Page 47 - Media Effects Advances in Theory and Research

36 ZILLMANN

tumors. After the passage of time, however, the threatening imagery
asserted itself by either elevating or sustaining the created concerns. After
exposure to the sanitized imagery, the concerns clearly dissipated with
time. More specifically, the influence of the program featuring sanitized
images diminished over time, whereas the influence of the program fea-
turing graphic, threatening images held steady and tended to grow.
Instead of using visual displays to evoke emotional reactions, Gibson
and Zillmann (1994) employed vivid verbal descriptions of emotional
happenings. A newsmagazine report about the crime of carjacking was
created and manipulated. All versions presented base-rate information
along with two detailed exemplars of carjackings. Base rates were given
as percentages of the severity of injury incurred during such crimes, or
they were suggested verbally in comparative terms. A survey was cited,
stating that 75% of all carjackings do not involve physical injury to the vic-
tims, that 21% of the victims get away with minor injuries such as bruises,
that 3.8% of the victims suffer severe bodily injury such as broken bones
and major lacerations, and that only an exceedingly small number of vic-
tims, 0.2%, is getting killed in the course of the commission of the crime.
The vague verbal parallel to these conditions stated that “most,” “some,”
“only a few,” or “almost nobody” would come to the harm in question.
The two exemplars detailed carjackings in one of these four injury condi-
tions. Focusing on the perception of the risk of getting killed in a carjack-
ing, the exemplar conditions thus can be arranged from minimal misrep-
resentation (i.e., two exemplars of no injury, being most consistent with
75% or “most”) through minor and severe to extreme misrepresentation
(i.e., two exemplars of fatal outcomes, being most inconsistent with 0.2%
or “almost nobody”).
The respondents’ fatality estimates are shown in Fig. 2.3. As can be
seen, the presence of only two exemplars readily influenced risk percep-
tion, overpowering the available correct base-rate information. More
important here, estimates of fatal outcomes of carjackings increased with
the severity of exemplified bodily harm. To the extent that the exemplars
that vividly detailed great suffering from severe injury elicited stronger
emotional reactions than the exemplars describing minor and minimal
harm, the estimates can be considered to accord with emotional reactivity.
Stronger exemplar-elicited empathic distress apparently fostered higher
estimates of severe harm from carjacking.
The investigation’s most significant findings concern the perception of
risk after the passage of time. Surely, the inappropriate exemplification of
fatal outcomes resulted in a substantial misassessment of fatality-produc-
ing carjackings shortly after exposure to the report. With the passage of
time, however, this misassessment grew to yet greater extremes. The
divergent interaction apparent in the figure suggests, in fact, that the
overestimation of risk increases with the degree of misrepresentation by