The potential effects of attrition in longitudinal research are addressed and a procedure for assessing its effects is recommended. We recommend that researchers assess the effects of subject attrition on their data by assessing: (1) the presence of non-random sampling using multiple logistic regression, (2) mean differences on the study’s variables between those who responded and did not respond to the subsequent data collection, (3) the restriction or enhancement of variances, and (4) changes in relationships among variables due to attrition. We demonstrate the procedure using data collected from a random sample of employed adults in the US regarding job satisfaction, job characteristics, demographics, and mood. In our data, subject attrition led to non-random sampling, affected the means and variances of some of the variables, but did not affect the relationships among the variables. The effects of subject attrition may be sample specific, but the procedure recommended for assessing its effects may be used in other data sets and substantive areas.

This study investigated the association of organizational characteristics with the percentage of management positions held by women in a group of medium-sized to large private sector workplaces. Results support hypotheses derived from resource dependence and institutional theories and illustrate that characteristics of organizations predict a substantial portion of the variation in the percentage of management positions filled by women across industries. Effects differ for organizations in manufacturing and nonmanufacturing industries.