Page 42 - Media Effects Advances in Theory and Research

2. EXEMPLIFICATION THEORY OF MEDIA INFLUENCE 31

(2000). The purpose of the overview here cannot be to duplicate such cov-
erage. It is, instead, to exhibit principal exemplification strategies and to
relate the findings of representative research demonstrations to the pre-
dictions articulated in the previous section.

The Base-Rate Issue

Research on the effects of news reports in which arrays of exemplars are
supplemented or juxtaposed by base-rate information shows with great
consistency that recipients form their assessments of the presented issues
on the basis of the exemplar sets rather than on abstract, quantitative
information.
Research by Brosius and Bathelt (1994) demonstrates that recipients
who learn about people’s likes and dislikes of particular products, or
about people’s support for and opposition to various civic issues, tend to
base their perception of the proportion of favorably disposed people on
the relative frequency of exemplars presenting favorably disposed peo-
ple, irrespective of the ratios that are explicitly stated in the base-rate
information. In these investigations, the ineffectiveness of presented base
rates was extreme in that, when ratios apparent from exemplar distribu-
tions were contradictory to those explicitly stated, recipients formed
their judgments nonetheless on the exemplars and totally ignored the
stated base rates. In later research (Brosius, 1995), base rates were espe-
cially highlighted to force attention on them. Even such efforts proved
inconsequential, however, and recipients continued to base their percep-
tions of incidence rates on sets of exemplars rather than on provided
ratios.
Corroborating evidence was obtained by Gibson and Zillmann (1994);
Zillmann, Gibson, Sundar, and Perkins (1996); and Zillmann, Perkins, and
Sundar (1992). In these investigations, the provided base-rate information
was either precise (i.e., expressed as a ratio) or vague (i.e., expressed lin-
guistically in comparisons like “most people”). Irrespective of the mode
of expression, the base rates proved inconsequential, with recipients
forming perceptions on the basis of relative exemplar frequencies.
These findings lend strong support to the initial part of Prediction 1.
However, the growing relative influence of exemplars over time, pre-
dicted in the subsequent part, was not in evidence in any of the cited
investigations. Instead, the exemplification effects proved stable over
time. Delayed effects were observed one and two weeks after exposure to
news reports. These findings suggest that base-rate information never
received the careful attention that it would seem to deserve. Exemplars
thus could exert their overpowering influence immediately after expo-
sure, and growing influence could not materialize.