Analysis of Coincidence-Time Loopholes in Experimental Bell Tests

We apply a distance-based Bell-test analysis method [E. Knill et al., Phys. Rev. A. 91, 032105 (2015)] to three experimental data sets where conventional analyses failed or required additional assumptions. The first is produced from a new classical source exploiting a "coincidence-time loophole" for which standard analysis falsely shows a Bell violation. The second is from a source previously shown to violate a Bell inequality; the distance-based analysis agrees with the previous results but with fewer assumptions. The third data set does not show a violation with standard analysis despite the high source quality, but is shown to have a strong violation with the distance-based analysis method.
Comment: 7 pages with 6 pages of supplementary information