It was discovered that there was a small bug in the polarized weighting calculations in StMCAsymMaker (see http://www.star.bnl.gov/HyperNews-star/protected/get/jetfinding/772.html for a more thorough discussion.)  I have rerun the 2006 simulation trees including the updated code, and show in the figure below the effect on the asymmetry for a sample of polarization models.  Red is the new result (offset by 0.5 GeV), black is the old result.  The difference is negligible.