The study involved a quantitative research technique called “difference in differences,” which mimics a controlled experiment with observational data. Difference-in-differences studies use variation in the timing and location of a “treatment variable” (such as BLM protests or police killings) to sort data into artificial control groups and treatment groups; researchers can then compare an event’s apparent effects in different settings or time periods. The new study compared police killings in cities that experienced BLM protests with those that did not.
“It’s really difficult to measure Black Lives Matter protests and lethal use of force by the police,” notes the study’s author, Travis Campbell, an economist at the University of Massachusetts Amherst. This is mainly because comprehensive data on police killings are lacking: the federal government does not track police officers’ lethal use of force,
ie they made it all up.