35

Most of us will have been taught the two-second rule when learning to drive.

The two-second rule is a rule of thumb by which a driver may maintain a safe following distance at any speed. The rule is that a driver should ideally stay at least two seconds behind any vehicle that is directly in front of the driver's vehicle.

But what research is there to lend weight to this rule of thumb?

I always liked to think that giving a 2 second gap between you and the vehicle in front was clearly a sensible idea.

Unfortunately, increasingly on British roads I see people driving just a few meters behind the vehicle in front. At 70 miles per hour, 3m puts you just one tenth of a second behind that vehicle. To put that into context, the Stanford press release Sleep deprivation shown to have as much impact on reaction time as alcohol, talks about reaction times in excess of 2 tenths of a second (200ms) for Navy Fighter Pilots, while the wikipedia mental chronometry page suggests a mean reaction time to a visual stimulus of around 190ms for college-age individuals.

Have there been studies looking at the relationship between following distance and accident rates? Between following distance and accident severity? What about between following distance and the number of accidents that would have been near misses if the following distance had been greater?

I have had a look, but it's not my field so don't have much of a clue as to where to start looking for such data or which journals would publish such studies.

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Mark Booth
  • 1,363
  • 13
  • 22
  • 1
    I would think that there are a lot of studies based on people's reaction times. Also, the automotive industry may have some published reports when talking about the safety features of their cars. – JasonR Dec 14 '11 at 17:03
  • 1
    I am not sure what you are skeptical of. You seem to believe in the 2 second rule and that it is unsafe to drive too close. I think there is a good question in here but I don't think it meets the skeptics standards which require a single notable claim to be skeptical of. This is almost the start of a good answer to "Is it unsafe to drive to closely?" – Chad Dec 14 '11 at 17:08
  • Reaction time for college aged individuals is about 160msec http://en.wikipedia.org/wiki/Mental_chronometry –  Dec 14 '11 at 18:31
  • A counter argument goes like this: The car in front of you isn't likely to stop instantly. More likely someone will tap on their brakes. Here's the numbers for that: A car will travel 3 meters in 160 msec at 67.5 km/h A car going 60 km/h will travel 2.67 meters per 160 msec The difference is 0.33 meters per second So if the driver in front of you slows down by 7.5 km/h (4.6mph, a tap on the brakes), you have about 9 seconds to react before you'll hit. That may seem like a lot, but you have to be paying close attention to use that time productively. Changing the track on your iPod to one you li –  Dec 14 '11 at 18:54
  • 9
    I assume that there is a probability curve describing the chance of an accident (and a smaller one describing the chance of a fatal accident) compared to the size of the gap at a particular speed. It would certain rise steeply as the gap reduces, but I don't expect any discontinuity at 2 seconds. My point is 2s is unlikely to be a magic number, but more a risk-assessment rule of thumb, about where society deems the risk-reward acceptable. – Oddthinking Dec 14 '11 at 23:58
  • Aside: I realised as a learner driver that the 2 second rule means all roads have the same bandwidth (maximum cars moved per hour) even if they have different lag (time taken to get to the other end) independent of the speed-limit. Then I read a book on traffic management (2 decades ago - don't ask me for a reference!), and found that wasn't true in practice, suggesting people don't follow the two second rule. – Oddthinking Dec 15 '11 at 00:01
  • I thought it was 3 seconds, assuming the velocity of the car in front of you is greater or equal to your own at any given moment. – Mateen Ulhaq Dec 15 '11 at 03:41
  • 2
    Think of it this way: if a oncoming car suddenly veered into the other car, causing them to be at a standstill, would you want to be 120ms*60kph away, or 3s*60kph? (Or maybe you would still hit, but at least it would be at <60kph.) – Mateen Ulhaq Dec 15 '11 at 03:45
  • @Oddthinking - Indeed a 2 second gap in Chicago on I-55 during rush hour - would quickly be filled by about 8 to 10 cars... I agree there is a probability curve I would like to see where 2 seconds falls compared to 1 second though. – Chad Dec 15 '11 at 21:59
  • 1
    I was taught 2 seconds at driving school, but my habit is a minimum of 3 to 5 seconds -- in dark conditions or rainy whether, I tend more towards what I call "the 5 second rule." _Prevention is, after all, one of the most effective cures._ – Randolf Richardson Dec 16 '11 at 03:24
  • 1
    @RandolfRichardson - The two second rule is intended for ideal driving conditions, good weather, you are alert, no-one is tailgating you. If there is anything is less than ideal then 3-4 or more seconds is recommended anyway. Two seconds is just the base-line. I have seen advice that if someone is tailgating you, you should increase the gap to 3 seconds, not 4 (which might seem logical) as the latter is more likely to encourage them to (possible dangerously) overtake you to get *into the gap*. – Mark Booth Dec 16 '11 at 11:29
  • 5 seconds... You could land a 757 in that gap at high way speeds. – Chad Dec 16 '11 at 19:32
  • 1
    http://psycnet.apa.org/journals/apl/52/1p1/49/ Taxi cab drivers who were slow to react, but could hit the brakes quickly when they did hit fewer cars from behind, but were hit from behind more often. An interesting reason to leave a larger gap even if you think you have fast reactions and good brakes - to give time for the person behind to stop too. – Oddthinking Dec 17 '11 at 03:46
  • 1
    http://www.scribd.com/doc/44530777/Chap-4 Discusses ANOTHER factor against driving too closely - the stability of the traffic flow is influenced by the buffer size. The ripple effect of one car slowing down and speeding up again can cause quite significant effects further back in busy traffic. – Oddthinking Dec 17 '11 at 03:54
  • 1
    Maybe `Is there anything special about 2 seconds as opposed to 1 second or even 1.9 seconds?` was the reason to vote to close. Of course, most humans could not estimate in the range of 1/10 s, so as a rule of thumb, 1.9 s would be as useless as 2.1 s. – user unknown Dec 17 '11 at 13:01
  • It does so, but it wasn't me, who voted to close - when I tried to do so, I was informed that this is not possible, if there is an open bounty. – user unknown Dec 17 '11 at 21:13
  • 1
    The reason I removed the "what support is there" part of the title is that every single Skeptics.SE question is asking for that. It is implied by context, so the title is pithier without it. – Oddthinking Dec 17 '11 at 22:15
  • @Mark Booth: Interesting. On a side-note, I do hope that the first course of action is to pull aside for vehicles "following-too-close" (_the term for tail-gating used by RCMP officers I've spoken with over the years_) so they can pass. This is what I prefer to do -- I figure "they're not going to back off anyway, so why be in their way if I can safely pull aside for them?" (and as I jokingly tell my wife "Look at that, another volunteer radar detector!" as they speed past). – Randolf Richardson Dec 19 '11 at 04:48
  • If the rule was 2.1 or 1.9 seconds, then many people would say "that's too complicated, I can't do that", and not even try to keep that distance. With a simple rule of two seconds, many will at least try. – gnasher729 Aug 25 '15 at 22:22
  • @RandolfRichardson I had someone really "following too close" and I took evasive action, turning suddenly and so on, and it turned out to be a police car (hidden lights), going to my apartment complex! I should have pulled aside, but I was a few hundred meters from home... I explained to him that his tailgating seemed aggressive and frightened me. He accepted that. When I maintain a larger gap ahead, people universally swirl past and insert themselves, forcing me to back off still farther. But I have never seen anyone actually drive faster than the stream of cars ahead of them. Do they learn? –  Oct 08 '15 at 23:28

2 Answers2

19

There's a good discussion of the issue of driver reaction time and the factors affecting it (with a couple of references) here: http://www.visualexpert.com/Resources/reactiontime.html

The best estimate is 1.5 seconds for side incursions and perhaps a few tenths of a second faster for straight-ahead obstacles.

Based on that article, I would say that the two second rule should be considered a minimum distance at which to follow and the following distance should be increased as factors dictate. For example, if you're tired, your reaction time will be longer. Allow yourself more distance. If it's dark or rainy or the road is slick, your ability to see an obstacle and stop will be degraded. Allow yourself more distance. If you have a talkative companion in the car distracting you, your cognitive load will be higher and your reaction time will be longer. Allow more distance.

While it probably doesn't have much relation to one's ability as a driver, a fun way to test one's raw reaction time is here: http://www.humanbenchmark.com/tests/reactiontime/

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Tom Barron
  • 452
  • 3
  • 3
  • Although Dr Green only cites two of his own papers at the end of his article, and the first now only seems to be available [behind a pay wall](http://www.tandfonline.com/doi/abs/10.1207/STHF0203_1), he does publish his second paper [on his website](http://www.visualexpert.com/Resources/realprt.html) which has some relevant analysis and this paper cites many more references, so thanks. – Mark Booth Dec 17 '11 at 18:07
  • 4
    The more you increase the distance from the car in front of you, the more the probability that some car squeezes in! – B Faley Jan 27 '12 at 19:02
  • @Meysam: ... and you have to slow down even more! :) So, you're safer. – Tom Barron Apr 13 '14 at 11:30
  • @TomBarron: so the most reasonable course of action is to stop. Very safe that way, with infinite following distance! If only everyone would realize this. (sigh) : ) –  Oct 08 '15 at 23:29
8

Iowa Driving Simulator studied reaction times in "Driver Reaction Time In Crash Avoidance Research: Validation Of a Driving Simulator Study On a Test Track". This study found that when a sudden threat appears, the timeline for driver reaction was:

  • t=0.00, threat appears
  • t=0.96, driver begins to release accelerator
  • t=1.64, driver begins to steer to avoid the threat
  • t=3.16, driver reaches maximum braking (0.96 + 2.2)

So suppose the driver in front of you suddenly slams on the brakes. If the reaction times found by this study are typical, you will reach maximum braking around 3.16 seconds later. Other things being equal, you will collide with the vehicle in front of you if you are less than 3.16 seconds behind it. The study cites other studies of collision avoidance behavior.

National Highway Traffic Safety Administration looked at the relationship between following distances and accident rates, which you specifically asked about, in "A Front-End Analysis of Rear-End Crashes". This study identified "following too closely" as a principal causal factor in 1 out of 5 rear-end crashes but did not say what their criterion of "too closely" was. If the data provided in this study are reliable, about 162,000 people are injured annually by insufficient following distance, and 600 people are killed.

MetaEd
  • 1,737
  • 1
  • 16
  • 13
  • 1
    As a police officer of many years experience, it's apparent to me that "following too closely" coupled with "driving too fast for conditions" is the causative factor on an inordinate number of accidents. – M. Werner Dec 20 '11 at 19:28
  • Excellent references, but only tangentially related to the problem of the following vehicle scenario. The first paper related to side incursions, which appear to significantly lengthen reaction times, while the second paper concentrates on RE-LVS (Rear-end Leading Vehicle Stationary) scenarios. It's a real shame that the RE-LVm (Lead Vehicle Moving) data set analysed is so small. It's difficult to draw any conclusions from that data other than that driver inattention is *significantly* more important than following-too-closely. – Mark Booth Dec 20 '11 at 19:50
  • @MarkBooth Thanks. What is your source for the claim that drivers take longer to react to side incursions than to sudden stops of lead vehicles? Also, my initial impression of the second study was the same as yours, that it was not relevant to collisions with moving lead vehicles. But then I realized that the LVS data are equally relevant, because "following too closely" was as much a factor in collisions with LVS as with LVM (1 in 5). In other words, those are the vehicles which stopped suddenly and then got plowed into from behind due to "following too closely". – MetaEd Dec 20 '11 at 21:24
  • That came from the references cited in [Tom's](http://skeptics.stackexchange.com/a/7269/2772) answer, which pointed at another of Dr Green's [papers](http://www.visualexpert.com/Resources/realprt.html) that mentioned it in one of the criticisms of the standard Olson & Sivak paper: *In contrast, many collision scenarios involve a lane incursion where a vehicle or pedestrian approaches from the side. The obstacle then first appears in peripheral vision, where visual sensitivity is lower and attention is weaker*. – Mark Booth Dec 20 '11 at 23:18
  • With regard to LVS data, it is quite a leap to assume that the LVS data is relevant to an LVM situation, just because there is a correlation in the statistics doesn't mean that the causes are the same, or the tactics to avoid those kinds of accidents. This was one of the problems with the study, they complained that their main source of information didn't distinguish between LVS and LVM rear-end shunts. – Mark Booth Dec 20 '11 at 23:25
  • @MarkBooth Yes, and it looks like the NHTSA study cites shorter driver reaction times, which also supports what you are saying about the relevance of the side incursion reaction time. I must not be getting what you're saying about the use of the LVS data in the NHTSA study. If the study specifically calls out "following too closely" as a cause of 1 in 5 LVS collisions, how are these collisions not directly relevant to the OP's question about a relationship between following distance and accident rates? – MetaEd Dec 20 '11 at 23:35
  • 6
    A little addendum: the vehicle in front started braking (even if at maximal breaking power) 3.16s ago... but he won't stop on the spot, giving the vehicle following it more than 3.16s before collision. Not much, but still – ADB Dec 21 '11 at 23:30
  • If you are at two seconds distance, you can't afford that your reaction clock would start when the car in front has reached maximum braking power. But then it usually does't - your clock doesn't start at the 3.16s, but either when the other car's brake lights go on, or when you notice the danger, which should happen only very shortly after the driver in front of you notices. – gnasher729 Sep 09 '14 at 15:07
  • @gnasher729 The 3.16 second gap that matters is the one after the brake lights ahead of you come on and before you yourself reach maximum braking power. – MetaEd Sep 13 '14 at 18:55
  • 5
    @MετάEd: That's wrong. Two seconds distance means you are fine if you copy everything the driver ahead does with two seconds delay. If your brake lights go on two seconds later than his brake lights, then your car reaches maximum braking power two seconds after the other car does. – gnasher729 Aug 25 '15 at 22:20
  • If you strike the car ahead of you, the law concludes that you were following too closely. This is the only sane way to look at things, unless some completely unexpected circumstance occurs, like the car ahead of you dumping oil on the road. Where I live, you will always get a ticket for rear-ending. It is true by definition. –  Oct 08 '15 at 23:32
  • @gnasher729: Unless a test involves a driver reacting to a car in front that engages maximum braking, I don't think it will accurately measure how long it would take the following driver to apply maximum brake. In most scenarios there is no need for immediate maximum braking, and a more measured initial response will be safer [among other things, it will give the following driver more time to react]. – supercat Apr 11 '16 at 18:27
  • No on the 3.16 seconds. You applied some braking before that 3.16 seconds and also the driver in front didn't instantly reach maximum braking. What's missing from this list is the point where you start to push the brake pedal. – Loren Pechtel Jan 21 '20 at 15:24
  • @user29285 Assuming the car ahead was already driving in the lane (not a swoop and squat) and that the car ahead didn't hit something. – Loren Pechtel Jan 21 '20 at 15:26