| Learning Objective #1: critically examine the positive impact of nonlinear modeling for possible applications in nursing research, particularly in nursing turnover predictions | |||
| Learning Objective #2: perform comparisons of linear and nonlinear modeling using key turnover predictor variables in a simple cusp catastrophe nonlinear model |
Prevention of turnover in a rising and dire international and sustained nursing shortage is crucial in managing shortage sequela. In order for administrators to intervene effectively, there is a need to target accurately the staff population at risk for turnover. Current forays into the realm of complexity sciences indicate that managerial decision-making may not be adequate in the existing healthcare systems environment. Nonlinear modeling methods offer promise for capturing more of the essence of human emotion and the impulsive aspect of nurses’ affective responses to issues leading to turnover, because they account for varying and apparently insignificant changes that linear representations simply do not portray. The purpose of this study was to examine the relationships between nursing turnover behavior and three known predictors of nurse turnover behavior: nurses’ perceptions of anticipated turnover, organizational commitment and job-related tension, in order to compare the predictability of a cusp catastrophe model and a traditional linear model of nursing turnover behavior. The results of this descriptive, correlational survey with a longitudinal cohort prospective data collection plan demonstrated that a highly predictive cusp catastrophe model could be generated from the convenience sample population involved (N = 1033; return rate 77.2% T1, 64.9% T2), with 80.4% correct predictions overall and 53.6% correct predictions of the actual terminations. An examination of a dynamic cusp model demonstrated that movements of participants across the bifurcation plane in the cusp model were indicative of retention or termination. However, the unique characteristics of the sample may have impeded the ability to design a more predictive model; thus, more research in this area is indicated. Nursing research, nursing administration and nursing practice should act on this evidence to benefit future studies and the profession of nursing.
Research Supported by a grant from NIH and NRSA, 5 F31 NR008461-02, 2003-2007