Thinking differently: Assessing nonlinearities in the relationship betweenwork attitudes and job performance using a Bayesian neural network

The relationship between work attitudes and individual job performance was investigated using artificial neural networks (ANNs). ANNs use pattern reco gnition algorithms that are well suited to capturing nonlinear relationship s among variables thereby providing a new perspective on research on this t opic area. Results from the neural network analysis provided strong evidenc e of nonlinearity suggesting that nonlinear models are needed to understand the work attitude-job performance relationship. In so doing, the neural ne twork model had greater predictive accuracy than did traditional OLS regres sion. Implications of this finding for theory development and future resear ch were discussed.