← All stories

Survey vs. outcomeseasy

When teachers want to leave, students stop showing up

The question

When teachers at a school report low trust and morale, do their students attend school less?

Across New York City schools, teacher-voice survey scores (would-recommend, trust, school climate) move together with chronic-absenteeism rates. Schools in the weakest fifth of the teacher-voice ranking (about 60% positive responses) average 43% chronic absenteeism. Schools in the strongest fifth (about 97% positive) average 33%. The 10-percentage-point gap climbs smoothly across each step from bottom to top — every quintile is higher than the one above it. Whether the strong teacher culture causes the better attendance, or engaged students cause teachers to feel better about the school, or a third factor (principal, building, demographics) drives both is something this data alone can't separate.

Read the full briefing

Data analysis

Investigation run 2026-05-31.

TL;DR

Schools in the weakest fifth of teacher-voice scores (about 60% positive) average 43% chronic absenteeism. Schools in the strongest fifth (about 97% positive) average 33%. The 10-percentage-point gap climbs smoothly across every step of the ranking. Whether one causes the other is the open question.

Teacher voice ranking vs. chronic absenteeism (2023-24, 1,503 schools)

Each step up in the teacher-voice ranking corresponds to 2 to 3 fewer percentage points of chronic absenteeism. The gradient is smooth — no group of schools breaks the pattern.

Chronic absenteeism by teacher-voice ranking (bottom fifth = weakest teacher culture, top fifth = strongest)
Bottom fifth (60% positive)
42.8%
Fourth fifth (77% positive)
40.5%
Middle fifth (85% positive)
37.0%
Second fifth (91% positive)
36.6%
Top fifth (97% positive)
32.5%

Which causes which?

Three plausible directions, none distinguishable from this data: strong teacher culture causes student engagement, which causes attendance; schools with engaged students give teachers more reason to feel positive; or a third factor (principal quality, building, demographics) drives both. A useful next step would be to test whether last year's teacher-voice score predicts next year's attendance, which would point toward the first direction.

Adjusting for student-body challenge

Schools in the weakest teacher-voice fifth also tend to serve higher-need student populations (higher challenge-group rankings on the 1-to-10 scale). Part of the gap is confounded by that. A challenge-group-adjusted regression would shrink the apparent effect — though national research suggests the school-culture-to-attendance link survives most controls. Worth running locally.

What's next

Run a regression where last year's teacher-voice score predicts this year's attendance, to test the direction of the relationship. Cross-reference with the city Department of Education's teacher-turnover data (a separate dataset).

Methodology & replication recipe — the data sources, queries, and steps behind this analysis.