ABSTRACT
Can phenomenological data such as reports of feelings, attitudes, goals and procedures be used as process variables in a psychological model? During the years when behaviourism was in its ascendency, the answer to this question was a clear, “No”. The experimental studies of fear communications that we began in the early 1960s were in partial accord with this bias. Their objective was to identify the conditions under which threatening communication about diseases elicited fear that led to both more favourable attitudes and to the performance of the action that was recommended for preventing the threat. The central hypothesis was that attitudes and actions would be reinforced when the rehearsal and performance of a recommended response reduced fear by eliminating the person’s sense of vulnerability to the threat. This hypothesis, derived from the revised drive reduction theory advanced by Dollard and Miller (1950), postulated that if the actions failed to remove fear and danger, avoidant denial and unrealistic reassurances would remove fear while leaving danger intact (see Janis, 1958; Janis & Feshbach, 1953). In keeping with the behaviourist tradition, we tested for the differential effects of high and low fear messages and of fear reduction on both attitudes and overt action, as we were sensitive to the fact that a subject’s post-exposure reports of fear and attitudes were verbal responses that might or might not be related to actual performance.
