National Motor Vehicle Crash Causation Survey

National Motor Vehicle Crash Causation Survey: Report to Congress DOT HS 811 059

Kindof interesting;  the 5,471 crashes examined in detail are supposed to be nationally representative as to type of crash (type of roadway, weather conditions, single vs. multi vehicle, etc).

Particular attention was given to the chain of events culimanting in the crash

Of particular interest is the “Critical Reason for the Critical Pre-Crash Event”, which is “the immediate reason for the critical pre-crash event and is often the last failure
in the causal chain”.

The events were attributed broadly to either a driver (5,096), vehicle  (130) or roadway/environment/weather (135). 110 were not able to be categorized. See tables 9a, 9b and 9c, pp. 23-26.

Clearly, the driver is the weak link with driver error being responsible for over 90% of crashes.

Speeding accounts for 15.3% of the driver errors (“Too fast for conditions” + “Too fast for curve”). By far the largest single category is something they call “Inadequate surveillance” (20.3%) which is described as “a situation in which a driver failed to look, or looked but did not see”. Note that this is distinct from distraction(14.5%) and inattention(3.2%).

This study is frequently cited as saying something to the effect that only 5% of crashes are due to speed. This is true only in the strict terms set forth; and ignores both combination factors (e.g. speeding and following too closely is not counted as a speeding collision) as well as increased severity that increases with speed.