Systematic Errors for APS W Cross Section

Ross has a very nice writeup of the W cross setion calculation at his blog page.  This page is intended to summarize and give a list of the systematic errors associated with the calculation, and (soon) the resulting estimated error for each one.

Vertex Efficiency:

Without embedding we need a method for calculating the vertex efficiency.  For the pure W MC sample vertex effic. is ~95% so lets consider this the nominal value.  For the L2W data sample passing trigger vertex effic ~85%.  If we consider the effic from the L2W sample a lower bound we can assign a ~10% maximum extent. uncertainty.

TPC Tracking:

Without embedding for the pp500 datat sample we need another way to look at the effects of TPC crowding.  We will use the track finding efficiency shown by Liaoyuan Huo at the Austin analysis meeting.  From embedding MC tracks in jets from Run 6 data Liaoyuan showed the average track finding efficiency was 0.880 ± 0.004.  To compensate for differences between this embedding procedure and track reconstruction in the W analysis at 500 GeV, we will use a more conservative error estimation and use a value 0.880 ± 0.01.  This will give an additional relative error of 0.0114 to be added with the reconstruction effeciency error from the pure W MC.

Jet Reconstruction:

We will use the results from the jet finder without scaling the raw TPC tracks and EMC hits.  Instead the charged and neutral energy's will be scaled by 6% (track finding efficiency and momentum resolution) and ??% (gain uncertainty) respectively.  The resulting upper and lower bounds on the reconstruction efficiency will be added in quadrature to find the resulting systematic error.

Gain Uncertainty:

Since we're using ideal gains, and the Z mass is high by ~??% we should assume a sizeable uncertainty in the gain.  Vary the gain by +/- ??% in the simulation and use the shift in the trigger and reco efficiencies as a maximum extent uncetainty.

Background Model:

Joe has some ideas in a previous e-mail.  The plan is to vary the normalization window, and inverted "awayside type" cut parameters.  This will take some trial and error after Jan has the final version of the optimized W algo.

Luminosity:

Ross computed a value for the BHT3 cross section.  He also generated a table of the luminostiy for each run. The total inegrated luminosity from our data set is

L=13.7403 pb^-1 +/- 2% stat. +/- 23% syst.  This is the value that will be used for the W cross section calcuation.

Yield Stability:

Check stability vs fill, luminosity, cut parameters.  Do we need a systematic error for this?