The construct validity of the Job Diagnostic Survey was examined in terms of its ability to detect expected mean differences in perceived job characteristics between people in: (a) large and small organizations, (b) large and small work groups and (c) high and low status jobs. These subgroups were compared using multivariate and univariate stepwise regressive analyses. Examination of significant effects and mean differences provided overall support of the construct validity of the Job Diagnostic Survey and suggest that macro variables such as organizational size and work group size are reflected in people's perceptions of their jobs. (Author)