Industrial hygiene air sampling and evaluation of controls for air contaminants is reviewed. At an operation where air contaminants are well controlled, there should be no significant difference in concentration between the worker's exposure, area samples collected near the emission sources, and background area samples collected away from emission sources. In an industrial environment, air contaminant concentrations are usually quite variable. To cope with this variability, the air sampling to evaluate the effectiveness of a control system should be carried out as a designed experiment. The first step is to formulate issues to be resolved, such as whether a hood contains the air contaminant in question, whether an uncontrolled emission source increases contaminant concentrations, whether the worker's involvement with a process that is a contaminant source elevates his exposure, and whether work practices affect exposure. Once the questions needing resolution have been specified, the experimental design and statistical analysis can be specified and the samples can be collected. To determine sample size, it is considered adequate to use power calculations for a t- test comparison of means, as if there were only two groups of data instead of a larger experimental design with more than two groups. For industrial hygiene data, it is recommended that measurements be repeated 4 times to ensure reasonable reliable detection of concentration of a factor of 4 for a type 1 error of 5 percent. This approach will ensure that the probability of type 2 errors for detecting the difference between two means is under 10 percent when the geometric standard deviation is under 2.0. A type 1 error is the probability of saying the difference in two concentrations is significant when there is no difference. A type 2 error is the probability of saying there is no difference when there is a real difference. Air contaminant data appears to be adequately described by a log normal distribution.