Four Perspectives on . . .
11
threat component of the appeal, immediately after participants are exposed to the threat
component, and again immediately after presentation of the recommendation component.
Author Withheld have recommended the use of message-component research designs.
That is, designs that break messages into their constituent parts, assess emotional responses to
each element of the message, and use the resulting emotional change measures as predictors of
persuasion. Clearly, the design called for in the current study is one type of message-component
design. Although well-suited to the questions that motivated this investigation, the design also
poses a threat to internal validity by introducing the potential for reactive measurement. In light
of this, we posed a research question:
RQ1: Will a design in which participants are interrupted after the threat component to
provide data yield different results from one in which they provide data only after
presentation of the entire message?
Method
Overview
Participants in the study read a message that warned of the dangers of influenza and
advocated that they obtain a vaccination. They also completed a questionnaire that measured
their tonic levels of BIS and BAS activation, their emotional reactions to the message, and their
estimate of the likelihood that they would get a vaccination.
Participants
The sample initially included 445 students enrolled in courses at the University of
Wisconsin-Madison. Each received a small amount of extra credit in return for their
participation. Following data collection, we applied a series of screens that reduced the overall
N. Four respondents were eliminated because they had received a flu vaccination prior to the
Convention