Does the school perform better or worse than expected?
The SALSA model (the National Agency for Education's tool for Local Correlation Analyses) calculates an expected merit value for each school based on three background variables: parental education level, share of recently immigrated pupils and gender distribution.
The chart places each school as a point where the x-axis shows the model-calculated (expected) merit value and the y-axis the actual. The dashed diagonal represents perfect prediction — schools exactly on the line perform precisely as the model predicts.
Points above the diagonal have a positive deviation: they perform better than what the pupil composition predicts. Points below the line have a negative deviation. The spread around the diagonal shows that the model explains roughly half of the variation (R² = 0.527) — the other half depends on factors such as leadership, teaching quality, resources and other circumstances.
The colour distinguishes municipal (orange) and independent (green) schools. Look for systematic patterns: does one group consistently cluster above or below the diagonal? The answer says something about whether operator type matters when you control for the pupils' backgrounds.
Statistics: academic year 2023/24. Source: Skolverket open data, Kolada and SCB. Processed by Skolkoll. Glossary · About the data.