Science Sunday; Genetic Drift, Placebos and Pot…and Questionable models..

Evolution the biochemists seem to be big on genetic drift as a major factor in evolution. Here is one article that talks about how genetic drift (observed at the organic chemistry level) affects human complexity. It makes us what we are, but might bode ill for our long term future as a species.

You want transitional forms? How about a reptile that makes a true placenta?

Science and Medicine
How about a cloned human embryo making a working stem cell? Of course, this is a very early result.

Why do placebos work? Evidently they activate the same receptor that marijuana does.

Earth’s History: a comet with water with similar chemical composition to the earth’s ocean was found. So is this how the earth got its water?

Evidently, doing PSA readings of asymptomatic men do NOT save lives; the number of false positives cause more harm than is offset by the early detection benefits.

The draft recommendation, by the United States Preventive Services Task Force and due for official release next week, is based on the results of five well-controlled clinical trials and could substantially change the care given to men 50 and older. There are 44 million such men in the United States, and 33 million of them have already had a P.S.A. test — sometimes without their knowledge — during routine physicals.

The task force’s recommendations are followed by most medical groups. Two years ago the task force recommended that women in their 40s should no longer get routine mammograms, setting off a firestorm of controversy. The recommendation to avoid the P.S.A. test is even more forceful and applies to healthy men of all ages.

“Unfortunately, the evidence now shows that this test does not save men’s lives,” said Dr. Virginia Moyer, a professor of pediatrics at Baylor College of Medicine and chairwoman of the task force. “This test cannot tell the difference between cancers that will and will not affect a man during his natural lifetime. We need to find one that does.”

Of course, such early testing is big business, so of course they will fight this. And, of course, the general public really doesn’t understand how a large number of false positives can render such an early detection test not only useless, but potentially harmful (in terms of unnecessary treatments and anxiety).

Speaking of shaky statistics, we see things like this in politics:

John King: The data is just simply not good for the president. And then on top of that, that’s looking backward on the data we already know, today markets are talking about a bear market, financial markets don’t look good, the housing market don’t look good, the president can’t look out at the horizon and see anything to be optimistic about, can he?

Fareed Zakaria: It looks pretty grim. There’s a famous model that was designed by a professor in economic at Yale, Ray Fair, which predicts outcome of presidential elections. He’s never been wrong in 40 years. And the basic inputs are economic growth, unemployment, inflation you know your basic economic data that the theory is that the campaign itself is actually irrelevant, that you give me the economic data and I will tell you whether the incumbent will get re-elected. Well, if you use those kinds of models and those numbers, the situation looks very tough for the president.

Oh, so you accuse me of putting my head in the sand because this model doesn’t show what I want it to show? Here is an example that goes the other way:

Allan J. Lichtman, a presidential historian at American University, has issued a prediction that, given an economy still teetering on the brink of recession and President Obama’s 40 percent approval ratings, looks awfully bold.

“Even if I am being conservative, I don’t see how Obama can lose,” Mr. Lichtman told Paul Bedard of U.S. News & World Report.

Mr. Lichtman’s prediction is based on his 1984 book “The Keys to the White House”. The book cites 13 factors that can work for or against the party of the incumbent president. If at least eight of the 13 keys are scored in favor of the incumbent, he will win the election, Mr. Lichtman says; if he gets seven or fewer, he will not.

The book claims to have called the winner of the popular vote correctly in each election since 1860. (It would not have gotten the winner of the Electoral College right in 1876, 1888 or 2000, when the popular and electoral vote split.) That’s 38 elections in a row! Superficially quite impressive.

But ‘superficial’ is the crucial term here. There are several problems with this model, and its results should be taken with a grain of salt.

So one of these “foolproof” models will be wrong.:)

What is going on? Think of it this way: suppose you had thousands of chimps flipping coins to predict each election for the past N elections. Given enough chimps, there will be a few “super chimps” that got every call right, so far. Would we view these chimps as gurus?


October 9, 2011 - Posted by | biology, evolution, health, health care, nature, science, space, statistics, stem cells

No comments yet.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: