Demographic Effects of Behavior Modeling in Seat Belt Use: Analysis of 15,000 Observations

Abstract

Automotive seat belt use is known to be highly effective in saving lives, but is still not universally practiced, with wearing rates around 60% to 70%. A series of studies, observational and experimental has established that behavior modeling occur when vehicles have more than one occupant. More times than would be expected by chance, both occupants show the same wearing behavior. However, the demographics of this modeling effect are not clear from the literature. The current study analyzed a large database, over 15,000 observations collected in Indiana in 2001 where the sample size was large enough to provide statistical significance for even small effects. Overall, a high degree of modeling behavior was observed. Modeling was found to be larger for males and larger for the under 21 age group as drivers. Modeling was also maximized when driver and passenger belonged to the same age or gender group. Comparing drivers with and without passengers, we found that the major effect of modeling was negative, i.e. there was a large difference between drivers wearing seat belt while alone (67%) and in the presence of an unbelted passenger (18%). Implications for public policy aimed at increasing seat belt use are presented.