Every year, secondary school graduates have to choose a university. This is a crucial decision for their future trajectories made under imperfect information. Therefore, quality indicators like university rankings and an excellence competition may provide valuable information for choosing a university. This paper analyzes whether prospective students in fact use quality indicators as a source of information within the application process and whether the influence of the indicators differs with respect to various quality dimensions - e.g. research quality, mentoring, faculty infrastructure, student assessment and excellence status. Therefore, I estimate the effect of different quality indicators from a German university ranking and an excellence initiative run by the German government on the university application decision of high-ability students. As identification relies on the variation in ranking indicators and excellence status over time, I can disentangle the effect of the additional information provided by the rankings from the common knowledge regarding university quality.

This paper contributes to the existing literature by studying the influence of ranking indicators with respect to various quality dimensions, while at the same time controlling for overall university attractiveness. I use a very comprehensive, administrative data set provided by the German central agency ("ZVS") administering the application process for medical schools. The data set contains individual information on all applicants at German medical schools for the years 2002-2008. The evaluation of the excellence initiative shows that in the course of the competition the share of applicants increased on average by 19% at the winning universities, which are today known as "excellence universities". The results regarding the different ranking indicators suggest that the non-research dimensions "student-professor ratio", the number of "clinic beds", and the "students' satisfaction" rather than the research-related indicators influence university choice of high-ability students. This may seem counterintuitive, but is plausible as research quality seems to be common knowledge within the group of high-ability students. In this case, the research related ranking indicators do not provide any additional information. Hence, university rankings are in action if they add new information to the common knowledge of university quality. Yet, the different quality indicators influence prospective student's university choice only to a moderate extent. Distance between a student's hometown and the university remains the most powerful determinant of university choice in Germany. Nevertheless, providing information on all quality dimensions separately instead of publishing university rankings in aggregated league tables widens the basis of information and thus supports a well-informed university choice. This in turn could reduce drop-out rates, increase human capital production and (depending on the social welfare function) may also increase overall welfare.

Furthermore, multidimensional rankings also induce incentives for the top research institutes to not only invest in research but also in the non-research quality dimensions such as mentoring, faculty infrastructure and the overall satisfaction of their students.

Keywords

higher education, university choice, college admission, conditional logit