4 ms·

If I may be pedantic: > Compared to men, women are 17 percent more likely to die in a car crash and 73 percent more likely to sustain serious injuries in a fro

If I may be pedantic: > Compared to men, women are 17 percent more likely to die in a car crash and 73 percent more likely to sustain serious injuries in a front-end collision. Yet despite these discrepancies, car companies are only required to test vehicle safety using crash dummies modeled after the average man. Guinea pigs are 2000% more likely to die in a car crash and 10000% more likely to sustain serious injuries in a front-end collision. We don't make crash dummies for guinea pigs. Since to objective is to reduce overall injury, the primary justification for a female crash dummy would be based not on the probabity of injury if you are a woman, but the probability of being a woman if you have received injury. This is of course also very high -- and it's great to see female crash dummies, long overdue. I'm just irritated at such sloppy writing.

LordDragonfang 4y ago
Guinea pigs don't drive cars, women do.
- SeanLuke 4y ago
  Did the paragraph say driver? Even so, when a guinea pig is driving, it is 100000% more likely to die in a car crash! The point is, I think, that the reporter needs a lesson in Bayes Law.
  - habinero 4y ago
    Thanks for valuing women as highly as guinea pigs.
  - LordDragonfang 4y ago
    >A crash dummy aimed at protecting women drivers The title of the article that we are currently discussing. Please don't pretend your comment doesn't exist in the context of the discussion it's a reply to, that's somewhere between dishonest motte-and-bailey and blatantly bad comment etiquette. (And beyond that, you're not even using a proper application of Bayes Law anyway, since "probability of injury if you are a woman" is what you should be concerned with if you're building safety devices)
scarmig 4y ago
It's a bit more complicated than that, not that the article you're responding to is particularly nuanced itself. The overall goal is to improve material and physical well-being for society in aggregate (itself a bit ambiguous, but bear with me). Suppose cars being optimized for the male case would minimize overall injuries. That will have the side effect, however, of dissuading women from driving at the margin, compared to a world where cars are optimized for some intermediate case. That's deadweight loss, and deadweight loss has a real economic and social cost. In the extreme, you could imagine a world where there's a net increase of one death in exchange for generating billions of dollars. In that non-realized world, that'd be an obvious trade off worth making. Of course, with that different safety optimization target we'd then have loss when the marginal male driver chooses not to drive. Do they cancel out? At least if we're starting from a male-optimized point, small changes in safety policies are likely to lead to minimal loss of male drivers (by definition of optima; linear terms are zero in the vicinity of a local optimum). At that male-focused optimum, however, female drivers can potentially drive large social gains by small changes in safety policy in their favor (since the safety policy is not already optimized for them). Which is all to say, what's socially ideal is likely somewhere between male-optimized and female-optimized, though not necessarily near the exact midpoint. And that's all assuming that car driving currently is currently male-optimized, which I'd question.

LordDragonfang 4y ago

Guinea pigs don't drive cars, women do.

SeanLuke 4y ago
Did the paragraph say driver? Even so, when a guinea pig is driving, it is 100000% more likely to die in a car crash! The point is, I think, that the reporter needs a lesson in Bayes Law.
- habinero 4y ago
  Thanks for valuing women as highly as guinea pigs.
- LordDragonfang 4y ago
  >A crash dummy aimed at protecting women drivers The title of the article that we are currently discussing. Please don't pretend your comment doesn't exist in the context of the discussion it's a reply to, that's somewhere between dishonest motte-and-bailey and blatantly bad comment etiquette. (And beyond that, you're not even using a proper application of Bayes Law anyway, since "probability of injury if you are a woman" is what you should be concerned with if you're building safety devices)

scarmig 4y ago

It's a bit more complicated than that, not that the article you're responding to is particularly nuanced itself. The overall goal is to improve material and physical well-being for society in aggregate (itself a bit ambiguous, but bear with me). Suppose cars being optimized for the male case would minimize overall injuries. That will have the side effect, however, of dissuading women from driving at the margin, compared to a world where cars are optimized for some intermediate case. That's deadweight loss, and deadweight loss has a real economic and social cost. In the extreme, you could imagine a world where there's a net increase of one death in exchange for generating billions of dollars. In that non-realized world, that'd be an obvious trade off worth making. Of course, with that different safety optimization target we'd then have loss when the marginal male driver chooses not to drive. Do they cancel out? At least if we're starting from a male-optimized point, small changes in safety policies are likely to lead to minimal loss of male drivers (by definition of optima; linear terms are zero in the vicinity of a local optimum). At that male-focused optimum, however, female drivers can potentially drive large social gains by small changes in safety policy in their favor (since the safety policy is not already optimized for them). Which is all to say, what's socially ideal is likely somewhere between male-optimized and female-optimized, though not necessarily near the exact midpoint. And that's all assuming that car driving currently is currently male-optimized, which I'd question.