Применение метода условных деревьев решений к моделированию выбора школы родителями
In the article we suggest a method for analyzing situations of choice, which is novel for sociology - the method of conditional decision trees. We describe the logic of the method, applying it to the case of parental choice of school in two urban districts of Saint-Petersburg. We show that decision trees work well for detecting groups that follow different decision making strategies. This can be an efficient tool for modeling and interpretation of the logic of choice. The method outperforms the traditional modeling by means of logistic regression, as it allows for assessing homogeneity of preferences (choices) of the groups detected, instead of simply finding the key factors related to choice. We recommend combining regression analysis with decision trees modeling in all kinds of academic and applied research studying complicated choices.