Our first study examined the combined effects of WOM
dispersion and taste similarity in a choice setting. Participants
were asked to choose between products characterized
by various rating distributions, presented in graphical form,
from a variety of product domains. Given that average ratings
are positive for the vast majority of products at realworld
platforms (Chevalier and Mayzlin 2006; Li and Hitt
2008) and consumers will tend to avoid low-rated options,
the stimuli for the study consisted of products with average
ratings well above the midpoint. Thus, prospect theory and
the broader principle of risk aversion suggest that individuals
will prefer consensus over dispersion. Over and above this
tendency, however, our model predicts that individuals will
be more likely to choose a high-dispersion option over a
low-dispersion option for domains characterized by dissimilar
tastes.
Choice pairs in the stimuli presented three different types
of trade-off between average rating and dispersion (described
below). These different trade-offs helped to disguise
our dispersion manipulation, by ensuring that distributions
in the pairs were not distinguished by variance alone and
also allowed us to explore questions related to the robustness
of our framework.