Here’s the most interesting chart IMO from the recent NTSB report on the connection between speeding and fatal crashes in the US. The report is lengthy, and tends to focus on automobile crashes on highways, but there are some interesting takeaways.
Here’s the chart itself [emphasis mine]:
Here’s how the report describes the data behind this chart. The key point here is that there are different solutions to speeding and safety for urban and rural roads. From the document:
Further, of the 6,369 fatal crashes involving speeding passenger vehicles, 3,469 occurred on rural roads (55%). According to the FHWA, 920 million vehicle miles traveled (VMT) occurred on rural roads, which represented 30% of the total VMT in 2014 in the United States. Among all of the rural road types, 18% of fatal crashes involving speeding passenger vehicles occurred on local roads while such roads comprised only 14% of all rural VMT. Similarly, in urban areas, it was local roads that had the largest over-involvement of speeding passenger vehicles (22% of fatal crashes involving passenger vehicles versus 15% of all urban VMT). These observations indicate that the risk attributed to speeding among passenger vehicles varies among road types and land uses.
Figure 3 shows the distribution of fatal crashes involving speeding passenger vehicles by land use and reported speed limit.20 On rural roads, most of these crashes occurred on roads with reported speed limits of 55 to 60 mph, whereas in urban areas most occurred on roads with reported speed limits of 35 to 40 mph. Eighty-two percent of all fatal crashes involving speeding passenger vehicles on rural roads (2,796 of 3,418) occurred at locations with reported speed limits of 45 mph and above. In contrast, these reported speed limits accounted for 40% of all urban fatal crashes involving speeding passenger vehicles, a total of 1,383 such crashes. Therefore, speeding as a contributing factor represented different percentages of fatal crashes involving passenger vehicles on roads that serve different functions, with different speed limits, and in different land use areas.
For “urban” roads, the data strongly confirms what we already knew about the relationship between speed and safety, namely that there’s a big safety threshold around 30 miles per hour under which fatalities are rare, and over which can be very dangerous. (Note how the fatalities shrink to near-zero underneath the 30 miles per hour threshold.) For me, the lesson here is that urban roads should be designed to reduce speeds.
Side note: the solution for rural highways and roads is a bit different… and I’m not sure what it is other than divided highways!
- The Manual on Uniform Traffic Control Devices guidance for setting speed limits in speed zones is based on the 85th percentile speed, but there is not strong evidence that, within a given traffic flow, the 85th percentile speed equates to the speed with the lowest crash involvement rate on all road types.
- Unintended consequences of the reliance on using the 85th percentile speed for changing speed limits in speed zones include higher operating speeds and new, higher 85th percentile speeds in the speed zones, and an increase in operating speeds outside the speed zones.
- Expert systems such as USLIMITS2 can improve the setting of speed limits by allowing traffic engineers to systematically incorporate crash statistics and other factors in addition to the 85th percentile speed, and to validate their engineering studies.
- The safe system approach to setting speed limits in urban areas is an improvement over conventional approaches because it considers the vulnerability of all road users
For a bureaucratic safety study, that’s some pretty hard-hitting language. If I was the MUTCD, I’d be hurting!
The NTSB report also includes a little table of solutions for speeding, which seem to be focused on urban situations. You can see a familiar list of remedies.
Deadly speeding is something that many people — traffic engineers included! — seem to take for granted. But, at least in urban settings, there are ways to solve this problem. Ditching the 85% rule, and starting to incorporate safety measures alongside traffic data, seems long-overdue.