Experimental Design Details
To analyze parents' preferences for school attributes, we will conduct a survey experiment in an online survey that is representative in terms of language regions. The survey will be carried out by a professional survey institute among Swiss residents who have compulsory-school-age children in 2024.
Given that students in Switzerland are assigned to schools based on their place of residence, we assess parents' preferences for i) school performance, ii) the share of students not speaking the school language, and iii) the share of students with special needs by calculating their willingness to pay a rent premium to be able to reside in the desired school district. Using a choice experiment, we ask respondents to imagine they are moving and must choose between two apartments that differ in price and school district and, thus, in the school to which their child will be assigned. The respondent must answer which of the two alternatives they prefer.
The alternatives differ by the following attributes:
- Monthly rent of the apartment (quantitative, 2 levels): +/- 10% of their current rent.
- Ranking of school in terms of academic schooling outcomes (quantitative, 3 levels): Best 25, Average, Bottom 25
- Share of students not speaking the school language (quantitative, X levels): 10%, 20%, 30%
- Share of students with special needs (quantitative, X levels): 10%, 20%, 30%
We create different choice sets that differ in attributes. Respondents will be randomly allocated to the choice sets of one of the three blocks. We stratify block randomization on language region.
We will implement a soft launch and, if necessary, revise the attribute levels after the soft launch. In the case that we need to adjust the attribute levels, we will not use the data of the soft launch for the main analysis.
The primary outcomes of interest are the coefficients corresponding to our four attributes and, in the case of school ranking, its interaction with the share of students not speaking the school language, and the share of students with special needs. With the dependent variable being the likelihood of an alternative being chosen, we estimate the previously mentioned coefficients using mixed logit models (MXL).