送交者: mangolasi 于 2006-1-28, 16:01:26:
回答: Correction: gravity won't make an AR process as it is fixed (hence 由 mangolasi 于 2006-1-28, 14:12:42:
1. I am entirely ignorant in the field of biology. So I don't know what are the probabilities the current researcher agree on (not entirely exclusive) (1) (2) (3) (4) and that radiation of cell phone is bad to the brain. However, it's difficult to believe there are no priority.
Even though those 5 cases are with the same probability, probability is not the only thing concerned (not even expected pay-off). One thing you need to consider is the reversibility (and if it is possible at all, the cost of it). Only (2) and (3) can not be eliminated by this criteria, (4) is definitely eliminated because we don't have big problem of forsaking some faint hope of tiny good in order to avoid (might be equally faint) risk of huge evil. (1) is more or less eliminatable given the current medical techology. Malign brain tumor is viewed as a bigger challenge than malign kidney tumor as far as I can tell. The same reasoning goes on when people in the neighborhood forsake the slight cost of bad cell phone signal (it's their home anyway, and people can always ask the caller call the landline at home if the signal quality is bad), for the huge risk with even a slight probability.
OK, as I don't know the detail, I am not really able to tell what one should do in this case. However, I don't believe the reason is so "linear" as that guy propose.
Besides cost of reversibility, we can also consider "externality": someone (e.g. the cell phone company, or a polluted factory) gains at the risk or expense of others (people living nearby). When something really bad happens, it's difficult to point a finger exactly to the company (especially on radiation pollution). Because the effect in the research, even confirmed, will be statistical rather than direct causal (another example is the smoking-lung cancer relation). It's difficult to "prove" that the company need to compensate. If the cost of legally identifying the culprit is huge, and there is a risk that this is a culprit, a rational response would be not letting it happen at the first place.
2. The author might know about things (even for that I doubt, see my comments on the coin), but doesn't know about people. It's natural to be desperate when you are seriously sick. It's this desperation makes people gullible. They are unlucky victims, rather than fools. (even they are fools, I hate the author's patronizing attitude).
"现在的专治糖尿病、肝炎、白血病的骗子，正是利用了人的这种错误潜意识，
我们设这些骗子的治愈率的均值为P，病人连找N个骗子后治愈率为1-(1-P)的N次
方，关键在于P值，如果能达到50％，则连看4个后其治愈率就会高达93.75%。但
从实际效果看，这个P值近似为0，N再大也没有用。"
His example of simple probability is laughable. In order for it to work, you must evaluate all docotrs you see as indentical. But when you are seriously sick, you simply think "the next one is different". This is not a sampling porblem, thinking the porpotion of the cured of EACH doctor. It is a searching problem, which you keep trying those with 0% cured rate until you hit the one, the one with almost 100% cure rate.
To start with, when you see 1000 patients going to 100 doctors once, and you observe only 10 patient get cured. 1% cure rate? No. For 99, the cure rate might be 0, and the remaining one is 100%. For 1%, you will not go to those 100 doctors, but if the doctors are different, a natural response (provided a fat wallet), is to search until you hit the one with 100%. And a large N is really useful in this case (provided bad treatment won't have disastrously negative effect): law of large number, so simple.
OK, we don't know by simplying looking at the aggregated data. But for low cure rate, there are 2 equally plausible explanation. Saying one with absolute term is really arrogant. And for the patients, they might be more inclined for the search model than the sampling model: there is no "rationality argument" in it as both are equally plausible (until you have extended data to prove otherwise).
What is really chilling in this article is its arrogance. The author doesn't respect the diversity of people, the diversity of people's perference on risk and other priorities. It's because this diversity of priorities we can enjoy what is called "freedom", which Amartya Sen said as the base of "rationality" (being an maximization machine according to some machnical rules is not rationality, it is machinary)--hopefully I can finish that book and write a review on that.