Header image

Matthew S. Bothner, Jeong-han Kang, and Toby E. Stuart

This article uses National Association for Stock Car Auto Racing (NASCAR) races to examine how competitive crowding affects the risk-taking conduct of actors in a tournament. We develop three claims: (1) crowding from below, which measures the number of competitors capable of surpassing a given actor in a tournament-based contest, predisposes that actor to take risks; (2) as a determinant of risky conduct, crowding from below has a stronger influence than crowding from above, which captures the opportunity to advance in rank; and (3) the effect of crowding from below is strongest after the rank ordering of the actors in a tournament becomes relatively stable, which focuses contestants’ attention on proximately ranked competitors. Using panel data on NASCAR’s Winston Cup Series from 1990 through 2003, we model the probability that a driver crashes his car in a race. Findings show that drivers crash their vehicles with greater frequency when their positions are increasingly at risk of displacement by their nearby, lower-ranked counterparts; the effect of crowding from below exceeds that of crowding from above; and the effect of crowding by lower-ranked contestants is greatest when there is relatively little race-to-race change in the rank ordering of drivers.

Administrative Science Quarterly, 52 (2007): 208-247

Matthew S. Bothner

When is a social actor most strongly influenced by its peers? This article addresses this question by clarifying when computer firms were most strongly affected by the choices of their structurally equiv- alent rivals to adopt a well-known technology: Intel’s sixth-gener- ation processor. The core hypothesis is that the effect of adoptions by structurally equivalent firms increases with the competitive pres- sure that a focal firm faces in its market position. The results show that a chosen firm is most strongly influenced by comparable others when it faces scale-based competition and is diversified. The im- plications of this study are twofold: a social actor’s sensitivity to the conduct of others may depend not only on its place in a hierarchy but also on the nature of its ties to an external audience; and a contingent theory of social influence may be necessary to charac- terize diffusion processes correctly, particularly when external and time-varying nonnetwork factors have significant effects.

American Journal of Sociology, 108 (2003): 1175–1210