SALSA (Skolverkets Arbetsverktyg för Lokala SambandsAnalyser — Skolverket's Tool for Local Correlation Analysis) is a regression model that compares schools' actual results with what can be expected given the student composition. The model answers the question: "How does the school perform compared to other schools with similar students?"
Background variables
The model takes into account three variables:
- Parents' education level — by far the strongest predictor (β ≈ 0.71)
- Proportion of recently immigrated students — students who have immigrated within the past four years
- Gender distribution — proportion of boys at the school
What does SALSA show?
SALSA does not show how "good" a school is in absolute terms. It shows how the results look relative to the conditions the school has. A school with a low merit value but a positive SALSA score performs better than expected — it "lifts" its students. See model-predicted merit value and residual for more details.
Explanatory power
The model explains approximately 53% of the variation between schools' merit values (R² = 0.527 for the 2024/25 school year). This means that nearly half of the differences are explained by other factors — the school's leadership, teaching quality, resources and other things that the model does not capture.
Practical use
SALSA is updated annually in January and covers schools with at least 15 students in year 9. Skolkoll uses SALSA data on each school page to show whether the school's results exceed or fall short of expectations.