Effects of workers’ behaviour on measurements made using pumped and diffusive sampling techniques

The act of taking a personal sample, to assess workers’ exposure to hazardous substances, may in itself bias the measurement. It has been suggested that the wearing of bulky, noisy, personal sampling pumps could disturb the workers’ behaviour sufficiently to alter their work pattern, and hence their exposure to the substances used. The present study compares a pump based system using charcoal tubes, and a diffusive badge sampler for monitoring organic vapours. Diffusive samplers are much less bulky and intrusive and are therefore less likely to have an effect. The study design was such that measurements were made on between 13 and 20 workers at five sites, on six working days. On each alternate day both sampling methods were used, side-by-side, on the workers and on the remaining days only diffusive samplers were used. Samples were analysed by gas chromatography and the results from different days were compared in a statistical analysis of variance.At one site the diffusive sampler concentrations on days when pumps were worn were approximately 50% higher than the corresponding samples on the other days. This difference could have occurred by chance one time in ten. At the other four sites investigated the differences were not statistically significant.The study design was not entirely satisfactory for separating daily variation from the effects of wearing pumps because the days and sampling systems were confounded. The observed differences in concentration cannot therefore be straight forwardly attributed to wearing pumped samplers. It was however evident that in the one site where a difference was observed the workers spent a large proportion of their time outside the workplace. This and other subjective observations made during the study suggest that wearing personal sampling pumps is more likely to affect the gross patterns of behaviour of workers rather than the fine details of their behaviour close to the source of the hazardous substance. Further research would be required to test this hypothesis.The present study also allowed a comparison of the two sampling systems, on the side-by-side samples. In all cases there was good agreement between the diffusive and pump tube systems. There was, however, some indication that the relationship was slightly non-linear, with the diffusive sampler tending to underestimate the concentrations at higher levels relative to the pumped sampler.