Just catching up on a few points -- too much on CSGNET recently to read everything.
Correlations first.
[From Bill Powers (2007.07.24.1339 MDT)]
Richard Kennaway carried this even farther, proving that group data can be made up of subsets of individual data organized in an infinity of ways that have nothing to do with the group data. There is simply no way to deduce an individual's internal organization from group data: Kennaway put the QED on that.
As I understand what happened, that paper has been rejected and was never published.
I only tried "Science", which turned it down on the grounds of not being the sort of paper they would deal with, which I think is accurate. I never submitted it anywhere else. If someone knowledgable in the area would like to suggest a journal, I could dust it off and give it another go.
[From Bill Powers (2007.08.02.0834 MDT)]
Brief comment. I keep forgetting this, but my initial question on this statistics business was very simple: given a certain correlation, what is the probability that a statement about an individual that is based on group statistics will be incorrect?
That depends on what the statement is. For the statement "for this individual, the Y variable is above the mean", and a bivariate normal correlation between X and Y, I answer this in table 2 of my correlations article.
For zero correlation it is 50% (of course). For any positive correlation it is higher: c=0.5 gives 67%, c=0.8 gives 80%, c=0.9 gives 85.6%.
For the statement "the Y variable falls in a specified decile", the answer is in table 3.
For zero correlation it is 10%. At c=0.9 it is still only 25%, at c=0.95 it is 47%, and c=0.99 gives 78%. In other words, if you're measuring correlations, forget about decile estimation.
The rest is politics.
[From Bill Powers (2007.08.02.0834 MDT)]
The second comment is similar: can anyone lay out the theoretical basis for libertarianism, representative democracy, rule of law, dog-eat-dog, or whatever system concept likes behind the political discussions? We're in a theory-based seminar here, so it seems appropriate to ask what theoretical bases there are for the various points of view.
For libertarianism, I would recommend David Friedman's "The Machinery of Freedom". It's not online, but as a sampler you can read some sample chapters and some of his other writings on the topic at his website, http://www.daviddfriedman.com/
[From Rick Marken (2007.08.01.1012)]
Great post Jeff. I think one of the great questions of the 21st
century will be why a wonderfully humanistic model of human nature
(PCT) has been so attractive to free market types whose approach to
society seems so inhumane, to me anyway.
There is no need to wait. All questions of this form can be answered immediately, because mutatis mutandis, they all have the same answer.
You judge them to have an inhumane view, because you believe the free market (a) has certain consequences, and (b) these consequences are inhumane. You then assume that everyone believes (a) and (b). However, other people draw their conclusions from their beliefs. They do not draw your conclusions from their beliefs. Wondering why they do not reject their beliefs on the basis of your conclusions is not to the point.
If, however, you read those advocating free markets (see the reference to David Friedman above, and the mentions I have made of his work before in this forum), then you would know that such people do not, in fact, predict the consequences you expect. They predict completely different consequences. They also predict consequences of the policies you favour that are completely different from yours: that is why they reject those policies. You may disagree with their predictions, but, having read their predictions, and their arguments for them, you cannot claim that *they* favour the consequences that *you* predict for *their* beliefs, nor that *they* oppose the consequences that *you* predict for *your* beliefs.
To me, PCT makes a very strong theoretical support for libertarianism. PCT says that people do what it takes to achieve their goals, and that their goals are inside their heads, where you can't get at them. As a result, if you do not like someone else's goals, there is very little you can effectively do about it. The only thing you can ever do to anyone is create disturbances to their perceptions. They may find a way around whatever you did instead of changing their goals. If the disturbance is sufficiently large, maybe they will change their goals, but possibly in a way that is even less to your liking. I have never attempted the herding of cats, but that is the phrase that comes to mind.
Hence such things as the Laffer effect, that above a certain point, raising taxes reduces revenue (because people rearrange their affairs to avoid the tax). Hence the ineffectiveness of performance indicators as a basis for salaries, promotions, and funding (because people perform to the letter of the targets instead of what management really wanted). Hence the phenomenon that passing a law against something does not stop it happening.
If one does not understand that (a) people have their own purposes and (b) you cannot control anyone else's purposes, then one falls inevitably into the pattern of demanding more and more government while seeing it work less and less.
I was going to write more, but drawing the straight road from minimum wage laws to hanging dissidents from meathooks is perhaps a little strong for the present audience. However, I do have a few questions for anyone who thinks that a free market is "inhumane":
1. If you have ever run a company and employed people, how did you decide who to hire and what to pay them? Someone who could do the job, and the going rate?
2. If you have ever sold a house or a car, did you try to get as much as you could for it?
3. If you have ever made a major purchase, such as a house or a car, did you look at many candidates and choose the one best fitting your needs and budget, and bargain with the sellers for whatever reduction or extras you could get?
4. Have you ever bought or sold at auction? eBay counts.
···
--
Richard Kennaway, jrk@cmp.uea.ac.uk, Richard Kennaway
School of Computing Sciences,
University of East Anglia, Norwich NR4 7TJ, U.K.