To explore the possible influences that work groups have upon their me
mbers' safety-related behavior, this research simulated small-group di
scussions of an industrial safety issue. The experimental task require
d the 96 subjects to select, from a range of seven graded options, the
level of personal protection they wished to adopt while performing a
series of mechanical operations under adverse conditions. Subjects mad
e their own private selections of personal protection, contributed to
a group consensus selection, and then made a final private selection.
As predicted, subjects' preferred self-protective behaviors shifted af
ter group discussion, becoming simultaneously more risky, more homogen
ous, and more confident. However, little support was obtained for a hy
pothesized polarization of preferences following group interaction.