Comic Genesis Forums

Actually, what my statistics class taught me is that while anything may be statistically calculable as you say, it will never give you a useful answer.<P>...sorry, it's just that I developed a strong hatred of statistics and its worthless imaginary numbers in eleventh grade.<P>------------------
I'm so sane it's driving me crazy.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by My name is Kenny:
<B>Actually, what my statistics class taught me is that while anything may be statistically calculable as you say, it will never give you a useful answer.
</B><HR></BLOCKQUOTE><P>Oh, that's absolutely true. I never said otherwise. But I feel that Guth would have wanted precision in his terminology.<P>Nicolas<P>P.S.: In light of the reply by Christopher, I thought I'd point out that I was agreeing that statistics would never provide a useful answer IN THIS SPECIFIC CASE, which is what I thought Kenny meant.<p>[This message has been edited by Nicolas Juzda (edited 05-25-2001).]

Hi,<P>The so-called "urgency factor" that Guth alludes to is, in fact, statistically calculable. Any behaviour is. You just need sufficient data to base a model upon.<P>In the case of chain letter forwarding, the most obvious variable to base the model upon would be the topic of the letter (pyramid scheme, lame jokes, petition, whatever). More complex models could include the presence of key words, or whatever else you want to include.<P>Nicolas<p>[This message has been edited by Nicolas Juzda (edited 05-24-2001).]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by My name is Kenny:
<B>Actually, what my statistics class taught me is that while anything may be statistically calculable as you say, it will never give you a useful answer.<P>...sorry, it's just that I developed a strong hatred of statistics and its worthless imaginary numbers in eleventh grade.
</B><HR></BLOCKQUOTE><P>
Actually, statistics quite often gives you useful answers. You just always have to calculate the uncertainty in your numbers, so that you know how trustworthy any given value is, and make sure that your assumptions are in line with reality.<P>Statistics are a wonderful, powerful tool. Unfortunately, Joe Average can't easily tell the difference between a valid statistic and an invalid one, which means that organizations publishing statistics feel free to post bad ones to augment their arguments, which means that most people believe that all statistics are a crock.<P>Learn a bit more about the subject, and you will start to see where the useful tools are.

Statistics too often means "98% of cats prefer Whiskas". The most important thing to ask, whenever a statistic is quoted to support an argument, is "what was the question", in other words how was the test performed. This could be the exact wording of the question, layout of the form, or what comparisons were offered (for instance I'm sure 98% of cats do prefer Whiskas to oil-soaked sawdust - - - ). Its also handy to know what the sample size was and how that relates to the population as a proportion. Then you get into esoterics like error distribution. <P>Properly applied statistics are a vital tool, used in industry from manufacturing to market research. Misapplied statistics are usually used when somebody's trying to sell you something, especially if its a politician<P>Ooops! Nearly forgot: - welcome to the board, Christopher.<P>Muttley.<P>Bureaucracy is a giant mechanism operated by pygmies
- Honore de Balzac<p>[This message has been edited by Muttley (edited 05-25-2001).]

All people should know basic statistics and logic. I think that critical thinking should be considered a vital part of education. It's important that people be able to recognize irrelevant or misleading statistics and logical fallacies when they come in contact with them, so they won't be swayed by bad arguments.<P>------------------
"Sun Ra? He's out to lunch, all right...same place I eat at!"
- George Clinton

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Muttley:
<B>Statistics too often means "98% of cats prefer Whiskas". The most important thing to ask, whenever a statistic is quoted to support an argument, is "what was the question", in other words how was the test performed. This could be the exact wording of the question, layout of the form, or what comparisons were offered (for instance I'm sure 98% of cats do prefer Whiskas to oil-soaked sawdust - - - ). Its also handy to know what the sample size was and how that relates to the population as a proportion. Then you get into esoterics like error distribution.
</B><HR></BLOCKQUOTE><P>Also, ask how the sample was chosen. In many cases, they draw from a pre-selected group, so a statistic like "98% of the 10 000 people we polled are against abortion" becomes meaningless if they only sent the question to people who had donated money to Pat Buchanan.<P>(That was just a hypothetical example, btw)<P>Nicolas