As far as measuring incidents, most of the car incidents are probably fender benders in a parking lot hardly the same seriousness as being non-fatally shot with a gun. Ideally you would want something like seriously injured with a motor vehicle versus seriously injured with a gun, but statistics are likely to be better on fatalities. Also since people rarely use more than one gun at a time but can own dozens I think a better comparison would be gun owners versus car drivers (more than twice as many car owners) or better yet, as Hentor suggested a measure of exposure; say man-hours spent shooting versus man-hours spent driving (don’t know for sure but would bet a difference of an order of magnitude)
With these statistics in place cars are a much safer bet.
I was wrong with the 3x number, but only because I forgot to take into account suicides. In 2011 it was just over 19,000 suicides and 11,101 homicides with approximately 500 deaths related to accidents.
That being said, the “incidents” number with guns included ALL incidents with guns, not just assaults, but anything that was reported to police. “He was waving a gun around and the ran off” would be considered an incident even though no fatality, injury, or even discharge of the gun happened.
Your point about exposure with cars, though is well taken. We can probably guesstimate what car use hours are fairly easily, but there’s a problem with guns. Each incident with a gun happens for moments. So how long do we consider “exposure”? The interaction time? How long the gun was carried with the person? How long the person owned the gun? Each of these assumptions brings a lot of uncertainty into the conversation because of how the guesses of time will have to be made.
I will stick with incidents, which isn’t “cars” it’s “number of accidents involving cars,” simply because the numbers are reported yearly in the same manner and doesn’t require the introduction of a lot of guesswork.
The probelm is that gun incidents aren’t directly comparable to car incidents. The fact that the relative number of incidents that result in fatalities is much higher for guns than for driving indicates that gun incidents are inherently more severe. So if you are interested in safety, fatality to fatality is the best way to compare the results, since they are easily defined and mean the same thing for both a car and a gun incident.
As far as exposure I was thinking of time spent handling a gun, (including cleaning, brandishing, shooting at a range, hunting etc.) versus time spent driving a car would be fairly equitable.
In any case the point is largely moot and we should probably get off this highjack but as a career statistician its hard for me to let these things go.
In 2011 it was just over 19,000 suicides and 11,101 homicides with approximately 500 deaths related to accidents.