These data are provided by Edwards and Mason (2014) and supplemented by data from the Bureau of Labor Statistics (unemployment rates), the Bureau of Economic Analysis (state-level real GDP per capita), and the Center for Disease Control (tabs of the 2010 BRFSS bringe drinking data).

The Edwards-Mason paper draws its data from 2009 traffic fatalities by type, age, and state drawn from microdata from the Fatality Analysis Reporting System (FARS) of the National Highway Traffic Safety Administration (NHTSA) and annual person miles traveled either by bicycle or by motor vehicle as defined above, drawn from the 2009 National Household Travel Survey (NHTS).

Note that the Edwards-Mason data excludes several states in which the 2009 NHTS did not measure any bicycling.

state = two-letter state abbreviation

logbikemr = natural logarithm of bicycling fatalities per billion km in 2009

logbikekmpc = natural logarithm of bicycling billion km per capita in 2009

logautomr = natural logarithm of auto fatalities per billion km in 2009

urate = average of monthly state unemployment rates in 2009

rgdppc = state real GDP per capita in 2009 in chained 2009 dollars

bingeprev = Age-adjusted prevalence of binge drinking among adults aged 18 years and older, by state, United States, 2010. Prevalence and intensity data were age-adjusted to the 2000 U.S. Census standard population. Binge drinking is defined as drinking 5 or more drinks on an occasion for men or 4 or more drinks on an occasion for women, during the past 30 days. Respondents were from 48 states (excluding South Dakota and Tennessee) and the District of Columbia

bingeint = Age-adjusted intensity of binge drinking among adults aged 18 years and older, by state, United States, 2010. Intensity is defined as the average largest number of drinks consumed by binge drinkers on any occasion in the past month