Four studies were carried out. In the first, it was hypothesized that individuals may have a predisposition for ‘reinvestment’ of controlled processing, which will lead to skill failure under stress as a result of disruption of the automatic functioning of the skill. Factor analysis of established personality measures related to the concept of reinvestment uncovered a reliable 20-item factor which appeared to be associated with reinvestment. In the second study a predictive validation of this ‘Reinvestment Scale’ was attempted. It was predicted that the motor performance of high scorers on the scale would be more likely to fail under pressure than that of low scorers, on the grounds that high reinvesters would be more likely to disrupt the smooth fuctioning of their own skill by investing it with controlled processing. High and low reinvesters learned a two-dimensional rod-tracing task to a level of performance approaching asymptote and were then required to perform under conditions of stress. Despite highly significant increases in the levels of stress exhibited, performance was unaffected in both groups, thus providing neither support for nor refutation of the prediction. An explanation of this was that the rod tracing task was not complex enough to present the kind of demands that would lead to reinvestment. Hence, a third study was carried out in which a more complex, golf-putting skill was considered. In this instance support was found for the prediction that the performance of high scorers on the Reinvestment Scale would be more likely to fail under pressure than that of low scorers. Finally, a fourth study was carried out in which further validation was sought by exploring whether a relationship existed between the reinvestment scores of university team squash and tennis players and the opinion of informed raters on their tendency to ‘choke’ under pressure. A correlation was again found, providing further evidence that high reinvesters are more likely to suffer from performance breakdown under pressure. It was concluded that the Reinvestment Scale does indeed assess a predisposition towards reinvestment of controlled processing, and may prove to be a valuable instrument in predicting skill failure in stressful situations involving complex, rule-bound skills.