The final major question before achieving final results is whether or not to use R = 0.6 or R = 0.5. There are arguements for each. One of the biggest questions is the change in total uncertainties for the move to R = 0.5. To inform this question I have performed inclusive-jet and Collins-like analyses for both radii and post below the comparison of results.

Changes For This Iteration

I have implemented a few changes for this iteration of the analysis. First, I found a mistake in how I was estimating the angular resolution. I was not taking the RMS for δφ over the full range; and, thus, the dilution was underestimated. This has been fixed. Second, I estimate the pion contamination, strictly, from the bin-by-bin fits to the n_σ(π) distribution. This is not applied as a correction, but the difference from the nominal and dilution-correction is assigned as a systematic. For the preliminary result I had looked into simulation, where the values were considerably worse. I believe this to be a bug, but to be conservative I assigned the systematic assuming the dilutions from simulation. I will investigate this further, but I think the current values are good enough for the present purpose. Third, I relaxed some cuts following the procedure outlined on a recent blog entry. Finally, I switched to the event-by-event polarization correction technique as outlined in the CNI documentation. Previously, I did a fill-by-fill polarization correction using the average polarizations for each fill, as given by the CNI group.

Table 1: Inclusive Jet Asymmetry for η > 0

So, the switch to R = 0.5 results in anywhere from 2-37% increase in total uncertainties, according to the current estimates. As expected, the effect is largest at low-p_T. In one bin (11.7-13.8 GeV/c), the total uncertainty actually decreases, however, this appears to be due to a fluctuation in the matching probabilities in the R = 0.6 sample. The same feature only in the R = 0.5 sample appears to account for the 37% incrase in the 9.9-11.7 GeV/c bin. These can be smoothed out for the paper.

Table 2: Collins-like Asymmetry for π⁺ at η > 0 and the Low p_T-Range

For Collins-like, the switch to R = 0.5 results in less significant changes, anywhere from an 18% increase to a 7% gain in total uncertainties, according to the current estimates.

Table 3: Resolution and Matching for Inclusive Jet Asymmetries

Table 4: Resolution, Matching, and Contamination for Collins-like Asymmetries