Four different class measures (social groups, a socioeconomic classification, a measure by Erik Olin Wright, and a structural class concept) were applied on a gross material of 3,252 persons registered by sickness insurance authorities. The measures were compared by description of absence days, average durations, sickness rates, and sex. No significant differences were found. By associating measures and absence days, differences were found only for doctor certified days, where the structural class concept diverged. Two conclusions are drawn: When dependent variables are of ideological character, practical considerations may decide choice of class measure. In other cases, caution is advised and further research needed.