Does Money Matter in Education
According to numbers released by the U.S. Department of Education, taxpayer investment in K-12 education is estimated at over $500 billion per school year. This makes the U.S. the world’s leading investor in education. Even during a period of heightened concern for homeland security, spending on education still exceeds that for national defense. This generous allocation of resources has caused many to demand accountability and evidence of academic improvement equal to the nation’s investment. Long term trends on the National Assessment of Educational Progress, however, have yet to reveal any measurable differences in performance for 17 year olds from 1971 to 2004 for any reporting metric. (NCES, 2004)
Common logic challenges the conclusion that our nation’s children have not been served by educational expenditures. It is intuitive to any parent, teacher, student or administrator that resources make a huge difference in the daily operation of a classroom. All other variables aside, it cannot be denied that impoverished schools struggling to acquire new textbooks and basic supplies begin the school year at a disadvantage. It is a logical hypothesis that school districts with a higher average SES population will experience fewer of these disadvantages than school districts with a lower average SES population. Repeated studies have confirmed this hypothesis by linking the socio-economic status of a school’s student body to academic achievement. (Coleman, 1966)
Given lackluster achievement scores and the need to sustain funding for education, the emerging question facing researchers is not “does money matter,” but rather in what ways does money matter the most? While research on general funding levels is extensive, there is a dearth of information detailing the specific allocation of funds and analyzing where additional funds have yielded the largest results. Such research is difficult to acquire because budgets for school districts are often inaccurate and use misleading district averages to estimate teacher salaries as a supplement when data on individual schools is unavailable. (Roza & Hill, 2003) In addition, diverse revenues streams comprised of federal, state and local governments as well as private donations make tracking funds a cumbersome process. There are even instances of school systems duplicating benefit payments because their accounting system was unable to catch the error. (Roza & Hill, 2003)
Further complicating education production studies related to the allocation of resources is the inability to compare school budgets even within districts. Many schools receive compensatory funds precisely because they serve under achieving populations. Hence, it should not be surprising to see schools with high expenditure levels with low scores on achievement tests. To more fairly evaluate schools, researchers have begun using a “value added” approach. This approach gauges the level of student achievement at the beginning of a school year and then attempts to measure a student’s progress over the course of the year as the actual academic value added by the school. The statistical requirements for this calculation are extremely complicated, however, and cast doubt on the validity of any numeric evaluation. (Ladd, 2001)
Despite these limitations, available studies have documented the impact of money on student achievement. While SES is a dominating influence, this analysis will show that school resources can positively impact student achievement independent from SES, especially when these resources are devoted toward instructional uses. Consequently, policies that target funding to the most disadvantaged schools and districts will not only be successful in promoting equity throughout the education system, they also are an important tool for enhancing achievement among low performing schools. The questions to be answered in this analysis are 1) does increased levels of school funding automatically translate into increased student achievement? and 2) what type of school funding is most likely to impact student achievement?
What Does the Research Say – Is Money Enough?
In 1966, James Coleman released one of the most widely known pieces of quantitative social science research in American history. It is known as the “Coleman Report.” As part of the 1964 Civil Rights Act, Coleman studied the disparities in funding between schools attended by a majority of black and white students. The surprising results revealed that funding was not closely related to achievement. Family socioeconomic status and the social composition of the school’s student body were more accurate predictors of student achievement. As student scores mimicked SES, funding levels appeared irrelevant for explaining the performance difference between schools. (Coleman, 1966)
Research by Eric Hanushek, a professor at the University of Rochester, builds upon the Coleman Report. Hanushek asserts that the majority of studies reveal no statistically significant relationship between test scores and school expenditures. Specifically, Hanushek looks at the impact of using resources to reduce class size – a common funding objective for many school districts. He focuses on the Project STAR study in Tennessee and maintains that the experiment failed to document gains in student achievement proportional to the increase in funding. (Hanushek, 1999) Hanushek does acknowledge, however, that class size reduction could be beneficial if schools attracted high quality teachers. Yet he quickly dismisses this caveat claiming that the actual expenditures required to create measurable improvements would be so exorbitant as to render them wholly unrealistic for any normal school budget.
Hanushek’s research was directly refuted in a response article by Greenwald, Laine and Hedges. (Greenwald et al, 1996) The Greenwald et al article criticized Hanushek’s research methodology and pointed to changing societal demographics as an explanation for stagnant achievement scores. Contrary to Hanushek’s assertions, the article finds it surprising that achievement scores have not declined considering the dramatic societal changes that have occurred over the last four decades – more children living in poverty, more single-parent households and more mothers joining the workforce. Greenwald et al credits a modest increase in educational spending for keeping achievement levels steady.
Despite reservations, both the Coleman Report and the subsequent research conducted by Hanushek have greatly influenced the debate on school finance. Their warning that money alone will not automatically translate into increased achievement has been well headed by fiscally conservative policy makers. Advocates for education, on the other hand, have rejected the implied conclusion to the research of Coleman and Hanushek that schools do not make a difference. They charge that it is too easy to simply assume that students are trapped in a SES pattern of achievement. To prove their theory, these advocates have accumulated a body of research which adds complexity to the school finance debate.
What Does the Research Say - Educationally Significant Use of Funds?
There is a solid and growing base of evidence suggesting that additional funds can enhance student achievement when targeted for educationally significant uses. As claimed by Ronald Ferguson, Associate Professor of Public Policy at Harvard University, “money matters when the real inputs that it purchases matters.” (Ferguson, 1991)
Ferguson conducted a study of approximately 900 independent school districts in Texas. He found that “differences in the quality of schooling account for between one quarter and one third of the variation among Texas school districts in students’ scores on statewide standardized reading exams.” (Ferguson, 1991) In particular, he found that money to support high quality teachers was directly linked to student achievement scores. Supporting high quality teachers was interpreted to mean everything from attracting young teachers with stronger literacy skills and advanced degrees, retaining experienced teachers, and reducing the student-per-teacher ratio. In contrast, spending on general operating expenses did not affect achievement scores. Virtually all of the differences in student scores observed by Ferguson were explained by instruction-related expenditures.
The reach of the Ferguson study is limited by its reliance on standardized test scores as the only measure of student achievement. Many scholars concerned with educational equality have expressed concerns about bias on standardized tests. In general, students who score well on standardized tests come from affluent families and benefit from many advantages in addition to higher quality schools and teachers. This makes it extremely difficult to isolate the effect of funding from raw scores on standardized tests. The study would be enhanced if it could have tracked the progress of student achievement through a value-added assessment.
Ferguson’s findings were supported by another Texas study conducted by Murnane and Levy in 1996. Murnane and Levy followed sixteen schools in Austin, Texas who received $300,000 per year for five years as a supplement to the normal school budget. These schools were awarded the additional funding to raise low achievement scores. In fourteen of the schools, the funding did not have a statistically significant impact. These fourteen schools used funding to decrease class size, yet they maintained the same instructional practices used by the larger classes. In two schools, however, the five years of funding saw achievement scores rise to the city average and attendance rates climb to the highest in the city. The two successful schools had a radical approach to the funding and revolutionized their instructional practices. They invested in professional development and orchestrated a campaign to involve parents and make them aware of low achievement scores in the school. The important conclusion to be emphasized from the Murnane and Levy study is not that fourteen schools saw no statistical impact, but rather that two schools were able to target funds in a manner that raised student achievement.
The conclusions drawn by Murnane and Levy are extremely promising, but must be approached with caution. The study’s methodology did not allow for a comparison control group which would prove that the results achieved by the two successful schools were, indeed, a product of funding and not other confounding variables affecting student performance. For instance, skeptics could suggest that there were inherent differences between the two successful schools and fourteen unsuccessful schools independent from the additional funding. Further studies on this topic which make methodological efforts to control for external variables among schools would be a welcome addition to the research literature.
Finally, persuasive evidence concerning the positive impact of resources when used instructionally was offered by Harold Wenglinsky in 1997. The Wenglinsky study used data from the National Assessment of Educational Progress, the Common Core of Data and the Teachers’ Cost Index to prove that certain school characteristics “influence achievement above and beyond SES.” (Wenglinsky, 1997) Wenglinsky specifically looked at per-pupil expenditures on instruction, per-pupil expenditures on central office administration and teacher-student ratios. The data revealed that instruction-related spending positively influenced the social environment of the school which in turn affected student achievement.
It is important to recognize that, while valuable, Wenglinksy’s study suffers from a methodological limitation shared by many studies using large, self-reported data sets. This limitation being the danger of unreported or misreported data. Quantitative studies are only as good as they data they manipulate and often districts do not have the technological ability to accurately track necessary information such as per-pupil expenditures. In this situation, districts may report averages as a best estimate, but district averages may or may not be accurate for individual schools.
In summary, the available research seems to suggest that “adequate funding levels are a necessary, but not sufficient condition for the development and maintenance of high quality school programs.” (Murnane, 1991) Studies have found that certain expenditures are more closely associated with student achievement than others. This is most obvious in the case of instruction-related items, but also holds true for non-instructional expenditures. While money for items such as transportation and building maintenance may matter, they are not directly related to student achievement. In fact, increased spending on transportation and maintenance may indicate a more scattered district and older facilities, both of which may bias achievement scores.
Should such evidence be needed, the two Texas studies and the Wenglinsky study cited above put to rest the argument that schools do not influence student achievement. The strong link between SES and student achievement revealed by the Coleman Report, however, remains supported by almost all available research. Although it may be beyond the ability of even the wealthiest school to entirely erase SES inequalities, schools can allocate resources to maximize the achievement impact for which they can legitimately be held responsible. If the research is to be believed, the largest impact lies in instruction-related spending.
Policy Implications
The policy implications of the research link between educational expenditures and student achievement are considerable and highly controversial. It is asserted that, “Educational policy should be based on the presumption that per pupil expenditures, smaller classes, smaller schools and more experienced teachers are all positively related to student achievement.” (Greenwald et al, 2006) Accepting this premise and supporting data, it becomes a school district obligation to ensure that all students are given the opportunity to benefit equally. If additional resources for instruction-related spending are important, than it is particularly important for students who already start at a disadvantage.
Part of the challenge faced by low performing schools is that district budgets are not uniformly distributed. Wenglinsky argues, “Because 47.4 percent of all school expenditures are financed by local property taxes, school districts with predominantly low-SES populations tend to have fewer economic resources than those with predominantly high-SES populations.” (Wenglinsky, 1997) This school finance system perpetuates a cycle in which high-SES schools continue to receive an academic edge over low-SES schools. Compensatory funding from state and federal governments is designed to offset SES inequalities, but districts have yet to achieve financial parity. (Roza & Hill, 2003)
Compounding the problem of resource allocation is the use of district averages to calculate teacher salaries. As the research has demonstrated that attracting and retaining high quality teachers is the single most effective investment a school can make, salary differentials can drastically influence student achievement even within districts. Marguerite Roza and Paul Hill conducted a study in 2003 in which they compared the real salary costs for individual schools to the district average. They study found that, “In each district, there were specific types of schools that routinely received fewer teaching dollars than the district budget claimed. In each city, high poverty, low performing schools were staffed with lower than average salaried teachers.” The use of salary averages masks the actual flow of dollars within a school district and supports a discriminatory financial structure
It is also documented that, given the opportunity, experienced teachers will migrate from low-performing, low-SES schools into high-performing high-SES schools. (Hanushek et al, 2004) In this instance, it can be assumed that salary differentials will not only mean unequal pay, but that the most qualified teachers will continue to be attracted by the highest-performing schools.
From a policy stand point, this data compels states to make more effective use of compensatory funds. Not only do states need to increase the amount of compensatory funding devoted to attracting and retaining teachers at low-SES schools, they should demand the reporting of real numbers on teaching salaries to calculate the distribution of compensatory funds. Policy makers need to be mindful that the purpose of compensatory funding is to provide resources to low-SES schools that supplement the general budget. These schools are in need of additional funding because they are charged with serving students who enter at a disadvantage. Education policy should be designed to alleviate this disadvantage, not institutionalize the inequalities that already exist.
As instruction-related spending is most likely to boost student achievement, schools also should ensure that compensatory funding is marked for instructional use. Even the term “instructional use,” however, is very vague and can justify a wide range of expenditures. Ultimately, how schools utilize an increase in instruction-related spending will rest with administrators. As each school situation is different, administrators will be faced with tough decisions. Some schools may choose to decrease the student-teacher ratio by hiring new teachers, some may need to raise salaries and retain experienced teachers, while still others may need to invest in professional development for teachers. Currently, research does not appear to offer any clear guidance as to which of these options, all related to instruction, may be the most effective. It appears to depend on the particular need to the school. As long as additional funding is accompanied by fair and rigorous accountability measures, the expenditure can still be justified.
The restructuring of education policy will be fiercely opposed on various fronts. First, teachers’ unions will fight to preserve collective bargaining agreements and preserve the privilege of seniority. Senior teachers now get the first choice of job openings and flock to high-paying schools. Discontinuing the use of district salary averages will erase this benefit. Also, the increased allocation of compensatory funding will temper the growth of general education budgets. High-SES schools who benefit from the status quo will object to any effort to redistribute school funding. Naturally, these schools will want to preserve their competitive edge.
Despite the objections, increased compensatory funding targeted for instruction is necessary to support performance among all students. The research documents that this allocation of resources will ensure that the most effective funds will be targeted toward the neediest populations.
Conclusion
It is undeniable that SES is the dominant factor in predicting student achievement. This fact, however, cannot be allowed to justify a defeatist attitude that questions that very ability of schools to “make a difference.” Research by Coleman and Hanushek wisely cautions us not to blindly throw money at schools even with the best of intentions. This does not mean that we should not be prepared to spend money on schools at all. Research by Ferguson, Murnane and Levy, and Wenglinsky have shown that schools can impact student achievement and that school resource levels dictate how large this impact will be. These researchers have also documented that it matters how money is spent. Instruction-related spending has been shown to have the greatest influence on student achievement. Additional research is needed to determine which expenditures related to instruction is the most promising for low-SES schools.
Education in the U.S. has been described by many as the “great leveler.” It is assumed that, given a fair start, students will achieve according to their merit. Inequalities within the school finance system casts doubt on this assumption. The current system of school finance permits a cycle of inequality between low-SES and high-SES schools. The inability to adequately fund certain schools from local property taxes makes compensatory funding increasingly important. In addition, the use of district salary averages makes it difficult to observe inequalities even when salary differentials are quite large. Policy solutions to these problems must be based on the premise that schools do matter.