Statisticians can prove almost anything, a new study finds

National Post ^ | November 20, 2011 | Joseph Brean

[marty60]

Most people learned this in Statistics 101.

A few Quotes:
# An unsophisticated forecaster uses statistics as a drunken man uses lamp-posts - for support rather than for illumination. ~ Andrew Lang
# There are three kinds of lies: Lies, Damn Lies, and Statistics.
~ Mark Twain ; Benjamin Disraeli
(Check Correct Attribution - http://www.york.ac.uk/depts/maths/histstat/lies.htm)
# Like other occult techniques of divination, the statistical method has a private jargon deliberately contrived to obscure its methods from non-practitioners. ~ G. O. Ashley
# I can prove anything by statistics - except the truth. ~ George Canning
# I could prove God statistically. ~ George Gallup
# Statistics are no substitute for judgment. ~ Henry Clay
# A single death is a tragedy. A million deaths is a statistic. ~ Joseph Stalin
# It has long been recognized by public men of all kinds . . . that statistics come under the head of lying, and that no lie is so false or inconclusive as that which is based on statistics. ~ Hilarie Belloc (The Silence of the Sea)
# It is commonly believed that anyone who tabulates numbers is a statistician. This is like believing that anyone who owns a scalpel is a surgeon. ~ Hooke R. (How to Tell the Liars from the Statisticians)
# The manipulation of statistical formulas is no substitute for knowing what one is doing. ~ Hubert M Blalock Jr (Social Statistics)
# Facts are stubborn things, but statistics are more pliable. ~ Laurence Peter
# I abhor averages. I like the individual case. A man may have six meals one day and none the next, making an average of three meals per day, but that is not a good way to live. ~ Louis D. Brandeis
# Fate laughs at probabilities. ~ Lytton E.G Bulwer

http://www.workinghumor.com/quotes/statistics.shtml

[blueunicorn6]

Psychological Science - there’s two words that don’t go together. They don’t like statistics, but they believe you can relive your birth experience.

[Hostage]

Well you are wrong and misinformed.

Statistical methods are indeed mathematical and proved to be unbiased, convergent, efficient, and so on. All of this is theoretically proven.

Random sampling designs are part of the theoretical framework and statistical design is also presented in terms of statistical ideals, all theoretically.

Because of Central Limit Theorems and Laws of Large Numbers which are indeed physical and provable, any number of hypotheses can be tested and theoretical arguments on testability can be made.

Where statistics is abused is ultimately revealed in conditioning and marginalizing, from where spurious non-zero correlations are induced. This phenomenon can be corrected by inclusion of variables and making adjustments.

All of the above are theoretically provable using our mathematical systems. The only way they can be wrong is if the physical laws of distribution are wrong or our mathematics is wrong.

So no, statistics is not wrong and those trained in it are not wrong. What is wrong is the way that others use and abuse it, sometimes knowingly abuse it, or ignorant.

And analogy would be to hear complain that there are too many lawyers in the world when in fact that statement is vacuous. A truer statement is “there are not enough good lawyers in the world” where ‘good’ is defined against a criteria of optimal justice and truth. The same can be said of statisticians, medical doctors, scientists, politicians, farmers, sports athletes, etc.

When you say that ‘good statisticians’ can’t be judged or defined adequately or objectively, you are wrong. It would be as wrong as saying the same thing when one substitutes the descriptor ‘plumber’ or ‘electrician’. Just as in these trades and others there is a factor of knowledge and experience that makes one contractor ‘good’ or meeting some industry standards, and there are indeed standards in the professional industry of statistics.

To read of an example, one need only read of what Demming did for the automotive industry of Japan to see how the ‘proper’ application of statistics can lead to vast improvements in quality and efficiency.

http://en.wikipedia.org/wiki/W._Edwards_Deming

[Scythian]

I once knew a statistician who drowned in a river whose average was only 1” deep.

[dagogo redux]

” . . . theoretically proven.”

I rest my case.

[Hostage]

No your case is not strong at all. First you need to read the link I left you about Deming.

Then you need to review physical theories that are proven in mathematics (special relativity) and not by conclusive experiment, but accorded experiment.

http://therandomfact.com/einsteins-law-of-special-relativity-still-in-the-dock/2210600/

Statistics as pertains to laws of distribution is physical and its theory is confirmed by experiment. So your quip about theoretical proof resting your case is nothing more than a chip on your shoulder. The fact is that theoretical physics is also PROVED by theoretical considerations and CONFIRMED by experiment, no different than theoretical statistics.

Your argument against statistics as science is not only shallow but flies in the face of the historical facts of eminent scientists who understood that physical distributional theory could only be confirmed in experiment by using statistical theory and its estimation methods.

Again, try telling the Japanese that their honoring of Deming for his statistical methods was misguided and wrong. The proof is in the results and the assertion remains valid that trained, experienced and objective statisticians are able to use statistics to improve our lives.