I read with interest and some amusement (on the mouse joke) the piece written here by Sascha Vongehr. I find his arguments wrong and decided to answer him in the comments thread of his post, but my answer got a bit too long and I did not want to hijack a nice discussion that was developing there; plus I found out that what I was writing could be suitable for this blog in its own right. So below I explain what I criticize about his arguments.
In short, Sascha takes the general scepticism by physicists to the 6-sigma result on superluminal motion found by Opera as an indication of them wanting to "protect the establishment" and the system of POP-science (publish-or-perish), and compares the high significance of that result to one-point-something-sigma effects used in other branches of science to "prove" something. He adds that the 6-sigma Opera result is actually a "confirmation" of the earlier "evidence" by Minos, which found a 1.8-sigma superluminal motion of neutrinos. Here is my reply.
1. A 1.8-sigma result, in particle physics, is nothing. Sascha should not compare to other fields of science, which are totally different things with different metrics. In clinical studies 1.8 standard deviations may grant a publication (with all the negative side-effects that Sascha correctly highlights), but in HEP nobody even considers it an effect. We test hundreds of quantities, and some always fall in the "two-sigma" bin without there being any effect to worry about. 1.8 sigma effects occur once in 20 trials of the null hypothesis, so they are usually nothing, really. Not the "evidence" that Sascha claims Minos was. One would be ridiculed if one took very seriously such an inconclusive result. The reason is largely based on what I discuss in point 2 below, namely the "degree of belief" of a theory or its falsification.
2. a 6.0-sigma result, in particle physics, is generally a significant observation of an effect. However, just how much one believes it depends on seventy years of practice in the field, when highly significant results have occasionally turned out to be due to unknown shapes of the probability density functions of the associated measurement errors. All but the most gullible of my colleagues know they should take with a pinch of salt a "significant observation", because we all know that our measurements are complicated and that we are not infallible.
In the case of the Opera measurement there are a number of things that may have gone wrong, and this alone could have lead anybody to doubt the correctness of the measurement, even if it had been giving v=c. Just as an example, there is a subtraction of the time offset due to the travel time inside a light guide which is 8.5 km long. The travel time of the signal there is 40 microseconds, and this is allegedly known to +- 1 ns, i.e. one part in 40 thousand. Now, that measurement was done in 2007, and since then the refraction index of the light guide might have increased by 0.001+-0.001, adding a 40+-40 ns offset and making the 60-ns effect seen by Opera a insignificant result. [Yes, plastic materials do change their refractive properties in certain cases].
I am giving this example just to explain how difficult these measurements are. Now, there is however another reason to doubt the correctness of the Opera measurement: indeed, physicists have learned to use the fact that "extraordinary claims need extraordinary evidence". No, it is not the same thing to find the top quark at 6-sigma or find that v>c at 6-sigma. The former was an expected result, the latter was not.
Personal bias ? Anathema ? No, just risk analysis. We are Bayesians deep within, and we know that if somebody comes by and tells us "there is a fire downstairs" we are less likely to doubt it than if she comes in and tells us "there is a 12-foot spider downstairs".
We need to remind ourselves that science progresses if we maintain a highly sceptical attitude toward anything we measure or find until we prove it and understand it; neither can be said of the Opera result. Science becomes chaos, instead if we accept any new measurement and its implications, however bold, as a new truth. So being sceptical is the right thing, and Sascha's argument is wrong. The scepticism on the Opera result is not due to physicists wanting to protect the establishment. It is due to their doing their job - requiring more evidence to accept more extraordinary claims.
- PHYSICAL SCIENCES
- EARTH SCIENCES
- LIFE SCIENCES
- SOCIAL SCIENCES
Subscribe to the newsletter
Stay in touch with the scientific world!
Know Science And Want To Write?
- Family Planning: Involving Men Empowers Women
- Some Poor People Have Better Health
- The Internet Of Things And Security: Will A Hacker Stop Your Heart Or Car?
- Mitochondria Signaling Pathway Links Energy Conversion And Cell Division
- Part I: Bee Deaths Mystery Solved? Neonicotinoids (Neonics) May Actually Help Bee Health
- Robert Langer, Ed Witten Awarded 2014 Kyoto Prizes In Medicine And Science
- Part II: Bee Deaths And CCD - Flawed Chensheng Lu Harvard Studies Endanger Bees
- "I was looking at The Winnower again this morning, which led me back here. I tried to comment on..."
- "There is no question they were created to cater to universities and are staffed by people from..."
- "And my second link should be not a copy of the first, but this: http://www.nature.com/news/half..."
- "Recently I saw The Conversation remove/censor legitimate comments arguing against an article written..."
- tags because there's no testing..." href="/robert_inventor/why_we_cant_backup_earth_on_mars_the_moon_or_anywhere_else_in_our_solar_system-148364#comment-182183">"Robert, you might want to reformat my post. I intended to highlight only the headlines for easy..."
- Behavioral interventions to prevent progression to diabetes equally effective in men and women
- Gender differences in adaptation to space flight
- Long-term complication rate low in nose job using patient's own rib cartilage
- Survival differences seen for advanced-stage laryngeal cancer
- Effect of once-daily, low-dose aspirin on heart attack deaths and other outcomes