Discussion

The main purpose of this paper was to explain the high correlation between the mean and variance of the reaction times in prolonged work tasks. The phenomenon was explained in terms of the parameters of the PE model, which was used as a limiting case of the PI model. A statistical test using structural equation modelling yielded corroborating evidence for the hypothesis that this correlation could be described in terms of the correlations between the PE model paremeters across subjects. However, one should take into account that the PE model was used as an approximation of the PI model. The PI model is sensitive to fluctuations in inhibition, whereas the PE model is not. Therefore, the PI model should be given preference over the PE model. The problem is that untill now no satisfactory methods have been available to estimate the parameters, which play a part. This was the very reason why in this study the PE model was used and not the PI model. It still would be more preferable to make use of the PI model in future research, instead of the PE model. This would require further theoretical research with regard to the PI model, especially, the development of proper estimators for the latent parameters having high priority.