## Higgs boson exists with 99.9999% certainty…

**Y**uk! Among the many articles celebrating this tremendous step in particle physics, there are many sentences like the one above, found in *Le Monde*. (This is actually the title of the article, with the additional sentence “*Il y a désormais plus de 99,9999 % de chances que l’observation soit correcte*.”) Both sentences being utterly meaningless, it would be nice if journalists and presumably physicists could understand the meaning of a p-value..! Other blogs have already pointed out the fallacy of the inversion of ** p(|x|>5σ)** into this meaningless

**so I will not fill many pages about the issue, however it sounds like there is an innumeracy issue there. Still, both my kids did basic confidence intervals in high school, where they were pointed the danger of the p-value inversion fallacy. (Of course, this presentation is fairly new in French high schools. In my days, days of yore, statistics was**

*99.9999%**definitely not*a high school subject!)

July 6, 2012 at 4:51 pm

But i would think it would be meaningful for a Bayesian, even if you disagreed with the calculation. The Higgs appraisals I have found construe the p-value or corresponding CI correctly, and they illustrate how scientific inference manages without invoking priors. They are based on numerous tests/checks of error taken together.

July 6, 2012 at 6:59 pm

Deborah: This was not a Bayesian vs. non-Bayesian comment, simply the much more basic remark that people are still unable to interpret p-values (and posterior probabilities) correctly when they state that the alternative hypothesis is .999999 true because under the null hypothesis the observed event [and not even that but a related tail event] is .000001 probable…