The increased opportunity to choose one’s school of preference has been raised as a key factor in many countries to promote equal opportunities and a higher quality of education. This has been endorsed by policymakers who assume that students make well-informed rational choices and that students stress only academic quality when deciding which school to attend. If this is true, it will benefit schools of high academic quality, rendering improved school quality overall. To date, little research has examined the validity of these assumptions despite the profound effects they have had for changing the school systems in many countries. This article describes an experiment to investigate the relative importance of factors in school choice. The aim is to test the validity of the theoretical assumptions that guided the school choice reforms in Sweden. Specifically, we draw on experimental data from prospective upper secondary school students in Sweden to contrast the principal school quality attributes behind the policy change (knowledge reputation and program feasibility), with the presence of friends and geographical attributes such as distance, location and accessibility, while controlling for individual characteristics. The results have important policy implications as they show that geographical factors are highly relevant for school choice preferences. In contrast to the reforms’ intentions, geography appears to have become more important than ever before. The findings thus reveal significant flaws in the assumptions that motivated the school choice policy reform.