ED Response to Public Comments

Study of School-Level Expenditures:

Response to Public Comments

Draft: 10/26/09

Overview of Comments

We received comments on this data collection from 17 individuals, most of whom were state fiscal data providers. Fifteen states submitted comments: California, Delaware, Florida, Illinois, Kentucky, Michigan, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, Utah, Vermont, Virginia, and Wisconsin. All of these states expressed concern about the proposed data collection, specifically about data accuracy and reporting burden. States also requested clarification about specific features of the data collection, including whether this is to be a one-time or ongoing data collection.

We also received comments from two researchers, Karen Hawley Miles at Education Resource Strategies and Marguerite Roza at the University of Washington. These two researchers were supportive of the data collection and made a number of suggestions for technical improvements.

Comments received about this data collection are summarized below, organized around four categories of comments:

• Concerns about data accuracy

• Concerns about reporting burden

• Requests for clarification about specific features of the data collection

• Suggestions for technical improvements to the data collection

The individual comments that were submitted are provided in Appendix A.

Concerns About Data Accuracy

1. Many states and districts do not currently have systems in place to track expenditures at school level, and collecting such data retrospectively is likely to result in inaccurate data. Because school districts have not been required to track local and state funds at a school level in the past, many have not retained that level of detail in their internal accounting records and there is no way to “conjure up” actual accounting records to support this reporting requirement. Auditors perform financial audits at the district level with minimal monitoring done at the school level.

PPSS Response: Although we understand that districts may not have comprehensive data on school-level expenditures, we believe they should be able to identify which staff were assigned to each school and to determine the salary expenditures for each school staff member. The proposed data collection asks districts and states to report other types of school-level expenditures if available, but does not insist that they report data that they do not have available.

2. Many expenditures that benefit schools are purchased and accounted for at the district level, often including such expenditures as professional development, curriculum development, instructional materials and supplies, computers, software, technology, contracted services, and library books and media. Similarly, instructional support and student support costs are often maintained at the district level and there may be no way under current systems to allocate these to the school level.

Also, in some states, regional entities such as supervisory unions play an active role in providing school-level resources. For example, in Vermont some supervisory unions buy entire curriculum programs and other support services for all of their LEAs, and some supervisory unions hire itinerant teachers in subjects such as art and music who serve schools in districts throughout the SUs; the VT DOE has determined that it would be difficult to allocate these expenditures to individual districts, and the problem would be compounded at the school level.

PPSS Response: As stated in Form A, page 3, this data collection is asking districts and states to report school-level non-personnel expenditures such as professional development, instructional materials, computers and other technology, contracted services, and library materials “if this information is available at the school level.” We are not requiring districts and schools to allocate centrally-provided expenditures to the school level if they have not already done so.

Similarly, the data collection does not require that states and districts identify all expenditures made by regional agencies on behalf of schools and to allocate such expenditures to individual schools. We have revised Form A to clarify that respondents may exclude expenditures made by regional educational agencies.

3. Data will not be consistent across states and may also be inconsistent across districts within states. Some states commented that districts develop their own procedures for allocating resources among schools, and the degree of variation may be considerable.

PPSS Response: We expect that these data will not be consistent across states and we do not intend to use the data to make comparisons across states. We recognize that there may be data inconsistencies across districts within a state as well, although we expect that this will be less of an issue than the across-state variation. The data collection form includes a checklist asking districts and states to indicate the types of expenditures included in the data that they submit, which will enable some empirical analysis of this issue.

4. Removing expenditures paid from federal funds is very problematic. States commented that they do not collect links between specific revenues and expenditures at the district level, and previous data collections have not required this, even for the district-level fiscal data collections.

PPSS Response: We recognize that removing expenditures paid from federal funds may be problematic for many districts and states, but this was specified in the statutory requirement for this data collection. The checklist on the data collection form asks districts and states to indicate whether federally-funded expenditures have been excluded from the data they submit, which should enable us to examine this data quality issue empirically.

5. Separating out special education expenditures is also difficult for some states. For example, one state agreed that including special education costs may distort comparisons of school-level per-pupil expenditures, but stated that the time and effort to remove them would more than double the cost of the entire project. Another state commented that they were unsure how to handle a special education expense funded by general fund dollars, something that does happen and is almost impossible to break out. Another state commented that even if special education expenditures could be extracted, this may not give an accurate representation of school level per pupil spending.

PPSS Response: It appears that excluding special education expenditures may be relatively easy for some districts and states and difficult for others. We have revised the language in Form A to clarify that this is requested “to the extent possible.”

Concerns About Reporting Burden

1. The data collection is being conducted with very short notice that was provided after the close of the school year for which data are being collected. Many states complained that the statutory requirement for this study was enacted in the middle of the school year in question and the specific reporting requirements were not drafted until September 2009, well after the 2008-09 school year was over. Some states noted that they have already begun or even completed collecting 2008-09 fiscal data and would have to mount a second 2008-09 fiscal data collection in order to comply with this reporting requirement.

PPSS Response: We understand that states may need to conduct a second 2008-09 fiscal data collection in order to meet the statutory deadline.

2. Reporting burden has been underestimated. Some states argued that the actual cost to produce these data will exceed the burden estimates in the OMB supporting statement. Some commented that they had already conducted their regular fiscal data collection from districts for 2008-09 financial information already, so they would have to undertake an additional data collection that would have to be developed with no new resources. For example, one state commented, “Part of the ability for us to report at the school level is an existing contract for part of this work which far exceeds the estimated costs in the document. Even the excess cost we will incur to comply with this requirement will exceed the estimate in the notice.” Another state commented, “Many school districts will have to resort to pulling individual invoices to try to piece together cost estimates that meet the criteria set out in the draft regulations.”

PPSS Response: After reviewing the comments that states provided about the time needed to respond to this data reporting requirement, we believe the Department’s burden estimates are reasonably accurate. We have estimated an average of 40 hours per district, based on an estimated 8 hours per district to review their data systems, plus 4 hours per school to compile the data for each school, plus 4 hours per district to actually submit the data to their SEA. The estimated average of 40 hours per district is based on a national average of 7 schools per district, but large districts with many schools may require additional time, and small districts with only 1 or 2 schools will need to spend less time compiling the data.

We do not expect that districts will need to pull individual invoices in order to respond to this data collection, because the data collection focuses primarily on salary expenditures and requires districts to submit non-personnel expenditures only to the extent that such data are available. Indeed, districts with good financial data systems may be able to generate the required data automatically without the need to review each school’s records individually. Also, states that already collect school expenditure data may have the necessary data in hand without the need to impose additional reporting burden on their school districts.

The estimated average of 741 hours per state to collect, compile, and submit the data to the Department also appears reasonable. Again, this is an average, and large states may need to take more time while small states should require less time because they have fewer districts. In addition, as noted above, some states may already have the necessary data available in existing state data systems and will need to spend considerably less time responding to this requirement.

We have added a request to Form A for districts and states to provide an estimate of the number of hours that they spent responding to this data collection, so that we may improve the accuracy of our burden hour estimates.

To reduce the burden of this data collection, we are changing the requirement for states to collect and report data for all Title I districts. The statute requires only that states report school expenditure data for districts that receive Title I funds under ARRA, i.e., those districts that receive Title I funds under either the Targeted Grants or Education Finance Incentive Grants formulas. The Department’s original data collection proposal had expanded the requirement to cover all Title I districts, including those that only receive Basic Grants, because of concern that these relatively low-poverty districts might show different patterns in the distribution of school-level resources. However, after weighing that concern against the burden issue, we have decided to ask states to report data just for the Title I ARRA districts, the specific districts covered in the statutory reporting requirement. We expect to be able to examine resource patterns in the lower-poverty Title I non-ARRA districts through the more in-depth five-state data collection in states that already collect school-level and personnel-level data from all of their districts.

3. The burden is compounded by the requirement to report the school expenditure data by March 31. Because existing state fiscal data collections may allow a longer time period for collecting and compiling fiscal data from districts, the March 31 reporting deadline contained in the ARRA may increase the burden of this data collection. For example, one state commented, “Our current contract usually gets the information to us by mid May. I have concerns about the additional work needed to break out parts of the data and comply by the March 31 timeframe, especially since the final regulations have not yet been posted.”

PPSS Response: The statute requires states to submit these data by March 31, 2010.

4. The additional reporting burden comes at a bad time since state agency resources are already stretched due to hiring freezes, employee furloughs, and other ARRA reporting requirements. For example, one state commented, “School district staff have been severely impacted by the economic down-turn. Support staff numbers have been reduced limiting their ability to respond to requests like this that will require hundreds of man hours to complete.” Another state commented, “It is extremely inappropriate to look only at the amount of time required for this task. In a time of rescissions, RIFs and furloughs this drain on our time and the time of school district staffs should be evaluated in the context of the entire set of additional ARRA requirements. For most state departments of education these tasks are unfunded mandates.”

PPSS Response: We understand that this data collection imposes additional burden on districts and states at a time when staffing may have been affected by the economic down-turn. However, this data collection is a statutory requirement under ARRA, and the Department does not have the authority to waive this requirement. In addition, ARRA has provided $10 billion in Title I funds and over $69 billion in other elementary-secondary education funds to assist school districts and states during the economic down-turn, and the amount of burden imposed by this and other ARRA reporting requirements is small in comparison to the amount of funds that are being provided. Moreover, 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 for these programs (such as the school expenditure data collection).

5. The Department should take more time to design this data collection using the procedures that NCES has developed over the years to ensure quality data. States noted that this data collection is a substantial change from the process that NCES uses for new and revised data collections. One state noted that “Notice, traditionally, has been given well in advance and US DOE and the Census Bureau have worked with CCD Fiscal Coordinators to produce quality collections, often introduced over time with those states most able to comply leading the way and offering others the benefit of their experience. The ‘emergency’ process employed in this case is guaranteed to produce unreliable results.”

PPSS Response: The statute requires states to submit these data by March 31, 2010, so we do not have time for a lengthy development process. However, we have designed this data collection to include two features that will enable the Department to examine data quality issues. First, we are asking states and districts to provide information about the characteristics of the expenditure data that they are submitting, so that we can assess the degree to which the specified expenditure definitions were feasible within existing district and state data systems. Second, the study will also collect more detailed data from five states that currently compile personnel-level and school-level data through existing systems, in order to examine the extent to which the aggregate data collected from all Title I districts appear to be consistent or inconsistent with the more detailed data available in the five states. Examining the characteristics and quality of the school expenditure data collected through this study may inform efforts to improve data quality for any future data collections.

6. The costs of conducting this data collection exceed the potential benefits. Several states expressed concern that the inconsistent data that they expect to result from this data collection make the effort not worth the considerable burden on states and districts. One state commented, “Decisions based on inconsistent data are bound to be flawed, and any benefit of having that data cannot possibly justify the cost of producing it.”

PPSS Response: The school-level expenditure data collected through this data collection will be used primarily to examine the extent to which school-level education resources are distributed equitably within school districts, and we expect that each district should be able to report school expenditure data in a way that is consistently defined across schools within the district. We recognize that there may be data inconsistencies across states and to a lesser extent across districts within a state, and we intend to examine this issue empirically through this data collection, in order to inform any future efforts to study the distribution of school-level expenditures.

Requests for Clarification About Specific Features of the Data Collection

1. Will this be a one-time or ongoing reporting requirement?

PPSS Response: The ARRA requirement for states to collect and report these data for all Title I districts applies only to the 2008-09 school year. However, the Department has also proposed to include school-level expenditures in the Civil Rights Data Collection to be collected from approximately 7,000 school districts in 2009-10 through the EDFacts system (see http://edicsweb.ed.gov/browse/browsecoll.cfm?pkg_serial_num=4127) for more information).

2. Does the phrase “all schools” include regional and district centers (Special Education, Alternative Education, Vocational Education, and Governor’s Schools?

PPSS Response: States should collect and report these data for all schools in regular school districts that receive Title I Part A funds. States may exclude regional districts that are not regular school districts, which may include special districts for special education, alternative education, vocational education, and governors’ schools.

3. Should expenditures for Preschool, Adult Education, School Nutrition, Summer School, or be included in the per-pupil amount calculation?

PPSS Response: In general, school-level expenditures to be reported should include salary expenditures and non-personnel expenditures associated with regular K-12 instruction, instructional support, pupil support, and school administration. Adult education programs should be excluded because they are not part of regular K-12 education. School nutrition programs should be excluded because they are generally classified in a separate category from instructional and pupil support and may be accounted for centrally. Districtwide programs or programs that serve students from more than one school should not be included. Summer school and preschool programs may be excluded because these programs are sometimes administered centrally and may use certain schools to serve students from multiple school attendance areas.

We have revised Form A to clarify how districts and states should handle the above expenditure categories, to the extent feasible based on existing data systems.

4. Should expenditures for special education programs from a non-special education or federal fund be excluded? For example, if a district paid a special education teacher or speech therapist from the general fund should that expenditure be included or excluded?

PPSS Response: In general we would prefer that districts and states exclude salaries for special education staff, regardless of funding source. However, we recognize that this may be challenging in some cases, and districts and states have the flexibility to handle this specific issue in the way that is most feasible for them given their existing data systems.

5. Should states and districts report these data only for Title I schools?

PPSS Response: No. School-level expenditures should be reported for all schools in districts that receive Title I Part A funds, including both Title I schools and non-Title I schools.

6. Is this a cumulative look at all of the expenditures for the 2008-09 school year or is it a point-in-time snapshot (like Title I comparability)?

PPSS Response: Districts and states should report all (cumulative) expenditures for the full 2008-09 school year.

7. Is it possible to collect this through the cost report, school financial report, or normal state reporting?

PPSS Response: Because of the statutory due date for states to submit these data for the 2008-09 school year to the Department by March 31, the Department has developed this special data collection to meet that deadline.

8. Form A, Table A-3 refers to an Appendix D, but there is no Appendix D attached.

PPSS Response: The reference to “Appendix D” should in fact refer to Table A-2, i.e., the preceding page of this form.  The form has been revised accordingly.

Suggestions for Technical Improvements to the Data Collection

1. In future data collections, define more clearly what a school-level expense is. In some districts, only 50% of dollars are coded to specific schools, while in others, 90%+ are coded to specific schools. Requiring districts to report a minimum percentage of district expenditures at the school level would improve the quality and consistency of the data so that school-level expenditures can be compared across districts and states.

PPSS Response: The Department will consider changes to future data collections at a later date.

2. Expand the personnel-level data collection over time to include all 50 states.

PPSS Response: The Department will consider changes to future data collections at a later date.

3. Add a “course schedule” component to this data request that shows which students (by protected ID) are associated with which teachers (by protected ID) so that the allocations of time (in English) and talent (better teachers however measured) can be tracked explicitly to student need (or past student performance).

PPSS Response: This suggested change would greatly increase the burden of this data collection by requiring districts and states to submit not only school-level data but also student-level and teacher-level data.

4. The data collection should include FTEs as well as salary expenditure data.

PPSS Response: School-level FTE data are already being collected through EDFacts and will be merged into the school expenditures dataset for analysis purposes.

5. Excluding those staff paid for with federal funds may not be feasible, so the data collection should ask states to expenditures for federal programs where possible and then note whether or not federal funds were excluded.

PPSS Response: We recognize that removing expenditures paid from federal funds may be problematic for many districts and states, but this was specified in the statutory requirement for this data collection. The checklist on the data collection form asks districts and states to indicate whether federally-funded expenditures have been excluded from the data they submit, which should enable us to examine this data quality issue empirically.

6. Some states will be able to meet the requirement relying on state systems without involving LEAs. This should be encouraged.

PPSS Response: States that have already collected the school-level expenditure data through existing data systems will not need to conduct a second data collection to meet this requirement.

7. For states that have more complete data than are being requested here, ask the states to submit their full expenditure files without stripping off any codes. This will enable a more complete analysis and enable future data collections to be designed more efficiently around states' current expenditure systems.  

PPSS Response: The study will collect more comprehensive data from five states. We will explore this option with those states once they have been selected.

8. Augment the analysis plan to examine funding patterns by school poverty quartile as well as the school poverty range categories specified in the Supporting Statement. There are many districts where all schools are Title I, and all schools have more than 75% poverty, but there are still relative inequities (e.g., if schools with 80% poverty receive more than schools with 90% poverty).

PPSS Response: We will consider this suggestion after the data have been compiled and we are able to review the characteristics of the dataset.

Appendix A:

Compilation of Public Comments Received as of 9/29/09

Comments from States

Brian Logwood, Virginia Department of Education

The Virginia Department of Education has reviewed the request for Study of School-Level Expenditures data collection, and has the following concerns:

• The request for data needed for this study was made in February 2009, and the reporting requirements were not drafted until September 2009, well after the 2008-2009 school year was over. Expenditure records were not maintained at the school level in Virginia during fiscal year 2009.

• Federal reports have not required school-level reporting in the past, therefore, the Commonwealth of Virginia and Virginia school divisions are not prepared for a retroactive, school-level data collection.

• Because local school districts were not required to record the school-level data needed for this data collection from the beginning of the school year, the data may not be readily extractable from existing accounting systems within the required time frame; especially considering the fact that the final regulations have not been posted.

• School districts in Virginia budget and account for expenditures on a district wide basis, as the Commonwealth of Virginia and local auditors have never required school districts to track expenditures to the school level. This means many of the requested data elements are not available to report on a school level. This raises the following issues:

• Many of the data elements requested by USED are not purchased at the school level. Rather, they are purchased for district-wide use. These include but are not limited to: professional development, instructional materials and supplies, computers, software, technology, contracted services, and library books and media.

• Generally, support level costs are maintained at the district level because they support the entire district.

• The requirement to remove all federal and special education expenditures from the school level expenditures is an incredibly complex task. In addition, if the federal and special education expenditures could be extracted, using the net expenditure amount in the per pupil amount calculation while including all students in the student count, even those receiving special education and Title I staff services, may not give an accurate representation of school level per pupil spending.

• The cost of creating, collecting, analyzing and correcting potentially flawed school level data for school year 2008-09 will be significant at a time when the state and localities have had to reduce budgets.

• School district staff have been severely impacted by the economic down-turn. Support staff numbers have been reduced limiting their ability to respond to requests like this that will require hundreds of man hours to complete.

• The estimated 40 hours per school district and 741 hours per state response may be unrealistic depending on how far school districts go to accurately calculate the required data. Many school districts will have to resort to pulling individual invoices to try to piece together cost estimates that meet the criteria set out in the draft regulations.

• Requests by the Virginia Department of Education for clarification of instructions from USED regarding this data collection have not been met. The Virginia Department of Education has requested guidance from USED on many issues to help school districts prepare for this new data collection. Detailed questions were submitted in April 2009, with three additional follow-up e-mails between April and August, 2009, concerning this data collection and a response was never received. Therefore, no additional guidance was sent to school divisions during that time to assist them in their preparations. Several questions still remain that the draft regulations do not answer:

• Does the phrase “all schools” include regional and district centers (Special Education, Alternative Education, Vocational Education, and Governor’s Schools)?

• Should expenditures for Preschool, Adult Education, School Nutrition, Summer School, or Non-regular Day School be included in the per pupil amount calculation?

• Confirmation that this is a one-time reporting requirement.

Vaughn Altemus, Vermont Department of Education

If statute requires Vermont to produce numbers, we will produce numbers. I would not use this data set for allocating financial resources within Vermont. I can't imagine using the data set to compare Vermont to other states.

First, notice of this requirement came 6 months into a fiscal year. We are just getting specifics now, after books are closed and districts have submitted reports.

We do not collect financial data at the school level. We do not collect links between specific revenues and expenditures at the district level. Neither the state of Vermont nor federal data collections have required this information.

Districts typically track spending at the school level. I do not believe they track revenues at the school level, which makes removing expenditures paid with federal revenues especially problematic. Districts develop their own procedures for allocating resources among schools. I have no idea how much variation exists, but expect that it is considerable.

Supervisory unions play a very active role in Vermont. The role varies substantially by SU. Some supervisory unions buy entire curriculum programs and other support services for all of their LEAs. Some supervisory unions hire itinerant teachers in subjects such as art and music who serve schools in districts throughout the SUs. VT DOE spent considerable effort trying to find a way to allocate these expenditures to districts. We discovered that producing adequate results would require collecting allocations by SU. The problem would be compounded at the school level.

This request is a substantial change from the procedures developed by the NCES over the years to ensure quality data. Notice, traditionally, has been given well in advance and US DOE and the Census Bureau have worked with CCD Fiscal Coordinators to produce quality collections, often introduced over time with those states most able to comply leading the way and offering others the benefit of their experience. (The Stats DC meeting where the Office of Planning, Evaluation, and Policy Development listened to complaints about the amount of effort and concerns over the quality of data is no substitute for the procedures that have been developed over time.) The "emergency" process employed in this case is guaranteed to produce unreliable results.

The emergency is justified by the need to monitor ARRA money. The ARRA money, like other federal grants, is subject to the requirements of the Single Audit Act. The revenue is also subject to the CBO requirements associated with all ARRA funds. These federal revenues will be the most closely monitored in history without this requirement.

Finally, it is extremely inappropriate to look only at the amount of time required for this task. In a time of rescissions, RIFs and furloughs this drain on our time and the time of school district staffs should be evaluated in the context of the entire set of additional ARRA requirements. For most state departments of education these tasks are unfunded mandates.

I urge the US Department of Education and Census Bureau to request a technical correction from Congress in which this requirement would be deferred so that the usual NCES process can be followed.

James Wells, Nevada Department of Education

We also will be able to report at the school level but are still trying to determine whether we will be able to differentiate between funding sources as well as General and Special Education. We are also unsure about what to do with a Special Education expense funded by General Fund dollars, something that does happen here and something almost impossible to breakout.

Part of the ability for us to report at the school level is an existing contract for part of this work which far exceeds the estimated costs in the document. Even the excess cost we will incur to comply with this requirement will exceed the estimate in the notice.

Our current contract usually gets the information to us by mid May. I have concerns about the additional work needed to break out parts of the data and comply by the March 31 timeframe, especially since the final regulations have not yet been posted.

Lastly, I have concerns about the comparability of the data not only between states but between districts within our state, especially as it relates to costs allocated from the district level to the schools.

Form A, Table A-3, Field 6: No Appendix D in any document.

Form A, Table A-3, Field 6: Should expenditures for special education programs from a non-special education or federal fund be excluded? For example, if a district paid a special education teacher or speech therapist from the general fund should that expenditure be included or excluded?

Burden: We have been informed by the contractor we work with to obtain school level data that the estimated cost is approximately $50,000 for the state level requirements for Form A. This is in addition to the amount we already pay them to collect school information, which is in the neighborhood of $250,000 per biennium. This amount will cover the costs to identify expenditures by revenue source in order to eliminate the Federal dollars. This is significantly higher than the estimate in the notice in the amount of $40,728 for the state and $960 per district.

Our contractor has estimated burden in hours at 2,576 (4 hours for each of 644 schools). This is significantly higher than the burden estimate in hours in the notice which for Nevada would be 1,421 (741 hours for the state and 40 hours for each of the 17 school districts).

Timeline: We generally receive our school level reports in mid May of the year after the reporting period. Finalizing reports by March 31 will cause additional burden at both the local and state level.

David Carlson, Wisconsin Department of Public Instruction

First of all, it should be noted that we believe that only a small portion of the Title I ARRA money was spent in the 2009-09 fiscal year due to the notification timing of allocations. Therefore, you may not have the data for the analyses you intend to undertake for Wisconsin.

The state of Wisconsin does not require school districts to report by school building. Although there is some merit to school building expenditure level data collections, it will be a significant effort for us if truly valid data is desired. Not only will our charts of accounts need be updated, they would need to be updated retroactively to the 2008-09 school year. In a time when Wisconsin state agency resources have been reduced by hiring freezes and employee furloughs, one could argue that it is bad timing given already existing ARRA reporting burden.

Additionally, it will take time to develop the calculations of "indirect" cost as outlined in Chapter 7 of the NCES "Financial Accounting for Local and State School Systems: 2009 Edition." There are many useful suggestions and generic models but it will take time for us to develop a comprehensive cost allocation plan. We believe it to be true that the "devil will be in the details" when it comes to cost allocations.

To summarize, a preferable strategy may be that going forward, NCES collaborate with the Wisconsin Department of Public Instruction and school district business office staff through such organizations as the Wisconsin Association of School Business Officials. The goal would be to develop the expenditure reporting requirements including cost allocation guidelines that meet the federal reporting needs while attempting to minimize the burden on local school district reporting.

Peggy O’Guin, California Department of Education

I hope that USDE will give serious consideration to changing, or at least interpreting, the law to remove the school-level reporting requirement so that LEAs are not required to produce school-level data they do not have. Decisions based on inconsistent data are bound to be flawed, and any benefit of having that data cannot possibly justify the cost of producing it.

Unresolved school-level reporting issues from California’s viewpoint:

• First and foremost, there exists no standardized protocol for uniformly attributing centralized costs to individual schools.

• Arriving at such a protocol would require careful consideration of a number of important issues with the involvement and input of stakeholders at the school, district, county, state, and federal levels. We believe that until we invest the time and resources to develop a standardized protocol, it is important that we give no on-the-fly guidance on the matter. We believe that poorly-formulated guidance developed in haste would be a disservice.

• We also believe that allowing LEAs to develop different methods to attribute costs to individual schools would compromise any uniformity and render the effort and the outcome useless.

• Note that protocols do exist for attributing costs to instructional programs. However, these program cost accounting protocols do not address issues relating to attribution of costs to school sites. Where states or LEAs use program cost accounting protocols to attribute costs to individual schools, their methods, and results, almost certainly vary from one state or LEA to the next.

• In California and in at least some other states, there is also an as-yet-unaddressed mandate issue in imposing a new requirement that LEAs account for their costs at the school level.

• To illustrate just two examples of issues relating to attributing centralized costs to individual schools:

• Centralized curriculum development costs. It would be easy to assume that the costs of developing a particular curriculum could be attributed to individual schools in proportion to the number of students enrolled in each school. However, not all students at each school necessarily benefit from the same curriculum; for example, ESL or Special Education students might not use a particular curriculum at all. Further, one or more schools might not use the same curriculum as the other schools; for example, a particular school operating as a magnet school or as a charter school might use an entirely different experimental curriculum than the other schools within the district.

• Maintenance and operating costs. It would be easy to assume that the costs of keeping schools open, clean, and in a state of repair could be attributed to individual schools in proportion to the square footage of each school. However, not all schools incur maintenance and operations costs proportionally to their square footage. For example, schools in some areas generate a disproportionate share of maintenance costs due to vandalism, or weather conditions, or the age or construction of the facility.

• In the absence of standardized protocols for attributing centralized costs to individual schools, any school-level cost data collected will be inconsistent. Analyses and decisions based on inconsistent data are bound to be flawed.

Glenda Rader, Michigan Department of Education

Auditors perform financial audits at the district level with minimal sample/program monitoring done at the school level. Many districts have not retained school level accounting records.

Title I Comparability studies are primarily performed by program staff and not by trained financial accounting/audit staff. Monitors have not necessarily completed these studies using district accounting records.

The ARRA Title I reporting legislative requirement was not issued until February of 2009. School districts were more than halfway through the fiscal year at that point in time. The draft guidance related to the required level of school level reporting wasn't released until September 2nd, 2009- Well beyond the school's 2008-09 fiscal year end. We still don't have final guidance with the actual required school level reporting elements.

Michigan public school districts have not been required to track local and state funds at a school level in the past and therefore most have not retained that level of detail in their internal accounting records. Michigan has not required the districts to keep the support level costs at a school level because many of the pupil support and instructional support functions are performed at the district and even regional level. (For example curriculum development, teacher consultants, reading recovery, teacher professional development, and instructional technology functions are often performed by Regional Educational Agency Staff with the costs born at that entity level rather than at the local district or school level.) Federal reports have not required states to report school level financial data in the past.

Because not all districts kept 2008-09 school level accounting records related to state and local fund expenditures., there is no way for them to "conjure" up actual accounting records to support this reporting requirement - NO MATTER HOW MUCH TIME USeD GIVES STATES TO PRODUCE THE DATA.

The cost of creating, collecting, analyzing, and correcting flawed school level data for fiscal year 2008-09 far outweighs the benefit. The stated benefit of the data collection is that it would provide data used to determine equitability among school buildings.

If this reporting requirement is not waived by USED, Michigan will be doing a major cost allocation of district level support expenditures to meet the requirement. The data WILL NOT BE accurate.

Von Hortin, Utah State Office of Education

I agree with many of [the points made by Glenda Rader, Michigan Department of Education] and my state will be able to do the school level. The one thing I found missing and will be a problem for Utah is trying to tease out the Special Education costs. I agree the Special Education costs may distort some costs but the time and effort to remove them, in my case, would more than double the cost of the entire project.

The incomparability and inconsistency make this effort not worth doing. More misinformation will come from this effort than benefit.

Ron Leclerc, New Hampshire Department of Education

I couldn't agree more with Michigan's comments. We in New Hampshire have never asked school districts to keep school level finance data. At this point in time any data collected would be considered suspicious if we were to ask them to collect at this level.

Debra Hemberger, Illinois State Board of Education

Even though this would be a hardship of time and cost, Illinois will comply if this requirement is approved in order to meet ARRA compliance.

Difficulties complying with this requirement: As is the same with many states, Illinois' current collection format of district financial data is not compatible for the school-by-school expenditure collection. Illinois currently collects expenditures at the district level, not at the school level. If we collect on an emergency basis, the data would not be audited. Also, Illinois will have to collect from districts the salaries that are paid from federal fund sources and net them out.

Suggested Process: Determine how many school districts receive Title I LEA funds. Determine those schools that receive Title I money and those that do not within those districts. Create a new form asking for the above expenditures by school building. Compile the data by district category and school subcategory and submit to USDE in the acceptable excel format. The estimated time to collect and submit this information to USDE would be approximately 200 hours.

Stephen Burrell-Gwydyr, New Mexico Public Education Department

New Mexico does not maintain school level data. I'm not sure we can get the school by school expenditure data from LEAs either. We collect information by LEA and they have all submitted their 08-09 financial information already. The only way that we could collect this information will be a separate data collection that would have to be developed with no new resources.

Marilyn Peterson, Nebraska Department of Education

We do not see anything in the proposed guidance about when districts should collect this data. Is this a cumulative look at all of the expenditures for the 2008-09 school year or is it a point-in-time snapshot (like Title I comparability)? We think it would help to spell this out.

Karen Conway, Kentucky Department of Education

Kentucky has been providing project numbers to be able to track these expenditures for ARRA. There has not been much direction from the beginning on the Federal level as to the standards that the SEA needed for reporting therefore directions to the districts at the school level may not have been clearly provided and could contribute to potential reporting errors.

We are concerned about the accuracy of the reporting and the additional time that it will take to report this information.

Kuforiji Ade, Delaware Department of Education

Delaware does not maintain school level data. I'm not sure we can get our hands on such data either.

Heather Leary, New Jersey Department of Education

New Jersey does not collect school level data and we have already begun the data collection of the 2008-09 audited financials by LEA.

Richard Birt, Florida Department of Education

We are going to be hard pressed to be able to provide this data. Is it possible to collect this through the cost report, school financial report, or normal state reporting? We have a staffing shortage and this is a definite burden to attempt to provide this data.

Comments from Researchers

Karen Hawley Miles, Education Resource Strategies

We at ERS are pleased to see that you are taking a step toward collecting more and better information about how much ARRA money schools receive.  In the short run, these efforts seem reasonable though we suspect they will ultimately be unsatisfactory in many ways.  In these comments, we focus on a longer-term agenda that we admit freely may not be appropriate for this emergency measure yet is critical for more effective data collection and use.  We encourage USED to consider the following long-term changes and whether any pieces of them might be considered in support of this data collection effort:  

 

1. Define more clearly what a school-level expense is. In some districts, only 50% or so of dollars are coded to specific schools, while in others, 90%+ are coded to specific schools. This can hide a multitude of sins. Requiring districts to report (for instance) at least x% of district expenditures or 100% of certain types of district expenditures to schools (eg:  guidance counselors, itinerant teachers, custodians, utilities, etc) would go a long way toward standardizing this and creating better school-level reports that can be compared across districts and states.

2. You are already getting personnel data from 5 states.  Expand this over time to include all 50 states.

3. Add a “course schedule” component to this data request that shows which students (by protected ID) are associated with which teachers (by protected ID) so that the allocations of time (in English) and talent (better teachers however measured) can be tracked explicitly to student need (or past student performance).

 

Until we get a data set that does all of these things, the field will be very hampered in its ability to see whether school dollars (a) are invested in students who need them most (within and across schools) and (b) are making a difference for specific students.

 

The current data request will help us better understand equity across schools.  But schools are still a pretty high aggregation when we’re most interested in student performance and whether students are getting what they need.

Marguerite Roza, University of Washington

Having done on a smaller scale the kind of data analysis sought in this collection, I have the following suggestions:

1. It is absolutely critical that FTEs be reported alongside salaries, but I don't see the request for FTEs listed in the materials.   If high-poverty schools spend less on salaries, we wouldn't know if it is because their teachers are lower salaried, or if there are fewer teachers at the school.   Salary files always contain FTEs so adding the requirement to report total FTEs for each category shouldn't add any burden.  This is a critical change that needs to happen if these data are to answer intended questions.

2. The reported expenditures are to exclude expenditures from federal programs.  In most SEA and LEA expenditure files, personnel salaries aren't coded by fund source.  In other words, where Title I might pay for 3 teachers at a school, no specific teachers are necessarily designated as the Title I teachers.  Rather, the Title I budget is deducted for the 3 times the average salary.  So excluding those staff paid for with federal funds simply may not be doable.   My recommendation on this would be to exclude expenditures for federal programs where possible and then note whether or not federal funds were excluded.

3. Some states will be able to meet the requirement relying on state systems without involving LEAs. This should be encouraged.

4. Many states will have more complete data than are being requested here.  For these states, I would suggest that they send their full expenditure files without stripping off any codes, etc. RTI can pull the info it needs.  Then researchers can fully explore the available data and the state expenditure structures.  Even more importantly, further data collections can be designed more efficiently around states' current expenditure systems.  

5. In the plan for analyzing personnel-level expenditure data, I'd add analysis of per pupil expenditures by poverty quartile (schools in highest poverty quartile, second highest, etc.). There are many districts where all schools are Title I, and all schools have more than 75% poverty, but there are still relative inequities (meanings schools with 80% poverty get a lot more than schools with 90%).  I can explain this analysis if needed.

17