Academic journal article Contemporary Economic Policy

What Do Schools Produce? Implications of Multiple Outputs in Education

Academic journal article Contemporary Economic Policy

What Do Schools Produce? Implications of Multiple Outputs in Education

Article excerpt


This article introduces the hypothesis that secondary schools create at least two competing "outputs"--standardized test scores and high school completion--which may be seen as substitutes. This hypothesis provides an explanation for the decrease in standardized test scores that occurred over the last 40 years along with an often overlooked increase in high school graduation rates. The article presents evidence supporting this hypothesis as well as evidence that different schools face different tradeoffs and discusses the policy implications of such trade-offs. (JEL I21, J24)


Public education in the United States over the last 40 years has been characterized by dramatically increased real spending coupled with falling test scores. [1] Naturally, the coexistence of these two trends has sparked considerable interest and concern. Many researchers view this combination as evidence of inefficiencies, and such inefficiencies certainly may exist. However, another important trend characterizes this time period--high school graduation rates increased dramatically. The percentage of the population having graduated from high school increased nearly 50% between 1960 and 1993 (U.S. Bureau of the Census, 1994, p. 157). The high school dropout rate, defined as the percentage of those between the ages of 25 and 29 with neither a high school diploma nor a GED, decreased by more than 50% between 1960 and 1975 (Jencks, 1992, p. 174); the high school dropout rate continued to fall until 1987 and has remained relatively stationary since that time (U.S. Bureau of the Census, 1993, table A-4). [2] At l east one account shows that SAT scores have changed very little over the last 30 years after correcting for class rank--scores for students in each quartile have been stable, but as more students from lower quartiles took the test, the average score fell (Berliner and Biddle, 1995, p. 21).

This research hypothesizes that falling test scores and increasing high school completion rates are directly connected--that, in fact, schools face a trade-off between test scores and graduation rates. Thus, we would expect that rapidly increasing graduation rates should be accompanied by falling test scores. In this case, graduation rates and test scores may be thought of as distinct, separate "products" of schools. Thus, schools using resources to increase graduation rates may be forced to accept a reduction in test scores. Initially this is counterintuitive; most people think of "good" schools as those producing high test scores and high graduation rates. Also, high test scores increase an individual student's probability of graduation. However, consider the school as the decision-making unit. Beyond a certain point, schools may be able to increase graduation rates only by accepting lower test scores (the students retained to increase graduation rates may lower average test scores), and may be able to increase test scores only by accepting lower graduation rates (allowing those students with the lowest test scores to drop out). The same argument holds if marginal students decide to stay in school due to increased returns to education. Simply put, test scores and completion rates may be substitutes rather than complements. [3]

A negative correlation between graduation rates and test scores provides some evidence of the trade-off between high school graduation rates and test scores. Table 1 uses data from the High School and Beyond (HS & B) survey (discussed in Section III) to show that the percentage of students completing high school is positively correlated with both the initial standardized test score and the standardized test score given during the first follow-up. In contrast, the percentage of students completing high school is negatively correlated with the gain in the standardized test score. This last measure is consistent with schools trading off gain test scores for graduation rates; however, this result could also be explained by a ceiling effect on the standardized test included in the HS & B data set. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.