Voucher systems have been vastly discussed on their ability to provide better quality education and more efficient school systems, but also on the role they play on segregation mechanisms through education markets. With free school choice, parents' preferences and the trade-offs they take into account become relevant, as they are free to decide between educational alternatives and are not restricted to their geographical location's school assignment. In this context, the Chilean case is a singular one due to its nationwide, unrestricted, system implementation. Chilean household's preferences have shown to be complex, heterogeneous, and tightly bound to socioeconomic attributes, as well as deeply intertwined with residential location, especially in Santiago, being a source for social and spatial inequities distributed around the city. Comprehending household's sensitivities on both of these choices and their interactions is essential for understanding the city's complex urban and social structure.

This paper seeks to understand the existing bond between school and residential location choices through discrete choice models. By constructing a dataset based on Santiago's 2012 travel survey, a latent class mixed logit model was estimated, capturing observed and unobserved factors that determine household's preferences on both choices. The latent class approach defines two different choice strategies: households that prioritize residential location, and households that prioritize school choice. This enables characterization of households' preferences and choice structures.

Main findings suggest both strategies are identifiable through household's characteristics, capturing heterogeneity in household's preferences. In addition, unobserved factors linked to access to office opportunities and neighborhood's socioeconomic level in both school and residential locations explain the correlation binding both choices, capturing heterogeneity. Results suggest that, for both segments, distance tends to be a relevant attribute when choosing. Also, households that prioritize school over residential location tend to be more elastic to socioeconomic characteristics in comparison to households that prioritize residence over school. These ‘school first’ decision-makers are characterized by higher incomes, possession of cars, and more residential mobility due to not being homeowners. Understanding this bond, and household's behavior regarding it, should allow policymakers to better plan for the reduction of social and spatial urban inequalities.

代金券系统在提供更好的教育质量和更高效的学校系统的能力以及它们通过教育市场在隔离机制中所发挥的作用方面得到了广泛的讨论。通过自由选择学校，父母的偏好和他们考虑的权衡变得相关，因为他们可以自由地在教育选择之间做出决定，并且不受其地理位置的学校分配的限制。在此背景下，智利案例因其全国性、无限制的系统实施而成为一个单一案例。智利家庭的偏好已被证明是复杂的、异质的，并且与社会经济属性密切相关，并且与居住地位置密切相关，尤其是在圣地亚哥，这是分布在城市各处的社会和空间不平等的根源。

本文旨在通过离散选择模型了解学校和住宅选址之间的现有联系。通过构建基于圣地亚哥 2012 年旅行调查的数据集，估计了潜在类别混合 logit 模型，捕获了确定家庭对两种选择的偏好的观察到和未观察到的因素。潜在阶级方法定义了两种不同的选择策略：优先选择居住位置的家庭和优先选择学校的家庭。这可以描述家庭的偏好和选择结构。

主要研究结果表明，这两种策略都可以通过家庭特征来识别，捕捉到家庭偏好的异质性。此外，在学校和住宅区，与获得办公机会和社区社会经济水平相关的未观察到的因素解释了两种选择的相关性，捕捉到了异质性。结果表明，对于这两个部分，距离往往是选择时的相关属性。此外，与将居住优先于学校的家庭相比，优先考虑学校而不是居住地的家庭往往对社会经济特征更具弹性。这些“学校优先”的决策者的特点是收入更高、拥有汽车，以及由于不是房主而导致的更多住宅流动性。了解这个纽带，