Archive for April, 2010

How to put someone in jail with statistics

April 15, 2010

BBC News – Dutch nurse Lucy De Berk acquitted of patient murders

The 49-year-old was first arrested in 2001 after the death of a baby in her care at a hospital in The Hague, which was thought to be a poisoning.

Afterwards, investigators found what they thought was a trend of suspicious deaths among 13 patients – all of whom were very young and disabled, or very old and in poor health – treated by Ms De Berk in the previous four years. Five others almost died in what investigators said were suspicious circumstances.

In 2003, she was convicted of four murders and three attempted murders, and sentenced to life in prison.

Part of the evidence against Ms De Berk was the testimony of a statistician, who said the odds were 342 million-to-one that it was a coincidence she had been on duty when all the incidents occurred.

Then in 2004, an appeals court convicted her of three additional counts of murder and upheld the life sentence.

Every time I see some number like “342 million to one” I cringe. When I see it from a statistician I growl. First, because it’s almost always based on false assumptions (like independence of all the things being looked at). Second, because it omits the universe in which these things are happening. In a world of 6 billion people, for example, events whose odds are 300 million to 1 against will happen, on average, to 20 people. (And if those are the odds for something happening on any given day, that means the thing whose odds are 300 million to 1 against will happen, on average, somewhere north of seven thousand times a year.)

This is also why I blenched when I saw that the idiots on Wall Street were using models that told them how much money they had at risk 95% of the time. Even if those models had been correct, that means one time out of 20 the risk would be bigger. If it was risk per trading day, that would mean a dozen times a year. If it was risk per trade, who knows? A dozen times a day?

People are lousy with large numbers.

The way medical research should be reported

April 14, 2010

Technology Review: RNA-Loaded Nanoparticles Fight Cancer

The researchers developed a nanoparticle carrying a molecular marker that binds to the surface of cancer cells, triggering the cells to absorb it. The siRNA carried within the particle was designed to silence a gene called ribonucleotide reductase M2 (RRM2), which regulates DNA synthesis and repair and is known to be an anticancer target. Because it was the first trial using targeted RNAi delivery for cancer, says Mark Davis, a professor of chemical engineering at Caltech and the study’s lead author, “we wanted to choose a gene that was suspected to be hugely upregulated in a broad spectrum of cancers” in order to increase the likelihood of being able to observe the novel therapy’s effect.

The researchers analyzed biopsy samples from three melanoma patients in the trial who had received different doses of the therapy. They tracked the particles in the different samples, finding that the amounts they could see in the tumor cells correlated with the doses the patients received. “That’s the first time anyone has seen that for any kind of particle delivery system, whether it’s a liposome, a nanoparticle, or anything,” says Davis. They also pulled out samples of mRNA cleaved exactly where the siRNA’s were designed to cut, showing that the RNAi did its job the way it was expected to. The study does not discuss the clinical effects of the treatment on the melanoma patients in the trial; that data, says Davis, will be presented at the meeting of the American Society of Clinical Oncology in early June.

As I was reading this, I kept thinking: yep, that’s all very nice, but does it actually shrink the tumors or help the patients live longer? And lo and behold, the next paragraph would bring up just that question.

Unclear on the concept

April 14, 2010

BBC News – Man fined for Bebo friend request

Barry Gillies, 25, admitted sending the request to Sally Ann Gray the day after a special bail condition had forbid him from approaching her.

He also admitted sending an e-mail to Miss Gray.

Solicitor Paul Ralph said: “I suppose it means there wasn’t any face-to-face contact.

“He seems to have learned his lesson about what the bail conditions mean.

“He now knows how he has to conduct himself if the relationship is to continue.”

[boldface added]

I’m not sure this solicitor is Gillies’s best advocate. If your client had been arrested for stalking a woman and immediately violated the no-contact order in his release from custody, would you talk about how he intends for the relationship to continue?

Of course, the Beeb deserves a slap too: he’s not being fined for a friend request, he’s being fined for violating a no-contact order. It’s like putting a headline “Man Arrested for Greeting” on a story about someone going up to a witness in their upcoming criminal trial and saying, “Hi. I know where you live.”

Atrios makes a funny

April 14, 2010

Eschaton

While Benen is correct about the odd discrepancy between the Post’s coverage of Ensign and Massa, there is one reasonable explanation for some of that discrepancy. The Massa story is more of a “local” story, in that it’s about the lives and experiences of Capitol Hill staffers.

Uh. Ensign was bonking one of his staffers, and paying off her husband, another one of his staffers, to keep him from blowing the whistle. Of course, both of them have since left washington, so maybe that doesn’t count.

I think what the shorthand really meant is “the lives and experiences of capitol hill staffers who are willing to have drinks or otherwise spend time schmoozing with Washington Post reporters”. In other words, the stories are because the people at the Post are lazy, and the government employees they like to hang out with are inveterate gossips. Coverage of the village, for the village, by the village

Why LED lights are too expensive

April 11, 2010

Because they last too long. No, seriously. If I put a 25,000-hour light in a typical fixture (4 hours a day) that’s 17 years until I replace it. Am I going to be around in this house to see if the claims pan out? And even if the light does last that long, why should I buy one now when in a couple of years it will cost half or a quarter the price? Early adopters always get screwed, but the longer the life of the product the worse they get it. An LED light is pretty much a lifetime investment, so the sooner you switch over, the more you pay.

Which is why even early adopters are having a hard time justifying them. Which is why, in turn, the prices aren’t dropping as fast as they otherwise might. Chicken, meet egg.

There are the power savings, but. Let’s say I replace that 60-watt bulb with a 6-watt LED. 54 watts at 15 cents a kilowatt-hour means about three and a quarter cents a day in savings for my 4-hour-a-day light. Payback time: about 3 years. Meanwhile, if I get a compact fluorescent for $5 or so, it might use more than twice as much electricity, but the payback time is more like six months. Sure, I’ll have to buy another in 5 years instead of 15, but both kinds will be even cheaper by then.

So this leaves the early adopters with longer time horizons (If I were building a house I’d probably go with LED fixtures) and the people who have money to burn in a good cause. Enough? Or not enough?