Explore how attendance, community adult BA+ rates, and income interact in a weighted school-level model.
Scope note: this preview currently runs the income-based specification. Poverty-aware companion analyses are published in Evidence Lab findings.
Model in use: participant-weighted linear regression with pairwise interactions (Education x Attendance, Education x Income, Attendance x Income), with 95% uncertainty intervals for scenario and delta.
Interpretation note: this simulator is associational, not causal. It reflects patterns in aggregated school data.
Private preview route
This page is currently private. Open it with ?preview=1 to use the prototype.
Each view shows the same evidence-based statistical model of associations, with different visual emphasis on how education, attendance, and income relate to predicted test outcomes.
Curves show predicted outcome versus education, with separate lines for low/median/high attendance. The income slider shifts all lines, so you can examine three-way interactions across education, attendance, and income.
Color scale range: —
Tip: in heatmap mode, click a point to set scenario education/attendance.
Rows used i
—
Participants i
—
Baseline prediction i
—
Scenario prediction i
—
Scenario delta i
—
Local sensitivity i
—
Uncertainty (95% CI) i
—
Confidence intervals reflect model-estimate uncertainty from aggregated school data; they are not individual-level prediction bounds and do not imply causality.