Too good to be true? Understanding change in organizational outcomes

When evaluating organizational outcomes, strong experimental designs are often not practical. As a result, assessments of change may be confounded by a number of threats to their validity. Therefore, attempts have been made to detect or control for many of these artifacts. This article explores a method for detecting response artifacts, using measurement and relational equivalence techniques. The large drop in the reported levels of sexual harassment in the U.S. military between 1995 and 2002 provides the context for testing this method. The results suggest that at least part of the reported decrease in harassment was artifactual. The implications for evaluations of organizational change are discussed.