As a related observation, one apparent bias that reflects the status considerations in the modern Anglospheric world in an interesting way is that while drunk driving has come to be considered as a heinous crime, the even more dangerous practice of driving sleep-deprived attracts no significant attention. It’s not even that much of an enforcement problem—if “sleepy driving” became as much of a public attention-grabbing buzzword as “drunk driving,” I’m sure our best technical minds would be up to the task of devising effective field tests for sleep deprivation, to be administered along with breathalysers.
I upvoted this to start with, then changed my mind. The statistics I’ve seen make alcohol clearly the leading cause of fatal car accidents.
Where did you get your data? I haven’t been able to quickly find a chart comparing car accidents in general by cause (e.g. alcohol- vs. cell-phone- vs. fatigue-related).
Also, how do you define “dangerous”? There are various measures we could use (total cost of crashes based on repair costs, total fatalities, risk increase per trip, etc.) and perhaps the two causes compare differently on different measures.
Some people claim that most crashes involves sleep deprivation, but I agree with Vladimir that we have no idea.
What is suggestive is two studies of one hour differences in school start times (americans drive to school at 7:30 at age 16). One compared Virginia Beach to Chesapeake. Better was when Lexington changed its start time. The student crash rate went down 16%. The state rate went up by 8%, so these are very volatile numbers and this could be data-mining. Also, day light savings causes crashes to spike. What’s great about the school studies is that they’re about practical interventions and not just assigning blame.
Where did you get your data? I haven’t been able to quickly find a chart comparing car accidents in general by
cause (e.g. alcohol- vs. cell-phone- vs. fatigue-related).
My comment probably wasn’t clear enough—I didn’t claim that sleep-deprived driving is a more frequent cause of accidents than drunk driving. I merely meant to say that it can pose even greater danger when it happens, and that it does happen often in practice. I would surely be far more dangerous on the road after a sleepless night than with a BAC of double the legal limit, though admittedly, there are probably people who handle sleep deprivation much better and alcohol much worse.
Now, obviously, there are no systematic data on how frequently it actually causes accidents. People who get into accidents won’t just go ahead and say that they were at fault for driving tired and sleep-deprived, and nobody will check them for this, unlike for alcohol. So we have to go by extrapolating from personal experience.
Another significant data point is that professional drivers are strictly regulated in this regard, with the rationale that sleep-deprived driving is extremely dangerous. I’ve never seen anyone challenge this consensus.
Finally, what the heck does “status” have to do with this, as opposed to merely “public perception of the importance of a given issue”?
The latter follows from the former. In North America, drinking alcohol is a fairly low-status activity (except perhaps for some sorts of connoisseurship, but even that is limited only to certain social circles). Drinking enough to cause any sort of impairment is considered extremely low-status. Thus, cracking down on drinking and driving is easy to sell as a winning political move in these times when promoting safety is among the most important strategic patterns in politics. In contrast, the image of industrious folk who rise early, work hard till late, and have no time for the lazy luxury of a good night’s sleep, invokes much more positive and high-status associations, even if their drowsy commutes impose as much danger as if they were drunk.
As another status-related observation—but note that this one is speculative and meant more as a suggestion for thought than a definite conclusion—it seems to me that the global trend for changing the nature of drunk driving from a peccadillo to a serious offense, even in places where boozing is looked upon much more favorably than in the U.S., is, like many other global legal and social trends, driven by following what the Americans do. Thus, it’s a reflection of the global high status of the U.S., which has been among the principal forces in the course of world history in recent decades.
the image of industrious folk who rise early, work hard till late, and have no time for the lazy luxury of a good night’s sleep, invokes much more positive and high-status associations
Like so many other status-related “explanations” this strikes me as a just-so story with no actual predictive power and no ready base of facts to check it against. For instance, the sayings in France “la France qui se lève tôt” (the early rising part of the country) and “la France d’en bas”) (the bottom tier) are nearly synonymous in political discourse. It’s hard to see that as a high status association.
Cracking down on drunk driving is easy to rally support for because—just as you said—doctors check for blood alcohol levels every single time a crash sends someone to the hospital. Facts are readily enrolled in support of the cause, whereas they remain more obstinately neutral in the case of non-professional fatigue.
This IMO is a much more fruitful line of inquiry than “status” if you’re genuinely interested in explanations for the dynamics of “hybrid” controversies where both nature and society play significant roles. I’ve found the writings of Bruno Latour a clear and effective antidote against simplistic thinking on such issues (see his The Berlin Key for a series of short pieces that illustrate his approach). Some of his stuff is apt to give Alan Sokal a smug smile, I’ll grant, but more of it is quite incisive.
Like so many other status-related “explanations” this strikes me as a just-so story with no actual predictive power and no ready base of facts to check it against.
One prediction is that advocating a more severe crackdown on drunk driving will elicit general approval, and may well be a good tactic for a politician, whereas advocating a forcible crackdown on sleep-deprived driving will make people think you’re a weirdo, no matter how good evidence you can show for the harm caused by it. This seems consistent with reality, as far as I see. If tomorrow a study appears making an airtight case that sleepy driving is actually as harmful as drunk driving in practice, do you think the government will rush to make it a criminal offence and start a project for devising practical field tests for sleep deprivation to be issued to cops along with breathalysers?
Another prediction is that if a prominent and reputable individual publicly admits to drunk driving, it will stir up controversy and damage his reputation—whereas if he instead tells a true story that involves him driving sleep-deprived in a manner that can be shown to pose a similar level of danger, this won’t raise any significant number of eyebrows. This also appears to be true. (For example, imagine what the reactions to a post like this would be if it involved drunk driving instead of what appears to be a reckless episode of sleep-deprived driving.)
Cracking down on drunk driving is easy to rally support for because—just as you said—doctors check for blood alcohol levels every single time a crash sends someone to the hospital. Facts are readily enrolled in support of the cause, whereas they remain more obstinately neutral in the case of non-professional fatigue.
However, in practice, the issue of drunk driving is given enormous public attention and the evidence for its harmfulness, though abundant and relatively clear, is often exaggerated (just look at the often ridiculous definitions of “alcohol-related” accidents in the commonly presented statistics). The evidence for harm from sleep-deprived driving might be less abundant and systematic, but the issue is given near-zero attention in comparison, which is way out of proportion even when we account for the scantiness of the evidence. Something other than a comparison of the available evidence must be invoked to explain this discrepancy.
“Something other than a comparison of the available evidence must be invoked to explain this discrepancy.”
Yes, but what leads you to think that a status-based explanation is helpful here? The two predictions you listed are both perfectly compatible with many non-status-related hypotheses. For example, it’s quite plausible that one meme happened to get a head start.
Hypothesis: Driving-while-drunk is a much bigger deal now than it was several decades ago, and it maintains its prominence primarily because of momentum: people keep talking about it because “everyone’s talking about it”. In order for driving-while-sleep-deprived to hold the same position, it would have to climb up that same hill, and the necessary PR work and/or meme pool churn hasn’t caused that to happen yet.
That, however, is not really an explanation, but merely a statement of what has happened. Why did one issue get a head start, and why is it being talked about in a self-reinforcing way, and the other not?
It seems clear to me that there must be something about these two behaviors that makes people react to them differently, independent of any objective evaluations of their dangerousness. But what is it? I honestly can’t find a better explanation than the associations people have with each behavior, where status considerations play an important role in shaping their response.
Another significant clue is that regulations intended to curb dangers from sleep-deprivation are brought and enforced against truckers, but not against medical residents, who regularly get behind the wheel after working 30-hour shifts (and not to even mention treating patients under such sleep deprivation). Do you believe that the dramatically different treatment of these professions has nothing to do with their status?
I agree that the differential treatment between professions is stronger evidence of status having an effect.
However, if I understand what you’re saying correctly, I don’t think this statement makes sense:
I honestly can’t find a better explanation than the associations people have with each behavior, where status considerations play an important role in shaping their response.
Alone, not being able to find a better explanation for something than X isn’t good support for X being the explanation. There needs to be significant positive evidence in favor of X, or there’s no reason to choose it over “Ayedunno” (aka the null hypothesis).
Alone, not being able to find a better explanation for something than X isn’t good support for X being the explanation. There needs to be significant positive evidence in favor of X, or there’s no reason to choose it over “Ayedunno” (aka the null hypothesis).
You’re right, of course. This statement certainly wasn’t well formulated. I do see significant positive evidence, however, even if the case isn’t airtight.
A less strong and more limited claim would be that the different treatment of these issues is—to some extent that can’t be determined exactly but is definitely significant—due to the fact that alcohol impairment evokes extraordinarily negative associations in the North American culture, which biases upward the amount of attention given to problems that involve this factor, as well as the intensity of public condemnation of people causing such problems, relative to similar problems that don’t involve this particular factor. It seems to me that this claim is evident from the facts mentioned earlier.
Saying that the cause of bias is that drinking is low-status is basically a shorthand for the above long-winded statement, with the added claim that the negative view of alcohol is to a large degree status-related, and that the apparent indifference towards certain dangerous behaviors similar to drunk driving is strongly reinforced by the fact that cracking down on them would mean condemnation and meddling into the lives of lots of high-status respectable folk. I agree that this latter point is harder to justify, and my personal view of it is influenced by personal observations and experiences that are hard to translate into formal written arguments.
That, however, is not really an explanation, but merely a statement of what has happened. Why did one issue get a head start, and why is it being talked about in a self-reinforcing way, and the other not?
A head start is sufficient to answer the second question. It is also at least as much as an answer as ‘status’. In both cases more detail is required. I am not taking a position here myself but if I wanted to find reasons for a ‘head start’ I would look at behvaioral trends at the time where this issue became a problem. Were people getting drunk or depriving themselves of 30 hours of sleep more often?
The statistics I’ve seen make alcohol clearly the leading cause of fatal car accidents.
But any time alcohol is involved in a fatal car accident, it will be identified as ‘the cause’, regardless of whether the accident would have happened anyway, or would not have happened if some other factor had been fixed.
If it were the case that drunk driving was fairly common and it was only slightly correlated with higher incidence of accidents, I would still expect folks to blame accidents where the driver was drunk on the alcohol.
You’re right, I was careless in speaking of “cause” when I really mean “present in a larger proportion of fatal car accidents than any other measured risk factor”.
an estimated 2.8% of drivers exceed the legal consumption, 5.8% when it comes to drivers involved in an accident causing death or injury and 16.5% for drivers involved in fatal accidents.
This relies on a method for estimating the “base rate” which isn’t disclosed, but if correct this seems on its face to support the causal inference. By comparison, from the same study,
drowsiness or exhaustion are factors in 2.8% of fatal accidents.
This is a report by professional accidentologists, and in France at least there is serious pressure on everyone involved to do a thorough job of analysis and reporting. Every single accident involving physical injury (severe enough for police to be involved) is reported in detail by the police on the scene on a standardized form which is then filed for central aggregation.
Reduction of driving related fatalities has been a major political objective in the past few years—moreover, one that is showing actual results. There are few enough instances of such efforts bearing fruit that this exception is worth noting. So, at present I have no particular reason to believe that sleep-deprived driving is “more dangerous” than drunk driving and that the focus on the latter reflects anything other than an appropriate evaluation of the relative safety gains.
I might change my mind if I came across a specific critique of the methodology used to collect accident data and likely causal factors.
I do wonder if ‘sleep-deprived’ is even suspected / reported in someone who is also drunk. I would imagine that a lot of drunk drivers are also sleepy, and I’m not sure you’d notice the symptoms of sleepiness and not attribute that to the alcohol.
You’d have to look closer at the statistics, I suspect that the clustering of fatalities would give a fairly clear picture: for instance, an altogether too frequent type of DUI fatality is young adults driving home from a night out, dancing and drinking, in the wee hours of the morning. Obviously fatigue is a factor on top of alcohol, but it’s not clear that anything would be gained by counting those as “sleep deprivation related accidents”.
I suspect that a kind of Pareto analysis is what’s really called for in limiting the costs of traffic accident; you want to target, not generic causes (e.g. speeding or drunk driving or fatigue) but specific clusters of behaviours that account for disproportionate numbers of accidents. If you can stamp out driving-home-drunk-and-tired-from-nights-out behaviours, you’re already reducing the total fatality count by a lot. (This is precisely how traffic safety campaigns have been organized lately here.)
The statistics I’ve seen make alcohol clearly the leading cause of fatal car accidents.
I’m not at all sure how sound these statistics are. What I usually see are claims that a very high percentage of the total number of car accidents are “alcohol-related.” However, whenever I try to find out how exactly an accident is determined to be “alcohol-related,” the definitions turn out to be murky and spurious, or even outright ridiculous. For example:
The National Highway Traffic Safety Administration (NHTSA) in the U.S. defines a fatal crash as alcohol-related if either a driver or a non-motorist had a measurable or estimated BAC of 0.01 g/dl or above.
So, these people take the percentage of all fatal accidents in which not everyone happened to have flat-out zero BAC, and later use this number as if it represented the percentage of fatal accidents actually caused by drinking. This is obscene even by the usual standards of lying by statistics.
As a related observation, one apparent bias that reflects the status considerations in the modern Anglospheric world in an interesting way is that while drunk driving has come to be considered as a heinous crime, the even more dangerous practice of driving sleep-deprived attracts no significant attention. It’s not even that much of an enforcement problem—if “sleepy driving” became as much of a public attention-grabbing buzzword as “drunk driving,” I’m sure our best technical minds would be up to the task of devising effective field tests for sleep deprivation, to be administered along with breathalysers.
I upvoted this to start with, then changed my mind. The statistics I’ve seen make alcohol clearly the leading cause of fatal car accidents.
Where did you get your data? I haven’t been able to quickly find a chart comparing car accidents in general by cause (e.g. alcohol- vs. cell-phone- vs. fatigue-related).
Also, how do you define “dangerous”? There are various measures we could use (total cost of crashes based on repair costs, total fatalities, risk increase per trip, etc.) and perhaps the two causes compare differently on different measures.
Finally, what the heck does “status” have to do with this, as opposed to merely “public perception of the importance of a given issue”?
Some people claim that most crashes involves sleep deprivation, but I agree with Vladimir that we have no idea.
What is suggestive is two studies of one hour differences in school start times (americans drive to school at 7:30 at age 16). One compared Virginia Beach to Chesapeake. Better was when Lexington changed its start time. The student crash rate went down 16%. The state rate went up by 8%, so these are very volatile numbers and this could be data-mining. Also, day light savings causes crashes to spike. What’s great about the school studies is that they’re about practical interventions and not just assigning blame.
Morendil:
My comment probably wasn’t clear enough—I didn’t claim that sleep-deprived driving is a more frequent cause of accidents than drunk driving. I merely meant to say that it can pose even greater danger when it happens, and that it does happen often in practice. I would surely be far more dangerous on the road after a sleepless night than with a BAC of double the legal limit, though admittedly, there are probably people who handle sleep deprivation much better and alcohol much worse.
Now, obviously, there are no systematic data on how frequently it actually causes accidents. People who get into accidents won’t just go ahead and say that they were at fault for driving tired and sleep-deprived, and nobody will check them for this, unlike for alcohol. So we have to go by extrapolating from personal experience.
Another significant data point is that professional drivers are strictly regulated in this regard, with the rationale that sleep-deprived driving is extremely dangerous. I’ve never seen anyone challenge this consensus.
The latter follows from the former. In North America, drinking alcohol is a fairly low-status activity (except perhaps for some sorts of connoisseurship, but even that is limited only to certain social circles). Drinking enough to cause any sort of impairment is considered extremely low-status. Thus, cracking down on drinking and driving is easy to sell as a winning political move in these times when promoting safety is among the most important strategic patterns in politics. In contrast, the image of industrious folk who rise early, work hard till late, and have no time for the lazy luxury of a good night’s sleep, invokes much more positive and high-status associations, even if their drowsy commutes impose as much danger as if they were drunk.
As another status-related observation—but note that this one is speculative and meant more as a suggestion for thought than a definite conclusion—it seems to me that the global trend for changing the nature of drunk driving from a peccadillo to a serious offense, even in places where boozing is looked upon much more favorably than in the U.S., is, like many other global legal and social trends, driven by following what the Americans do. Thus, it’s a reflection of the global high status of the U.S., which has been among the principal forces in the course of world history in recent decades.
Like so many other status-related “explanations” this strikes me as a just-so story with no actual predictive power and no ready base of facts to check it against. For instance, the sayings in France “la France qui se lève tôt” (the early rising part of the country) and “la France d’en bas”) (the bottom tier) are nearly synonymous in political discourse. It’s hard to see that as a high status association.
Cracking down on drunk driving is easy to rally support for because—just as you said—doctors check for blood alcohol levels every single time a crash sends someone to the hospital. Facts are readily enrolled in support of the cause, whereas they remain more obstinately neutral in the case of non-professional fatigue.
This IMO is a much more fruitful line of inquiry than “status” if you’re genuinely interested in explanations for the dynamics of “hybrid” controversies where both nature and society play significant roles. I’ve found the writings of Bruno Latour a clear and effective antidote against simplistic thinking on such issues (see his The Berlin Key for a series of short pieces that illustrate his approach). Some of his stuff is apt to give Alan Sokal a smug smile, I’ll grant, but more of it is quite incisive.
Morendil:
One prediction is that advocating a more severe crackdown on drunk driving will elicit general approval, and may well be a good tactic for a politician, whereas advocating a forcible crackdown on sleep-deprived driving will make people think you’re a weirdo, no matter how good evidence you can show for the harm caused by it. This seems consistent with reality, as far as I see. If tomorrow a study appears making an airtight case that sleepy driving is actually as harmful as drunk driving in practice, do you think the government will rush to make it a criminal offence and start a project for devising practical field tests for sleep deprivation to be issued to cops along with breathalysers?
Another prediction is that if a prominent and reputable individual publicly admits to drunk driving, it will stir up controversy and damage his reputation—whereas if he instead tells a true story that involves him driving sleep-deprived in a manner that can be shown to pose a similar level of danger, this won’t raise any significant number of eyebrows. This also appears to be true. (For example, imagine what the reactions to a post like this would be if it involved drunk driving instead of what appears to be a reckless episode of sleep-deprived driving.)
However, in practice, the issue of drunk driving is given enormous public attention and the evidence for its harmfulness, though abundant and relatively clear, is often exaggerated (just look at the often ridiculous definitions of “alcohol-related” accidents in the commonly presented statistics). The evidence for harm from sleep-deprived driving might be less abundant and systematic, but the issue is given near-zero attention in comparison, which is way out of proportion even when we account for the scantiness of the evidence. Something other than a comparison of the available evidence must be invoked to explain this discrepancy.
“Something other than a comparison of the available evidence must be invoked to explain this discrepancy.”
Yes, but what leads you to think that a status-based explanation is helpful here? The two predictions you listed are both perfectly compatible with many non-status-related hypotheses. For example, it’s quite plausible that one meme happened to get a head start.
Hypothesis: Driving-while-drunk is a much bigger deal now than it was several decades ago, and it maintains its prominence primarily because of momentum: people keep talking about it because “everyone’s talking about it”. In order for driving-while-sleep-deprived to hold the same position, it would have to climb up that same hill, and the necessary PR work and/or meme pool churn hasn’t caused that to happen yet.
That, however, is not really an explanation, but merely a statement of what has happened. Why did one issue get a head start, and why is it being talked about in a self-reinforcing way, and the other not?
It seems clear to me that there must be something about these two behaviors that makes people react to them differently, independent of any objective evaluations of their dangerousness. But what is it? I honestly can’t find a better explanation than the associations people have with each behavior, where status considerations play an important role in shaping their response.
Another significant clue is that regulations intended to curb dangers from sleep-deprivation are brought and enforced against truckers, but not against medical residents, who regularly get behind the wheel after working 30-hour shifts (and not to even mention treating patients under such sleep deprivation). Do you believe that the dramatically different treatment of these professions has nothing to do with their status?
I agree that the differential treatment between professions is stronger evidence of status having an effect.
However, if I understand what you’re saying correctly, I don’t think this statement makes sense:
Alone, not being able to find a better explanation for something than X isn’t good support for X being the explanation. There needs to be significant positive evidence in favor of X, or there’s no reason to choose it over “Ayedunno” (aka the null hypothesis).
You’re right, of course. This statement certainly wasn’t well formulated. I do see significant positive evidence, however, even if the case isn’t airtight.
A less strong and more limited claim would be that the different treatment of these issues is—to some extent that can’t be determined exactly but is definitely significant—due to the fact that alcohol impairment evokes extraordinarily negative associations in the North American culture, which biases upward the amount of attention given to problems that involve this factor, as well as the intensity of public condemnation of people causing such problems, relative to similar problems that don’t involve this particular factor. It seems to me that this claim is evident from the facts mentioned earlier.
Saying that the cause of bias is that drinking is low-status is basically a shorthand for the above long-winded statement, with the added claim that the negative view of alcohol is to a large degree status-related, and that the apparent indifference towards certain dangerous behaviors similar to drunk driving is strongly reinforced by the fact that cracking down on them would mean condemnation and meddling into the lives of lots of high-status respectable folk. I agree that this latter point is harder to justify, and my personal view of it is influenced by personal observations and experiences that are hard to translate into formal written arguments.
A head start is sufficient to answer the second question. It is also at least as much as an answer as ‘status’. In both cases more detail is required. I am not taking a position here myself but if I wanted to find reasons for a ‘head start’ I would look at behvaioral trends at the time where this issue became a problem. Were people getting drunk or depriving themselves of 30 hours of sleep more often?
Literally, “The France that lifts itself early” and “The France of the base”?
Not quite. “Se lever” in this context directly means rising, as in rising from bed, “en bas” is a direction, literally “below”.
But any time alcohol is involved in a fatal car accident, it will be identified as ‘the cause’, regardless of whether the accident would have happened anyway, or would not have happened if some other factor had been fixed.
If it were the case that drunk driving was fairly common and it was only slightly correlated with higher incidence of accidents, I would still expect folks to blame accidents where the driver was drunk on the alcohol.
You’re right, I was careless in speaking of “cause” when I really mean “present in a larger proportion of fatal car accidents than any other measured risk factor”.
In my neck of the woods,
This relies on a method for estimating the “base rate” which isn’t disclosed, but if correct this seems on its face to support the causal inference. By comparison, from the same study,
This is a report by professional accidentologists, and in France at least there is serious pressure on everyone involved to do a thorough job of analysis and reporting. Every single accident involving physical injury (severe enough for police to be involved) is reported in detail by the police on the scene on a standardized form which is then filed for central aggregation.
Reduction of driving related fatalities has been a major political objective in the past few years—moreover, one that is showing actual results. There are few enough instances of such efforts bearing fruit that this exception is worth noting. So, at present I have no particular reason to believe that sleep-deprived driving is “more dangerous” than drunk driving and that the focus on the latter reflects anything other than an appropriate evaluation of the relative safety gains.
I might change my mind if I came across a specific critique of the methodology used to collect accident data and likely causal factors.
I do wonder if ‘sleep-deprived’ is even suspected / reported in someone who is also drunk. I would imagine that a lot of drunk drivers are also sleepy, and I’m not sure you’d notice the symptoms of sleepiness and not attribute that to the alcohol.
You’d have to look closer at the statistics, I suspect that the clustering of fatalities would give a fairly clear picture: for instance, an altogether too frequent type of DUI fatality is young adults driving home from a night out, dancing and drinking, in the wee hours of the morning. Obviously fatigue is a factor on top of alcohol, but it’s not clear that anything would be gained by counting those as “sleep deprivation related accidents”.
I suspect that a kind of Pareto analysis is what’s really called for in limiting the costs of traffic accident; you want to target, not generic causes (e.g. speeding or drunk driving or fatigue) but specific clusters of behaviours that account for disproportionate numbers of accidents. If you can stamp out driving-home-drunk-and-tired-from-nights-out behaviours, you’re already reducing the total fatality count by a lot. (This is precisely how traffic safety campaigns have been organized lately here.)
Oh, and I forgot to address this part:
I’m not at all sure how sound these statistics are. What I usually see are claims that a very high percentage of the total number of car accidents are “alcohol-related.” However, whenever I try to find out how exactly an accident is determined to be “alcohol-related,” the definitions turn out to be murky and spurious, or even outright ridiculous. For example:
So, these people take the percentage of all fatal accidents in which not everyone happened to have flat-out zero BAC, and later use this number as if it represented the percentage of fatal accidents actually caused by drinking. This is obscene even by the usual standards of lying by statistics.