Skip to content

School location

At a glance

Engine Inside SILO (Java), every simulated year
Code landuse/schools
Models Three multinomial logits: K–8, high school, college
Sensitive to Car time, attendance boundaries, enrolment and capacity, school composition
Produces The school of every student

Every simulated year, inside SILO, students are assigned to schools by three multinomial logits over the zones with enrolment of the student's band:

Band Zones Observations
K–8 1,144 1,509
High school 719 668
College 141 673

Each model has a size term in which public enrolment inside the home's attendance boundary is a separate component, car time, boundary terms and the own-race share of the zone's enrolment of the band. The K–8 and high-school models add the share of enrolment in charter or magnet schools, and the K–8 model the test-score quality of the zone's public schools (with a weaker effect for incomes under $40k). The attendance boundary dominates public K–12 choice: an in-boundary seat weighs about 470 times an out-of-boundary public seat in K–8 and 72 times in high school.

Band Parameters Hold-out log-likelihood per student
K–8 13 −1.953
High school 6 −1.810
College 9 −2.789

The own-race share is recomputed every year from the current students; quality and the charter-or-magnet share are fixed inputs. Enrolment is an upper limit, enforced by shadow prices, and student status is aligned to the ACS enrolment rates. Full estimates: School location.