ISSN : 1229-0696
It has been well confirmed among the scholars that attitude toward the ad(A<sub>ad</sub>) substantially influences the formation of brand attitudes (A<sub>b</sub>). 'Salient attribute hypothesis' and 'affect transfer hypothesis' have been generally accepted to explain the mechanism by which A<sub>ad</sub> affects A<sub>b</sub>, but they conflict with each other. While 'salient attribute hypothesis' maintains that the ad itself works as a brand attribute, 'affect transfer hypothesis' insists that feeling responses to the ad is associated with the brand. In this study the authors tried to examine the relative validity of each hypothesis by investigating the correlations between A<sub>ad</sub> and A<sub>b</sub>, which changes with the level of repeated ad exposure. The total of 50 university students participated in the experiment as a group of 5 to 7 subjects. The participants of each group watched 40 TV commercials which were selected randomly from the pool of 400 commercials. After the exposure to the ads, A<sub>ad</sub>, A<sub>b</sub> and level of ad exposure were measured. This experimental procedure repeated twice with one week interval. The results reveal that the correlations between A<sub>ad</sub> and A<sub>b</sub> diminishes during the low ad exposure period and recovers through the period of high ad exposure. This forms an U-shaped curve and it is interpreted as follows: when the ad exposure frequencies are low, the advertisement itself works as 'a salient attribute' of an ad to affect the formation of A<sub>b</sub>, but as the frequency of ad exposure goes up, the effects of salient attributes are diminishing and the 'affect transfer' effects are taking the major role in the formation of A<sub>b</sub>.