The question whether schools would be better if they had more money has been a subject of debate among educators, politicians and economists since the Coleman Report was published in 1966. According to Kevin J. Payne and Bruce J. Biddle, University of Missouri/Columbia these arguments have been fueled by a lot of not very good research. They recently reviewed past studies and the conclusions drawn from them, and compared them to a new study of the independent effects of both child poverty and school funding on achievement in math. In addition, they compared math achievement in the United States to achievement in other industrialized nations.
Counter-intuitive findings about funding and school achievement
When money buys improvements in most things in society, why should it not improve education? Most research evidence has been negative. Many studies find little relationship between student test scores and either school expenditures or the smaller class sizes and other things that money can buy. Payne and Biddle provide a well-stated summary of the issue. They remind us that the United States differs from most other industrialized nations in that it funds the bulk of public education through local taxes.
The result is great variations in school funding from state to state and even within states. Per-pupil funding ranges from about $3000 to $15,000. Despite these disparities, there are only a few studies that suggest that differences in funding levels significantly affect school achievement.
There are two possible explanations for these counter-intuitive findings. The first is that the net effect of funding on achievement is really quite weak. If differences in students’ achievement are caused by their home environment, then it doesn’t really matter whether a student enrolls in a well-funded or poorly funded school. The second possibility is that most studies of the impact of school funding on student achievement have flawed research designs.
The Coleman Report was the first serious attempt to examine the effects of school characteristics on student achievement. It was commissioned by the National Center for Education Statistics in response to the Civil Rights Act of 1964. In brief, the report found that factors related to the students’ home backgrounds and the characteristics of other students in their schools had the most impact on achievement, and that “schools bring little influence to bear on a child’s achievement that is independent of his background and general social context.”
Payne and Biddle state, however, that there were major errors in the study that reduced its estimates of the size of the school effect. In fact, Coleman corrected some of these errors in a later analysis (1972). This reanalysis generated larger estimates for the effects of schools, but it did not attract much critical attention. These researchers report that many later studies found positive but not statistically significant net effects for school funding. Recent analyses of this past research suggest that although the impact of funding may be questionable for short-term achievement measures, studies of long-term results indicate stronger effects.
According to Payne and Biddle, many studies done after the Coleman Report reveal similar design flaws. For example, most of these studies did not examine school funding directly but instead measured some of the things that funding can buy in schools. Also, many of these studies provide few details about the sample used, the ways in which the data were collected, the ways the measuring scales were constructed, and the reliability and validity of those scales.
Most of these studies show evidence of serious problems with the data, such as the use of achievement data that are not based on random samples of student populations. These researchers conclude that because of the widespread prevalence of design flaws in the research literature, most of these studies can not be used to reach valid conclusions about the net effects of school funding in America.
Payne and Biddle agree that home environment is an important factor in student achievement and therefore needs to be considered separately from school funding to determine the independent effect of each on school achievement. Child poverty is an important measure of home environment and is another way in which the United States differs significantly from other industrialized nations.
Between 20 and 25 percent of U.S. children live in poverty. Poverty means these children are likely to suffer from chronic dental and health problems and inadequate diet. The Luxembourg Income Study, which has measured income distributions in 18 Western nations since 1990, shows that the U.S. child-poverty rate is about 45 percent higher than that of any other country studied and between five and eight times higher than rates in most industrialized nations.
The influence of child poverty on achievement is well documented. However, most studies do not separate the effects of poverty from the effects of other risk factors likely to be correlated with poverty (such as race, ethnicity or immigrant status). Since poor children are also likely to attend badly funded schools, it is necessary to separate out the effects of these different variables.
A new research design
Because of the problems evident in prior studies, these researchers sought a new and better design to study the net achievement effects of poor school funding and child poverty in the United States. They wanted research that was based on a national sample and that used good measures for school funding, child poverty and achievement while controlling for the possible effects of other variables. In addition, they wanted to compare the net effects with achievement results from other nations.
For their study, they used data collected by the National Center for Educational Statistics (NCES) and the International Association for Educational Achievement (IAEA). The NCES data provide details about funding, poverty rates and other information for school districts in the United States.
The IAEA data provide math achievement data for 23 countries, including the United States. They used data for eighth graders in these countries and found that math achievement was closely linked to the level of curriculum (remedial, typical, pre-algebra and algebra) in the classroom.
These researchers included NCES data from 205 classes in 136 schools in 67 districts in 32 states. The variables compared were a measure of school funding and the percent of child poverty. The percent of nonwhite persons in the district was included to separate race or ethnicity as a factor from poverty.
Significant correlations were found linking student achievement in U.S. schools to differences in curriculum level, school funding and child poverty at the district level. Further analysis demonstrated that funding and poverty have significant effects independent of one another, and that, taken together, the effects on achievement are very large.
When students were compared by race, the effects of funding, child poverty and level of curriculum all remained substantial and significant. The effect of race, by itself, was smaller than the effects of the other three variables and not statistically significant in predicting achievement in this sample.
Payne and Biddle conclude that these results suggest that the level of school funding and child poverty have statistically significant net effects on average student achievement in the United States. They believe that their study, despite some limitations, clearly demonstrates that the effects are large and relatively independent of one another and of the impact of other factors such as curriculum and race.
Three factors taken together — curriculum, funding and child poverty — account for one-third of the achievement differences between school districts. These researchers suggest that reducing child poverty and funding disparities might improve student achievement more than many of the current reform programs, and they believe the current research data should dispel the myth that funding does not matter in public education.
“Poor School Funding, Child Poverty, and Mathematics Achievement” Educational Researcher, Volume 28, Number 6, September 1999, pp. 4-13.
Published in ERN November 1999 Volume 12 Number 8