This study examines the suitability of Sanders' (1983) Model of Stress and Human Performance as a tool for investigating the effects of different stressors upon the cognitive processes involved in choice reactions. A general discussion of past and current approaches to stress research is followed by a detailed discussion of Sanders' model. This model integrates linear stage and resource allocation frameworks based, respectively, on Sternberg' s additive factor logic and Pribram and McGuinness' identification of the three neurophysiological systems of arousal, activation and effort. In the model, the basal energetical mechanisms of arousal and activation are viewed as affecting perceptual and response preparation processes respectively, whilst the upper, coordinating effort mechanism affects decision-making. The fundamental prediction of the model is that stress will arise whenever the effort mechanism is unable to restore imbalances i n the basal mechanisms and, furthermore, that different stressors can be identified as affecting one or both of the basal mechanisms. By manipulating the experimental variables identified in the model as specific to a particular stage, this frame-work was used to i nvestigate the locus of noise effects on the model. Three experiments are described together with a discussion of the findings, which show that noise could not be clearly demonstrated to affect either arousal or activation. This is followed by a report of a further experiment which examined another stressor, anticipatory fear, in which subjects were required to jump from a balcony 15 feet into a pit below. Despite a main effect of fear upon RT, Sanders' hypothesised interaction between fear and activation was not revealed. The subsequent discussion centres on the failure of these experiments to demonstrate interactions between the stressors concerned and either of the basal mechanisms, thereby casting doubt upon the suitability of the model for examining stressors such as noise and fear. Recommendations for revision of the model and further research are proposed.