Quote:
from Gilda
No. You're comparing apples to oranges. I was comparing caregiver molestation to caregiver molestation, and my reasoning is sound.
|
Again your use of statistics is flawed. Who says that the victims molested by men weren't molested by
the same man or a very small group of men. Whereas it is very unlikely that the children involved all have the same caregiver. Secondly, the statistics point to the "caregiver" scenario as being the preferred modus operandi for female perpetrators and "stranger" scenarios being preferred by males.
Quote:
|
Female perpetrators were more likely to be caretakers than male perpetrators, whereas male perpetrators were more likely to be strangers than female perpetrators.
|
The actual study was suggesting something in the way of a Modus Operandi. That is women committing this crime feel more comfortable doing it to someone close -such as a child under their care. Males committing this crime feel safest committing it to a unknown child.
Thus talking about male caregivers as being molesters is not supported by the data presented here.
However, lets take your hypothetical 100 molested children and re-analyze the statistics so it says something
that I prefer it say. (See
How to Lie with statistics -particularly the chapter on Statisticulating for more information on the games that we are playing).
From your population of 100 -22% (we won't talk of confidence levels or variation) were molested by females and 78 were molested by males.
Since it is possible that one male can molest 78 children and it is unlikely that 22 children have the same caregiver. One can conclude that it is 22 times more likely for a perpetrator to be a female than a male -this from analysis of the perpetrators perspective. Of course from the victims perspective they are still more likely to be molested by a male than a female.
Yes, again this is comical. There are some underlying assumptions when studying statistical data. Some of these statistics are taken from the perception of the victims and is limited with what you can determine about the perpetrators. What I'm trying to say is that you are repeatedly making assumptions that may or may not stick. To do a proper statistical analysis requires more use of the numbers than what we are given here. We need to know N. We need to sample randoms from populations and not just pick out the ones we are studying. If we are applying the data to a predictor model then we should perform
General Discriminant Analysis and not just ad hoc a 22/100 without a concern for the variance among different populations.
On a more speculative note: I'm willing to gamble that women who are caught committing this crime get a less severe punishment than men.