|Title||Essays on the impact of education on misclassified civic outcomes: studies of Italy and the UK|
In this dissertation I examine the impact of education on a range of civic outcomes in Italy and the UK which embody two of the main dimensions of social capital: civic engagement and social trust. The central aim of this thesis is to attain a credible re- lationship between education and civic outcomes, accounting for diverse issues which may obscure it. Namely, unobservables driving education choices (i.e., endogeneity), and the tendency to under-report sensitive topics and over-report civic opinions (i.e., misclassi�cation). This approach allows me to ascertain the extent to which the
causal e�ect of schooling on the civic indicators is either genuine or is driven mainly by endogeneity and a systematic misreporting by educational levels. I also investigate how these elements vary by contextual factors of the two countries. The contribution in this area is given by utilizing data from these two countries, considering a distinct group of civic outcomes (i.e., civic opinions and civic behaviours) and by dealing with misreporting. Previous research does not explicitly control for misclassi�cation and focuses on civic engagement, one aspect of social capital. Furthermore, I con- tribute by introducing a hurdle ordered probit with misclassi�cation to account for two issues regarding the distribution of a self-reported ordered outcome, its skewness and its misclassi�cation. The main �ndings are: (i) for Italy, qualitative overall con- clusions regarding the causality of education on civic outcomes are indeed a�ected when accounting for misclassi�cation: education turns out to be insigni�cant across civic behaviours, (ii) for the UK, on the contrary, education has signi�cant positive e�ects on all civic outcomes due to upward biases induced by endogeneity, (iii) both Italy and the UK, however, do not di�er substantially overall with regards to misre- porting: most civic outcomes are misclassi�ed for either country, and misreporting is more severe for civic behaviours due to a larger in
uence of social desirability.