Skip Navigation

Collection of Private School Finance Data: Development of a Questionnaire


In 1993, Sherman and O'Leary noted that the lack of data on expenditures by private elementary and secondary schools compromised the ability of the United States to submit complete and accurate finance data to the international Indicators of Education Systems (INES) project of the Organization for Economic Cooperation and Development (OECD). In a paper describing gaps in finance data, they concluded that:

    The absence of comprehensive data on private elementary and secondary school finances represents a critical gap in information both for domestic policy and for international comparisons. There are, however, data collections both in the United States and abroad that provide models for a private school finance collection. (Sherman and O'Leary, 1993).

Since that time, staff from the Pelavin Research Center have worked closely with various members of the private school community, including association representatives, individual administrators, and academic researchers, to develop methods for collecting private school finance data. As described below, initial efforts focused upon the expenditure data collected by the associations themselves. A second line of inquiry explored the feasibility of collecting data through a new instrument developed with the assistance of private school administrators and association representatives. The concluding task was to refine the proposed instrument, through continued consultation with private school representatives and through field testing in a variety of private school settings. These efforts are described below.

Estimates of Expenditures Based on Data Collected by Private School Associations

In 1994, the (NCES) asked the Pelavin Research Center to investigate the extent to which private school associations routinely collect finance data; and whether or not the associations' surveys provided an adequate basis on which to generate valid national estimates of private schools' expenditures.

Three associations were identified as regularly collecting expenditure data from member schools - the National Catholic Education Association (NCEA), the Lutheran Church-Missouri Synod (LCMS), and the National Association of Independent Schools (NAIS). Garet, Chan, and Sherman (1995) attempted to develop national estimates of private school expenditures by estimating expenditures for various sectors of the private school universe. This task was hampered, however, by a lack of finance data for religious schools other than Catholic and Lutheran schools, and for nonsectarian schools not affiliated with NAIS. After deriving alternate estimates of total operating expenditures ranging from $16.4 to $17.7 billion, Garet et. al. concluded that:

    NCES cannot obtain precise national estimates of private school expenditures by relying solely on data provided by private school associations; most associations do not collect data on school finance (Garet et al., 1995).

In a second study based on the associations' data, Garet, Chan, Isaacs, and Sherman (1997) attempted to refine the national estimates further by linking expenditure data collected by the private school associations with information on school characteristics collected as part of the NCES Schools and Staffing Survey (SASS). They found that several school characteristics are clearly associated with per-pupil expenditures, including teacher salaries, the teacher/student ratio, and the ratio of support staff and maintenance staff to teachers. However, the models were not able to fully explain the large differences in per-pupil expenditures among the three sectors of private education under study: Lutheran day schools, NAIS day schools, and NAIS boarding schools. It therefore was concluded that:

    to obtain improved estimates of the total amount spent by private schools in the United States, it will be necessary to collect data on expenditures directly from a national sample of private schools - either through the SASS or through a new special-purpose study. While much can be learned about the determinants of expenditures through models of the kind estimated [in this report], such modeling efforts cannot, we believe, substitute for the collection of new data (Garet et al., 1997).

Strategies for Collecting Finance Data from Private Schools

Concurrently with efforts to link the private school associations' expenditure data with the national data on school characteristics, Pelavin Research Center began consulting with members of the private school community, to determine the feasibility of directly collecting finance data from private schools through a nation-wide survey. Private school association representatives attending a Private Schools Meeting sponsored by NCES in November of 1994, expressed interest and support for such an undertaking. After an initial review of the scant literature on private school finance, and after telephone interviews with private school researchers, Isaacs, Garet, and Sherman (1996) conducted 3 focus groups and 16 site visits in the fall of 1995 - consulting with a total of 28 private school administrators about their school budgeting and accounting practices, as well as about their views concerning a possible national data collection effort. Three major conclusions regarding the design of a national strategy for collecting financial data emerged from the literature review, focus groups, and site visits:\3

  • Any questionnaire must be sensitive to a great deal of variation among private schools. Some private schools are essentially autonomous organizations that operate as free-standing, not-for-profit or proprietary institutions, while others are deeply interdependent, sharing revenues, expenditures, and services with closely linked "parent" organizations (e.g., local churches). Schools also differ in terms of size, administrative capacity, sources of revenue, and components of expenditures.
  • Any data collection effort must be based upon the premise that schools vary in accounting practices. Some schools, especially larger schools and those affiliated with NAIS, tend to share accounting categories and sophisticated accounting systems. Others, especially smaller schools and schools affiliated with churches, have idiosyncratic systems that vary from school to school. In addition, the school principals, business officers, and bookkeepers at private schools vary in their technical expertise. The diversity in accounting practices and expertise makes the choice of a common reporting framework quite challenging.
  • Most of the private school administrators expressed an initial skepticism about the value of a national collection of finance data. However, once engaged in conversation, administrators noted several potential benefits of a private school finance survey, ranging from the direct benefit of receiving finance information that could help an administrator compare his or her school to "similar" schools, to the indirect benefit of educating the public about private schools. The initial resistance and ultimate interest suggest that future work in developing a financial data collection strategy must involve close consultation with representatives of the private school community.

Three preliminary survey instruments were appended to the final report submitted to NCES in June of 1996. All three questionnaires used terminology familiar to private school administrators; were sufficiently flexible to incorporate schools with diverse accounting practices; and collected expenditure data that could be put on a sufficiently common footing, to permit the reasonable comparisons of data collected for various types of private schools, as well as comparisons between public and private schools.

Two of the alternate versions of the preliminary questionnaires were designed primarily to obtain valid data on total operating expenditures for private schools. The third version was designed to collect much more detailed data across a simplified version of the "function by object matrix" used in the National Public Expenditure Finance Survey. That is, private school administrators were requested to report expenditures across four major sets of functions: (1) instruction-related activities; (2) administration; (3) the physical plant; and (4) other services, and across the three major object categories of salaries, benefits, and supplies and purchased services. The private school association representatives assembled at the Private Schools Meeting convened by NCES in March of 1996 endorsed this third preliminary questionnaire as the preferred alternative, because of the value of the detailed data that could be collected under this approach.

The remainder of this report concerns the development of this third preliminary questionnaire into the Private School Finance Survey questionnaire presented in Appendix A. The first step was to clarify, in a meeting with NCES officials in September of 1996, the scope of the project - and, more specifically, issues related to survey administration. Second, a meeting of a Private School Finance Survey Technical Work Group was convened in November of 1996, bringing together NCES officials and private school association representatives to critique the proposed questionnaire and to discuss the overall benefits and costs of collecting finance data. Third, two separate pilot tests of successive versions of the survey were conducted in January and March of 1997. The first two steps are discussed below; the third step is discussed in section II.

Issues of Survey Administration

Before we could proceed with refining the survey instrument, certain decisions had to be made regarding the overall data-collection strategy. For example, was the goal to be limited to collecting estimates of total expenditures, in order to answer the first two sets of questions laid out on the first page of this report, or was the goal to collect more detailed information about spending across different functions and objects to address a much broader range of research questions? Should it be a universe survey or sample survey? Was it to be a mailed questionnaire or an in-person field-collection instrument? Was the survey to be free-standing, or linked with the existing Schools and Staffing Survey? Staff from Pelavin Research Center and NCES met at the outset of this final task in the project to discuss these issues and to resolve a number of other issues related to survey administration.

Goals and Scope of the Survey

The first key decision made in September of 1996 was to attempt to collect sufficiently detailed finance data to answer questions about how private schools allocate resources (that is, how much is spent on instruction vs. administration, how much on salaries vs. equipment, and so forth). In other words, NCES officials concurred with advice of private school association representatives to base the final instrument on the third preliminary version of the questionnaire, rather than upon either of the two simpler alternates. The many analytical advantages of collecting data across different categories of expenditures were judged to outweigh the disadvantages of increasing the burden of respondents, who would be required to consult their financial records and write down actual expenditure amounts across a number of items. The goal guiding the development of the questionnaire was, therefore, to design a document that was simple enough to encourage administrators to complete it, yet complex enough to collect data of sufficient accuracy and level of detail to address a multitude of questions related to the cost, efficiency, resource allocation decisions, and sources of support for private schools.

Sample Size

A second set of issues relates to sample size. On the one hand, a small sample size is preferred, in order to reduce costs and burden both to the Federal government administering the survey and to the individual administrators who must take time from their many responsibilities to respond. On the other hand, a large sample size, or even a collection from the whole private school universe, is needed, if the data are to be analyzed and reported for fine-grained subsets of private schools.

More specifically, NCES could administer a private school finance survey to a Fast Response Survey System (FRSS)-type sample of 800-1000 schools to get reliable national estimates. The larger Schools and Staffing Survey (SASS) (sample of 3,360 schools in 1993-1994) would yield estimates across 19 affiliation groups, as well as national estimates. Finally, the Private School Universe Survey (PSS) could provide a sampling frame for a universe study or for drawing a larger sample than the SASS sample.

The repeatedly stated opinion of private school administrators is that, because of the diversity of private schools, the data must be reported by type of school, if the data are to be meaningful. For example, association data suggest that per-pupil expenditures range from $2,200, in Lutheran day schools, to $8,300, in NAIS day schools, and $19,200, in NAIS boarding schools.\4 Private school administrators told us that overall averages across such different types of schools would not be very meaningful to administrators in any of the types. Their preference is that the estimates be made for many different sectors of the private school universe, not just national estimates. They are not fully satisfied with the level of detail provided in the estimates from SASS, despite the fact that some estimates are provided across 19 affiliation groups. Estimates that involve cross-tabulations of school type by grade level, or school-type by size are usually done according to a nine-category typology of schools, and the nine categories in the NCES typology are too broad to permit most schools to compare themselves with what they define as "like" schools.\5 Ideally, they would prefer that the data be reported by association, and in even more fine-grained categories, such as "elementary Lutheran schools," or "NAIS boarding schools."

The difficult choice of an appropriate sample size has not been fully resolved at this time. Although SASS does not have a large enough sample of private schools to permit reporting of school expenditures in the detailed categories sought by some private school associations, it does permit analyses by at least nine sectors of the private school universe, as well as some analyses by association. This makes the SASS sample of higher value than a FRSS-type sample, which might be limited to providing estimates across all private schools. At any rate, the expense of mounting a survey across a sample size larger than SASS may be prohibitive. One possibility is for NCES to administer the survey to a nationally representative sample of schools, through SASS or another mechanism, while making arrangements for supplemental administration of the survey by associations that wish to sample all of their members. This would provide participating associations with the more detailed analyses they desire. Both NCES officials and some association representatives have indicated enough preliminary interest in such a strategy that it merits further consideration.


Because of the interest in securing national estimates and estimates for particular sectors within the private school universe, if possible, most of our efforts focused upon designing a questionnaire that could be administered across a fairly large sample of schools, which, because of cost constraints, resulted in a mailed questionnaire. Some administrators in the focus groups and site visits, however, noted how much easier it is to respond to a face-to-face interview than to fill out a paper survey. A few administrators suggested that in-depth case studies or field studies of a few private schools would provide data of more use to administrators than national summary statistics on private schools. Another alternative that could be explored is the use of computer technology - such as providing respondents with questionnaires written on diskettes, to facilitate the reporting of accurate data elements that sum to reasonable totals and are checked, as entered, for obvious mistakes.

Linkages to Schools and Staffing Survey

As suggested by the discussion of sample size, one of the possible vehicles for survey administration is the Schools and Staffing Survey. The advantages and disadvantages of linking the survey directly to the SASS were explored at the outset of this final task. From the viewpoint of the education researcher, a finance data set has much higher value, if it is linked to school characteristics - and, ideally, to educational outcomes. At a minimum, any analysis of the finance data requires good information about enrollment (number of children served). In addition, it is useful to have data about staffing (numbers of staff, and information about salaries), types of benefits provided to staff, the curriculum, and the types of programs and services offered (e.g., whether or not the school provides special education services), and school organization. These types of data allow analysts to view the quantitative data on school finance within the broader context of school type, size, staffing, curriculum, organization, and so forth.

There are, however, potential disadvantages to adding the finance survey to the Schools and Staffing Survey. Of primary concern is whether adding questions on finance to the SASS will lower the overall response rate. This could become an issue simply because of the sensitive nature of finance questions and the length of a finance survey that collects data by function and object. Two factors, however, may alleviate this concern. First, the finance survey could be administered separately from the SASS, as much as one year later. That is, if the SASS is administered in the fall of 1999, with questions about school characteristics in 1999-2000, the finance survey should probably be administered in the fall of 2000, when financial records of actual expenditures for 1999-2000 are available. In this way, the finance data would cover the same school year as the data on school staffing and characteristics. Furthermore, the negative effects of a potentially low response rate to the finance questions would not contaminate the overall SASS administered a year earlier. Second, the private school association representatives have been quite supportive of a finance data collection. If such support is enlisted for the finance survey, school administrators will be much more likely to complete the questionnaire.

After carefully weighing the pros and cons, a tentative decision was reached in September of 1996 to proceed with a finance questionnaire that would be linked to the SASS. The major implication of this decision was that the finance survey appended to this report contains no questions about school enrollment, staffing, or other school characteristics. Additional questions would need to be added to the finance survey, if it were to be revised to serve as a free-standing questionnaire. Linkage to the PSS would allow access to enrollment data, although the time needed to draw a new sampling frame might cause a lag of two or three years between the PSS enrollment data and the finance data. The PSS, of course, would not provide the same level of schools and staffing data as in SASS.

Time of Year for Survey Administration

Almost all administrators interviewed in the focus groups and site visits thought that late October or early November would be the best time to fill out a finance survey. The hectic activity around school opening is over by late October, and the financial reviews of the preceding fiscal year are generally completed by that time. The majority (four-fifths) of the 28 schools in the focus groups and site visits operated on a July to June fiscal year. Most of the remaining schools had a fiscal year that began on August 1st; one school's fiscal year began on September 1st, and one began January 1st, following the calendar fiscal year of the sponsoring church. The NAIS finance survey is due around November 1st, and business managers of NAIS schools thought it preferable to do both surveys at the same time.

Interests and Concerns of the Technical Work Group Regarding Finance Data

The collection of data on private school income and expenditures is of interest to various groups, including private school administrators and teachers, the parents of students enrolled in private schools, education policy makers and researchers. The second major step in developing the final questionnaire was to draw together a Technical Work Group of representatives of private school associations and research analysts from NCES to represent some of these constituencies. At its December 5, 1996 meeting, the Technical Work Group expressed interest in seeing the development of a Private School Finance Survey in order to:

  • Gather information about aggregate spending on private school education, thereby drawing a more accurate picture of total spending (public and private) on elementary and secondary education spending in the United States;
  • Learn, and make the public aware of, the true cost per child of a private school education. Knowledge of accurate per-pupil expenditures is important for determining the financial needs of private schools, for helping private school administrators and parents compare spending in their school with spending in similar schools, for making a comparison of per-pupil expenditures in public and private schools, and for informing discussions of vouchers for private school education;
  • Identify how resources are allocated in private schools - that is, how much of the budget is allocated to instruction, administration, and so on. Such information could be used by private school administrators to compare themselves to similar schools, and to help find more efficient ways to administer their own schools. It also could be used to compare the resource allocation decisions made in public schools to those made in private schools, and to compare the efficiency of public and private schools;
  • Improve our understanding of the role of resources in education. To the extent that expenditure data can be linked with other aspects of private school organization (e.g., information regarding services provided, curriculum, and student achievement), these data can permit both school staff and policy makers to assess the implications of alternative resource allocation strategies; and
  • Improve upon the surveys currently administered by private school associations.

Members of the Technical Work Group also identified three major concerns regarding a private school finance survey.

Of primary concern is the administrative burden on respondents - particularly administrators in small schools who may carry a teaching load in addition to their administrative responsibilities. If the administrative burden is too high, large numbers of respondents may not complete the survey, resulting in a low response rate. A related concern from the perspective of NCES is that, as discussed above, the finance survey will add to the overall respondent burden of SASS - and by doing so may reduce overall response rates to the SASS. Representatives of private school associations advised that the success of the survey will depend upon how it is presented to administrators, and what is done to persuade them that it is worthy of their time. These same representatives warned that response rates are likely to be particularly low among schools that do not belong to any association.

A second major concern is that a government-sponsored finance survey may be viewed as "government intrusion," inconsistent with the "independent nature" of many private schools. Private school administrators have strong concerns about the confidentiality of the data and the uses to which they will be put. Again, the extent to which the survey is endorsed and supported by the private school associations may be critical in determining how it is viewed by individual administrators.

Finally, it is difficult to gather data that are truly comparable across schools. Differences in terminology can be as basic as defining "school," "income," and "expenditures." For example, are income and expenditures from preschool programs considered "school" income and expenditures? Furthermore, the inter-related financial relationship of many private schools with sponsoring religious organizations increases the difficulty of gathering data in way that is meaningful. Private school representatives were concerned that if the data were not collected and interpreted carefully, misleading comparisons could be made between different types of private schools, and between public and private schools.

In conclusion, the Technical Work Group on a Private School Finance Survey was eager to see the development of a private school finance survey, believing that the potential benefits outweighed the costs. At the December 5th meeting of the Technical Work Group, it was decided that the goal of the survey is to collect information that is relevant to the private school community, researchers, and policy makers, and is accurate and comparable between diverse types of private schools, and between private and public schools. A further goal is to design a questionnaire that minimizes the burden of school administrators by being short, clear, oriented toward the terminology of private school administrators, and attuned toward the diversity of private schools in the United States.

Efforts to meet these goals involved drafting four versions of a Private School Finance Survey: the November 1996 version, which was presented for review to the Technical Work Group; the January 1997 version, which was sent to eight schools in the first pilot test; the March 1997 version sent to nine schools in the second pilot test; and the May 1997 version, submitted with this report in Appendix A. All four versions share an overall framework. Schools are asked to report actual income and expenditures for their most recently completed fiscal year. Introductory items guide respondents toward a common definition of "school income" and "school expenditures" by clarifying the treatment of preschool programs, programs operating outside the regular school (i.e., extended-day programs and summer camps), and financial aid. A relatively straightforward item collecting information on income, by source, follows. The next several items collect information on various categories of expenditures, discussed in more detail in Section III. Finally, the survey concludes with an item on non-cash contributions, such as donated computers, publicly provided transportation, and free space. The ability of various types of school administrators to complete such a questionnaire is discussed in Section II; and details about individual items are discussed in Section III.

3/See Isaacs, J., Garet, M., and Sherman, J., Strategies for collecting finance data from private schools. NCES Working Paper No. 96-16, June 1996, for a full report of these activities.

4/Garet et al. (1997), p. 11.

5/The nine categories in the NCES typology are Catholic parochial, Catholic diocesan, Catholic private, conservative Christian, other religious (affiliated), other religious (unaffiliated), regular non-sectarian, special

IntroductionPrev Contents NextMajor Findings of Pilot Tests