The Ghost of Paley
Posts: 1703 Joined: Oct. 2005
|
Quote | The table on page 93 is about receptions into prison on sentence, the table on 136 is about prison pop. This is why I presented the relevant data from BOTH tables, otherwise it's not comparing like with like. Geez Gimpy, I did expect that you could read just a little.
|
Yes, I didn't read the heading carefully, and I apologise. Notice that Louis claims to be looking at the evidence objectively, yet assumes any mistake I make must be due to dishonesty and bigotry. But wait, how could he assume motives unless he thinks his belief is beyond criticism?
Anyway, I didn't use the right table, so my percentages were a little off. But I'll trust Louis's division for now....let's compare his numbers:
Quote | [3.74% of all prisoners were 15 -17 year olds 11.696% of all prisoners were 18-20 year olds - my edit] 21.8% of all prisoners were 21-24 year olds. 22.5% of all prisoners were 25-29 year olds. 33.7% of all prisoners were 30-39 year olds. 13.8% of all prisoners were 40-49 year olds. 5.6% of all prisoners were 50-59 year olds. 2.6% of all prisoners were 60+ year olds. [....] [2.84% of muslim prisoners were 15 -17 year olds 11.099% of muslim prisoners were 18-20 year olds - my edit] 24.3% of muslim prisoners were 21-24 year olds. 25.9% of muslim prisoners were 25-29 year olds. 33.2% of muslim prisoners were 30-39 year olds. 12.7% of muslim prisoners were 40-49 year olds. 2.6% of muslim prisoners were 50-59 year olds. 1.3% of muslim prisoners were 60+ year olds. |
So this makes 59.736% for the under 30 general pop, and 93.436% for the under 40s.
For Muslims, this makes 64.139% of the under 30, and 97.339% for those under 40. Around a 5% difference. Now this shows that the Muslim prison population is skewed young, but here's a couple of problems that Louis has yet to address:
1) As I've said two times already, comparing age profiles among prisoners can be deceptive, because it might reflect cultural regression rather than the true demographic profile. To his advantage, Louis seems to realise this, since he reposts his links to the general UK population. More about this below.
2) The gap between the expected and actual Muslim prison population is not small -- it's 250% times too high. Pointing to a mitigating factor (age structure) without attempting to quantify by how much this factor reduces this discrepancy won't demonstrate much . So far, he has not made the slightest attempt to quantify this factor. Why not? Louis keeps reminding us how stupid I am, and I know how to carry these tests out, so it should be a trifle for him. I can think of several possibilities:
1) The data won't allow it (in which case, Louis can't support his claim);
2) Louis doesn't know how to do the proper adjustments (nothing wrong with that, but you'd think he could find someone who could. Why hasn't he?);
3) Louis doesn't have the time (then why not find someone who can do it, or use his time more constructively -- I can't believe he would rather insult me than demonstrate his claim).
I'll let Louis confide which option is correct, because make no bones about it, without a proper analysis, the relatively small discrepancy he found, even if it represents the true demographic profile, does not begin to overturn that whopping 250% discrepancy that I found. The discrepancy was about 5% -- look at the totals above.
So let's look at page 6 on this link. What do we find? That most of the skewedness in the Pakistani/Bangladesh category is in the "under 16" / "65 and older" categories, which is not the high-crime age group! .[Edit: Louis's spreadsheet makes this point crystal clear. The distribution is skewed in the non-crime-prone direction.] Given that 27% of Muslims are of the non-subcontinental, non-white variety, there's a good chance that there isn't much of a difference in that all important 15 - 39 age group (if there is, it might even favor non-Muslims), which if true absolutely crushes his claim. Even if there is a difference in that all important age group, it appears to be minor.
Sorry Louis, this is why I won't be content with a link dump and hand-waving: everytime I investigate your little stats, they collapse under their own triviality. You've had plenty of time to build your counterargument. Either admit that the stats can't survive a serious analysis, or take some more time to collate them. Your barking doesn't work with me, cause I can see your arguments have no bite.
Quote | P.P.S. And Gimpy, THIS is the hasty generalisation fallacy, not your twisted version.
|
Next time try looking up the full definition:
Quote | This fallacy is committed when a person draws a conclusion about a population based on a sample that is not large enough. It has the following form:
Sample S, which is too small, is taken from population P. Conclusion C is drawn about Population P based on S. The person committing the fallacy is misusing the following type of reasoning, which is known variously as Inductive Generalization, Generalization, and Statistical Generalization:
X% of all observed A's are B''s. Therefore X% of all A's are Bs. The fallacy is committed when not enough A's are observed to warrant the conclusion. If enough A's are observed then the reasoning is not fallacious.
Small samples will tend to be unrepresentative. As a blatant case, asking one person what she thinks about gun control would clearly not provide an adequate sized sample for determing what Canadians in general think about the issue. The general idea is that small samples are less likely to contain numbers proportional to the whole population. For example, if a bucket contains blue, red, green and orange marbles, then a sample of three marbles cannot possible be representative of the whole population of marbles. As the sample size of marbles increases the more likely it becomes that marbles of each color will be selected in proprtion to their numbers in the whole population. The same holds true for things others than marbles, such as people and their political views.
|
Your little Buddhist group of prisoners (.9% of the total male prison population after the increase) is too small to be of any use, because small changes in the actual numbers will represent a huge percentage increase. Tiny fluctuations like this cannot be used to infer much about Buddhist crime tendencies over time.
-------------- Dey can't 'andle my riddim.
|