East Greenwich School Committee
School Building Oversight Committee


Enrollment Forecast 2022-2031

From the 12/08/2021 Demographic Study presentation

School YearPupils

The Cohort-Component Method Forecast

The McKibben forecast uses a different methodology than previous enrollment projections (the Cohort-Component Method)1

Expected future populations of age-gender cohorts are computed using a base-year population combined with age-specific fertility, mortality, and migration rates.

These are combined with historical enrollments by grade.

The required datasets are:

  • The 2010 Census population for the attendance area
  • A set of age-specific fertility rates for the district
  • A set of age-specific mortality rates for the district
  • A set of age-specific migration rates for the district
  • Historical enrollment figures by grade

Actual enrollments for this graph were obtained from the RI Department of Education 2 and the U.S. DOE Public Elementary/Secondary School Universe Survey Data 3

Cohort Survival Ratio Method Projections

All previous projections used the Cohort Survival Ratio method
  • Dr. McKibben described this as "the only game in town" until the mid-1990s
  • Not a Forecast, but projections based on mechanical extrapolations of observed enrollment counts
  • Very sensitive to noise in the birth counts
  • Long-term projections are evidently not very stable

East Greenwich Birth Count History 1853-1930 and 1990-2020

  • 1853-1930 counts obtained from public birth records at East Greenwich Town Hall 4
  • 1990-2020 counts obtained from various enrollment projections
  • Interestingly, the data shows sharp drops in births, usually around wars and economic downturns
  • There are no really big positive increases that are not recoveries from downturns

A Probability Model for Recent EG Births

A negative binomial probability model fits the last 31 years of births quite well.
  • The theoretical interquartile range is 102 to 119. We expect the number of births to fall in this range 50% of the time. In the data, 15 of 31 data points are in this range.
  • The theoretical 95% probability interval is 87 to 137. We expect the number of births to fall in this range 19 times out of 20. In the data, 2 of 31 data points are outside this interval. We expect 1.5.
  • In this model, the expected number of births is neither increasing nor decreasing. The fact that it fits the data pretty well indicates there is no strong evidence for a trend up or down in births.

Modeled Birth Count Probabilities

This is a histogram of the birth count probabilities for a single year.