0
$\begingroup$

At the TAUP conference in 2011, the Borexino collaboration presented an analysis of electron-neutrino elastic scattering events in their liquid scintillation detector to claim the first observation of neutrinos produced by the pep reaction in the Sun. They quoted a value for the total flux of solar pep neutrinos of 1.6±0.3 (units of 0.01c) and a significance level for the detection of 97%.

I'm wondering is the confidence level justifiable? I'm wondering does anyone know whether it is distributed as a normal distribution and, as result, is the significance level valid?

$\endgroup$
0
$\begingroup$

For normal statistics 97% confidence is between 2 and 3 sigma.

When quoting a 'detection' of a small signal in a large background you are claiming to have statistically distinguished the signal from zero, so for normal statistics tat would be a claim that the centroid is nearly three times as far from zero as the 1-sigma uncertainty (and $0.3$ is less than a fifth of $1.6$).

However, this kind of bounds-setting measurement doesn't generally follow normal statistics and there is likely a more involved theory at work here. Possibly Feldman-Cousins which has seen a lot of use in the field in recent decades.

In any case, that level of confidence does not seen unreasonable given the quoted measurement, but the details depend a lot on exactly how the analysis was done. There should be a pre-print or paper either now or soon.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy