When the DCA temperature sensor was replaced on August 10th, the psychro-dyne indicated that it was reading 1.7 degrees too high. From that date forward, the 2015 DCA-IAD average temperature differential shrunk from 5.06 degrees through August 9th to 3.45 degrees thereafter -- a reduction of 1.61 degrees. So, I think that's a pretty good indication that the old sensor was reading about 1.7 degrees too high for many months. If it was doing so beginning no later than New Year's Day 2015, the official yearly average temperature at DCA was inflated by a full degree by the faulty sensor.
However, there have very likely been past sensor issues at many sites -- including DCA and IAD -- that received little attention or even went undetected. For example, in 1976, the DCA-IAD annual temperature differential was 5.9 degrees, whereas in 2009 it was only 1.7 degrees. Were the DCA and IAD sensors both accurate in those years? So, I think the bottom line is that there was a problem with the DCA sensor for much of 2015, but attempting to adjust the numbers would open up a new can of worms.