<< Do I infer correctly that (if A = the 100-digit number, and W = the 98-digit number) the number of "right" i.e. representative samples would be, according to Gilgames, their difference R = A-W ? Then the proportion of "right" to "wrong" samples is R/W = a number of order 100: that is, there are about 100 times as many "right" samples as "wrong" samples. Donald Burill >>
I only said, that if the number of misleading samples is enormous and there is high monetary interest to influence the political/economical processes through statistics, then one should not suppose, that the samples are random. Hundred years ago this principle was named as original sin, twenty years ago the name was Murphy's law, today I hear about it less and less. laszlo kiss . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
