Occupational exposure to formaldehyde was evaluated at a large chemical facility. Of particular interest was the validity of so-called Homogeneous Exposure Groups (HEGs) in which it is assumed that all individuals are exposed, on average, to the same level of contaminant. As a preliminary step a field study was conducted comparing the use of diffusion monitors to sorbent tubes for personal air sampling. Results from 26 matched pairs showed the two methods to be comparable [mean difference (badge - tube) = 0.03 ppm; mean (badge) = 0.17 ppm]. Six HEGs were formed based on job tasks and location. Using a randomized design, multiple full-shift samples were collected (with monitors) from representative workers in each HEG (127 measurements from 44 workers). Application of an analysis of variance (ANOVA) model indicated that the total variation in exposure across the entire sample population was partitioned as follows: 79% within-worker, 8% between-worker (within HEG), and 13% between HEG. Thus, assignment of HEGs in this case had only a marginal impact (13% reduction in variance) on the exposure assessment and substantial effort could have been avoided by randomly sampling the entire population prior to grouping.