In “Beyond Spending Fetishes” (November 1995), Eric Hanushek argues that while the financial investment in education has doubled during the last 25 years, the returns have not been equal to those of previous eras. He asserts that dramatic increases in per pupil expenditures and the reasons for them (for example, smaller pupil/teacher ratios and more experienced, more educated, and better compensated teachers), have resulted in stable—not increased—student achievement levels.
- More than 20 percent of children today are living in poverty (Sternberg and Colman 1994), and the ratios are even higher in urban centers and minority populations.
- Approximately 25 percent of all children live in single-parent households.
- The percentage of mothers in the work force has more than doubled during the last generation.
These factors have forced schools to meet a greater variety of demands and to provide more services than ever before. Expenditures have increased to support this broader mission. However, the common perception of the magnitude of increases in real spending may be incorrect.
The Funding Debate
Hanushek argues that expenditures per pupil have “more than doubled in real terms” during the last 25 years. Another recent report (Rothstein and Miles 1995) argues that while total spending did increase during this period, the increase in real spending is only about two-thirds as large as Hanushek asserts. This study, which used a more appropriate means of adjusting for inflation, also found that a significant portion of the increase was allocated to special education. The increase in inflation-adjusted per pupil expenditures for regular education was 28 percent (slightly more than 1 percent per year) during the period from 1967 to 1991.
Given this relatively modest annual increase and the rather dramatic changes in society, one might expect significant declines in student achievement. Yet evidence to the contrary may be found in the National Assessment of Educational Progress (NAEP).
NAEP may provide the best data concerning the academic achievement of high school students during the last two decades. These data show that the achievement levels of white, black, and Hispanic students have increased significantly in reading (Mullis et al. 1994). While the scores for white students have remained stable in mathematics, the scores for blacks and Hispanics have increased significantly. In science, the achievement levels of whites have declined, while scores for blacks and Hispanics have remained stable.
Thus, NAEP data demonstrate increases in overall achievement levels and significant gains in the achievement levels of certain student populations. To determine whether these gains are enough to justify the increased expenditures, one would have to evaluate the impact of the societal changes noted earlier. At a minimum, it suggests the possibility of positive relationship between resource allocations and student achievement.
A Question of Methods
In his article, Hanushek reiterates a statement he has made frequently during the last 15 years. He says that “there is no relationship between expenditures and the achievement of students” (Hanushek 1981). He focuses on one resource variable—class size—and argues something which most teachers would find difficult to believe: that the ratio of pupils to teachers does not affect learning.
Hanushek cites the Project STAR study in Tennessee, and—contrary to the conclusions of a recent review (Mosteller 1995, Finn and Achilles 1990)—maintains that this program fails to demonstrate that student achievement increases significantly in smaller classes. In our reanalysis of Hanushek's data (Hedges et al. 1994), we found substantial evidence of a positive relation between a number of school resources (including class size) and student achievement.
A more recent analysis (Laine et al. 1995) conclusively demonstrated that a variety of factors (including per pupil expenditures, class size, school size, and teacher characteristics) are positively related to student outcomes. Hanushek apparently misunderstands the methods employed in both the re-analysis and the subsequent work. These methods—known as meta-analyses or research syntheses—are widely accepted statistical techniques. They are designed to aggregate data from a wide variety of studies that test the same conceptual hypothesis, although they may differ in the details of their design or their measurement techniques.
Rather than pool the conclusions of the individual studies as Hanushek does (the method he uses—vote counting—is now rarely used in empirical research, where sophisticated syntheses are expected), meta-analysis combines the information from individual studies, then utilizes it to derive a conclusion based on the entire data set. When the evidence—including Hanushek's own data—is analyzed in this rigorous manner, it supports the conclusions that resources do matter.
We do not advocate the simplistic policy of “simply adding resources.” We have never argued that a simple increase in expenditures will necessarily result in increased achievement. We do recognize, however, that sufficient resources are essential if schools are to provide the educational opportunities we desire. Educational policy should be based on the presumption that per pupil expenditures, smaller classes, smaller schools, and more experienced and educated teachers are all positively related to student achievement. Proceeding under misperceptions makes it more difficult to convince legislators and voters that increases in school resources are necessary to meet increasing student needs.
