At last week's meeting it was mentioned that the current method of determining the E/p mean may not be the best way since it is dominated by the low momentum tracks. As a reminder, the current method involves creating a distribution of E/p for all tracks passing all cuts and then fitting the E/p distribution for each ring with a Gaussian and an exponential and taking the mean of the Gaussian as the mean E/p. Since E/p vs p should be flat, I investigated a couple different ways to fit E/p vs p with a constant and compared that to the original method. Attached are PDFs showing the plots for this for every ring.
 
The two "MeanComparison" files show the E/p mean calculated a few difference ways and compared graphically. Each file has two pages for every ring of the calorimeter. The first page shows the E/p vs p distribution for that ring. On top of that plot is plotted the fitted mean E/p for each momentum bin. Then there are the three different mean lines plotted. 
The blue line is a constant fit to the binned E/p means (each bin given equal weight).
The red line is a constant fit to the 2D E/p vs p histogram
The magenta dotted line is the event weighted fitted E/p mean (the original method of getting the mean)

For all three fits the momentum range used is the same.

On the following page is the E/p distributions for each of the bins, which is used to plot the black points (which is useful for determining which points have good fits).