Stats Made Easy

Practical Tools for Effective Experimentation

Thursday, February 28, 2008

Not always right, but never in doubt

This is the motto of those fearless few, such as successful surgeons, who forge ahead with never a look back. As they progress, by-standers bullets just bounce off these never-doubters impervious armor. Assuming it’s true that such confidence (if not outright arrogance) is not misplaced, this a highly desirable trait for doctors, lawyers and other professionals (such as statistical consultants) that others rely upon for good advice on critical matters. Umpires and referees mustn't ever waver in their calls, but as all fans would doubtless agree, bad judgments are made every game, especially against their home team. Also, consider the current race for the American Presidency – do you hear any candidates saying that they doubt the country can ever be put back on the proper path (right or left, depending on party)?

Trouble comes when an expert in a specific area cannot acknowledge incompetence in other endeavors. If you ever run into such a maddening individual who dismisses your greater experience in an area of primary interest, consider this premise of neurologist Robert A. Burton: Certainty arises out of involuntary brain mechanisms that function independently of reason, which I paraphrased from this author's web page for the newly-published book On Being Certain.

Obviously ignorance is bliss in my case, because I am not the least bit interested in reading Burton’s book – it will undermine my confidence in the few things I really feel certain about. Ever since I saw someone at an American Statistical Association conference wearing a shirt proclaiming that “Statistics Means Never Having to Say You’re Certain” it’s been hard for me to develop 100 percent confidence in anything. To be told that I am manifesting an ephemeral mental state like anger or other emotions when I leave no doubt in pillars of certainty such as ‘2 plus 2 equals 4’ would be too much for me to bear.

Can Burton possibly be right? No way! I doubt it very much.

Sunday, February 24, 2008

Models snowed by vagaries of winter weather

When warned of winter storms, my wife must pay attention because she teaches preschool, which might be canceled if enough snow falls. The weathercasters like to tell us where their models predict the heaviest bands of precipitation, but often these fall all over the map. The TV meteorologists then say the "the models don't agree" as a hedge against being blamed for a bad forecast. My standard joke to my spouse, who just wants a simple answer on the amount of snow to be expected, is that of course you cannot get models to provide consistent insights on such complicated natural phenomena – they are far too busy primping themselves for their next photographic shoot! (In my mind I always picture at this point the shallow character Ben Stiller played in the movie Zoolander and his inane arguments with fellow models.)

On a more serious note, it came to my attention (a bit belatedly) that the American Statistical Association (ASA) issued a Statement On Climate Change change several months ago. It included this statement: “The design and analysis of computer experiments [DACE] is an area of statistics that is appropriate for aiding the development and use of climate models. Statistically based experimental designs, not currently used in this field, could be more powerful.” I added the acronym DACE as the shorthand for an approach that seems to be getting more-and-more attention as simulations increasingly complex. The objective of DACE is to produce a transfer or surrogate function that provides an adequate approximation of what the simulation actually predicts. Because these computer programs often are very costly to run, a model of its output can be very valuable for taking short cuts to areas of primary interest by researchers. For example, see this DACE done by Canadian hydrologists studying the Smokey-River watershed in the Edmonton, Alberta area.

I also found a website offering Postdoctoral Opportunities in Statistics at the National Center for Atmospheric Research that includes a number of nuggets for further data mining. The abstract for a project on Estimation of Climate Model Parameters, notes that this work is “a novel implementation of new methodology called ‘Design and Analysis of Computer Experiments’ (DACE)” and that “some computer experiments will always be too expensive to run, so one must be judicious in the experiments that are run.”

In any case, I liked all the colored pictures of contour maps, radar images, hurricanes and tornadoes. It looks compelling enough to distract even a real model from their mirror.

Monday, February 18, 2008

Counterintuitive finding: Sugar substitute correlated to weight GAIN

Purdue University researchers revealed earlier this month that the artificial sweetener saccharin caused rats to put on more weight than others fed sugar. Manufacturers of the sugar substitute responded that this study oversimplified cause for obesity, which involve many factors. See this report by Los Angeles Times Staff Writer Denise Gellene for both sides of the story.

I have no idea if this result will extrapolate to humans, but I would not be shocked if it did. As I gain life experience (that is, get older), my skepticism about generally-held, but never tested, assumptions grow stronger. I find myself more and more reluctant to jump on board with what most people come to accept as irrefutable. Thus a counterintuitive result like this, that something thought to reduce weight actual induces it, does not surprise me.

If you find this result to be difficult to swallow (ha ha), think of how hard it must have been to give up the obvious fact of the earth being flat. Famed physicist Stephen Hawking opens his classic book A Brief History of Time with the story (probably apocryphal) of a flat-earth believer who says to a cosmologist that the earth is supported on the back of a tortoise. When asked how this can be supported, she triumphantly declares that it is turtles all the way down.

“The path of sound credence is through the thick forest of skepticism."
– George Jean Nathan

Saturday, February 09, 2008

Smoker vindicated: Saves $90,000 in health care by not quitting

In my last blog I mildly mocked a colleague who defended his smoking habit by claiming it would kill him younger and thus save on health care. Dutch researchers recently released a study that supports this non-intuitive repercussion of an unhealthy lifestyle. They calculate the cost of care to be $90,000 less for the tobacco junkies versus those who lead a healthy life. They also threw cold water on the idea that obesity weighs down healthcare systems: By eating as much as they like, the eager eaters save nearly $50,000 in lifetime medical costs. Unfortunately these savings in caring for smokers and obese people come at the cost of shorter life spans. See the study’s stats at this article by Maria Cheng of Associated Press.

This begs the question as to whether it’s worth spending a great deal of money on government programs aimed at prevention of smoking and the like. Some might wonder whether these expenditures make much of a dent in the rates of tobacco addiction. However, a new report issued by the American Journal of Public Health claims that U.S. states that spent more on anti-smoking programs had the fewest smokers. (One wonders if this is correlation or causation.) Furthermore, the Centers for Disease Control and Prevention (CDC), a co-sponsor of the study, say that all but four states devote less than they recommend on the hazards of tobacco use. The CDC calculates that there would be 7 million fewer smokers in the USA had more money been spent. Evidently the state health officials came to their own conclusions on the costs versus benefits of attempting to dissuade smokers from shortening their lives as a desirable tradeoff for the “comfort” ( as the statistician Cochran put it) of their cigarettes.

Friday, February 01, 2008

The smoking statistician

The other day my daughter Carrie sent me this interesting bit of trivia on the field of statistics:

“Today in English we discussed some stuff that you should put in your blog. We are reading a book about the history of cigarettes.* It is pretty dry, but there are some interesting things. We discussed the first Surgeon General’s panel, which was set up like a jury. Each side (tobacco and anti-tobacco lobbyists) was allowed to eliminate possible panel members as they saw fit. The final panel included a statistician. The man was also a smoker, and after seeing the results of the studies he announced that based on the statistical evidence, despite pleas from his friends and family, the emotional benefits outweighed the health risks.

Here is the relevant passage:
‘In spite of the findings, as well as the urgings of his fellow committee members and the entreaties of his wife and daughter, Cochran relied on his own statistical analysis to support his decision to continue smoking. Having smoked for a long time, he could not become a statistical non-smoker, only a former smoker. Quitting now, he reasoned, would reduce his chances of succumbing to lung cancer from 40 percent higher than a non-smoker's to 24 percent. 'I think the comfort of my cigarettes is worth that 16 percent chance,' he explained. He nonetheless conceded that he would probably cut down, and he noted that, 'I certainly intend to see that my children never start.'

It sounds like the kind of crazy statistician logic that you would use! Cochran is William Cochran, and the panel was during the Kennedy administration.”

Cochran co-authored a book on Experimental Designs, so I'd heard of him (the American Statistical Association offers this detailed bio). I like this observation of Cochran about response surface methods (RSM): "... polynomials are notoriously untrustworthy when extrapolated." Before Carrie's heads-up I remained unaware of Cochran’s connection to the U.S. Public Health Service Advisory Committee formed in 1963 to research the effects of smoking on lung cancer. His cavalier attitude about being addicted to tobacco reminds me of a very sharp programmer who smoked. When hassled about the potential costs in terms of taxpayer-funded health care, his comeback was that smoking kills people at a younger age and thus reduces the ultimate medical expense!

*A Cigarette Century: the Rise, Fall, and Deadly Persistence of the Product that Defined America by Allan M. Brandt.

"THE single most shattering statistic about life in America in the late 1990s was that tobacco killed more people than the combined total of those who died from AIDS, car accidents, alcohol, murder, suicide, illegal drugs and fire."
-- Lead statement from The Economist article An evil weed which features Brandt and his book.

PS.As much as smoking turns me off, I must admit to enjoying Jason Reitman’s movie Thank You for Smoking, which featured an astute lobbyist (played by Aaron Eckhart) who could turn any shot right back at the anti-tobacco forces. It is very amusing to see him skewer the self-righteous Senator (actor William H. Macy).