Democracies depend on citizens who are willing to take part, and the attitudes that shape that willingness often begin to form in adolescence. Knowing which factors make young people more likely to expect an active political life could help schools and policymakers nurture the next generation of engaged citizens.
New research from Germany suggests that teenagers’ confidence in their own civic abilities is one of the strongest signals of whether they expect to be politically active. Surprisingly, however, students with greater civic knowledge tended to have lower expectations of such involvement.
The study, published in Psychology International, was conducted by Purya Baghaei, Nurullah Eryilmaz, and Rolf Strietholt of the International Association for the Evaluation of Educational Achievement in Hamburg. They analysed data from 1,487 students in the German state of Schleswig-Holstein who took part in the International Civic and Citizenship Education Study in 2022.
The students were around 14 years old on average. The researchers used an explainable machine learning technique known as Random Forest to examine 31 possible predictors, ranging from age, gender, and family background to civic knowledge, attitudes, and experiences at school.
They tested three versions of their model. The most complete version explained about 43% of the variation in students’ expectations, but much of that accuracy came from closely related measures, such as whether students expected to vote or take part in legal political activities.
When those overlapping measures were removed, a clearer picture emerged. In the most restrictive model, citizenship self-efficacy, meaning students’ belief that they can take part effectively in civic affairs, ranked as the strongest predictor, followed by beliefs about the importance of conventional citizenship and then civic knowledge.
Willingness to take part in school activities was the leading predictor in the intermediate model. Civic learning at school also became more prominent once the most closely related participation measures were set aside.
The most unexpected result concerned civic knowledge. Although it consistently ranked among the most important predictors, higher knowledge scores were linked with lower expectations of active political participation, a pattern confirmed by several different analyses.
This contrasts with earlier international research, which has generally found positive links between civic knowledge and expected participation. The authors say the reason cannot be determined from their data, and they suggest the specific outcome measured or the Schleswig-Holstein context may play a part.
They also stress that the finding does not show that learning more about politics causes young people to disengage. The data were collected at a single point in time, and the outcome reflected students’ expectations rather than their actual behaviour.
The researchers add that Random Forest offered only a modest advantage over simpler statistical models, and in one version a simpler method performed slightly better. Its real value, they argue, lies in helping researchers see which factors matter most and how they relate to the outcome.
Because only two German states took part in the 2022 assessment, the results are not nationally representative. The team calls for replication in North Rhine-Westphalia and other countries, alongside longitudinal research to explain the puzzling role of civic knowledge.
