Doug,
At a risk of being wrong...again.. I would like to take an issue with
your binomial distribution statement. What we are doing at BLAST, to me
at least, is conducting six different experiments each with a set beam
and target helicity state. For each of these experiments we collect
yields, related to the cross section and thus follow Poisson
distribution. After we collect the yields from these "experiments" we
compare them to see if one experiment gives us an "excess yield". We go
to great length by randomizing (or pseudo randomizing) the target spin
state in order not to know which "experiment" comes next. I don't see
any correlation between these six "experiments". In fact one can
argue that every time we flip the polarization state we start a new
"experiment". We then combine many mini experiments into six big ones
assuming that the running conditions are normally distributed, which is
at the center of combining-different-polarization-data discussion. Let
me know what you think, and where I'm wrong :-) .
Cheers, Vitaliy
P.S. In fact there are other "excess yield" experiments out there that
don't deal with spin, like NMC or "color transparency".
Douglas Hasell wrote:
> Hi,
>
> Following on from Michael's initial discussion last week about
> combining asymmetries and calculating the uncertainties I derived a
> few equations which someone should check.
>
> But my general conclusion is that the way we are combining
> asymmetries has some limitations and particularly I would like to
> suggest that everyone checks how they calculate the uncertainties as
> the statistics are binomial and not normal or gaussian.
>
> Anyway, as usual I got carried away and generated 6 pages of text
> and equations. Please check it and let me know if I made any mistakes
>
>
>
> Cheers,
>
> Douglas
>
> 26-415
> M.I.T. Tel:
> +1 (617) 258-7199
> 77 Massachusetts Avenue Fax: +1
> (617) 258-5440
> Cambridge, MA 02139, USA E-mail:
> hasell@mit.edu
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:32 EST