Replication and validation of anticipated turnover model for urban registered nurses.

Replication of the Anticipated Turnover Model Among Nurses for urban registered nurses provided substantial support for the stability and generalizability of the theoretical model. In both the original and replication studies, causal modeling was used to test the theoretical model predicting job satisfaction, anticipated turnover, and actual turnover. The replication study included 385 full-time nurses in two public and two private urban hospitals. The replication validated, with younger, more educated staff, the major findings from the original study. Group cohesion and job satisfaction effectively predicted anticipated turnover in the replication and the original study. Anticipated turnover was a good predictor of actual turnover, with discriminant analysis yielding 73.2% successful predictions in the replication study and 76.2% in the original study. Job satisfaction effectively buffered job stress. The replication also substantiated the position that job satisfaction strategies need to be targeted specifically to the types of clinical services.