“If it saves the life of a single child…” and other nonsense

This post is by Phil Price.

An Oregon legislator, Mitch Greenlick, has proposed to make it illegal in Oregon to carry a child under six years old on one’s bike (including in a child seat) or in a bike trailer. The guy says “”We’ve just done a study showing that 30 percent of riders biking to work at least three days a week have some sort of crash that leads to an injury… When that’s going on out there, what happens when you have a four year old on the back of a bike?” The study is from Oregon Health Sciences University, at which the legislator is a professor.

Greenlick also says “”If it’s true that it’s unsafe, we have an obligation to protect people. If I thought a law would save one child’s life, I would step in and do it. Wouldn’t you?”

There are two statistical issues here. The first is in the category of “lies, damn lies, and statistics,” and involves the statement about how many riders have injuries. As quoted on a blog, the author of the study in question says that, when it comes to what is characterized as an injury, “It could just be skinning your knee or spraining your ankle, but it couldn’t just be a near miss.” By this standard, lots of other things one might do with one’s child — such as playing with her, for instance — might be even more likely to cause injury.

Substantial numbers of people have been taking their children on bikes for quite a while now, so although it may be impossible to get accurate numbers for the number of hours or miles ridden, there should be enough data on fatalities and severe injuries to get a semi-quantitative idea of how dangerous it is to take a child on a bike or in a bike trailer. And when I say “dangerous” I mean, you know, actually dangerous.

The second problem with Greenlick’s approach is that it seems predicated on the idea that, in his words, “If I thought a law would save one child’s life, I would step in and do it. Wouldn’t you?” Well, no, and in fact that is just a ridiculous principle to apply. Any reasonable person should be in favor of saving children’s lives, but not at all cost. We could make it illegal to allow children to climb trees, to eat peanuts, to cross the street without holding an adult’s hand…perhaps they shouldn’t be allowed to ride in cars. Where would it end?

Finally, a non-statistical note: another state rep has commented regarding this bill, saying that “this is the way the process often works: a legislator gets an idea, drafts a bill, introduces it, gets feedback, and then decides whether to try to proceed, perhaps with amendments, or whether to let it die.” If true, this is a really wasteful and inefficient system. Better would be “a legislator gets an idea, does a little research to see if it makes sense, introduces it,…” Introducing it before seeing if it makes sense is probably a lot easier in the short run, but it means a lot of administrative hassle in introducing the bills, and it makes people waste time and effort trying to kill or modify ill-conceived bills.

17 thoughts on ““If it saves the life of a single child…” and other nonsense

  1. In the 1960s, there was a single-panel comic called, "Halto's There Ought To Be a Law." Bills like this may be why Texas only allows its legislature to sit in regular session for 140 days, every other year. So many legislators, so little time!

  2. Phil:

    Your last paragraph reminds me of the strategy of submitting the first version of a scientific paper to a mediocre journal, from which you get lots of free comments and advice, after which you can submit the revision to a better journal.

    I've never chosen this strategy on purpose, but sometimes I've done it by accident!

    P.S. I think you can correctly impute my opinion of the proposed Oregon law.

  3. A law confiscating all of Mitch Greenlick's wealth, and all future income beyond that needed for a sustenance level, will provide the resources to save the lives of hundreds of children in Africa.

  4. "…it means a lot of administrative hassle in introducing the bills…" has to be balanced with "I want to be seen as the person who thinks of the children." The proposed bill is bound to
    die rapidily
    get free airtime in the news while
    not getting any poor review since most of his bikers constituents are likely not going to work with their 4 yo kids in the back of their bikes.
    What is the downside of doing this ?

  5. I second Bob O'H, there is a definite generalizability issue here. We're trying to generalize from people riding their bikes to work (presumably often in rush hour traffic, late themselves, etc.) to a parent riding with their child on their back (I bet this usually happens in less traffic, with less time pressure, and more awareness on the parents part). Regardless of what you characterize as "injury" we're looking at the wrong sample.

  6. Andrew,

    My name is Chris Harvey and I am a research assistant in a cancer lab at Oregon Health Sciences University in Portland Oregon. I am a budding statistician (applying to biostats MS programs presently; fingers crossed)and follow your blog regularly. After reading this post, I sent an email to Mitch Greenlick asking his thoughts, and the response was, well, interesting.

    I am tempted to share our email dialogue, but am hesitant to because I am unsure if it is ethical. Maybe its unethical if I do not share his response! Let me know what you think.

    Regards,
    Chris

  7. As part of this law, adults might be prohibited from driving with a child under 6 unless absolutely necessary for the health of that child — no trips to the grocery, no trips to the park, no riding in the car to pick up older sister from school, etc. Any trips taken would need the prior approval from Social Services. This would likely have a much bigger impact on the number of young children dying in vehicle accidents.

  8. Chris,
    If neither you nor Greenlick stated or strongly implied that your exchange was confidential, either of you can share it. If you and he were friends, I could imagine some ambiguity about whether privacy was implied, but that's not the case here. You can freely share it.
    That said, it would be appropriate and just generally nice of you to send him another email, perhaps something like "Thank you for your response, I found it interesting and I know that others will as well so I am posting it to [this blog]; if you want to add additional comments of your own I know they'd be welcome." But there's certainly no ethical requirement for you to do such a thing.

    –Phil Price

  9. Phil,

    You answered my initial question exactly; thank you. Correspondence between myself and Mitch Greenlick follows,

    Chris to Mitch:

    The article may seem a bit harsh, but the author brings up some interesting points which I should have stated more clearly.

    1: In the present study "Bicycle Commuter Injury Prevention: It is Time to Focus on the Environment," the distinction of what is classified as an injury might be considered too broad to be credible. What are your thoughts on this claim?

    2: I am going to infer that Price believes the best way to "semi-quantitate" the danger of a child on a bike or in a trailer is to use real data, which "should" be available. If that data is not available or does not exist, to me, it seems hasty to purpose a law that may or may not address an actual problem. Without the data how can we know that the problem "children in bike seats/trailers are getting injured" is real?

    3: "The second problem with Greenlick's approach is that it seems predicated on the idea that, in his words, "If I thought a law would save one child's life, I would step in and do it. Wouldn't you?""
    Do you think his slippery slope argument against your claim has any merit? How do you respond to the argument that the purposed law is unnecessarily meddlesome?

    4: In response to, "We've just done a study showing that 30 percent of riders biking to work at least three days a week have some sort of crash that leads to an injury… When that's going on out there, what happens when you have a four year old on the back of a bike?"
    We have yet to confirm, with any amount of certainty, that commuters crashing with a young child on board is a significant problem which warrants legislation. Might it best to find the data or conduct a study before we claim, either way, that it is or is not a significant problem?

    If you feel your words have been misinterpreted, please let me know where I have gone wrong.

    Chris

    Mitch to Chris:

    YES, my words are and have been totally misinterpreted. I never said the OHSU study had data indicating it was about kids on bikes at all. Never. What I said was listening to the report of that study got me thinking if riders that skilled had a relatively high level of accidents what happens to less skilled riders having an accident with a kid on the seat behind the driver or in a trailer following the bike.''

    And it turns out the answer is there is no data indicating it is safe or not. I agree with Price's statement that there should be data, but there is NONE.

    In my life as a researcher or as chair of preventive medicine at the medical school I address a problem like that by applying for a grant. As a state legislator I address a problem like that by drafting a bill. In the public health precautionary mode the bill says if we don't know it is safe stop it until we know. That begins the discussion, and when not dealing with bikers or anti-fluoride nuts a rational discussion begins. Not so in this case, at least not yet.

    I am under pressure, including from many bikers, to eliminate baby bottles made with bisphenol-A, because it might not be safe for some kids — the precautionary principle. Not put fluoride in the water system because it might not be safe for some kids — the precautionary principle. Not put three-year olds on the back seat of a bike because it might not be safe — are you kidding me? The bikers response — we know it is safe because we do it and we certainly would not do it if it was not safe.

    Is that how you work as a scientist, or how Price works in warning of the dangers of Radan? Of course not.

    HELP!!!

    Mitch

  10. Chris, thanks for getting Greenlick's response and for passing it along. It is interesting.

    Greenlick says:
    1. "YES, my words are and have been totally misinterpreted. I never said the OHSU study had data indicating it was about kids on bikes at all. Never."

    I think casual readers can be forgiven for assuming that, since Greenlaw mentioned the study when justifying his proposal to ban carrying kids on bikes, the study has something to say about carrying kids on bikes.

    For myself, I had noticed that the study said nothing about kids on bikes, but didn't see a need to make that point. However, several commenters noted the point and discussed it. Other people may be misinterpreting Greenlick's words, but I haven't seen it here.

    2. Greenlick says "I agree with Price's statement that there should be data, but there is NONE."

    Considering how many data sources there are, especially for things like serious injuries and fatalities, that's a remarkable claim, almost unbelievable. And in fact, it's not true. It's true that it was hard to find such data; by "hard" I mean that I didn't find it on the first page in any of my first three Google searches, although I did on the fourth one. There are, in fact, relevant data available — of course — and some research papers and other analysis. For instance, the National Highway Transportation Safety Association has data on "pedalcyclist" (as opposed to motorcyclist) deaths among children nationwide, by year, and has a publication that presents some of the relevant data in a plot. The older children, age 8-14 years, are presumably mostly riding their own bikes, but the data on under-1-year-olds and 1-3-year-olds presumably represent kids in trailers on in bike seats, with the 4-7-year-olds a mixture of kids in seats and kids riding their own bikes. (The plot is Figure 3. Unfortunately for the purpose of reading data off the plot, the number of deaths among the two lower age groups is too low to read off precisely. The number appears to be 0 in the under-1-year age group in all years except for 1 or 2 fatalities in 2006, and of the order of 1 to 4 deaths per year in the 1-3-year-old age group, in most years, although it appears to be 0 in 2006, 2007, and 2009).

    There's also an old (1991) paper that specifically looks at "injuries associated with bicycle-mounted child seats", using data from the US Consumer Product Safety Commission. Although one might expect that some aspects of this paper aren't relevant anymore, such as the numbers of injuries per year, since child seat use has grown and seat design may have changed, one might expect other aspects to still apply. For instance, a large fraction of injuries occurred, and presumably still occur, when the child is in the bike seat and the bike is stationary, without the adult straddling it, and falls over. I would expect that type of accident to still be fairly common.

    Anyway, data are available.

    3. Greenlick says "As a state legislator I address a problem like [the lack of data] by drafting a bill. In the public health precautionary mode the bill says if we don't know it is safe stop it until we know."

    As I mentioned at the end of the blog post, drafting a bill before knowing anything about your subject is a really inefficient way of proceeding. If this is really the way state legislators in Oregon (and perhaps elsewhere) deal with the lack of data, that's sad.

    But in fact, there do seem to be enough available data to provide at least some basis for a decision on whether banning child bike seats and trailers is justified.

    4. "Not put three-year olds on the back seat of a bike because it might not be safe — are you kidding me? The bikers response — we know it is safe because we do it and we certainly would not do it if it was not safe."

    It's funny-amusing and funny-peculiar that Greenlick would claim that people are misinterpreting him (by assuming that the study he cites actually has data on injuries to children), but then would misrepresent or misinterpret the objections of "the bikers." I suppose it may be true that some cyclists claim there is NO danger in carrying a child on a bicycle, but I am not one of them, and I think almost all cyclists would agree that there is SOME danger. But — a major point of my post, and the reason for its sarcastic title — the idea that anything that is at all hazardous should be banned is ridiculous. The relevant question is not "is there any danger," it is "how large is the danger."

    Few things in life are perfectly safe. Assuming a law banning kids in bike seats and bike carriers would save about 3 lives per year nationwide — a rough estimate based on the aforementioned NHTSA data — is this something that, as a society, we think should be banned?

    5. Finally, I'd like to make a point that hasn't been mentioned before. Being a state legislator isn't easy or fun. It doesn't pay well, there's a lot of work involved, there's a lot of responsibility, and no matter what you do you will make a lot of people unhappy. I almost feel bad about piling on in this way. Almost.

Comments are closed.