Three experiments were conducted to examine the hypothesis that attitudinal affect is a monotonic function of frequency of exposure. In the first experiment while supporting the frequency-affect relationship a non significant difference between males and females and positive and negative set conditions were obtained. In the second experiment low association value stimuli produced significantly greater increases in affect over exposure than relatively higher association stimuli. Duration of exposure at 2 and 5 second intervals only were shown to enhance evaluative rating while frequency of exposure again showed the predicted increase in affect, in a third experiment. Further inspection of the results of the first and third experiment indicated that whilst most showed greater preference for familiar stimuli there was a group of subjects preferring low exposure stimuli. In the second part of the study subjects were presented with stimuli at different exposure frequencies and stimulus complexity to test the hypothesis that affect was an inverted U shape function of stimulus complexity with subjects preferring stimuli of medium complexity regardless of exposure frequency. Although the hypothesis was supported marked individual differences were obtained by examining the complexity level at which the increase in positive affect as a function of exposure was most pronounced for each subject. The results were interpreted as supporting the theory of optimum levels of arousal and it was suggested that not only do the differences in the point of maximum increase in affect over exposure frequency provide a measure of an individual’s optimal level but also they reflect a cognitive style of search for novel as opposed to familiar environmental contingencies. It is suggested that the need for familiar versus novel stimulation is not only a crucial determinant of an organism’s affect arousal but also the distinction underlies many of the more well known and researched cognitive styles. Some of the factors determining the development of orientations towards familiar versus novel stimulation are discussed together with directions for future research.