What Parents Want – Applying Machine Learning to Predict Preferences and Support for School Choice Policies in K-12 Education

School choice policies allow parents to act on their preferences when choosing a school for their child(ren). These policies, however, can lead to planning problems for policymakers . We contribute to a better empirical understanding of parental school preferences and the factors shaping these preferences. Using survey data (n = 4,574), we contrast five machine learning algorithms with logistic regression. We illustrate the predictive superiority of machine learning algorithms and that parents’ attitudes on time spent on standardized tests, and school funding are the most decisive predictors of their preferences and support for school choice policies.