Talking Michigan Transportation

A veteran researcher talks about why fatal crashes continue to rise in Michigan

December 16, 2021 Michigan Department of Transportation Season 3 Episode 85
Talking Michigan Transportation
A veteran researcher talks about why fatal crashes continue to rise in Michigan
Show Notes Transcript

Preliminary numbers show fatal crashes increased again in 2021 over 2020, up by about 10 percent, continuing a disturbing trend reflected in national data.

Earlier this year, the National Highway Traffic Safety Administration released a report detailing factors contributing to the higher death count, including speeding and a decline in seat belt use.

This week’s Talking Michigan Transportation podcast features a conversation with Peter Savolainen, a Michigan State University foundation professor in the department of civil and environmental engineering, who has conducted extensive research on driver behavior.

Savolainen talks about the most recent Michigan numbers that found that 1,067 people died on Michigan roads this year. As of Dec. 7, compared to a year ago, there are 101 more fatalities and 369 more serious injuries.

While travel was down nationally anywhere from 20 to 30 percent on average over calendar year 2020, Savolainen observes that most crashes were down by similar proportions, except for the most severe. He says there were pronounced increases in the number of fatal traffic crashes.

“We've been trying to understand exactly what's been driving that, and there's been a lot of discussion nationally that speed is playing a role in that to some degree,” says Savolainen.

He also talks about the troubling increase in pedestrian deaths and the possibility that, for a time, more people were walking instead of using public transit because of the pandemic. The Governors Highway Safety Association reports that drivers struck and killed an estimated 6,721 people on foot last year, and “a shocking and unprecedented” 21 percent increase in the pedestrian fatality rate from 2019 to 2020 was the largest-ever annual increase as a result of traffic crashes since the government’s tracking system was established in 1975.

Among other related topics, Savolainen discusses automated enforcement technology. He says data shows conclusively that awareness of the enforcement brings down speeds.


Jeff Cranson: Welcome to the Talking Michigan Transportation podcast. I'm Jeff Cranson.


Cranson: This week, I’m pleased to have a repeat guest, Dr. Peter Savolainen who is a professor at Michigan State University. He does a lot of research and study in the area of traffic safety and driver behavior and studies crashes and trends and has some fairly updated information that he can talk about today. Thank you for taking time to be here.

Peter Savolainen: Sure, happy to join you, Jeff.

Cranson: So, why don't you talk about what you've seen in comparing things over the past, you know, 22 months now, I guess, since the pandemic began, and what's been going on with speeds and crashes and seatbelt use and pretty much all those things we've talked about before in Michigan.

Savolainen: Yeah, so we've been doing some work dating back to the onset of the pandemic, in essence. And I’m sure many people are familiar that travel was down nationally anywhere from 20 to 30 percent, on average, over calendar year 2020. And most of the crashes were down by similar orders of magnitudes except for the most severe, where we actually saw very pronounced increases in the number of fatal traffic crashes and result in fatal injuries in those crashes. So, we've been trying to understand exactly what's been driving that, and there's been a lot of discussion nationally that speed is playing a role in that to some degree. So, here in Michigan, we've actually been looking at speed for quite a while because we've actually had some MDOT work that we've been assisting with to try to ascertain what the impacts of the speed limit increases from 2017 were, for example. And at this point, we've got some early returns on that project where we looked at crashes in 2018 and 2019 compared to the three years prior to when those increases had occurred. And we see crashes are basically up across the board both on the freeway and non-freeway systems, and those increases have been most pronounced in terms of fatal and severe injuries. So, unfortunately, when we look at the pandemic, we've seen that same general trend is still holding and that people are going faster. And it's not necessarily the highest speed roads, which going into this, we weren't necessarily sure what the data would show us, for example. But as we look at the various speed limit categories, for example, at the high-speed roads, so 70, 75 and so forth, fatalities are up between 6.3 and 6.5 percent. And then we saw increasingly larger increases in fatalities as the speed limits got lower, interestingly. So, if we look at 40 and 45 mile per hour roads, we're around a 10 to 11 percent increase. Then as we get down to 30 miles per hour, we saw a 30 increase in fatal. Interestingly, 25 mile per hour, our lowest speed roads, fatalities were up by 98 percent, roughly.

So, you know, part of that is being driven, we suspect, by traffic congestion. So, obviously, there are fewer vehicles on the road and as a consequence, people are going faster. Then there are when crashes do occur, they tend to be more severe as a consequence of that. But with that being said, it's also been quite a bit of difference in terms of who's on the road. Unfortunately, that's where our level of detail in terms of the data we have is not quite as precise because in looking at some of those crashes, we see a lot more involvement with pedestrians and bicyclists. So, for example, pedestrian fatalities are up 18 percent yearly, and bicyclist fatalities were up more than 80 percent as compared to 2019. And, you know, part of that's probably explainable by the fact that we've got some, you know, transit systems that weren't probably being utilized as heavily, there were some closures and reductions in service, so part of it is you have more people walking and bicycling. Then actually if you think about the nature of the people that we're driving, the drivers we've seen at the onset of the pandemic and moving forward are probably not the exact same population of drivers we saw beforehand. We've got certain groups of people who telecommuted and are continuing to telecommute. You've got other groups who were probably less impacted by the travel restrictions that were introduced, so they may have kept driving as normal. And when you dig a little deeper into the crash data, we do find some troubling trends, and this is consistent with a lot of our prior work in that a lot of the highest risk groups and those that are subject to disproportionate numbers of traffic fatalities also engage in other high-risk behaviors. So, as an example, if we look at the increases in fatalities where drinking's involved, we saw an increase of roughly 14 percent as compared to only 11 percent where drinking wasn't involved. One of the stark differences is restraint use, and for what whatever reason, the group of drivers out there now has been less likely to wear their seat belts. So, as we compare 2019 to 2020, we actually saw a one percent reduction in the number of fatalities among drivers who were belted but a 13% increase in fatalities among those who were not belted. So, you got kind of an interesting—

Cranson: Yeah, do you have any personal theories on that in particular of why seat belt use declined during the pandemic? And I don't even know, in a modern vehicle, would people be going to, like, extra efforts to bypass all the things that your car does to make you want to wear a seatbelt? I mean, how is that even possible?

Savolainen: Well, that's a great question. I’ve recently purchased a newer car, and I get irritated because I’ll have my dog in the front seat, and if I don't put a seat belt in, I’ll have just an incessant beeping sound. And I haven't gone to the effort of disabling those systems. But I think a lot of this reverts back to that fundamental question of who's actually on the road and who's doing the driving, potentially. So, if you think about people who would be less inclined to, you know, follow the state home orders that were introduced and some of the other mandates, and I’ll try not to delve too much into politics here, but inherently, I think a lot of those people would also potentially be less receptive to being told what to do in terms of belt use and speeding and some of these other forms of more aggressive driving behavior. So, that's one of the things that we're really trying to investigate broadly here at the traffic safety community is just trying to understand what's been different. If we look at traffic volumes, I had noted the big reductions in 2020, but by the time we got to the fall, traffic volumes were almost back to normal. So, now, you know, it's more or less business as usual in terms of the number of vehicles on the road, but we've seen the sustained elevated fatality risk. And I know the most recent numbers I saw were that fatalities were up by over a hundred as compared to last year, so we're going to have our largest number of traffic fatalities in quite a number of years now again in 2021.

Cranson: Yeah, a 10 percent increase. It's terrifying and nobody knows this better than you as somebody who studies traffic safety, but, you know, we make these comparisons. And if 40,000 people die on the nation's highways, if 400 planes carrying 100 people went down imagine how society would react to that, yet somehow, we've become numb to these numbers on the roads. But going back to what you were talking about on the lower speeds and the lower posted speed areas where those fatalities have also gone up, anecdotally, I’ve seen that. I’ll bet you have too. In two lane roads around the city, you know, whether it's Lansing or Grand Rapids or Detroit, people are driving really fast, and obviously, you'd have to be for to have a fatal crash in a 25 mile per hour zone. You'd have to be going much faster than 25 miles per hour, right? I mean, given how our cars are built now.

Savolainen: Well, that would be true with the exception of, again, the most vulnerable groups: pedestrians and bicyclists.

Cranson: Well, yeah.

Savolainen: In that case, going from 25 to 30 or 35 has huge implications, but you're correct if we focus just on motor vehicle crashes and the crash worthiness of those vehicles. We do have to see quite some pronounced increases. And it's been somewhat interesting as a part of our speed limit evaluation, we haven't seen very drastic changes in most drivers. But it's really that highest speed subset of drivers, so those who are going 10 or 15 or 20 or more miles per hour above the speed limit where we've seen a lot of those of increases. So, I think those are some of the major problems, and then those problems are exacerbated by the fact that we still have issues related to driver distraction and things of that nature. And, you know, by and large, if you just get back to the fundamental physics, you know, speed kills, unfortunately. And for various reasons we're tending to drive faster than ever. And I think that's one of the things we may have been cautiously optimistic about is as traffic started to rebound to more typical levels that we'd see a reversion back to some of those lower fatality rates, but while traffic has started to become more congested again, the speeds haven't come down in a measure that would be, you know, what we would expect historically. So, we're all kind of, you know, scratching our heads really at what's been driving this at scale. So, I think that's one of the areas where, you know, we have really good data on how many vehicles are on the road, we have very good data on crashes, but we don't have great data in terms of who's in those vehicles and doing that every day driving by and large, so—

Cranson: I did talk to some troopers with the State Police last year, and they were optimistic, confident then, like you said, that when traffic volumes returned to pre-pandemic levels that those fatal crash numbers would go down, and I think they're a little surprised at what we're seeing. But you make a really good point about the cyclists and the pedestrians, especially in those lower speed zones. Do you think, I mean, long term would you be advising DOTs and cities, you know, transportation directors in various cities across the state where you have more people maybe commuting to work on a bike that they need to be thinking about design more? I mean, they already do. I mean, every traffic safety engineer and everybody who works in design takes very personally the idea that the road that they've designed could be as safe as possible. Roundabouts are obviously part of that, so, I mean, what do you think big picture needs to happen? Is there going to be a rethinking of some of the things that we take for granted now?

Savolainen: Well, if we step back and look at things from a 50,000-foot perspective, there was recently an article in the Atlantic actually where your podcast was cited and had to emphasize that a lot of these increases have been attributed to some form or fashion to driver behavior. So, the argument is well, what are we doing as a safety community to address things? So, there have been a number of important initiatives that have been picked up really over the last 10 to 15 years. If we go back, in 2010 we had the first edition of the Highway Safety Manual, which is basically a best practices document that uses a data-driven approach to improving traffic safety. In addition to that we've started going towards the track of what's referred to as a Safe System approach in design of these roadways. What that means essentially is we're operating under the assumption that drivers are going to make errors. And as a consequence, we try to design our roadways to accommodate those errors. And realistically, it's always been an implicit part of the design process, but there's been quite a bit of change, especially in urban areas where we have organizations like NACTO, which is the National Association of City Transportation Officials. And in these environments where we have large numbers of pedestrians, bicyclists, transit use, etc., we want to try to design those facilities so that we can manage speeds and we can try to keep those vehicles going within what speeds are reasonable given those contextual environments. And that's also a big change that's occurred with the most recent edition of our National Geometric Design guidance document, which is a policy for the geometric design of streets and highways for the AASHTO Green Book, so we've seen a real paradigm shift. Historically, we designed roads, you know, largely to get people from point A to point B as quickly as possible, and oftentimes, we'd find that would be at the sacrifice of safety. But what's happened since then is we have a lot better evidence now. We've got ammunition that helps us to understand if we vary certain design characteristics, so if we make lanes narrower or wider, if we make shoulders narrower or wider, if we introduce other differences from the driver's perspective and what the impacts of those are on safety and on crash risk and on the severity of crashes when they do occur. So, you know, I’m cautiously optimistic. And prior to the pandemic we had been seeing some positive momentum and some downward trends in some of these circumstances, but unfortunately, it’s not as if there's a silver bullet here. I think as we look at some of the high-risk behaviors even when we introduce some of those improvements, it's still largely going to be those same groups of drivers that are going to go significantly over the speed limit, that are going to drink and drive, that are going to engage in these other risky behaviors. So, enforcement I think certainly plays a role as well. That's been challenging because I’m sure we're all familiar that over the course of the pandemic we also had some interesting political issues that had risen with respect to law enforcement, and they were in kind of a difficult scenario. And, you know, there's so much that can be done. We look at things like distracted driving as an example, and we use some of the dynamic message signs to try to help curb that behavior. Research has shown that has, you know, marginal impacts unless you pair that with targeted enforcement, for example. So, we had actually just wrapped up a study for the National Safety Council where we worked with the Office of Highway Safety Planning, and we found that the use of those dynamic message signs showed some, you know, minor decreases in the rate of distracted driving, specifically cell phone use. But when we paired it with targeted enforcement, we saw much more pronounced decreases, and actually, after those campaigns had ended, we continued to see lower rates at least in a few weeks after they had completed. So, you know, we talked about what are referred to as the four Es of traffic safety, which are engineering, education, enforcement and emergency medical services, and really all those different roles are important. In order to really realize reductions in crashes, injuries and fatalities it takes a very broad coalition of stakeholders working together largely.

Cranson: Stay with us. We'll have more on the other side of this important message.

Narrator: [Phone ding] You looked. You looked again. Put the phone down and pay attention when you drive so you arrive alive. Remember: don't drive distracted.

Cranson: I want to get back to one of those Es, enforcement, in a minute. But since you mentioned that op-ed in the Atlantic, and I’m a huge fan and reader of the Atlantic, but it was an op-ed. It wasn't an article, and I think it was, you know, really kind of specious and thinly premised. I mean, it talked about foggy conditions, for instance, one of the anecdotes that cited and tried to suggest that's, like, somebody else's fault. And I would say—I think anybody, police, law enforcement, traffic safety engineers would say that's the very definition of driving for conditions, whether it's ice or snow or fog. I mean, how do you even argue that that's not a driver decision?

Savolainen: No, I agree, so just to clarify, that wasn't the first time there's been dispute to the use of the 94 percent statistic we see frequently cited. And that basically states that human behavior is a critical reason for 94 percent of crashes. Well, that doesn't mean that 94 percent of crashes are entirely the fault of the driver, but if you look at the actual crash reports and the circumstances leading up to these events ultimately, there's a chain of events that that occur. There are, you know, several times the number of crashes that occur on a daily basis that are near crash events where someone happens to break quickly enough, or they notice something just in time to stop and avoid that collision. So, as designers, you know, we're trying to make the roadway environment as accommodating as possible. We're very conservative from a design standpoint in several examples, so we assume that drivers react relatively slowly. We assume that they decelerate at relatively moderate rates of speed. So, as a consequence of that there's a considerable factor of safety that's built in, but we can't design away crashes. And I think when you when you look at that op-ed, I think the main point they were trying to get to is that, you know, it isn’t all on the driver, which I would agree with that. But it also, you know, in my opinion it had suggested that the education and enforcement and engineering communities weren't really doing anything about the problem, and that's where I think I would take a little bit of issue there. But it has actually led to quite a bit of discussion within the profession. So, I’m a vice chair of the Education Council of the Institute of Transportation Engineers, or ITE, and we've had a lot of discussion board posts. And we've got daily mailings via email where there's been considerable discussion and debate amongst the community about things we could be doing better here. But as I had alluded to previously, we really have taken more of a scientific and data-driven approach, And I think that op-ed also talks about the auto industry and the role that it can play here. And there's been considerable advances, you know, over the last 15 to 20 years, but especially things like adaptive cruise control, automatic emergency braking.

Cranson: Sure.

Savolainen: There's a lot of things that are being done by all these various stakeholders, but at the end of the day, you know, until and unless all these vehicles are fully self-driving, and we could have a much longer debate about if and when that will happen, the impetus is really on the drivers to be vigilant when they're behind the wheel. Every day when I’m driving, I see numerous people who are distracted by cell phones and various other things inside and outside of the vehicle. So, you know, I think there are certainly things that these transportation safety professionals can do to help to design for and mitigate these risks when drivers do make mistakes, but ultimately a lot of it does fall on the driver. So, I don't want that point to be lost, certainly.

Cranson: Well, yeah, exactly. I guess that's what bothered me so much about that. It tried to frame it as if we were trying to take everybody else involved in a transportation system off the hook and we're not. But this whole thing, this whole term “accident” originated with 19th century industrialists who knew that people in manufacturing were getting limbs lobbed off, and they wanted it to be an accident. And that carried over into the manufacture of automobiles and the industry wanting to say it was an accident. And every time you use that word you take everybody off the hook. Nobody calls a plane crash an accident, so I guess that's the part that frustrated me. I felt like that was kind of lost in that diatribe. But let's get back to enforcement for a second because you mentioned that as one of the four Es. What do you think, I mean, setting aside laws that would have to change and policies, what do you what does your research tell you about automated enforcement and what a difference that really can make in driver behavior?

Savolainen: So, there's really two general areas where automated enforcement has been used at large scale, and the results have been I would say somewhat mixed for various reasons. Those two areas are first of all, automated enforcement in speeding violations where we're using things like video cameras to try to track vehicles and automatically issue citations when people are traveling over the speed limit. What the data from those studies conclusively suggests is that those types of laws and the ability to automate enforcement definitely bring speeds down. And I know from personal experience when I’ve driven through communities that have automated enforcement, I definitely do reduce my speeds as a consequence. So, what the data generally suggests there is that speeds go down, and when crashes occur, they tend to be less severe. The other area though of automated enforcement would be red light running cameras as an example. And there have been some fairly highly publicized stories that have been written, especially in Chicago about those systems being used, the argument is, as a money-making machine for law enforcement. There's actually been some evidence that suggests organizations and municipalities may set those the yellow intervals at these signals artificially low so that there's a disproportionate number of red light running. So, when you look at the automated enforcement research there, it typically shows dramatic reductions in the most serious crashes where vehicles would run the red light. But what tends to happen there is that you'll also have a significant increase in rear end collisions because drivers don't want to get ticketed. So, from that perspective, I think automated speed enforcement is an area where we certainly see some potential and some tangible benefits. Automated enforcement of red light running, I think you've got to be somewhat careful in terms of when and where you apply that, but I think by and large, the general results have been positive there. As you had alluded to, we don't really have that capability presently in Michigan. It would take changes to state law in order for us to introduce automated enforcement, and I’m sure there would be a number of groups that would be displeased with that. But I can say certainly from my own perspective, I would be likely to drive slower if I was under a continual threat of being cited as a consequence of that as well because without that enforcement measure, you know, you're just assuming and relying on people to behave in the best interest of everyone. And from our own personal and selfish perspectives, one of our objectives is to get from point A to point B as quickly as possible. So, you know, you're working against the prevailing needs and interest of many drivers in trying to reduce speeds in this manner.

Cranson: Well, you mentioned when we do get to fully automated vehicles at some point, like you said, I think that's a discussion for another day. But the things that stair-step us there along the way like adaptive cruise control and lane assist are in their own ways revolutionary already. I think that's one of the other puzzling and frustrating things is that we're inventing this technology to keep us all safer and even while we're doing it it seems that we're finding, you know, more ways to speed and distract ourselves. And I also know, as somebody who has adaptive cruise control, that there's nothing more frustrating than the people who come from behind you and think they have to fill in that gap because you must be going, you know, too slow. So, it'd be nice if we don't have all automated vehicles soon, if everybody just had adaptive cruise control just imagine what that would do. So, back to messaging for just a second. We talked about the dynamic message signs and some of the study and research that Michigan State has done on those. I want you to listen to this from a conversation last year with Jonathan Atkins at the Governor's Highway Safety Association and a question I posed to him about what does make a difference.

So, what does alter behavior? What do you think, you know, can actually break through and make people think twice, whether it's slowing down in a work zone or not using their phone or at least not texting while they're driving I mean, any of these things?

Jonathan Adkins: Well, it's evolving a little bit, you know, before the pandemic, I would have said a strong enforcement message, you know, to remind people that if they break the law they're going to get a ticket. We know that that works. Click It or Ticket has been effective, but the way we think about law enforcement in this country is certainly changing. So, there's some indications and there have been some studies as well that indicate that a softer message can be more effective. For adults, a message of slow down so you get home safely to see your kids tonight can be effective. We're seeing some signs of that. For young drivers, they don't fear death, but they fear disfigurement. So, we've had some studies and some evidence that show that that fear of you may be in a traffic crash and be hurt, you may not be beautiful anymore. That can resonate, but just a simple “don't do it” message isn't going to have much impact.

Cranson: So, in their adolescent mind there really is a fate worse than death?

Adkins: Yes, you know, not being beautiful, being disfigured in some way. We've seen that in some studies, but also in presentations with students against destructive decisions and other youth groups. There's not a fear of dying, but there is that fear of disfigurement.

Cranson: So, what do you think about that? Does that track with anything that you've seen either in data or anecdotally?

Savolainen: Yeah, no those were a few important points that were brought up there because as a part of the aforementioned MDOT study we recently completed, we basically looked at these messaging strategies from two perspectives. Now, in one case we had conducted statewide surveys to try to understand what types of messages were most likely to resonate at least based upon people's self-reported perceptions and how they would respond to the use of these messages. So, to that end, we did actually see that, to reinforce the preceding discussion there, if the message was on a more personal level, like it talked about the risks posed to a person's family or, you know, there was some local connection, there were some advantages to that. There was less support, I guess, for more punitive messages. But with that being said, when we look at the use of those messages in practice the situation kind of shifts to a degree, and it really varies probably across different groups of the population. So, one of the challenges to the safety community is to try to understand how to message these things. I mean, that's something we could talk about broadly and this would be a whole nother rabbit hole. We've got such a drastic divide politically where we agree on 95 percent of issues. But I think just the way that things are messaged, you know, we've got certain groups that are at polar opposites in areas where we should have, you know, clear consensus, and traffic safety is certainly one of those areas. So, I think trying to raise awareness amongst the public, you know, you had cited the 400 plane crashes per year, essentially that is what we're experiencing on our roads. But the public I don't know if they're desensitized or just ignorant to these issues, but it's a real public health dilemma. So, I think it's going to take creativity from everyone involved. Then, you know, to the point of the self-driving vehicles, what I’m really curious to see is during the transition period. To your point about, you know, adaptive cruise control and then you've got somebody tailgating you, there's been research that shows with just a few percent of the fleet being autonomous you can eliminate traffic congestion. But that's a by-product, you know, of other people adapting their behavior as well. So, you know, there's a lot of complexity of the issue here, certainly. I think, you know, we're really just scratching the surface. We're continually trying to strategically think through what some of the efforts are that could help us to drive these numbers down, but it's been it's been a struggle this year, certainly. We're hopeful that we can start pulling some of these numbers back down, but 2021 has certainly not been a banner year from that standpoint.

Cranson: Yeah, well that's probably a good place to end it. As always, I really appreciate you and your colleagues at Michigan State University and others involved in this kind of research. I think that it gets frustrating. Sometimes it seems to defy logic or explanation, but the important thing is to keep thinking about it and keep trying to figure it out and try to make the improvements that we can. So, I really appreciate you and the work you do.

Savolainen: Thank you very much, Jeff. I appreciate the opportunity.


Cranson: Thank you again for listening to this week's edition of the Talking Michigan Transportation podcast. I would like to thank Randy Debler and Corey Petee for engineering this week's podcast. To subscribe to show notes and more, go to Apple podcasts and search for Talking Michigan Transportation.