In "Tornado Alley", the weather hazard that is far more likely to injure or kill is not the supercells and tornadoes in the spring.
The map above shows where fatal icy road crashes are most likely to occur based on two seasons of data (2008-2010). The data shows that the Midwestern and Plains states have the highest incidence of road ice-related deaths. These are regions that traditionally place the majority of their weather awareness on convective severe weather (hail, tornadoes and high winds) that inflict comparitively much less human casualties:
Comparing 2008-2009 Average Annual Fatalities from Tornadoes [1] and Road Ice [2]
Probable explanations for the Plains/Midwest region's higher road ice fatality numbers are the fact that travel often takes place on rural highways at high speeds, the average trip/errand requires more mileage (increasing the risk of ice encounters), and all travel must be done by automobile (public transportation or walking/biking is not an option). In addition, unlike in northern states, icing events in the Midwest/Plains tend to be intermittent and widely spaced in time rather than season-long. Intermittent icing events cause more accidents simply because motorists are caught unprepared. In general, the fewer road icing events a location expereinces per winter, the greater the rate of fatalities per event results.
Areas where roads are snow and ice covered most of the winter (such as the Rockies and the extreme northern regions) tend to have a lower concentration of road ice deaths due to the population being well-adjusted to the hazard. Most road ice fatalities in the Rockies, for instance, occur with out-of-state travelers passing through on the major highways. Major urban areas - such as Chicago and much of the New England region - also tend to have lower fatality rates, most likely due to shorter distances on the average trip/errand, available public transportation and slower speed limits on most roadways.
ADDENDUM: I have received some questions about the sample size of the data in the two-year period shown, specifically if higher-fatality tornado years like 1999, 2011 and 2013 would change this statistic. For further consideration, see this table by NOAA on the U.S. Reported Tornadoes and Average Number of Deaths per Year from 1961 to 1990. Note that for many of the states, the average tornado fatality figures for this time period are actually lower than the two-year sample shown in the comparison table for 2008-2009. Missouri's 1950 to 2008 tornado fatality average is 3 per year. Adding Joplin (158) to that total brings the average up to 6 per year, less than the 2008-2009 average and still less than the 2008-2009 average from road icing.
It's safe to say that there will be anomalous years where a single tornado or outbreak will surpass road icing in total fatalities for that year. But looking at the average year, road icing is by far the greatest danger to the average person in these regions.
