Category Archives: Other science

aka stamp collecting

Bad Name for a Medical Partnership

Posted on by

Scott and Scurvy

How the cure for scurvy was found and lost again.

They had a theory of the disease that made sense, fit the evidence, but was utterly wrong. They had arrived at the idea of an undetectable substance in their food, present in trace quantities, with a direct causative relationship to scurvy, but they thought of it in terms of a poison to avoid. In one sense, the additional leap required for a correct understanding was very small. In another sense, it would have required a kind of Copernican revolution in their thinking.

I find this fascinating: the application — almost — of the scientific method, only to fail at the crucial falsification stage. And how the wrong answer propagates because of this failure.

Back to the Woodshed with You

Posted on by

Last week I took George Will to task for his scientific illiteracy and misrepresentation of the “no statistically significant warming” statement that has given every global warming denier a naughty tingly feeling during the past few weeks.

I missed something.

I was going to include a graphical example, and I should have, because I would have found one more problem with the statement. I was reading a post over at Skeptical Science, where graphs were included, and did a mental reconstruction and realized my error of omission. I’ll grab the GISS graph from that post (slightly different slope, but the concept is the same), and add in two lines: one representing no increase in T, and one representing twice the amount slope of the best fit.

temp change

Now, one can see here that even though it’s obviously not the best fit to the data, the “no increase” line is a semi-plausible fit. It’s possible. But here’s the problem: look at the temperature in 1995 based on the two scenarios. If one is going to claim that the temperature has not increased in the last 15 years, one also has to admit that it’s about a tenth of a degree warmer than we thought it was. So all of the global warming that “didn’t happen” before is even worse, and harder to explain away.

Personally, I think not distorting the science in the first place is probably the best way to proceed.

What Are the Odds?

Posted on by

DNA’s Dirty Little Secret

In Puckett’s case, where there were only five and a half markers available, the San Francisco crime lab put the figure at one in 1.1 million—still remote enough to erase any reasonable doubt of his guilt. The problem is that, according to most scientists, this statistic is only relevant when DNA material is used to link a crime directly to a suspect identified through eyewitness testimony or other evidence. In cases where a suspect is found by searching through large databases, the chances of accidentally hitting on the wrong person are orders of magnitude higher.

The reasons for this aren’t difficult to grasp: consider what happens when you take a DNA profile that has a rarity of one in a million and run it through a database that contains a million people; chances are you’ll get a coincidental match. Given this fact, the two leading scientific bodies that have studied the issue—the National Research Council and the FBI’s DNA advisory board—have recommended that law enforcement and prosecutors calculate the probability of a coincidental match differently in cold-hit cases. In particular, they recommend multiplying the FBI’s rarity statistic by the number of profiles in the database, to arrive at a figure known as the Database Match Probability. When this formula is applied to Puckett’s case (where a profile with a rarity of one in 1.1 million was run through a database of 338,000 offenders) the chances of a coincidental match climb to one in three.

It’s scary that a judge didn’t find these statistics relevant, and scarier still that the FBI is denying database access to scientists who might confirm that their statistics are right or wrong.

I wonder, though, if the defense asked if the police had tracked down the other ~300 suspects who would have fit the DNA profile in the US, and if that would have raised reasonable doubt?

FOOF is not for the Faint of Heart

Posted on by

Things I Won’t Work With: Dioxygen Difluoride

aka FOOF

Sulfur compounds defeated him, because the thermodynamics were just too titanic. Hydrogen sulfide, for example, reacts with four molecules of FOOF to give sulfur hexafluoride, 2 molecules of HF and four oxygens. . .and 433 kcal, which is the kind of every-man-for-himself exotherm that you want to avoid at all cost. The sulfur chemistry of FOOF remains unexplored, so if you feel like whipping up a batch of Satan’s kimchi, go right ahead.

George Will is a Boulder

Posted on by

Global warming advocates ignore the boulders

He’s certainly not a scientist, nor, seemingly, is he scientifically literate.

In his latest steaming pile of op-ed on global warming, Mr. Will attempts to call into question the “settled science” of global warming by discussing virtually no science at all. Seriously — a bunch of politicians not being able to agree on a course of action does nothing to question the science. And likewise for businesses making a business decision. But it is the claim that there has been no recent warming that is what really bugs me. George almost gets it right earlier in the op-ed, when he says there has been no statistically significant warming in the last 15 years, but here he (and many others) sin by omission. If one follows the link back to the BBC interview with Phil Jones, one gets a better picture

Do you agree that from 1995 to the present there has been no statistically-significant global warming

Yes, but only just. I also calculated the trend for the period 1995 to 2009. This trend (0.12C per decade) is positive, but not significant at the 95% significance level. The positive trend is quite close to the significance level. Achieving statistical significance in scientific terms is much more likely for longer periods, and much less likely for shorter periods.

In other words, if one has a sufficiently noisy data set, it is always going to be possible to pick a subset of the data where the noise masks any statistically significant trend. It doesn’t mean the trend isn’t there, or that the best fit is a zero slope. When Jones says “Yes, but only just” he’s telling us that one can just draw a flat line through the data, but this means that one could also draw a line with a slope of 0.24C per decade through the data, and it would have the same importance — you can’t exclude warming at that rate, either.

Imagine this question being asked instead:

Do you agree that from 1995 to the present it’s possible there has been global warming at a rate as high as 0.24C/decade?

The answer would have to be essentially identical, i.e. it would have to be yes. You can only statistically exclude warming at a higher rate than that!

What one certainly can’t do (that is, with any intellectual honesty) is conclude that this is an absence of warming. Statistically speaking, if the best fit to the data were a line with no slope, one could rule out neither an increase nor a decrease — one could only quote a limit on those trends. That’s one of the things about science — we try and quantify our results, rather than bandy about generalities. You might force a sound-bite answer out of a scientist (or worse, get there by ripping a quote out of context), but the instinct is to properly qualify the result.

So what if you don’t want both of the above scenarios to fit Dr. Jones’ answer? If you want a statistically significant answer you have to do as he suggests and look at a longer set of data in order to beat the noise down (random noise will average out with the square root of the number of data points). Anyone who does experimental science knows this, and is one of those things that a scientifically literate person should know. So the choice of a short data set is a form of cherry-picking — selecting a data set in such a way as to present a misleading result. If one looks at a longer data set, a statistically significant trend does emerge, and it is one of warming.

George, you’re not a scientist. I had some respect for you in the days I used to read your opinion pieces, because you could and did make cogent arguments, even if I did not agree with you. But science is based on facts, not opinions, and when you have to misrepresent those facts to make your point, your conclusions aren’t worth the paper on which they are printed.

Update: there’s more

Feel the Burn, Baby

Posted on by

Pictured: Incredible gravity-defying ant that can carry 100 times its body weight

The photo shows an Asian weaver ant hanging upside down on a glass-like surface and holding a 500mg weight in its jaws. It was captured by Dr Thomas Endlein of Zoology department at the University of Cambridge who was investigating the sticky feet of ants and other insects.

No mention of results of any steroids testing on the ant.

Lots and Lots

Posted on by

Physics Buzz: How Much Snow did Washington DC REALLY Get?

Washington DC looks like a 10 x 10 mile square with a bite taken out of it. All together the city is 68.3 square miles. It’s not too hard to figure out the total volume of snow dumped on the city so far.

55.6″ depth of snow x 68.3 square miles roughly equals 8,820,000,000 cubic feet of snow, or 249,000,000 cubic meters. If you were to build a giant cube of snow that big it would be 2,066 feet, or 629 meters on each side. That’s almost two fifths of a mile or two thirds of a kilometer per side. That’s the volume of about 238 Empire State Buildings. That’s a lot of snow.