Education Spending

Read this before the first chart: the per-pupil figure on this page is derived, and it is not the state’s published per-pupil figure. The six spending charts that follow are built from two statewide sources — the CT Office of Policy and Management’s Municipal Fiscal Indicators and OPM’s estimates of statutory aid — plus US Census income estimates on one chart. Four further charts sit below those six: they count students and reported positions rather than dollars, they come from the Connecticut State Department of Education’s EdSight reporting rather than from either OPM dataset, and they carry their own note where they begin. What the two OPM sources do and do not support:

• Per-pupil spending here is a division this site performs, not a figure the state publishes. It is a town’s reported education_expenditures divided by its reported school_enrollment_average, both from the same OPM dataset. The CT State Department of Education publishes its own per-pupil measure, Net Current Expenditures Per Pupil (NCEP), which excludes categories such as capital, debt service and certain grant-funded costs and is compiled on a district basis rather than a town basis. NCEP is not carried in this repository and is not shown anywhere on this page. The two measures are built differently and will not agree; a figure from one is not a substitute for a figure from the other.
• The denominator is average membership, not a headcount. school_enrollment_average is the average number of students on the rolls across the year, not a single-day October enrollment count of the kind the state Department of Education reports. A per-pupil figure built on one denominator will not match a per-pupil figure built on the other.
• Regional school districts change what a town’s education number covers. Where a town belongs to a regional district for some or all grades, its reported education spending covers the grades it runs itself, and the regional district’s spending is reported separately. Its per-pupil figure is therefore not on the same basis as a town that runs its own district across every grade. Several towns in the peer sets on this page are in that position, and the charts do not adjust for it.
• These are audited actuals across all governmental funds, not the adopted budget. The education figure includes state and federal education grants that flow through the town’s books and that the town’s own budget never appropriates. It is a different, and usually larger, number than the education appropriation in a town’s own adopted budget, and the education share shown here is a different measure from the education share of an adopted general-fund budget.
• Coverage. The six spending charts cover every fiscal year the state publishes town by town; the exact span appears here once the data loads. The four charts below them come from a different source on a school-year basis and cover a span of their own, named in their own note.
• Two peer groups, each summarized by its median. The surrounding towns are the municipalities sharing a land border with this town. The District Reference Group (DRG) is the group of Connecticut school districts the state judged to serve similar student and community populations. The median is the middle value of the group in each year, taken across the towns’ own figures; it is less sensitive to one unusual town than an average, and it is not necessarily the same town from one year to the next. The subject town is excluded from its own group medians. The District Reference Groups were last assigned in 2006 and have not been redrawn since, per the state’s own assignment document, the CSDE Research Bulletin: District Reference Groups, 2006.
• The two state datasets on this page cover different years. The spending and enrollment figures end in the most recent fiscal year OPM has released town by town; the Education Cost Sharing figures are forward-looking estimates covering a separate, later set of fiscal years. Each chart names the years its own numbers come from rather than assuming they line up.

The four charts below count students and positions, not dollars, they come from a different source than the six charts above, and they count students on a different basis: the student figure below is the state’s single-day October 1 headcount, while the six spending charts divide by average membership across the school year, from the OPM dataset. The two measures give different student counts for the same town in the same year. Every figure on this page uses one measure or the other, and no figure on this page mixes them. These four charts are built from the Connecticut State Department of Education’s EdSight reporting — the enrollment dashboard and the FTE staffing report — and were collected one school year at a time, because the reports’ own trend view returns only a few years at once. They are reported by school year rather than by fiscal year, so their years do not line up with the spending charts above, and no chart on this page places the two sources on one axis.

• Coverage and collection. The school years covered, and the collection method the file records for itself, appear here once the data loads.
• The source’s own caveats, carried word for word: these appear here once the data loads.
• Which district each chart covers. The districts this file carries for this town appear here once the data loads. Staffing is reported by the district that employs the staff. A town that sends some grades to a regional school district therefore does not have those teachers and administrators in its own district’s count — they are counted in the receiving district. Every chart below names the district each line covers, and no chart adds two districts together.
• Grouping the assignment categories is this site’s editorial decision, not the state’s. The state reports positions under assignment categories whose names are its own and whose wording has changed over the years covered. Two charts below combine those categories into named groups so the legend stays readable; the groups, and the exact source categories in each, appear here once the data loads. A different grouping would produce different lines, and the groups are this site’s, not the state’s.