Prior studies were generally consistent with the findings of Moore, et al., who found that per pupil special education costs are approximately 2.3 times that of general education. However, a great deal of change has occurred in education over the past ten years, and CSEF has identified a number of important questions about special education costs that remain unanswered. For example, what are the comparative costs of individual types of special education programs, such as public versus private placements for comparable students? What are the fiscal implications of alternative special education program policies? For example, are general classroom placements in neighborhood schools more or less costly for students with low-incidence disabilities? To what extent, if any, do special education costs exceed revenues, thereby affecting the resources available for students in general education programs? 3
Unfortunately, existing budgeting and expenditure records are not of much assistance when estimating the actual costs of educational programs. As described by Levin (1983), they generally do not include all of the cost information that is needed, usually are organized by line item rather than by type of program, and due to varying interpretations of cost information and accounting conventions are generally not comparable across districts. As a result, "the costs of any particular intervention are often embedded in a budget that covers a much larger unit of operation" (page 50). This approach to gathering educational cost data focuses on individual classrooms, such personnel measures as numbers of staff and student counts, and such non-personnel resource measures as materials and equipment. 4 The conceptual design for the last major national special education expenditure study, developed by the current CSEF co-directors, was based on this type of resource cost methodology. 5
A major CSEF research goal is to further develop basic special education cost analysis methodologies and procedures and make them more useful for and applicable to special education policy development. This goal is comprised of three related research objectives. First, it is important to have a better understanding of the types of special education cost questions that local, State, and Federal policy makers need answered, so that solutions address the appropriate questions. Second, more cost-effective methods of data collection need to be developed. One reason more special education cost analyses are not conducted is that they tend to be time-consuming and labor-intensive. Finally, ways to make the best possible use of existing data must be found.
To achieve these three research objectives, which are related to the overall development of enhanced special education cost analysis capabilities, CSEF has formed collaborative arrangements with Kentucky, Oregon, and Massachusetts. CSEF chose these States because they had recently enacted, or were on the verge of enacting, special education finance reform; had specific special education cost-related policy questions to be addressed; and had an interest in forming a relationship with the Center for the purpose of developing an ongoing statewide education cost analysis capacity.
The special education cost analysis projects in Oregon and Massachusetts are currently underway. In the Oregon project, an examination of the costs and benefits of inclusive special education placement practices in a selected group of districts that are actively implementing such policies is under way. In Massachusetts, CSEF is conducting a statewide special education cost analysis of recently enacted special education finance reform.
As an example of the kinds of cost questions that are of interest to States and the information produced through these types of studies, the results of the Kentucky Special Education Finance Study are presented below as a case study. This study was completed in October 1994.
In order to address these questions, a series of statistical analyses were conducted on a combination of data sets derived from existing State data files and data collection efforts carried out by CSEF staff during the 1993-94 school year. The data gathered by CSEF came from a sample of 63 schools in 17 districts. The districts were stratified according to size and the percentage of children identified as eligible for special education services.
Total expenditures for special education services in Kentucky were estimated at $218.5 million, compared to total State and Federal funding of $218.8 million (table 5.2). These results suggest that, overall, public school districts in Kentucky are spending approximately the same amount of money on special education as that generated by State and Federal funding.
Total Amount (millions) --------------------------------------------------------------------- Revenues (Federal and State add-on) $218.8 Expenditures Personnel $204.8 Non-personnel 4.2 Tuition for out-of-district placements 9.5 ------ Total $218.5 Excess of revenues over expenditures $ 0.3Source: Center for Special Education Finance (CSEF).
Determining whether this funding level is adequate requires subjective judgments to be made about the overall quality of services being provided. However,State and Federal revenues appeared to be sufficient to support current levels of special education across the State.
CSEF's best overall estimate of special education costs versus special education revenues showed a very high degree of alignment statewide ($218.5 million versus $218.8 million). However, because of the difficulty in obtaining accurate information on the cost of non-personnel special education resources,our upward bound estimate of statewide expenditures is $247.6 million, which would equal a statewide funding deficit of $28.8 million.
Which types of districts exhibit systematic differences in the relationship between expenditures and revenues for special education?
The data show a fairly wide range of variation in the ratio of expenditures to revenues among districts. The following patterns of variation in the ratio of special education expenditures to revenues were found.
How do the three funding weights currently in use in the State compare with the actual costs of serving these various categories of students?
Table 5.3 presents an approach to comparing the study cost estimates to the State funding weights currently in place. Special education expenditures are based on data collected on samples of special education students. The costs of instruction in the general education program represent a combination of the basic costs of instructional personnel, and is derived from the data CSEF collected at the school and district level. The value in the estimated weight column is calculated by subtracting 1 from the ratio of total per pupil special education costs to total per pupil general education costs. In theory, the estimated weight value is what should be used to calculate the add-on special education revenues. The KERA weights actually used to calculate the add-on revenues are presented in the last column for the purpose of comparison. Only for the high-incidence special education students are the estimated weights lower than the KERA weights. The overall average estimated weight for the speech or language students is 50 percent more than the current KERA weight, and the weight for the low incidence students is only .22 points (or 9.4% = 100 x .22/2.34) higher.
Estimated Student Category Estimated Total Per Ratio of and Percentage Total Per Pupil Special to of State Special Pupil General General Education Special Education Education Estimated KERA Population Cost Cost Costs Weight Weight -------------------------------------------------------------------- Elementary school Speech or language (18. 7) $3,172 $2,398 1.32 0.32 0.24 High incidence (47. 7) $4,756 $2,398 1.98 0.98 1.17 Low incidence (5. 3) $7,511 $2,398 3.13 2.13 2.34 Middle/junior high and high schools Speech or language (0.2) $3,662 $2,250 1.63 0.63 0.24 High incidence (24.9) $4,579 $2,250 2.04 1.04 1.17 Low incidence (3.2) $9,468 $2,250 4.21 3.21 2.34 Overall average Speech or language (18.9) $3,177 $2,324 1.37 0.37 0.24 High incidence (72.6) $4,695 $2,324 2.02 1.02 1.17 Low incidence (8.5) $8,278 $2,324 3.56 2.56 2.34a/ The data in this table do not include non-personnel costs (e. g. , books, supplies and materials, furnishings and equipment, utilities, travel, fees, and school and district buildings). In addition, the costs of transportation and food services are excluded from these calculations.
Sizable differences between these cost estimates and the revenues generated by the State's current three special education funding categories were also observed. Special education funding weights based on the results of this study suggest that students in the speech or language category were underfunded by approximately 50 percent (0.24 versus 0.37), that high incidence students were overfunded by about 16 percent (1.17 versus 1.02), and that low incidence students were underfunded by about 13 percent (2.34 versus 2.56).
4 For a more thorough discussion of the conceptual underpinnings of this methodology, see Levin's description of the "ingredients" approach (1983), or for discussion and examples of its operationalization, see Chambers and Parrish (1982, 1984, 1993).
5 Jay Chambers and Tom Parrish, co-directors of CSEF, created the basic design for the last national special education cost study completed by Moore, et al. , (1988) under a subcontract to Decision Resources Corporation.