Quote:
Originally Posted by Shiraz
sure its is too low to qualify as a detection under the 5 sigma rule that the particle physicists have embraced. However, if you are 95% confident that you have found something, you cannot ignore it either. The particle physicists would just plough on with more runs until they had more confidence, but that is not an option with a strictly one-off gravitational wave detection. I would think that co-incident 2 sigma events at each of the two detectors would have to be quite significant - although who knows how they assess confidence when they are looking for something from a group of "sort of predictable" structured signals, at 2 sites and against noise that presumably has both random and structured components.
Let's see what they do, but it is interesting that that have already begun to publicise the extra candidate events, even those below 2sigma. My guess is that, now they have one hard detection and a lot of real noise to study, they will be able to increase the confidence estimates on any other candidate events.
|
Interestingly enough the neutrinos travelling faster than light fiasco was a six sigma event but with no confirmation test on different apparatus.
History has shown it was a nice strong signal of a systematic error.
Dark matter has been "detected" at three sigma using the SuperCDMS but the calculated mass/collision cross section are inconsistent with the exclusion limits of the Xenon10 and 100 detectors.
http://www.science20.com/science_20/..._matter-109299
Steven