Salivary cortisol is often used in occupational field studies when measuring stress reactions. For purposes of precision and accuracy in measurement, and interpretation of results, it is crucial to know the sources of variability that exert systematic influence on sampling. Variability can be both biological and methodological in origin, and failure to identify its sources may induce erroneous interpretations of Type I and Type II. This review aims to increase our knowledge and provide an overview of the biological and methodological variations of relevance for field measurements of salivary cortisol. It is concluded that: (i) time of sampling has to be carefully registered and included in the statistical analysis; (ii) samples have to be collected at the same time of year in longitudinal designs; (iii) food intake has to be avoided in at least the 2 h before sampling; (iv) vigorous exercise has to be avoided in at least the 2 h, preferably longer, before saliva is collected for measurement of cortisol; (v) variation in results obtained by different laboratory techniques emphasizes use of the same, or otherwise made comparable, laboratory techniques; (vi) concentration of cortisol is dependent on the material of the tampon; (vii) despite the absence of hard evidence, it is recommended that information be collected and results possibly statistically controlled for alcohol consumption, medication, such as oral contraceptives, and treatment for mental diseases; (viii) saliva samples can be stored at -20 degrees C for at least 1 year; (ix) cross-comparisons of absolute concentrations across studies might be difficult and therefore the establishment of reference intervals for the population studied and method used is recommended.