There is an international debate about the possibility of lowering the current occupational exposure limits (OELs) for respirable crystalline silica (RCS). Part of the discussion centers on knowing whether the OEL adequately protects workers. Exposure sampling using low-flow-rate pumps forms the basis of epidemiological studies, and thus the Technical Department of the National Silicosis Institute has carried out a study into the precision of workplace measurements with personal respirable dust samplers. The study consisted of taking two samples (one from either side of the worker’s breathing zone) during different types of activities (quarries, marble workshops, slate workshops) and in semi-controlled conditions (dust tunnel). A total of 366 sample pairs were obtained, of which 186 were acquired in the dust tunnel and 180 came from the workplace. The statistical analysis of the difference between the pairs of results, expressed as the percentage error, showed the following results for RCS in workplace measurements: the median was 27.4 % (range: 0 % to 181 %), the 75th percentile was 51.7 %, and the 25th percentile was 12.7 %. The workplace measurements were divided between two industry sectors (marble and slate workshops), which were examined separately. Results obtained in a dust tunnel showed that for RCS, the percentage error of the difference between the pairs of results had a median of 23.9 % (range: 0 % to 90 %) with a 75th percentile at 40.1 % and a 25th percentile at 11.1 %. The results from the dust tunnel were not found to be statistically significantly different from the field results. Several important conclusions were obtained: firstly, a single sample is not enough to adequately evaluate exposure in order to avoid errors stemming from differences in “valid” measurements as shown in the statistical analysis, and secondly, the main source of error is the variability of the dust concentration during real measurements.