Road safety concerns are a legitimate concern when promoting increased bicycle use. Currently, the share of bicyclist traffic fatalities and injuries is not proportional to its mode share, and presents both a public health concern and a disincentive to people taking up or continuing to bicycle for transportation. Bicycling is not an inherently dangerous activity; automobile drivers pose the most risk of harm in crashes with bicyclists. Despite that, drivers' attitudes and behaviors toward bicyclists have not enjoyed much systematic study, particularly in the United States. This research explored the dimensions of drivers' attitudes toward bicyclists, including implicit bias and social attitudes, and examined the relationships between these attitudes and drivers' self-reported behaviors. The online survey included a cognitive test of respondents' implicit preference between drivers and bicyclists. The research questions are detailed in the introduction (Chapter 1), followed by a review of selected literature (Chapter 2) and detailed methodology (Chapter 3). The first set of results (Chapter 4) explores the potential usefulness of the implicit method and the attitude measures developed for this research, and presents an analysis of drivers' attitudes and what predicts more positive attitudes toward bicyclists. The second set of results (Chapter 5) extends the analysis to drivers' self-report behaviors, and how demographics, individual travel behavior, attitudes, and the built environment predict drivers' behaviors related to bicyclist safety. The dissertation concludes with a discussion of the contribution to the literature on driver attitudes and behaviors, and the implications for both practice and research.