ASSESSING THE SAMPLES OF PRIOR ORGANIZATIONAL RESEARCH

The sampling designs used in organizational research have been less than consistent across different reported studies. In this analysis we examine the reported relationships among several key organizational variables in ten separate previously published studies to determine the degree to which major differences in sample designs have influenced findings. We isolate that portion of reported associations due to unique characteristics of the samples and report the association these ‘sample design effects’ have with particular sample designs. Results indicate homogeneous samples of organizations inflate reported relationships yet leave significant sources of variation uncontrolled in sample selection. an alternative to sampling of homogeneous organizations is suggested by the fact that larger and probabilitistically selected samples are also associated with larger reported relationships among organizational variables.