In the past 6 years, over 1,000 tracer gas measurements have been made to determine the air exchange rates in the EPA Indoor Air Quality Test House. The major factors that affected the air exchange rate include: weather conditions, operation of the heating and air-conditioning (HAC) system, and closing or opening the interior doors. An empirical model that correlates air infiltration rate to indoor/outdoor temperature difference and outdoor wind speed can predict the air exchange rate with an average predictive error of 10%. The model also makes it possible to construct a continuous air exchange profile, with which the diurnal and annual patterns can be examined. With the HAC system on, the air exchange rate showed diurnal and annual patterns can be examined. With the HAC system on, the air exchange rate showed a moderate increase as compared to that with the system off. In a striking contrast, the interior doors had much greater impact on air exchange: under similar weather conditions, closing all interior doors doubled the air exchange rate when the HAC system was on. Type III mixing difficulty was observed in the test house. Results from extensive computer simulations showed that, for tracer decay method, air exchange rate calculated by using the data in the first hour and after 4 or 5 hours contained much greater error than that calculated by using the data in between. Recommendations are made for minimizing such error.