Group-mean exposure categories have previously been shown to result in less biased exposure estimates in the wood processing industry [8, 13, 16].
Nonetheless, in the present study, the exposure categories were relatively crude, with a relatively large variance in concentrations within each exposure group (table 1).