Page 100 - Media Effects Advances in Theory and Research

4. MEDIA CONSUMPTION AND PERCEPTIONS OF SOCIAL REALITY 89

much television they watch prior to providing their prevalence and likeli-
hood estimates. This order of data collection was sufficient to eliminate
the cultivation effect. As Morgan and Shanahan (1996) note, a number of
studies that have reported finding no evidence of a cultivation effect
either measured television viewing prior to measuring social perceptions
or introduced the study as one pertaining to television. Although Morgan
and Shanahan’s meta-analysis did not find support for such source-
priming as a moderator, their results showed that the effect sizes for the
non-source-primed studies tended to be slightly higher than the effect sizes
for studies in which source was (inadvertently) primed. Thus, it seems pos-
sible that the inability to observe cultivation effects in previous studies may
have been due to the inadvertent priming of source information.
Note also that it is not necessary to prime source characteristics
through data collection. Priming simply refers to making a construct more
accessible in memory. For some people, particular constructs may be
chronically accessible (Higgins, 1996). So for whom might the construct of
television, and its potential effects, be particularly accessible? One group
may be communications majors, or for that matter, any student who
might have had a course that deals with potential effects of television; in
other words, people who may often comprise the subject pools that aca-
demics (and especially those in communications departments) use in their
studies. Thus, it is plausible that null findings for cultivation effects in
some studies may be due to the special characteristics of the sample.

Involvement Explanations. There are a number of factors that may
relate to level of involvement with constructing judgments. For example,
level of involvement may differ as a function of sample composition. Col-
lege students may be less intimidated than older adults or younger peo-
ple by the university setting that may be used to collect data (Shrum,
1997). Alternatively, individual differences may exist that relate to
involvement, such as interest in the topic (e.g., crime by those with direct
experience with it) or general interest in solving problems (e.g., those high
on need for cognition; Cacioppo & Petty, 1982). Involvement may also
vary as a function of data collection method. Data that are collected
through anonymous questionnaires may induce less accuracy motivation
than data collected in, say, personal interviews (Shrum, 1997, 2001).

Time Pressure Explanations. Shrum (1999a) showed that simple dif-
ferences in data collection methods, presumably related to differences in
time pressure, can have a significant impact on the magnitude of cultiva-
tion effects. In that study, the difference was whether the data were col-
lected via a phone or mail survey. Other situations can contribute to time
pressure, whether real or imagined. Although not entirely independent of