Dark Matter Particles May Have Been Detected 156
During two seminars at Stanford and Fermilab on Thursday, researchers described signals for two events detected deep in an old iron mine in Minnesota that might mark the first detection of dark matter — or not. The presenters said the chances that the signals they detected were caused by something other than "neutralino" dark matter particles was 23 percent. "One source indicates that we'd need less than 10 total detections within the CDMS' range in order to have a high degree of confidence in the results." The NY Times describes the Cryogenic Dark Matter Search methodology: "The cryogenic experiment is nearly half a mile underground in an old iron mine in Soudan, Minn., to shield it from cosmic rays. It consists of a stack of germanium and silicon detectors, cooled to one-hundredth of a degree Kelvin. When a particle hits one of the detectors, it produces an electrical charge and deposits a small bit of energy in the form of heat, each of which are independently measured. By comparing the amounts of charge and heat left behind, the collaboration’s physicists can tell so-called wimps from more mundane particles like neutrons, which are expected to flood the underground chamber from radioactivity in the rocks around it." Here are the research team's summary notes of the latest results (PDF).
Re:It must be true! (Score:2, Insightful)
Re:1:4? (Score:5, Insightful)
Re:Meanwhile, Rome burns (Score:4, Insightful)
Re:Meanwhile, Rome burns (Score:1, Insightful)
Re:Meanwhile, Rome burns (Score:4, Insightful)
It appears from a quick Google search that the University of Minnesota [uppermidwestherc.org] is funding it. I guess you're against NASA, too, but in favor of pouring trillions down the Iraqui quagmire?
WTF are you doing on slashdot? Trolling?
Re:It must be true! (Score:2, Insightful)
Re:It must be true! (Score:1, Insightful)
Joe Public: "Well, I don't understand it, so it can't possibly be right."
Re:It must be true! (Score:3, Insightful)
They never said that "it's only 77% or so likely to be a positive signal", or that "the chances that the signals they detected were caused by something other than 'neutralino' dark matter particles was 23 percent". What they said is that there is a 23% chance that, in the total absence of DM particles, the background would generate those two events they found. Or: If they did the same experiment, with the same exposure and analysis (and if DM doesn't exist) hundreds of times, 23% of those experiments would show two or more events.
The chance that you flip a coin three times and get the same side up all three times is 25%, if the coin is honest. This doesn't mean that if you flip a coin three times and get the same side up three times, there is a 75% chance that the coin is not honest.
Going a bit deeper in the statistics, the probability that A happens with the hypothesis B is NOT the same as the probability that hypothesis B is true given that A happened. To "flip" those 23% probability that the background gives you 2 events to a probability that the two events events are caused by the background, you need to apply Bayes' theorem [wikipedia.org]. You can only do that if you use, as prior knowledge, the probability that a dark matter particle exists. This prior can't be defined precisely by anyone, each physicist would give you a different value; so you can't apply Bayes' theorem here without being heavily biased.
What you can infer is that there is a pretty good chance that this was just a background fluctuation, specially since their previous results had zero events (with a similar background expectation). The REAL point that physicists got from the talk was that CDMS reached its limit, and has to be upgraded to SuperCDMS to stay relevant.
Re:It must be true! (Score:5, Insightful)
The only comment I have to make in the other direction is that I am uncomfortable with the probabilities that scientists have suddenly started to give - "there's a 77% chance we are completely correct".
Except it's not even that.
There saying there's a 77% probability that the result was not due to random noise, and that they actually did detect particles that are within the range predicted for neutralinos by Supersymetric Theory. Does that means it's a neutralino? Not necessarily, but it is a pretty strong argument of the "hypothesis -> experimentation -> verification" variety. Does it mean that everything they predict for neutralinos is true, or that Supersymetric Theory is "completely correct"? No.
I wish for some good old scientific conservatism, and the need to put percentages on the proportion of 100% correct you are feels a bit dubious.
Again, they're only putting a percentage on this not being a null result. Your characterization is wrong.
They're being conservative. But they're excited. And when you take a theory as ridiculously successful at making predictions as the Standard Model, make a logic extension to it and then that theory quite possibly has had its first verified prediction, that's not unreasonable.
I remember when scientific skepticism on slashdot involved people taking issue with specific aspects of the experimental procedure. Not people complaining that they don't like the result or how snooty the scientists are using statistics to measure their success.
Re:It must be true! (Score:3, Insightful)
In the eighth grade, when I was told that an argument that could be summed up as "Respect my authority!" with no factual backing was a fallacy.