Orton's Arm

Right, and I didn't blame immigrants for infrastructure decay. They are, however, responsible for the bulk of the infrastructure expansion that will be needed over the next 20 years.

I noticed the line: I doubt very much of that money's because people will be taking longer showers or living in rainier places over the next 20 years. And while much of that money is probably needed to replace existing (and deteriorating) infrastructure, a lot of that $300 - $400 billion undoubtedly stems from the need to accommodate the next twenty years' worth of immigrants. Allowing the U.S. to be absorbed into the Third World is a more expensive proposition than most people realize!

And it's this attitude which causes you to misunderstand and misrepresent my views so consistently and so completely. Dude, if you're going to argue with me for over 50 pages about regression toward the mean--and if you're going to drag the whole PPP board into that debate--the least you can do is get some clue about the ideas you think you're arguing against. Which, quite frankly, you don't have right now.

Such a long post, and yet with so very little comprehension of anything I've been writing for well over 50 pages. Let me give you an example with which you're now no doubt sickeningly familiar. (Ramius, the fact that I've given this example many times before means you don't get to accuse me of flip-flopping.) Consider a population with 1000 people with true I.Q.s of 130, 100 with true I.Q.s of 140, and 10 with true I.Q.s of 150. They're given an I.Q. test in which there's a 20% chance of getting lucky and scoring 10 points too high, a 20% chance of getting unlucky and scoring 10 points too low, and a 60% chance of being scored correctly. Of the people who score a 140 on the test, 200 will be lucky 130s, 60 will be correctly scored 140s, and 2 will be unlucky 150s. When those who scored a 140 are retested, the lucky 130s will (on average) receive a 130, the correctly scored 140s will, on average, receive a 140, and the unlucky 150s will, on average, receive a 150. The pool of people who scored a 140 on the first test contains more lucky 130s than unlucky 150s. When that group is retested, its average score on the retest will be lower than 140--closer to the population's mean. What on earth about the above paragraph could possibly justify you in thinking that I'm incapable of distinguishing between the distribution of the population and the distribution of the error term? Not only is the distinction itself absurdly easy to understand, I've repeatedly demonstrated ample grasp of it. I've always been very clear about the distinction between measured I.Q.s and true I.Q.s; and that the observed changes were strictly due to changes in the measurement error term, not the underlying I.Q.s for individual people. Other people aren't as dumb as you think! Get that through your head! You don't need to go around repeating "Things that are different aren't the same." The reason you don't need to go around repeating that is because other people actually have functioning brains. I did not say that error causes the underlying population to regress toward the mean. For you to be so totally ignorant of my position after 50 pages of repeating it is simply inexcusable. The presence of measurement error is necessary for test scores to regress toward the population mean in a test/retest situation. If you had an error-free test, nobody would get lucky or unlucky when taking it. Everyone would be measured correctly the first time around, so their scores wouldn't change upon being retested. It's the dynamic of a) people being normally distributed, b) being given an error-prone test, c) being placed into a subset based on their test scores, and d) being retested which causes the movement of the test scores toward the population mean. The underlying values don't change, and I've never even hinted that they do. Dude, seriously. Have you seen me go around saying, "Joe was a true 130 who got unlucky when he took the test. In fact, he got so unlucky he became a true 120 and thereby regressed toward the population mean." I mean, seriously. If I'd been saying stuff like that, you'd be perfectly justified in your characterization of my views. As it is, you're arguing against a straw man. I don't know whether you've deliberately created the straw man, or whether you honestly didn't bother to take the time to understand my posts. The phrase "master race" is one you've introduced into the discussion, and is nothing more than cheap and easy propaganda on your part. And maybe you think your intellectual dishonesty is justified because you're fighting what you at least claim to believe is "Nazism." For my own part, I believe intellectual dishonesty is never justified, but that's just me. My actual view is that the average stupid person has a lot more person than the average smart person; and that financial incentives should be used to encourage a somewhat saner reproductive outcome. If that's Nazism, then John Kerry is Karl Marx. In an attempt to discredit my position, you pointed out that if two parents both get, say, a 140 on an I.Q. test, their children will, on average, have I.Q.s somewhere between 140 and the population's mean. My purpose in discussing the test/retest phenomenon was to demonstrate that the population of people who obtained test scores of 140 contains more lucky 130s than unlucky 150s. Therefore, while those parents may have had measured I.Q.s of 140, their true I.Q.s were lower. So at least some of what appears to be a case of children regressing toward the mean is actually just that their parents weren't as far away from the mean as their test scores indicated. On the other hand, if you have two very tall parents, their children will generally also be tall; but not as tall as their parents. This regression toward the mean apparently holds true for just about every trait. On the surface, it would seem regression toward the mean would keep species from changing very much over the long term. Whether parents are the fastest or the slowest, the smallest or the biggest, the smartest or the dumbest, their children will generally be closer to the average than their parents. Nevertheless, both natural and artificial selection can produce powerful, long-term change. While regression toward the mean may slow the pace of change, it can't change the direction, nor can it dampen the ultimate magnitude of it. Darwinistic forces were strong enough to gradually change single-celled organisms into human beings, despite the fact that the dampening influence of regression toward the mean was presumably relevant every step of the way. To imagine that regression toward the mean has somehow locked our species into one particular long-term outcome for intelligence is just plain stupid.

If I did mention the college I went to, you and Bungee Jumper would be the first off the line to make jokes disparaging it. The college itself wouldn't matter--it could be the University of Michigan, Notre Dame, whatever. Just as the articles from Stanford, Duke, etc. didn't stop the two of you from making jokes about regression toward the mean, the school's actual pedigree wouldn't stop the two of you from denigrating it. And those foolish enough to listen to the two of you would falsely assume there was some truth to the jokes. What harm are jokes on a message board? None, except that some of the people reading them--hey, maybe some of the people actually writing them--would be in a position to hire and fire others. If, say, GG learned that an applicant graduated from the same school as me, do I trust him to do the right thing and not hold that commonality against the applicant? Absolutely not. If I'm keeping my mouth shut about specifics, it's to protect my classmates from people just like you.

McDonald's laid you off? As for your comments about oil, know that production is already declining in many parts of the world. American oil production peaked in 1970 at over 4.0 gigabarrels per year, but has been gradually declining since. Today it's at about 1.8 billion gigabarrels per year. Oil production is such that a given nation's or region's reserves don't run dry all at once. Instead, production increases while reserves are located and drilled, then it peaks, then gradually falls. Production peaked in Canada in 1973, it peaked in Mexico in 1977, it peaked in Venezuala in 1970, and it peaked in Saudi Arabia in 2006. The larger OPEC countries' reserves are, the more their cartel allows them to produce. Hence, these countries have a strong incentive to exaggerate the size of their . . . reserves. Dr. Ali Samsam Bakhtiari, a former senior executive of the National Iranian Oil Company, has stated that Iran's oil reserves in particular, and OPEC's in general, are wildly exaggerated. He believes world oil production is at its peak, and will fall 32% by 2020. But I'm sure you know more about all that than he does.

Let's just hope Ramius figures out he's not supposed to eat his prize.

Implying, of course, that our supply of fossil fuels will last through your respective lifetimes. I sense you're being overly optimistic about things staying the same (continued fossil fuel availability) and overly pessimistic about the possibility for change (the creation of a viable electric car).

It's from a top-50 school, which makes it better than whatever "Degree in a Cracker Jack box" school Ramius is currently embarrassing. And I received a more rigorous education in statistics than did Ramius. Then again, Ramius's statistics education consisted of the following classes: Degree in a Cracker Jack Box U course schedule Statistics 101: How to parrot the phrase "correlation does not imply causation" Statistics 102: How to call other people idiots for using statistics

How on earth do you think I'm guilty of "treating the behavior of the error as being behavior of the population"? As you pointed out, I acknowledged the probabilistic error distribution is distinct from the distribution of the underlying population. Nor is this any new revelation, as you can see for yourself if you go back and look at the methodology for my Monte Carlo simulation. Initially, I created a normally distributed population. Then I gave each member an I.Q. test with an error term (mean of zero and a SD of 1/4 the amount of the SD of the population distribution). So you can stop pompously repeating that kindergarten-level phrase about things which are different aren't the same. We both (I hope) agree that in a test-retest situation, with a normally distributed population and a normal or uniform error distribution, those who obtain extreme scores the first time around will tend to score closer to the population's mean upon being retested. You point out that this isn't because the underlying values are moving toward the population mean (they're not) but because of what's going on in that error distribution. Fine. I understand that. I've understood that from the beginning, as you can see if you go back and reread my posts. You seem to feel that "regression toward the mean" ought to mean "regression toward the mean of error" when referring to a test-retest situation with an imperfect test. I feel that whether you're right or wrong about that, "regression toward the mean" does mean "regression toward the population's mean" even in the test/retest situation. But given the fact that we are now (as far as I'm aware) in agreement about the underlying phenomenon and its effects, it would be stupid to spend another 50 pages arguing about a mere definition.

Given the stupidity of just about everything I've ever seen Ramius post, I'm quite comfortable with my conclusion about his intelligence.

You know what? I'm tired of this endless debate. So I'll type out a nice, polite, intellectually rigorous post, in hopes of getting things squared away once and for all. So yes, the proximate cause of the movement on the retest is the fact that the error term will have an average value of zero when people are retested. The expected value of the retest movement is in the direction of the population mean. Hopefully we can shut down this over-long argument by finally coming to an agreement on what's happening. Givens: - A normally distributed population - A measurement mechanism with a normally distributed error term with a mean value of zero - An initial test - A subset being retested. All members of the subset must have obtained either above-average or below-average scores on the initial test. Given the above, you'll have the following: - A normal distribution for the entire population - A probabilistic error distribution for each member of the population. (By that I mean that each member's most likely score is his or her true score, with a 64% chance of scoring within 1 error-dist SD of his or her true, etc.) - When you retest the subset, their scores will, on average, move in the direction of the population's mean. This is due to the following factors: * Any subset consisting of exclusively below-average members will have a disproportionate number who got unlucky on the test * Any subset consisting of exclusively above-average members will have a disproportionate number who got lucky on the test * In retesting the subset, each member's probabilistic error distribution will, on average, reset to zero. * In simpler terms, the subset was initially selected in part based on luck the first time around. Upon being retested, that luck goes away, and the subset's scores move in the direction of the population's mean. This is not to say that every member of the subset will move in that direction. On the contrary, some members will move away from the population's mean upon being retested. But the effect of those few members who move away from the mean will be more than offset by the many who move closer. The average member of the subset will move closer to the mean upon being retested.

What was it like riding the retard bus to school every morning?

As I said earlier, you just looooooove to accuse me of flip-flopping. Even though there's no justification for doing so. But this time around, I'll add that you loooove to accuse me of a lack of rational thought. Lenin said to always accuse your enemies of that which you yourself are guilty. You've been surprisingly good about following his advice.

There is variation in die rolling from one roll to the next, and it's accurate to describe that variation as luck-based. Whenever you have a test/retest situation, with the outcomes of both determined by random chance or luck, extreme scores on the initial test are generally followed by regression toward the mean on the retest. My quote you provided from earlier is therefore accurate. Measurement error is one possible method for an element of random chance to be introduced into a test/retest situation. For example, your measured height could equal your true height + (Normdist, SD 1, Mean 0). With the introduction of such an error term, the correlation between test and retest falls below 1, and hence there will be regression toward the mean.

With respect to Ramius?

Biden's near the top of my list, because his plan for Iraq is probably the best available. Newt Gingrich has a strong track record with welfare reform and federal spending discipline, so I could see him in the Oval Office. He's also very bright, and could probably give us outside-the-box, rigorous solutions to the problem of healthcare. My concern with Gingrich is that, just based on his family life, he lacks the maturity and character needed to be a good president.

It's always disappointing when someone uses fake credentials to lie his or her way into a position of influence or power. But at least nobody that the article mentioned was questioning the accuracy of Jordan's work. That's more than can be said about the NY Times' Jayson Blair.

Oddly enough, I'd been expecting an intelligent response from you. I guess you like to be unpredictable.

The 80% narrow-sense heritability means that if you were to try to predict the height of tall people's kids before those kids were born, your best prediction for those individuals is to conclude that they'll be 80% as much above-average in height as are their parents. So if the parents are 1 SD above average, the kids will, on average, be 0.8 SDs above average. What you seem to be pointing out is that some of the kids will may be 0.6 or 0.7 SDs above average, while others might be 0.9 or 1.0 SDs above average. It's always amusing whenever you point out something perfectly obvious as though nobody else was aware of it. I have a master's degree, and it's from a better school than whatever Cracker-Jack box degree program you conned into taking you on.

Regression toward the mean takes place whenever you have a test/retest situation with a correlation between test and retest of less than one. The presence of measurement error will generally cause the correlation between test and retest to be less than one. But hopefully you knew that already. The truth is there are two similar--but somewhat different--phenomena, that for whatever reason both got labeled "regression toward the mean." Phenomenon A is what Sir Francis Galton observed--the children of tall parents are generally closer to the population's mean. Phenomenon B is where those who obtain extreme scores on an imperfect test score closer to the population's mean upon being retested. The reason these two phenomena got lumped together under the same label is because they both refer to the following: 1. An underlying population in which an average value is more likely than an extreme value 2. An initial measurement 3. Taking some subset of the population based on the value of the initial measurement 4. Some form of remeasurement, where results of the remeasurement have a less than 1:1 correlation with the results of the initial measure. Suppose I were to tell you that, say, 80% of a person's gender-adjusted height could be predicted by knowing the height of his or her parents. If I told you that the parents were 1 SD above the mean, then statistically speaking, the most likely value for their children's heights is 0.8 SDs above the mean. Or let's say I told you that 80% of a person's score on an I.Q. retest could be predicted based on this person's score on the initial test. Someone who scored 1 SD above the mean on the first test would be expected to score 0.8 SDs above the mean on the retest. It's this similarity--caused by correlation coefficients less than one--which got the two phenomena lumped together. This, despite the fact that phenomenon A has to do with children being closer to the population's mean than their parents, while phenomenon B is about measurement error's involvement in the misleading appearance of movement toward the population's mean.

The point about two different distributions is obvious. Carry on . . . Okay, I think you're saying that the change in test scores from test to retest is a result of what's going on in #2. If that's what you're saying, you're correct. Your statements in the above paragraph aren't as difficult to understand as you seem to think. In fact, the underlying logic is rather simple and intuitive. There's a strong--but less than 1:1--relationship between your score on I.Q. tests #1 and 2. There's no relationship between the direction or magnitude of the measurement error on the first test and the error on the second test. But the phrase "regression toward mediocrity" was originally coined by Sir Francis Galton, and refers to the regression toward the population's mean. You're certainly welcome to point out whatever sloppiness you feel may exist in that shorthand phrase. But whether you like it or not, the population mean is what's referred to in the phrase "regression toward the mean."

The hot girl analogy seemed pretty clear to me. As you say, the Bills should be focusing on building their own team, not nervously looking over their shoulder at what everyone else in the division may be doing.

At least based on this paragraph, we're in agreement about the consequences of a test/retest situation, where the correlation between test and retest is less than one. While I can't see any major flaws with your statements about error, the mean of the error isn't what's referred to in the phrase "regression toward the mean." Your use of the word "causes" is clumsy and overly simplistic. Regression toward the mean will take place under the following circumstances: 1. A distribution where an average value is more likely than an extreme value 2. An initial test where the results are due in whole or in part to random chance 3. Select a subset of members based on their initial test scores. If your subset's scores are above the mean, it will have more lucky members than unlucky ones. If it's below the population mean, the opposite will be true. 4. Retest the members of the subset. The population variance to which you're referring allows condition #1 to be fulfilled. The presence of measurement error allows condition #2 to be fulfilled. Did Ramius write that for you? While I wouldn't expect him to have heard of the Gordian knot, everything else in that paragraph is consistent with his level of stupidity.

Just when I think this is the absolute worst level of idiocy Ramius will ever reach you come out with something like this.