STUDIES:   

138/

For those who don’t want to read the 136 papers above, @dean_c_knox, @conjugateprior, @jonmummolo

Intuition for the problem: Suppose perfect as-if experimental conditions: a set of police encounters (i.e. sightings) identical but for race of civilians. Suppose racial bias leads police to stop white civilians if engaged in serious crime (e.g. bank robbery), stop black civilians regardless of behavior. Comparing white bank robbers to black civilians committing no crime. If we then found no disparity in rates of force against black/white civilians, that should be alarming! Intuition for the problem: Suppose perfect as-if experimental conditions: a set of police encounters (i.e. sightings) identical but for race of civilians. Suppose racial bias leads police to stop white civilians if engaged in serious crime (e.g. bank robbery), stop black civilians regardless of behavior. Comparing white bank robbers to black civilians committing no crime. If we then found no disparity in rates of force against black/white civilians, that should be alarming! Prior work ignoring this feature substantially underestimates racial bias in use of force (Fryer, 2019) Prior work ignoring this feature substantially underestimates racial bias in use of force (Fryer, 2019)
Administrative Records Mask Racially Biased Policing
presentation slides
Knox, Dean / Lowe, Will / Mummolo, Jonathan
6 March 2020
https://scholar.princeton.edu/sites/default/files/jmummolo/files/klm_slides_2020.pdf
Slides for study 139/
Thoughts?
Comment on this study on its Twitter thread.