This QA was performed on the individual jets comprising a di-jet event. All the attached pdfs show the same data for the individual triggers.  Even after the QA was completed there were still runs that fell outside the green line(s), but these runs were left in the runlist since they did not meet the removal requirement.  The trigger that examplifies this the most is JP0, this doesn't bother me since it has the lowest event statistics due to high pre-scale.  In contrast JP1DiJet (small pre-scale) and JP2 (take-all) look much cleaner.

Also the last time I present the di-jet QA, there was a comment that for my eta and phi plots that the average was slightly above zero.  Attached are plots showing phi VS eta for the di-jet triggers.  (Other triggers are avaible, but these two show the most dramatic effect).  If you look at phi around -1, this area seems slightly "colder" than the rest of the detector.  Since the Barrel isn't completely symmetric this is leading to the fact that the average for phi or eta aren't exactly zero, but still prettly close.

I think at this point since most of the quantities show a luminosity dependence, it would be advantageous to QA versus luminosity and see how the data looks after that has been completed.

Also note that these plots were made with an older trigger algorithm, not the latest one that has been shown in the blog previous to this one.  I would guess that with the design of the new algorithm; that JP0/JP1 might look worse, but JP0DiJet/JP1DiJet would better due to having more events.