A Turn For the Worse?
April 4th, 2008My friend Dan and I are a bit obsessive about traffic. We may not discuss “three phase traffic theory” over lunch, but we’ve definitely discussed this traffic site on more than one occasion, which describes how to break up traffic jams through careful driving. I’ve tested out its advice a few times, and even by yourself it definitely works. (However, if you try out that method be prepared for some angry drivers who think you’re just driving like a grandma. It may lower your speed even more than usual, so it’s not for the impatient.)
Anyway, that’s not the point of my post. I just wanted to give you a little background about me and my little traffic obsession. The real question I’ve wanted to answer for a while now is: how many people use turn signals, and does the type of automobile involved matter? Everybody seems to think that no one uses their turn signals. I’ve talked to people from all over the U.S. and no matter where they’re from, they’re either convinced that nobody uses their turn signals here (in Tampa, FL) or where they’re from, yet until now I’ve never seen a lot of data on the topic.
After digging around a bit, however, I found some good data out there. There’s an excellent story on turn signal use here, based on the Response Insurance National Driving Habits Survey. That survey has some really fascinating findings, but as usual the data is not published freely. (There is a sad underutilization of data-sharing sites like Swivel, which I blogged about last week.)
With just a few statistics and no direct source of data (how honest are people in traffic surveys?), it was time for Dave the “data vulture” to go back into action. For the last few months I’ve been watching automobiles on the road, seeing whether people use their turn signals or not in real-life situations. (If you didn’t think I was weird already, I may have just cemented that fact in your mind.)
I broke out turn signal use into three categories: Used signal early (A), used signal late (B), and did not use signal at all (C). I decided that anyone who used their turn signal before the absolute last second was doing an “A” (which shows evidence of preparation), and any other usage was a “B” (their heart was in the right place). If there was no turn signal, I called it a “C”. Since I believed that vehicles of certain types behaved differently, I also recorded the type of vehicle, which I broke out into 8 categories: Bus, Economy, Motorcycle, Semi, Sports Car, Sports Utility Vehicle (SUV), Truck, and Van (includes mini-vans). So, for example, if a saw a sports car use their signal ahead of time, I recorded “Sports A”.
Sounds simple enough, right? Actually, it wasn’t. First of all, I had to decide what constituted a turn. Should I just count turns on public roads? How about pulling into parking spots? Or changing lanes? In the end, I just figured when someone had a choice to turn, that was a turning “event”. It didn’t matter whether they were in a parking lot, a turn lane, or a neighborhood. If they turned, I counted it. I left out many other events that people use turn signals for, though, like parking, lane changing (a lost cause, I assure you), and on/off ramps to freeways. Still, that was a pretty strict criterion. I mean, who uses turn signals in their own apartment complex? (Besides me, that is.) I thought it was the only fair way to go, though, so that’s what I decided on.
Second, not every vehicle fit into my eight categories. What kind of vehicle is a Jeep? If you saw the old-school military-style model, you might think it’s basically a covered truck. If you see newer models like the Jeep Cherokee, you might think it’s an SUV. Or you might even think it deserves its own category. In the end, I called them SUVs. I found lots of examples of other borderline cases as well - is a Mini Cooper an economy car or a sports car? How about a turbo Volkswagen Beetle? Or a four-door Acura Integra (now the TSX)? Most of the time I had to rely on my own best judgment, obviously.
Third, some people really pushed the barrier between an “A” and a “B” turn. In a few cases I had to throw out data because I honestly couldn’t tell. What if someone uses their turn signal at the last minute, but then has to stop before they turn? Or if they turn very slowly but still use their signal very late? Again, these were judgment calls.
Lastly, I tried to make sure I was watching the road as much as possible. Let’s face it, it’s hard to keep one-hundred percent attention on road all the time. I’d say I got data on 85-90% of potential “turning events” that I saw (more on this in a second), but I definitely let some go by when my mind was wandering or when traffic was tight. Is this a potential source of bias in the data? Yes, but I don’t think it was too big of a deal, since I was as vigilant as possible.
I had another big concern to deal with, however. Driving has a strong emotional component to it. I use my turn signal 95-99% of the time, which puts me in the distinct minority. I use it at much as humanly possible. That means I have an obvious bias against drivers who don’t do so. (Just being honest here.) Thus, I had to try harder than usual to collect data in an impartial fashion. If I just collected whatever data I could find on the road, I could easily single out drivers I didn’t like or those who didn’t signal, and then I wouldn’t have a representative sample.
So I came up with a strategy: I only looked at the automobile directly in front of me. If they happened to turn, I would use that event as data. I ignored everyone else. That meant I had little control over what data I was recording or what automobile I was watching. Plus, it allowed me to focus on what they were doing so that I could get the best possible data.
However, that also meant I missed a LOT of data. It took me several months to get 350 observations of 299 vehicles, which is the most time I’ve ever spent getting data for a post here. To minimize other sources of bias, I also only recorded data when I was driving alone, and I didn’t go out of my way to collect any data. I just simply drove how I normally do, which meant I got the majority of data driving to and from work.
Here were my findings, in a nutshell:
(You may notice that there are only 297 vehicles on that list, in only 6 categories. I left out the lone bus and motorcycle observations I had, which meant axing those two categories and data points. Also, I left out the “B” data since there were only 9 “B”s, but I included that data in the totals, so some of them may not add up like you expect.)
Conclusion? People use their turn signals just over half the time. (54.6% if you leave out the few “B”s that signal at the last possible second.) Is that good? Well, to be honest, I thought the percentage would be lower. And remember, I was counting all turning events, even in extremely casual situations. Depending on your point of view, then, people use turn signals more or less than you may have thought. And when they do use their signals, it tends to be in advance, usually 2-3 seconds before the turn. I didn’t really expect that, but that’s what I saw. All those last-second turn signals I saw before I started collecting data were probably vehicles changing lanes instead (something I observed much more often).
Another shocker was that SUVs used their turn signals slightly more than average, even more than economy cars. I’d read in this Malcolm Gladwell article (and elsewhere) that SUV owners were typically more insecure and less skilled drivers. I also read there that SUV owners were some of the most likely to die in accidents. Turns out that despite what I read, they were better signallers than I thought (maybe because more women drive SUVs). In a further twist, economy cars use their turn signals at a slightly below average rate (even though they are the most common type of car on the road), and those drivers are much less likely to be killed in a crash (according to the Gladwell article). Who knew? I always figured that economy cars, semis, and vans were the best turn signal users on the road. Two out of three ain’t bad, I guess.
The real kicker is that sports cars were the second best turn signallers on the road, at about 63% turn signal use (though they were a small part of the overall dataset). Ironically enough, my friend Dan (who I mentioned earlier) is also an obsessive signal user like me, and he drives a sports car. Maybe I should’ve guessed.
I’ll tell you one thing that didn’t shock me a bit: Trucks were absolutely dead last in turn signal use by a huge margin. (8.7% below average in turn signal use!) I guessed beforehand, based on my experience, that truck drivers (the four to six wheel kind) were the least considerate drivers on the road. And why is that? Because they’re mostly driven by young, aggressive, male drivers like myself. Takes one to know one, I suppose. If you don’t believe me, you can read that link I posted earlier, which confirms that young men are maybe the least likely group to signal on the road. Many of them are driving trucks.
And, if you read that Gladwell article, you’ll remember that the vehicle with the worst ranking for fatalities was the Ford F-Series truck, which ironically is maybe the top-selling vehicle in the nation. That means those plentiful Ford F-Series drivers are also some of the most likely to injure themselves and others while also being the least likely to use turn signals. It’s a deadly combination of less-safe vehicles and young, impolite, aggressive male drivers. If you had to pick a kind of vehicle to stay away from on the road, trucks would be your best bet by far.
But enough about stereotypes. Let’s delve a little bit deeper. Are there obvious patterns to turn signal use? Here’s what I found on turn signal use for the 41 vehicles (out of 297, or about 13.7%) that turned more than once:
As you can see, people are just as likely to be inconsistent in their use of turn signals as not. My guess was that people were like me, where they were always using their turn signal or not bothering at all, but it turns out that I was mistaken.
If there’s such inconsistency, why do people signal, then? I have a theory. A while back, my buddy Dan had an excellent observation that stuck with me while I was collecting data: “There’s only one spot in America where everyone uses their turn signal - claiming a parking spot.” Now obviously that quote is a bit tongue-in-cheek, but I think there’s a lot of truth to it. What I saw over and over again was that people were using turn signals in uncertain and ambiguous situations. The turn signal helped them assert themselves and define the situation.
In fact, that’s probably the only commonality I ever saw in turn signal use. Despite my general findings, I was often wrong about particular vehicles. I would size up someone driving an SUV (or truck) aggressively and think, “They’ll never signal here”, and then they would. Or I’d see the most courteous driver on the road, someone who let 3 people go in front of them, and watch them come up on a clear signalling situation - like turning off a busy road onto a side street - and they wouldn’t signal.
I figured people would be too distracted to contemplate whether they should use their turn signals, but it turns out they did just that all the time. I think they’re constantly subconsciously sizing up situations. After a while I got extremely good at guessing when vehicles would signal just by the circumstances. Even the least considerate driver on the road will use their turn signal when they have to dart out in front of cross traffic while turning. It’s a weak situation to be in, and I figure the turn signal helps them feel more in control.
Thus, even though the type of vehicle involved (and indirectly, the type of driver) matters, what matters much more is the circumstances in which they’re turning. The type of vehicle merely helps fine-tune the guess in borderline situations. You’d probably do just as well by noting the driver’s age, gender, and whether or not they’re on a cell phone, to be honest. (In fact, as many of us know, car insurance actuaries do consider all of those, except cell phone use.)
I should mention that I have reservations with my data. I had a small sample size, which was a little less than perfect. And while the percentages of vehicles in my sample are fairly comparable to what people in the U.S. drive, I still think the topic is worth revisiting. There could be a seasonal bias here, for example. During the holiday season, Tampa gets legions of unexperienced drivers from up North that we often call “snowbirds” (or “damn tourists”).
Thus, I’d like to get some data during the “off-season”, when less tourists are here. Look for a followup post on this topic in a few months (if I can bother myself to collect that much data again). Besides, I’d love to be proven wrong. I can’t wait to see some 80-year-old lady in a mini-van speed past me into my lane and, tires squealing, make a quick turn without a signal while flipping me the bird. It’s ok, lady - you’re just another “Van C” to me.
(Though I covered most of my dataset pretty thoroughly, here it is in case you want it. As usual, it’s not very cleaned up but certainly usable.)
| | | del.icio.us |
April 4th, 2008 at 7:22 pm
I was surprised at your finding that people are so consciously involved in each decision to use their turn signal and that there weren’t a lot of other strong correlations.
As usual, another great article. I seriously don’t know how you do it.
April 13th, 2008 at 2:54 pm
Before I get into the full swing of my comment, I’ll preface this with the fact that I do believe I’m breaking my “datamineshaft commenting cherry”. Therefore, you’ll have to excuse the excessive length of this comment.
That being said, I found this entry to be one of the most interesting articles I’ve read in quite some time. Not just because of the inherent bias that comes along with being a friend of yours, but because of the fascination I have with traffic and the ire and stress that I feel nearly every time that I’m in traffic.
First off, the link towards the top of the entry regarding experimental ways to ease traffic congestion was a great resource to look at and think about. Early on in my “driving career”, I realized that traffic is definitely a fluid entity. And I’ve so often been perplexed by the “invisible accident”, the traffic jam that seems to have appeared out of nowhere, and for no reason whatsoever. It was very insightful to me that with just a little bit of extra thought, anyone could come to the same conclusion as to why this occurs all the time. So, hat’s off to you for including that, and I’ve bookmarked the site so that I can do a little experimentation myself.
Also, the findings from the data you crunched were indeed fascinating. Most surprising to me was the data on SUV’s. Of course, when you discussed some of the latest demographics on who drives SUV’s nowadays (and the demographic of those who drive pickup trucks), this surprise was tempered.
I’m interested in the fact that you are looking to revisit this topic in a future post. I was both very impressed and interested in the topic and the findings that came with your observations. If you do choose to revisit further down the line, perhaps I can offer a few ideas.
For instance, it would be interesting to see a breakdown of turn signal behavior not just on public roads and highways, but also in neighborhoods/apartment complexes and in private parking lots. You mentioned your reasons for excluding these categories in your post, and I understand that to include these areas would undoubtedly take much more time and effort to make these extra observations, so I’ll leave that in the “it would be interesting to see” category and leave it at that
Also, you mentioned that you’re interested in making observations in the off-season of tourism. Another thing I’ll throw into the “it would be interesting to see” category would be what turn signal habits are like during the winter holiday season. As we all know, that magical time between Thanksgiving and Christmas sure brings out the worst of everybody when it comes to driving.
In conclusion (heh, finally), I thought this was an excellent read. I gained a lot not only from the entry itself, but from the numerous links peppered throughout. That was definitely the most open tabs I’ve seen on my web browser for quite some time. Also, as I’m exceedingly interested in the topic, I would gladly volunteer some time if you need any help making observations, should you choose to revisit this topic in the future.
All in all, bang up job, man.
Wow, that’s some great feedback Dan. Thanks. You’re right, it’d be quite hard to add extra features to the data, though I considered a few things like you’re talking about, where you take into account context, gender of the driver, number of occupants, etc. As you read, though, it was quite a challenge just to get the data that I ended up with. I thought of some interesting analyses you could do with more data, but then again you might get in an accident trying to collect them!
I’m with you about the seasonal data, obviously. If you want to help, feel free man. As long as you’re as anal about data collecting as I am, I could definitely use the help. I can show you the methodology in a few minutes or less (though you probably understand already from reading this post). A smartphone or some kind of portable recording device is a must, though. Pen and paper won’t cut it here. I’m thinking of starting up here in a month or so, when I get back from my lengthy vacation. I’ll probably ask you then if you’re down to help. Thanks again!
- Dave
July 5th, 2008 at 9:07 pm
Grrreat article. I often wonder these kinds of things to myself when I drive (and I drive a *lot*), so it’s nice to see someone write about it!
But really, throughout pretty much all of the article, I couldn’t help but worry about if I used my signal enough that day that you guys rode with me to Freebird’s…. @.@ Now I’m REALLY going to be self-conscious when I drive…thanks, Dave.
Like you forget you’re breathing (until just now).
Thanks for the kind feedback, Lindsey! And you might be self-conscious when you drive, but we humans are good at adapting. Unless you were already obsessive to begin with, you’ll forget all about this stuff soon enough.
- Dave