The second major type of study is an experimental or intervention study. In one or more buildings or spaces, the ventilation rate is set sequentially at two or more values and the health outcomes are measured at each ventilation rate. Typically, a few days or a week elapses between the change in ventilation rate and the health assessment. Much of the potential confounding is eliminated with this type of study; for example, personal, job, and most building characteristics are unchanged when ventilation rates are modified. However, there is still a possibility of residual confounding by some parameters, such as indoor temperature, outdoor conditions (temperature, pollen, daylight hours) or job stress, which may vary among the experimental periods, as well as bias from various effects of being studied: placebo effects or repeated questionnaire effects, (Menzies, 1993a; Jaakkola, 1995). Most confounding and bias effects, however, may be in either direction. The stronger experimental studies measure potential confounding factors and control for them in analyses and also simultaneously study a similar control group which is expected to have comparable levels of confounding factors, but experiences no change in ventilation rate. The changes in health outcomes in the control group are used to ‘‘correct’’ the measured changes in the experimental group. Repeating the experiment several times in a single group reduces the limitations of a study without a control group.