Jan 17 2014

When are anecdotes helpful?

In the discussion following my post about the use bicycle helmets, many spoke about their own personal experiences with its use and how it influenced their decisions on whether to use them. An interesting side discussion then ensued about the use of anecdotal information in making decisions.

When it comes to making public policy, one clearly should not rely exclusively on anecdotes because those can be all over the place. One needs statistically significant patterns in the data before one makes up and enforces rules on the population at large. But anecdotes can be useful in that process in providing concrete examples of abstract ideas and it is also true that anecdotes are often the first indicators of whether there may be something worth investigating further. For example, clinicians often write up interesting cases for publication and these can turn out to provide important insights that lead to controlled studies.

On the other hand, when it comes to personal decisions about how to act, I think anecdotes can be very powerful determinants. We all make our own cost/benefit analyses, however informally, before we take any action and especially when the cost is very low, our estimate of the benefit can be hugely influenced by anecdotes.

In the case of helmets, the cost is low and having had an accident yourself or hearing of the death or brain injury of someone who did not have one can convince someone to wear one based on simple plausibility arguments, even if there are no studies showing benefits, and this is perfectly fine. (I don’t know if such studies exist or not. I am simply using this as an argument.) But when it comes to governments imposing mandatory helmet rules, then they need something more than anecdotes. They need real evidence of the safety and public health benefits.

When I am driving, I always insist that everyone wear seat belts, even if they are in the back seat, although the law does not rehire it. Why? I haven’t actually looked at studies on the safety of back seat passengers but the cost of doing so is so low, and the benefits so plausible, that it seems silly not to do so.

Incidentally, Baxter the Wonder Dog also always wears a seat belt. We put a special harness on him and the seat belt passes through the loop. This gives him some limited movement on the back seat that allows him to lie down but in the event of an accident he will not be flung and injured. Also I have read (anecdotes!) about how dogs allowed to move freely can distract the driver by sudden movement. The writer Stephen King, for example, was seriously injured when walking along a road when he was hit by a truck whose driver had an unbelted dog that suddenly moved onto his lap, causing him to swerve and hit King. I have no idea if there are studies that show that seat belts for dogs reduces their injuries in collisions or reduces driver distraction. But anecdotes and plausibility about the potential benefits were enough for me to decide.

And I don’t see any problem with that.


Skip to comment form

  1. 1

    Given the doubtful (and possibly biased and agenda laden) research in that helmet post

    Always good to remember how many studies cannot be replicated, especially medical studies. And if it sounds to good to be true it is.

    19 people. 30 minutes.Not cycling, but sitting in an 81F room with no wind completing some sort of test.

    They did this sitting down test twice: same room, just with or without helmet.

    “These results resolve previous disparate studies to suggest that, although helmets can be uncomfortable, any effect of wearing a helmet on cognitive performance is at worst marginal.”

    19 people, 30 minutes, sitting down, completing tests.

    it would seem anecdotes and personal risk preferences were indeed better ways for people to decide than the (unreplicated) research.

    However, I was swayed in that discussion by mnb0 who demonstrated with a yt video how in the Netherlands a friendlier societal attitude towards bikes makes riding without helmets much safer, even when cycling in the snow!

  2. 2

    If you are riding 25mph in a paceline, you had better be wearing a helmet.

    Also road racers in mountains hit 50 and 60 mph.

    Studies of scalp temps with properly vented helmets can show cooler with helmet than not, but it probably varies with speed and wind.

    I road ride here South Texas and wearing a helmet is the least of my heat problems!

  3. 3

    I was once convinced by an anecdote, and it was helpful.

  4. 4

    “I haven’t actually looked at studies on the safety of back seat passengers but the cost of doing so is so low, and the benefits so plausible, that it seems silly not to do so. ”
    These are tricky subjects. There have been studies on this that show the opposite effect. There is even a theory on it: it’s called risk-homeostatis. In layman language this means that measures like safety belts give drivers a false feel of safety, which increases the chance of accidents, including grave ones. A spectactular example is given by Armen Alchian from the University of California. He (tongue-in-cheeked) proposed to replace airbags by sharp bars and calculated that the amount of accidents would enormously decline.
    On the subject of biker helmets one should compare statistically say The Netherlands with the USA. I haven’t found any data, but I’m absolutely not sure that biking in The Netherlands is more risky than in the USA, for the reason I already gave.
    In cases like these I never rely on anecdotes. What does it say that I never had an accident, despite riding a bike without ever wearing a helmet for 45 years now?

  5. 5

    When I am driving, I always insist that everyone wear seat belts, even if they are in the back seat, although the law does not require it.

    In California, the law does require it. When it first went into effect, there was the usual bellyaching about ‘Big Brother”, and people would try stupid tricks like looping one arm only through the shoulder harness to make it look like it was being used, but eventually everybody realized it was really no hassle to just wear the darned thing and nobody even talks about it anymore.

  6. 6

    What does it say that I never had an accident, despite riding a bike without ever wearing a helmet for 45 years now?

    Absolutely nothing.

  7. 7

    Texas, of all unlikely places, has laws requiring back-seat passengers to wear seat belts.

  8. 8

    @6 Nathan: exactly. The same applies to anecdotes of other people about accidents.

  9. 9

    Jimthompson, is right, riding a pace line, or any high speed riding, or mountain biking, you better be wearing a helmet. In fact most clubs and event organizers require it. But when I’m riding 2 miles to the ice cream shop on side streets at 12 mph, I don’t want to be stopped and given a ticket for not wearing a helmet.

  10. 10

    The validity of anecdotes is often a matter of their credibility and usability.

    If I tell you it rained yesterday where I lived, you’d probably believe me even if I were lying because it’s an everyday event. If I said the moon appeared brown, you might doubt me but if could be verified through meteorological sources. If I said it rained frogs, nobody would believe me except the most gullible rubes.

    Credibility determines usability. A story might be worth hearing if it matches something already proven, or has no critical value and won’t swing opinion or policy. But usability is usually about misuse, about people using stories (true or not) to forward an invalid argument. Even if a story is true, single or rare cases do not negate or disprove the majority of cases, though some would like you to believe otherwise (e.g. racist and bigoted politicians blathering about isolated cases of welfare fraud, poor people failing drug tests, a dozen instances of voting fraud, a criminal act by one person from a group of ten million, etc.).

    As for bike helmets, this item is anecdote and opinion, not fact, but still helpful. And of course, many of the commenters are anti-bicycle, not just anti-helmet.


    I’ve lost count of people who wear helmets improperly (like a skull cap, with their forehead bare). They might as well not be wearing one. What part of your head is most likely to hit the ground in a crash as you go over the handlebars, your forehead or the back of your head?

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite="" class=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>