Research on threat appeals has yielded conflicting findings concerning the relative effectiveness of high threat versus low threat in persuasion. Studies reviewed here have investigated the effects on persuasion of variables in each of five areas: (a) the nature of the recommendations, (6) personality characteristics of the recipients, (c) source credibility, (d) learning of the message content, and (e) the interest value of fear. Few variables have been found which consistently interact with fear. One conceptual consideration and four methodological considerations are suggested as possible sources of the inconsistency in the findings. A postulated curvilinear relationship between fear level and persuasion may help reconcile some of the conflicting findings on highversus low-threat appeals.

A the time of the formation of the Department of Social Psychology at Columbia University in 1961, Richard Christie (1965) presented a paper in which he discussed several significant changes in research from 1949 to 1959. These trends included (a) a considerable increase in experimental manipulation, especially the use of multiple experimental groups; (b) an increase in the use of parametric statistics, especially the analysis of variance; and (c) an increase in the use of "captive samples," especially college students. Christie lightly noted that, An extrapolation of the preceding documented trends would indicate that within a few years the number of published articles in the Journal of Abnormal and Social Psychology should reach an asymptote with all articles reporting experiments on college students using analysis of variance designs [pp. 150-151].On the other hand, Sargent (1965), speaking on the same occasion as Christie, said, I would predict a nonexperimental trend in social psychology (or perhaps several nonexperimental trends) in addition to the established and accepted experimental emphasis. As I see it, this would include other types of empirical study, such as controlled observations, case studies and personality analyses-some of which are easy and some difficult to quantify [p. 35].

No abstract

No abstract

No abstract