I appreciate the response, but I don't think you really took the time to investigate and cherry-picked a single quote to prove your point. The source you listed is actually disproving your premise.
> ... the risk of police victimization, police homicides and assaults were found to be very infrequent occurrences during traffic encounters. The results of this study cast doubt on the Court’s assumption of danger during the routine police–citizen traffic encounter.
> Despite anecdotal accounts that show particular instances of danger confronted by law enforcement officers, there are few research findings or statistics to support the assumption of police dangerousness during traffic encounters.
> Most, if not all of the assertions that routine traffic stops are dangerous rely on the Federal Bureau of Investigation(FBI)’s data on police homicide and assault victimization during traffic encounters. All of these assertions share one crippling methodological flaw. They fail to appropriately account for the frequency of the police activity in the risk calculation
So yes, there are going to be a number of fatalities during traffic stops, but there are also a very large number of total traffic stops. Police work is inherently dangerous, and if you're stopping a large number of people you're going to run into dangerous people with a much higher chance of a fatality.
To assert that the average person be fearful of a traffic stop because some people were killed is just completely false. Police work is dangerous, I imagine they go into any situation with the assumption the person is dangerous.