ED Response to OMB Comments

Study of School-Level Expenditures:

Response to OMB Comments

Draft: 10/26/09

1. We appreciate that ED is excluding special education expenditures and including non-personnel expenditures as discussed over the summer, and we recognize the limitations in addressing these issues. What are the implications of excluding special education expenditures “if possible”? Will the data vary by district within States? ED points out that some of the biggest variations are between districts within a state. Will this limit between-district comparisons? Same questions for “non-personnel expenditures at the school level (if available)”?

PPSS Response: Regardless of whether we offer states and districts the flexibility to exclude special education expenditures “if possible” and to report non-personnel expenditures at the school level “if available,” it seems likely that the data we collect will not be consistent across states and may also be inconsistent across districts within a state. Indeed, states that submitted comments in response to the Federal Register notice expressed numerous concerns about the consistency of data that would be obtained through an emergency data collection conducted retrospectively on very short notice. See, for example, the comments submitted by Virginia, Vermont, California, and Michigan.

Certainly the data would be more valuable if they were consistently defined across all school districts, but accomplishing this would a much greater amount of lead time in order to work with states and districts to revise their accounting systems to define, track, and report this information in a consistent manner. Because this is a retrospective data collection that is asking for data for a school year that has ended, we will have to collect the best data that districts and states are able to provide based on their current accounting systems. If we were to assume that states and districts will be able to retrospectively report school expenditure data using the exact definitions that we specify, and do not give them the opportunity to provide information on the definitions that they actually used, then we might still have data that are inconsistent but we would have less information about the nature of those inconsistencies. We think a more useful approach is to recognize the challenges inherent in this first-time-ever national data collection on school-level expenditures, specify the definitions that we hope districts and states will use, and collect information on the extent which they do or do not follow those definitions.

Because of the concerns about the consistency of the data across states and also across districts within a state, the most valid use of these data will be to examine the distribution of school-by-school expenditures within districts, where definitional inconsistencies should be minimal, if any.

2. ED has made a decision not to include LEA functions? Does this apply to LEA staff that serve multiple schools? How difficult is it to apply school-related district expenditures on a school-by-school basis? What is lost in not including expenditures on LEA functions?

PPSS Response: It is extremely difficult if not impossible for districts to retroactively allocate school-related district expenditures to individual schools. If fiscal records were not already kept on a school-by-school basis for such expenditures, they would not have records to suballocate the expenditures to individual schools.

3. How will ED choose the five states for a more in-depth analysis? Will ED take into account the extent to which States collect school-level data or will it be random? Will these states be able to spend additional ARRA funds on this effort?

PPSS Response: The selection of the five states will not be random; we intend to select five states that already have school-level and personnel-level data available. We have identified six potential states that appear to have at least some of these more detailed data available; we intend to first contact these six states to determine more specifically the types of school-level and personnel-level data that each has currently available. If we find that there are not five states from among this group of six that can provide the data we are seeking, then we will need to investigate whether there are other states that could provide such data.

Because this component of the data collection is intended to collect data that already exist in state data systems, we have estimated an average of only 18 hours per state to respond to this portion of the data collection (compared with 741 hours per state for the main “universe” data collection) – see page 13 of the OMB Supporting Statement, Part A. Therefore, we do not believe states would need to spend additional ARRA funds on this component of the data collection. However, the Department has proposed adjusting the statutory cap on State administration under Title I, Part A to help defray the costs of data collections that are specifically related to ARRA funding, and the five states would be able to use this to cover this supplemental data collection, if necessary.

4. Is there a common definition of what to include in non-personnel expenditures at the school level? ED lists professional development, instructional materials, computers, contracted services, and library books/media center materials. Is that a common and complete list?

PPSS Response: Our approach attempts to provide a common definition, but districts will not necessarily have school-level expenditure data available for these items. The list of possible types of school-level non-personnel expenditures is based on our previous work in studying resource allocation. If reviewers would like to suggest additional expenditure categories that should be considered, we would be happy to discuss this further.

5. The five State analysis will compare school per pupil expenditures by poverty quartile. Can any of this kind of analysis be done with the national data set?

PPSS Response: We do not intend to conduct national summary analyses of the school-level expenditure data by national poverty quartile, because it is unlikely that the data will be consistent across states and may also be inconsistent within states. Ensuring consistent data would require a much longer development process that should include substantial opportunities for state and district input.

6. In the five State analysis, would ED consider additional comparisons of Title I vs. non-Title I schools, for example by poverty quartile and across the state, as opposed to focusing primarily on within district comparisons?

PPSS Response: We intend to use the data collected to examine the consistency and comprehensiveness of the school-level data. The data collection form includes a checklist asking states to indicate the types of expenditures included in the data that they submit, which will enable some empirical analysis of this issue. We can also examine total reported school-level expenditures as a percent of total district expenditures; if this percentage varies considerably within or across states, that may suggest that the school-level expenditures were not consistently defined. If this analysis finds that the school-level expenditures appear to be consistent within a state, then it may be warranted to use the data to conduct analyses within a state as a whole, such as the within-state poverty quartile analysis that you have suggested. We think it is better to defer a decision on this issue until we have the data in hand and are able to examine the apparent consistency of the data.

7. How does ED expect to report the data – in aggregate, or also by subcategory, such as teacher expenditure per child, by poverty level of school, etc?

PPSS Response: At a minimum we expect to be able to report on total salary expenditures for school-level personnel, as well as on the two personnel subcategories of instructional staff and teachers. In addition, for those districts that report on non-personnel expenditures, we will be able to also examine and report on total personnel and non-personnel expenditures, as well as non-personnel expenditures alone. Further information on our data analysis plans is provided on pages 7-8 of the OMB Supporting Statement, Part A. The analyses described in that section could be repeated for each school expenditure category.

8. If ED did not have to worry about reporting burden, what additional components would you propose to include in a school-level per pupil expenditure calculation to most effectively reflect whether school resources are distributed equitably?

PPSS Response: The concern is not just reporting burden but also data quality and consistency. The proposed data collection will collect vastly more detailed information than has previously been available on school-level expenditures. We believe it would be most prudent to examine these data and then consider whether a further expansion of the data collection is needed.

3