Re: Question on errors.

From: Rene Brun (Rene.Brun@cern.ch)
Date: Tue Mar 28 2000 - 22:06:23 MEST


Hi Patricia,
Why do you call Sumw2 if you are happy with the sqrt(contents) ?

Rene Brun

On Tue, 28 Mar 2000, Patricia Fachini wrote:

> Hi,
> 
> I define my histograms as:
> 
>   typedef TH1D StHbt1DHisto;
>   mNumerator = new StHbt1DHisto(TitNum,theTitle,nbins,MinvLo,MinvHi);
>   mDenominator = new StHbt1DHisto(TitDen,theTitle,nbins,MinvLo,MinvHi);
>   mDifference = new StHbt1DHisto(TitDif,theTitle,nbins,MinvLo,MinvHi);
> 
>   mNumerator->Sumw2();
>   mDenominator->Sumw2();
>   mDifference->Sumw2();
> 
>   mNumerator->Scale(1./NEvents);
>   mDenominator->Scale(1./NEvents);
>   mDifference->Scale(1./NEvents);
> 
> and after running my macro:
> 
> root4star [17] hh2->GetBinContent(50)
> (Stat_t)8.04982038234593347e+02
> root4star [18] hh2->GetBinError(50)
> (Stat_t)1.26884362824517383e+00
> 
> root4star [19] hhh2->GetBinContent(50)
> (Stat_t)7.85032637287047692e+03
> root4star [20] hhh2->GetBinError(50)
> (Stat_t)3.96240501355956098e+00
> 
> You see that the error in the bin is not the square root of the bin
> contents, as I would expect... What am I missing here?
> 
> Thank you,
> Patricia
> 
> __________________________________________________________________
> Patricia Fachini, Ph.D.                   1-313-577-5419 -> WSU
> Universidade de Sao Paulo                 1-631-344-7635 -> BNL
> Wayne State University                  pfachini@physics.wayne.edu
> 
> 



This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:22 MET