Four Perspectives on . . .
12
experiment. An additional five were eliminated due to missing data on one or more of the
variables. Next, four respondents, three who were allergic to eggs and one who was pregnant,
were removed because flu vaccinations are contra-indicated for persons with either condition
(and the message informed them of this). Finally, to ensure variance in the dependent measure,
70 participants were removed who reported, prior to the message, that the likelihood that they
would obtain a vaccination was greater than 90%. Application of these screens resulted in a final
N of 361.
Design
The experimental design can be conceived of as a between-subjects 2 X 2 with one
repeated measure. The first between-subjects factor, level of threat, was included to ensure
variation in one of the variables of interest, that is, fear. It was expected that variations in the
level of threat in the message would yield corresponding variations in fear. Threat was
manipulated by creating two versions of a message that varied in severity (described below). The
second between-subjects factor, measurement condition, was included because it would enable
us to assess the extent to which our measurement procedures influenced the outcome of the
study. Participants in the interrupted condition provided emotion measures prior to the message,
immediately after reading the threat component of the message, and immediately after reading
the recommendation component. In contrast, participants in the non-interrupted condition
provided emotion measures prior to the message, then read the message in its entirety before
responding to emotion measures that focused on each of the two message components
individually.
Fear Appeals
Convention