Student Enrollment Forecasting, Explained

How We Prepare Enrollment Forecasts

At FLO we use a variety of data sources, analytical tools (e.g., Esri ArcGIS, Alteryx), and methods to prepare enrollment forecasts. In addition to enrollment data from school districts or state education agencies, data sources include U.S. Census and American Community Survey demographic data, live births from state departments of health, regional or state-published population estimates and forecasts, and residential land use information provided by planners from local jurisdictions.

District-wide

After compiling the necessary data, our first step is to produce district-wide enrollment forecasts by grade for each year of the forecast horizon, as these forecasts represent the largest population and are, therefore, expected to be the most reliable. To achieve this, we prepare forecasts by individual grade, starting with kindergarten (K). We use several years of live birth counts by age of mother and K cohort (e.g., September to August when K eligibility is based on age five on September 1st) from state health departments and project future births by applying expected fertility rates to projected five-year age group populations of women of childbearing age.

After compiling historical births, K-to-birth ratios are calculated by dividing K enrollment by births five years earlier. These ratios represent a combination of net migration of district residents between birth and age five and the district’s K capture rate, which is the share of residents enrolled in district K classes. Recent observations influence assumptions for future K-to-birth ratios, but factors such as recent and expected trends in net migration (due to accelerating or slowing housing development); recent trends in or specific knowledge of changes in private, home school, or charter school market share; and any recent or expected changes to district policies may necessitate adjustments to baseline ratios. 

Forecasts for grades 1–12 rely on GPRs—the rates at which student cohorts progress to the next grade level (e.g., the ratio of 3rd graders in a given year to 2nd graders the prior year). During the COVID-19 pandemic, districts often saw GPRs for most grades below 1.00, accounting for families choosing options other than public schools. Depending on the district, several years of more stable GPRs following the COVID-19 disruption may be used as a baseline for GPR forecasts. Still, similar to K-to-birth ratios, adjustments may be needed to account for known or expected future events, trends, or policies.

Resident Forecasts

Districts often plan at a more granular level than district-wide enrollment, so the next step in our process is to develop forecasts of student residents by grade for school attendance areas, also known as catchment areas, as well as the number of students that enroll with a district but live outside its geographic boundary. These resident forecasts are prepared using many aspects of the district-wide approach but with a more granular analysis of recent historical enrollment trends and current and planned residential housing development, leveraging our expertise in Geographic Information Systems (GIS) software and spatial data analysis. Sums of initial forecasts will likely differ from the forecasts of district-wide enrollment; therefore, final forecasts for attendance areas and out-of-district residents are derived by proportionally adjusting initial forecasts to match the district-wide forecasts by grade, a process referred to as controlling.

School Forecasts

Although resident forecasts are essential for understanding transfer rates and modeling attendance area boundary changes, enrollment forecasts for individual schools are typically more valuable to districts for capacity and staffing planning purposes.

Starting with elementary schools, we prepare initial forecasts for new K classes at each neighborhood elementary school using ratios applied to the final forecasts of K residents in the schools’ attendance areas. Initial enrollment forecasts for grades 1–5 use GPRs unique to each school, with students added as needed for new residential development. Forecasts for entry grades for secondary schools initially use GPRs based on the historical relationship between the entry grade and each school’s feeders, while initial forecasts for other grades use GPRs specific to each school and grade. Each K share and GPR is reviewed and adjusted if necessary to account for outliers that may influence the average rates or to ensure that the school’s total enrollment forecast remains consistent with its resident forecast. Initial school forecasts are then controlled to match the district-wide neighborhood school forecasts by grade.

Many districts offer alternative educational pathways, such as dual language schools that serve the entire district (and do not have attendance areas) and whose enrollment is less dependent on the district’s demographic trends and more influenced by the program's popularity or the facility’s capacity. We forecast these schools in collaboration with district staff to ensure that capacities, program changes, and class sizes are considered.

Results

The resulting forecasts are valuable resources for school districts undertaking planning efforts, from projecting staffing and capacity needs using school enrollment forecasts to school siting and attendance area boundary changes using resident forecasts. Forecasts for geographic areas smaller than school attendance areas may even be developed to facilitate boundary realignment or to establish boundaries for new schools. Depending on district needs, forecasts can also be prepared for individual grades or custom grade group configurations, such as modeling change from a K–8 structure to K–5 and 6–8 schools.