I know by experience that, when a test does not say
what practitioners would like, other tests are
tried - like when one goes around looking for someone
that finally says one is right.^{7}
Indeed, in the statistics
practice there is much freedom and arbitrariness
about which test to use and how to use it. This is
because *hypothesis tests* of the so called
*classical statistics* do not follow strictly from
probability theory,
but are just a collections of *ad hoc prescriptions*.
For this reason I do not want to enter on what
CDF finally quotes as *p-value* (with the only comment
that it does not even seem a usual
p-value).
Let us then just stick to the paper, reporting here
the claim, followed by a reminder about
what a statistician would understand by that name:

- ``we obtain a p-value of , corresponding to a significance of 3.2 standard deviations'';[9]
- ``the p-value is the probability of obtaining a test statistic
at least as extreme as the one that was actually observed,
assuming that the null hypothesis is true.''[11]
[The
*null hypothesis*() is in this case ``only standard physics, without contributions from new phenomena''.]

Giulio D'Agostini 2012-01-02