Page 34 - Media Effects Advances in Theory and Research

2. EXEMPLIFICATION THEORY OF MEDIA INFLUENCE 23

In order to function as exemplars, events must share attributes with oth-
ers to a degree that makes them classifiable as members of the same popu-
lation of events. They must share a number of primary, defining attributes.
However, because the events are not likely to be identical in every other
regard, they may differ on any number of secondary attributes. These sec-
ondary attributes define potentially significant variation within the popu-
lation. This variation often justifies the specification of subpopulations. In
such cases, the initially secondary attributes become primary ones, and
variation in the subpopulation is defined by the secondary attributes spe-
cific to the events of this subpopulation. Finally, there may exist lower-
order variation in attributes of minimal relevance. Such attributes, if incon-
sequential for all practical purposes, may be disregarded.
Events manifest in exemplars can be abstracted in specifications of
entities that have certain properties, that perform certain actions, or that
cause certain consequences. Conditions under which the indicated events
occur may be stipulated in addition.
These conceptional considerations, as well as those concerning repre-
sentation by exemplar displays that are discussed in the following section,
are further elaborated in the initial presentation of exemplification theory
(Zillmann, 1999).

EXEMPLAR DISPLAYS AND REPRESENTATION

Phenomena are exemplified under two distinct sets of conditions. In the
first, essential parameters of the distribution of events in an event popula-
tion of interest are known. These parameters can derive from assessments
on all events in a population. More likely, however, the parameters are
estimates based on the assessment of a representative sample of the event
population. Representativeness is achieved by adherence to selection pro-
cedures that ensure that every event of the population has the same
chance for inclusion in the sample. In the second set of conditions, a pop-
ulation of events is presumed to exist, but parameters of the event distri-
bution are not known.
In case population parameters are known, the degree to which selected
exemplars represent the variation of secondary event attributes can be
ascertained. A high degree of representation is achieved by honoring the
condition of equiprobability for the inclusion of exemplars in a sample.
Random selection of exemplars is impartial to secondary attributes and
thus prevents the inappropriate aggregation of specific event groups at
the expense of others. The greater the violation of random exemplar
extraction by the arbitrary selection of cases that are deemed “of particu-
lar interest” according to idiosyncratic criteria, the lower the degree of
representation by the resulting exemplar sample.