RTI International is a research institute headquartered in North Carolina that has created a tool for identifying “evidence of racial disproportionality” in traffic stops. Racial disproportionality is measured by comparing traffic stops throughout what they call the “intertwilight period”.

This is the timeframe in the evening between when the sun begins to set and when it’s finally dark. The basic idea is to compare the race and/or gender of people pulled over for police traffic stops when it is light out at 5 PM versus who gets pulled over when it’s 8 PM when it’s dark.

If a department is racially profiling it would theoretically show up through this method. Officers can see who they’re pulling over easier when it’s light then just an hour or two later when it’s dark, so if the racial disparity shifts at this time then there might be racial profiling going on.

To apply this tool to New Orleans we first need the city’s Field Interview Card database. I isolated just traffic violations that occurred from the start of 2014 through the present and tested to see whether there was racial profiling going on. That provides a dataset with just over 102,000 rows.

The result for each group looks something like this:


I ran black males, black females, white males and white females individually to see if any of the groups saw a statistically significant racial disproportionality in traffic stops. I also ran all four groups with a function in the test which restricts the dates to those which fall 30 days before or after daylight savings time (DST). This is done to ensure “that the periods analyzed have roughly equal amounts of daytime and nighttime.” It also restricts the number of cases analyzed meaning shifts must be bigger to reach statistical significance. Anyhow, this is what was found:

Race/Gender Racial Disparity Light Dark DST Restrict Disparity DST Restrict Light DST Restrict Dark
B/M No 40.28% 41.16% No 40.34% 43.27%
B/F No 26.32% 27.06% No 28.46% 27.43%
W/M Maybe 15.56% 14.19% No 15.79% 13.13%
W/F No 8.35% 8.52% No 6.91% 7.72%

White men are slightly less likely to be pulled over when it’s dark then when it’s light and white women, black men and black women are slightly more likely to be pulled over. But these differences are relatively small and only the white male difference could be evidence of racial disproportionality.

White men and black women are less likely to be pulled over in the dark, and black men and white women are slightly more likely to be pulled over when adjusting for daylight savings time. None of these measures are statistically significant and there is no concrete evidence of racial disproportionality.

The “Veil of Darkness” is interesting though absence of evidence is not inherently evidence of absence. Ultimately this provides another means of seemingly confirming NOPD’s commitment to constitutional policing. It’s another piece of evidence, nothing more, nothing less.