Uster uses algorithm to create more balanced school classes

The Swiss public broadcaster SRF reports that the municipality of Uster is using an algorithm to distribute children more evenly among schools and classes, drawing on anonymised tax data to help identify social inequalities.

Children attend nine primary schools and numerous kindergartens in the Uster school district. However, some schools have particularly high proportions of children who speak other languages or come from socially disadvantaged families, which can make everyday teaching more difficult.

“Children’s backgrounds and where they live are decisive for their educational development,” Patricia Bernet, Uster’s primary school president and a city councillor for the Social Democratic Party, told SRF at Pünt school.

Uster has used the computer-assisted allocation system for two years to distribute children from different social and linguistic backgrounds as evenly as possible among schools and classes. Oliver Dlabac, founder of the research company Ville Juste, developed the tool. The system is now also used in other municipalities in the cantons of Zurich, Thurgau, St Gallen and Lucerne.

The system first considers available places and the distance to school, followed by social integration. It does not use information about individual families. Instead, it works with anonymised data for entire sections of streets or neighbourhoods, which indicate the proportion of socially disadvantaged households in an area.

The approach is based on what education researchers call a “tipping effect”. Analyses of PISA data and education data from Zurich suggest that learning conditions can deteriorate when the proportion of children facing multiple disadvantages in a class or school reaches 30 to 40 per cent.

“Above this threshold, it is not only the affected children who face challenges,” Dlabac said. The performance of other pupils could also suffer. The system is therefore intended not to move individual children, but to prevent a strong concentration of social and linguistic disadvantage at particular schools.

After two years, Uster’s school authorities have drawn a positive interim assessment. Classes are now more evenly composed than before, Bernet said, and social and linguistic differences can be distributed more effectively.

The issue remains sensitive, especially when children enter kindergarten. Parents regularly question school distances, class compositions and the allocation of their children, and objections have been filed in some cases.

Short and safe journeys to school remain the highest priority, Bernet said. Social integration is considered only after all other criteria have been met. She said criticism often came from parents with strong educational backgrounds, who feared their children would have to travel across the city to achieve social integration. She rejected that concern.