It has been demonstrated that there is a strong relationship between occupant behaviour and the thermal performance of dwellings. At the same time, some aspects of this behaviour, especially with respect to natural ventilation, constitute some of the most important sources of uncertainty in the field of building energy simulation. A survey about perception of thermal comfort and occupant behaviour was carried out in Santiago de Chile during December 2009 and January 2010 in a pilot case study corresponding to an apartment building. This paper proposes a methodology based on the systematic application of multivariate statistical techniques which were applied to the collected data of the survey. The results of the analyses show that daytime ventilation is not strongly correlated to the perception of thermal comfort, probably because it is mainly oriented to hygienic purposes. On the contrary, nigh ventilation appears as a very significant predictor for the same dependent variable. The final objective of these models corresponds to the definition of occupant behaviour profiles which can be used as hard data to make calculations of thermal behaviour of dwellings more accurate and reliable. At the light of the results of the energy building simulations, night ventilation presents a high potential as passive cooling technique, considering also the climatic conditions of Santiago de Chile.