Loading school data…

Independent vs municipal schools — what does the data say?

The differences between independent and municipal compulsory schools — before and after adjusting for pupil background.

Use this for

Get the main pattern quickly, find a local angle and jump straight to the charts that carry the story.

Data & format

Statistics story · 3 visualisations · copy reviewed 17 October 2026 · Skolverket, Kolada and SCB.

The debate over independent versus municipal schools is one of the most polarised in Swedish education policy. But what does the data say? We take you through three perspectives — geographic distribution, raw merit scores and SALSA-adjusted results — and also set out what the comparison cannot show.

Where are the independent schools?

Independent schools are unevenly distributed across the country. Among active compulsory school units, the independent share is 32% in municipalities with more than 200,000 residents but 7% in municipalities with fewer than 15,000, and more than 100 municipalities have no independent compulsory school at all (school unit register, October 2026).

The map below counts all school forms except preschool — the independent share on the right, and on the left all other school units, nearly all of them municipal. Click around and see how it looks in your county.

Share of independent (right) and other, mainly municipal (left) school units, all school forms except preschool.

Open Independent vs Municipal fullscreen

Merit scores — do independent schools have higher grades?

Comparing raw merit scores in year 9, independent schools as a group have a higher average: 239 against 223 merit points, with medians of 245 against 221. The figures mainly refer to final grades in the 2025/26 school year and are calculated per school unit. The distribution for independent schools is shifted to the right.

But interpret with caution. According to the SALSA data, pupils at independent schools on average have parents with more education, and fewer of them are newly arrived. Part of the difference therefore reflects pupil composition, and the comparison says nothing about which schools teach best.

Distribution of year 9 merit scores per school unit, split by municipal and independent operator.

Open Merit value year 9 fullscreen

SALSA — results adjusted for pupil background

SALSA is Skolverket's model for comparing results with pupil background. It calculates an expected merit score from three background factors: parents' level of education, the share of newly arrived pupils (counting pupils with unknown background) and the share of boys. The deviation between the actual and expected value shows whether results are above or below what the model predicts. Schools with fewer than 15 pupils in year 9 are not included.

In the 2024/25 school year, the groups differ by just over 24 merit points on average in the SALSA sample. That is more than in the section above, mainly because the schools not included in SALSA, among them the smallest, often have low merit scores. This applies especially to independent schools. After adjustment the gap shrinks to 9.5 points: independent schools are on average 6.9 points above the model's expectation and municipal schools 2.6 points below.

Distribution of SALSA deviations per school unit: municipal (orange) and independent (green).

Open SALSA score distribution fullscreen

The conclusion is nuanced: independent schools have higher raw merit scores, and the gap shrinks to less than half but does not disappear when the model accounts for pupil background. The greatest variation exists within each group. One in six municipal schools has a higher merit score than the independent median, and almost a third of independent schools are below the SALSA model's expectation.

Operator type therefore says little about the individual school. Look at the school's own results, teaching staff and pupil environment instead.

What the data does not show

The copy was reviewed on against final grades for the 2025/26 school year, SALSA for 2024/25 and the school unit register as of October 2026. The charts load the latest published data and may show a newer period than the copy.