Page 114 - Media Effects Advances in Theory and Research

5. MEDIA PRIMING 103

that we use in making judgments of the president (Iyengar & Kinder,
1987; Iyengar, Kinder, Peters, & Krosnick, 1984; Iyengar et al., 1982; Iyen-
gar & Simon, 1993; Krosnick & Brannon, 1993; Krosnick & Kinder, 1990;
Pan & Kosicki, 1997), and various stereotypes that influence how we
make judgments of people from the stereotyped group (Hansen &
Hansen, 1988; Hansen & Krygowski, 1994; Malamuth & Check, 1985;
Pechmann, 2001; Power et al., 1996; Wyer et al., 1985).
Unfortunately, there has been little focus on understanding the nature
of the media priming phenomenon. Further, the few explanations of the
mechanism by which the media acts as a prime vary from one domain to
another. There have been no attempts to integrate the research on media
priming across the different areas, let alone the models of how media
priming work. However, as we shall see, what these models have in com-
mon is a reliance on the priming research from psychology. In the next
section, we provide a brief background of the psychological research on
priming and then discuss the current models of media priming within
each domain.

MODELS OF PRIMING

Priming procedures were first used in cognitive psychology to explore the
structure and representation of information within network models of
memory (e.g., Anderson, 1983). Network models of memory assume that
information is stored in memory in the form of nodes and that each node
represents a concept (e.g., there is a “doctor” node in memory). Further-
more, these nodes are connected to related nodes in memory by associa-
tive pathways (e.g., “doctor” is linked to “nurse” but not to “butter”). An
additional assumption of network models of memory is that each node
has an activation threshold. If the node’s level of activation exceeds its
threshold, the node fires. When a node fires, it can influence the activation
levels of other, associatively connected nodes. For example, if the “nurse”
node fires, activation spreads from the “nurse” node to related nodes,
such as “doctor.” One consequence of spreading activation is that the
related node now requires less additional activation for it to fire. The
additional activation may accrue as a result of spreading activation from
other related nodes, or it may result from environmental input (i.e., read-
ing the word doctor). A typical behavioral outcome of spreading activation
is that a judgment about or pronunciation of a word (e.g., doctor) is faster
when it is preceded by a related word (nurse) than an unrelated word (but-
ter). A final assumption of network models of memory is that the activa-
tion level of a node will dissipate over time if no additional source of acti-
vation is present. Eventually, given no more activation, the activation