The headline
The conventional wisdom that "urban schools underperform rural schools" or vice versa is wrong in both directions. NCES locale codes (the federal system that classifies every public school as urban, suburban, town, or rural) show that average outcomes within each category are more similar than the rhetoric suggests, and that within-category variance dwarfs between-category variance.
What the data says
Across the US public school universe (NCES 2024-25), average student-teacher ratios by locale type are:
- Suburb (large): 16.3
- City (large): 16.3
- Rural (fringe): 15.8
- Rural (remote): 12.8
- City (mid-size): 17.2
- Town (fringe): 15.9
The largest gap is not urban-vs-rural but rural-fringe vs rural-remote: schools more than 25 miles from an urbanized area have meaningfully smaller staffing ratios, reflecting the structural floor on staffing a small, isolated school.
On test outcomes
State-assessment proficiency comparisons are harder because each state uses a different scale. But within-state, the locale effect on proficiency is small once you control for free/reduced lunch share. A high-poverty suburban school performs similarly to a high-poverty urban school of comparable demographics. A wealthy rural school performs similarly to a wealthy suburban school.
The honest summary: locale type is not a strong independent predictor of school outcomes once you condition on the student population the school serves.
Why the framing fails
Education policy debates frequently invoke "rural schools" or "urban schools" as monolithic categories. The data does not support that framing. Resource allocation, teacher recruitment, and curriculum decisions are better made at the school level than at the locale-type level.
Methodology
Source: NCES Common Core of Data 2024-25 locale codes (the 12-category system: 11/12/13 City, 21/22/23 Suburb, 31/32/33 Town, 41/42/43 Rural by remoteness). State assessment proficiency joined where available.