Further. ATUS does not allow to test one of the main predictions of the model: low types are more responsive than high types to macroeconomic conditions. The information on ability contained in the survey is very limited. Therefore. the next. section of the paper investigates the cyclicality of education by looking at PSE enrollment from CPS. On one hand. This dataset has several advantages over ATUS. It is available for a longer period of time and it allows to distinguish between ability types. On the other hand. CPS does not include information on the time spent studying, making the quantitative comparison between the model and the data difficult. Nevertheless. since there is a correspondence between the time devoted to education and enrollment decisions. analyzing enrollment is still useful to test the qualitative predictions of the model. If economic contractions encourage PSE enrollment. the time that individuals spend studying will necessarily increase.