Dual standards of school performance and funding? Empirical searches of school funding adequacy in Kentucky and Maine

Abstract

This study examines potential consequences of the discrepancies between national and state performance standards for school funding in Kentucky and Maine. Applying the successful schools observation method and cost function analysis method to integrated data‐sets that match schools’ eight‐grade mathematics test performance measures to district funding, this study conducts empirical searches for adequate instructional expenditures per pupil to meet desired proficiency targets on national versus state assessments. While Kentucky (high‐stakes testing state) had a lower performance standard than Maine (low‐stakes testing state), this study reveals a relatively lower level of school funding adequacy and a weaker relationship between school expenditures and performance for Kentucky than for Maine. The study suggests that state educational accountability systems and policies may influence the level of state performance standards and the proficiency gaps between national and state assessments, which in turn lead to potential gaps in school funding. Implications for policy and research are discussed to address problems with dual standards of school performance and to improve school funding adequacy and efficiency.

Keywords:

• accountability
• adequacy
• school funding
• performance standards
• high‐stakes testing

Notes

1. Reschovsky and Imazeki (2001, p. 378) discuss the study by Augenblick, Myers, and Anderson (1997) that excluded all districts with particularly high and low levels of property wealth and of per‐pupil spending: ‘Although this method provides a norm level of per‐pupil spending, no systematic attempt is made to measure the likely variation in the costs of adequacy due to characteristics of individual school districts and their students.’ Verstegen (2002, p. 778) also points out several potential problems with this type of study: ‘If the funding system is inadequate for most districts in the state, as was the case in Kentucky, then a study that attempts to correlate current spending in select districts to student outcomes will generally fail to capture actual costs. Moreover, what a district spends and what resources/standards may cost can be separate issues that are muddled in these types of analyses. Still, another difficulty with this approach is that it defines efficient district as neither high nor low spenders when outliers are removed, leading to the possibility of recommendations that underfund education.’

2. Category labels and brief generic definitions of performance standard for each assessment are as follows:

•	NAEP Proficient: students demonstrate competency over challenging subject matter and are well prepared for the next level of schooling.
•	KIRIS Proficient: the student understands the major concepts, can do almost all of the task, and can communicate concepts clearly.
•	MEA Advanced: Maine students successfully apply a wealth of knowledge and skills to independently develop new understanding and solutions to problems and tasks.

3. The F‐33 survey is part of the US Census Bureau’s ‘Survey of Local Governments.’ It is a universal survey of school funding in the nation and states (see O’Leary and Moskowitz 1995). The district level is the lowest level of aggregation in the F‐33 data, and no nationally representative database exists that measures different types of expenditures at a lower level of aggregation. Because the school funding measure from the F‐33 data is available only at the district level, the measure of PPE would simply vary among school districts within each state but not within districts (among schools and classrooms).

4. In fact, the study’s 30% proficiency rate as the minimum school performance target was higher than the states’ current annual measurable objective in eighth‐grade mathematics to meet NCLB as of the 2004/05 school year : 26.93% in Kentucky (Kentucky Department of Education 2004) and 13% in Maine (Maine Department of Education 2003). In order to reach the ultimate target of 100% proficiency by the 2013/14 school year , both states plan to take a stair‐step approach that will allow schools to improve slowly in the first few years as improvements are phased in, and then more dramatically once improvements are fully implemented.

5. SES is a standardized measure of student’s family and school’s SES level, and it is derived from the NAEP data. It is a factor composite of parents’ education (PARED), availability of reading materials at home (HOMEEEN2), and school median income (MEDINC). Student‐level factor loadings are as follows: PARED, 0.77; HOMEEN2, 0.68; MEDINC, 0.63. The factor has an eigenvalue of 1.46 and explains 49% of the combined variance. The underlying assumption is that these variables are related to voters’ willingness to pay for education and only indirectly affect the cost of education through their effects on school performance (eighth‐grade mathematics). Hausman test of endogeneity rejects the null hypothesis that there is no simultaneity regarding the relationship between instructional cost and school performance. This supports the use of instrumental variable estimation method.

6. Obviously disadvantaged groups of students are more expensive to educate but currently there are no national standards for determining the cost of educating different student groups. One common practice in school funding formula involves adjusting fall student enrollment by weighting poor students by 1.4, special education students by 2.3, and limited English proficiency students by 1.2.

7. The Employment Cost Index of the US Bureau of Labor Statistics was used to convert 1996 dollars into 2005 dollars. The Employment Cost Index for ‘Total compensation: Elementary and secondary schools: State and local government’ has increased 1.3 times from 1996 to 2005. It needs to be noted that the study’s estimates of instructional costs are not directly comparable with those from previous studies of school funding adequacy in Kentucky and Maine because they used different criteria, measures, and methods. Previous studies used evidence‐based or professional judgment methods to estimate base costs of education (Quality Counts 2005). A study that used an evidence‐based method estimated an average of $6893 per pupil as the basic cost in Kentucky to reach the state’s goal of having all students perform at or above ‘proficient’ on the state assessment by 2014 (Odden, Fermanich, and Picus 2003). A study of funding adequacy in Maine that used a professional judgment method estimated the per‐pupil operating cost of $4543 for middle‐school level regular students (Silvernail 2000).