Selective exposure to a biased sample of others is supposed to play an important role in explaining the false consensus effect (FCE). Selective exposure is generally assessed by asking subjects to estimate the prevalence of certain behaviours among their friends. We argue that this measure may easily lead to erroneous conclusions about the mediating role of friends' positions, since subjects are also likely to overestimate the prevalence of their own responses in a sample of friends. In the present study a process tracing task was used in which subjects were asked to think aloud when estimating the prevalence of specific opinions about 20 issues. Three predictions were tested. First, subjects were expected to refer to similar others more often than to dissimilar others. Second, referring to similar others was expected to lead to larger FCEs. Third, referring to dissimilar others was expected to lead to smaller FCEs. Results provided direct evidence of selective exposure, but this was most pronounced for issues for which opinions were relatively heterogeneous, i.e. for which there was no clear majority or minority opinion. The availability of similar others was further shown to be associated with larger FCEs, whereas the availability of dissimilar others was associated with smaller FCEs.
No takes yet. Share an insight, caveat, or question.
Bosveld et al. (1994) studied this question.