Presents the results of simulations of test houses using a nodal simulation infiltration code. The two detached houses are three-level family houses consisting of a basement where technical rooms and instruments are placed; the first floor, which is the test level; and a loft. The paper also presents the fractional factorial experimental design used to estimate parameter and interactive effect coefficients; by varying input parameters, such as air- tightness and wind speed, for example, it is possible to make obvious which are the critical parameters and what level of confidence is required for input parameters to guarantee a given confidence in the output. The capability of the sensitivity analysis and the mean age-of-air to serve as tools for air quality evaluation are discussed. Comments on parameter interaction, as well as recommendations for a more general sensitivity study including several types of buildings are also provided.

KEYWORDS: Analysis, models, calculating, air infiltration, housing, air quality, testing, indoor, computer programs.