Sampling Method: Coefficient of Variation

Sampling Method: Coefficient of Variation

---

The coefficient of variation (CVt) of a sampling method is a measure of the inaccuracy of the result. The CVt is determined by random errors in sampling, reprocessing and/or analysis. In the European standard EN 482 (in the Netherlands: NEN-EN 482) the maximum values of the CVt for a sampling method are given for a specific concentration range and measurement time. The CVt values have te be used when assessing the exposure in the workplace atmosphere, f.i. with a measurement strategy according to (NEN-) EN 69.

In DOHSBase Compare the presented values of CVt are the values that may be expected under ideal conditions. Normally greater values of CVt are seen, mostly due to uncontrollable factors, such as changes in concentrations, disturbing substances, temperature, humidity or human activities. The most random errors are largely independant of the concentration: pumpvelocity and readability. NIOSH (Leidel & Busch, 1977, pag. 50) calculates the 90%-tile confidence interval of the results as follows:

1.64 * H * CVt

In which is H the concentration at which the CVt is validated. When H is unknown, the best estimate of H is the limit value.

The influence of the measurement accuracy (CVt) in the assessment is especially large when:

• the exposure is near the limit value;
• the is only one result;
• dust measurements have been carried out by a gravimetrical method; fibres are counted or another sampling method is used with a large inaccuracy in reading the result.