We integrate latent attitudes of individuals into a mode choice model through latent variable and latent class models with the help of psychometric indicators that enable us to measure these attitudes. The aim of the inclusion of attitudes is to better understand the underlying choice preferences of travelers and therefore increase the forecasting power of the choice model. We first present an integrated choice and latent variable model where we include attitudes towards public transportation and environmental issues explaining the utility of public transport. Secondly we present an integrated choice and latent class model where we identify two segments of individuals having different sensitivities to the attributes of the alternatives resulting from their individual characteristics. Calibration of these types of advanced models on our sample has demonstrated the importance of attitudinal variables in the characterization of the population heterogeneity.