seleff.tex

Lensing pipelines are not able to return 
a shear measurement for all galaxies in an
image. If a lensing pipelines preferentially rejects 
objects based on their ellipticity and their orientation relative of the ellipticity of the PSF, then the assumption that the population of galaxies used to
measure shear have $< \epsilon > = 0 $ is no longer true. 
This selection bias effect has been discussed in STEP1 
where it was determined that results for some lensing pipelines  
improved when the non-zero mean intrinsic ellipticity of the surviving 
objects was corrected for.

The fraction of successful shear measurements
 for each pipeline is shown for sources with SNR $>$ 20 in Table \ref{table:sur_sel}.
The pipelines with the highest efficiency of successful measurements
are PK, DE and KM, which are all moment based methods. 
As implemented in this challege the GM method
rejects a portion of the returned shear measurements and 
uses a weighting system in determining the average shear. Before the
flagged objects were rejected, the efficiency of GM was 0.93 before and 0.24 after rejection. These values are
therefore not directly comparable to the other lensing pipelines. 

Pipelines may reject objects in a ellipticity dependent way. To test
for this selection bias, the $\gamma$ measured is corrected for
the intrinsic ellipticity of the sources as measured by each lensing
pipeline. I3 and DE do not appear to have a selection bias as their
accuracy is similar, before and after correction for $< \epsilon > \neq 0 $. The
pipeline that is most affected by selection bias is PF. Q, M and C as measured for each
lensing pipeline after correcting for selection effects is included in
Table \ref{table:QMC_sel} and Table \ref{table:MC_sel}. \green{references to tables are orphaned here} 

%%%%%%%%%% The refs above are still a problem

%%%%%%%%%%%%%%%%%% ************************
\begin{table}
        \centering
        \begin{tabular}{|c|c|c|c|c|c|c|c|}  
          \hline
          Pipeline  &  efficiency \\
          \hline \hline
          DE &  0.92 \\
          \hline
          PF & 0.86 \\
          \hline
          GM & 0.24 (0.93) \\
          \hline
          MJ & 0.81 \\
          \hline
          PK & 0.90 \\
          \hline
          I3 &  0.79 \\
          \hline
          IM &  0.55 \\
          \hline
          KM & 0.91 \\
          \hline
        \end{tabular}
        \caption{Percentage of galaxies at SNR$>$20 for which the shape measurement pipelines return a successful measurement.}
    \label{table:sur_sel}
\end{table}



\begin{figure}
\centering
\includegraphics[width=0.5\textwidth]{fig/QMC_main_sel_f.pdf} \\
\includegraphics[width=0.5\textwidth]{fig/MC_main_sel_f.pdf} 
\caption{Average Q and M results for all pipelines for objects 
SNR $>$ 20 after correcting for selection effects in the top
panel and average M and C results for all pipelines
for objects SNR $>$ 20 in the bottom panel.}
\label{fig:QMC_main_sel}
\end{figure}