International comparisons of sickness absence rates are difficult to make because of different national social security arrangements or recording procedures. Therefore a cross national study of sickness absence in Belgium, West Germany, and the Netherlands focused firstly on "contextual" aspects of sickness absence such as work incapacity benefit schemes, job security regulations, and the role of occupational physicians. Substantial differences were observed in regulations, local definitions, and available data. Those differences provide hypotheses for possibly divergent absence levels as well. For instance, Belgium shows the most diversified control procedures, the lowest level of sickness benefits, and the most rigid qualifying criteria for invalidity benefits. Sickness absence data were obtained from companies of four different clusters, each consisting of a comparable Belgian, German, and Dutch organisation. Conceptual, administrative, and statistical sources of bias were accounted for by selecting companies which resemble each other as to their product, production process, size, and geographical location: by using standardised definitions, absence data, and indices (observation period 1 January 1980 to 1 January 1982); and by measuring population variables to eliminate obvious explanations in terms of workforce composition. Controlled comparisons in these multiple clusters showed considerable and consistent differences in sickness absence rates. Whereas Belgian employees had 20.3 days of sickness absence (standardised for sex, age, and occupation) a year, German and Dutch employees had 28.5 and 39.1 days off a year respectively. Factors that might account for these differences are discussed.