned about the types of districts that benefit and