While the various trigger, track, and tower distributions from the Run 9 dijet embedding agree well with our Run 9 pi0-triggered data, there's a significant excess in the embedding for towers with a raw (before hadronic correction) 'pT' between 2 and 3 GeV/c.  See slide 7 of the presentation in the link below for details:

https://drupal.star.bnl.gov/STAR/blog/dmawxc/jetcorr-update-may-22nd-2018

It was suggested that this might be a weighting issue: low 'pT' tracks might get smeared upwards into our trigger 'pT' range and the corresponding tracks / towers might get an unreasonably large weight (especially if that track came from a low partonic 'pT' bin).  One way to filter these out is to require that your trigger falls within 50% or so of the event's partonic 'pT'; e.g. if your event has a partonic 'pT' between 7 and 9 GeV/c, then your trigger 'pT' has to be between 4.5 GeV/c and 13.5 GeV/c.

Below are the trigger, track, and tower 'pT' distributions requiring that the trigger be within 50% of the partonic 'pT'.

So, as you can see, the excess remains.  I tried restricting the trigger 'pT' further to within 25% of the partonic 'pT', and below you can see the result for the RFF configuration.

So while the embedding is within 20% of the data for 'pTtwr' ~ 2 - 3 GeV/c, it still shows a clear excess relative to the rest of the distribution.  To me, this suggests that the issue isn't with the weighting... The next thing I plan to look at is how the embedding compares to data after the hadronic correction.