?
Регрессия с категориальными предикторами: критика применения фиктивных переменных и логлинейный анализ как альтернативный подход
The focus of this article is the methodological aspect of protest behavior determinants identifying; specifically variants of handling with categorical (nominal and ordinal) predictors which hypothetically explain this or that protest potency level. Thus, when using regression for explaining the protest potency levels one may transform categorical predictors into dummy variables. Such a solution makes the model bulky and generates troubles with assessing this model. The authors suggested loglinear analysis as the alternative to the mentioned solution. There search’s aim was parallel implementing the two mentioned methods and comparing them on the base of i) the suggested a priori criteria and ii) gained empirical results. The raw data were extracted from Elektoral'naya Panel 2011–2012of WCIOM. The dependent variable was Protest Potential; the hypothetical predictors belonged to the socio-economic bloc of the Panel. The results show that there are serious statistical reasons to reconsider methods of determinants identifying when working with categorical variables, and loglinear analysis worked better comparing the quality and the set of determinants.