Author Topic: #CHAZ  (Read 9407 times)


  • Hero Member
  • *****
  • Posts: 4913
« Reply #180 on: July 07, 2020, 10:09:55 PM »
This was one such post (you may need to read the post above it for full context):!/msg310835/#msg310835

But IIRC there was a really long thread about this topic a year or so ago where for a few pages some posters went back-and-forth using various datasets to draw their respective conclusions. Like I said I remember compiling WaPo data and overlayed with FBI reporting data, then used a few city/state datasets to factor out environmental variables. If you're really interested, that would be the way to reperform the analysis to see what factors are impacting the overall result and from there to draw your own conclusion.

Additionally I took a look at one such study referenced (

The authors conclude significant racial bias in any-use-of-force encounters; but lack of bias in shootings (as distinct from any lethal engagements). The study admittedly suffers from data quality issues. 1- Data suitability (self reported data also used by FBI through 2015, the year in which the FBI admitted data quality problems and determined to re-define their DQ standards) and 2- Data collection (was hand-tabulated by grad students albeit with some 2nd layer control, but again self-reported from POs themselves). Finally, the main dataset used to conclude no racial bias in shootings (Houston data Table2c - was not available in the paper and frankly I am not digging around online to find it)

This is not to dismiss the conclusions outright, but to understand we have (1) a narrow definition of police killings i.e. shootings only; and (2) compromises made to determine suitable datasets which I personally would disagree with.

This was 1/4 studies mentioned - as seen here the choices particularly in how you build your dataset are very important, hence my earlier suggestion to seek out these sources yourself, understand each source's reporting methodology, and perform your own analysis. This way you at least understand the implicit compromises which must be made to draw any conclusion.
« Last Edit: July 07, 2020, 10:45:40 PM by LC »
"Lethargy bordering on sloth remains the cornerstone of our investment style."
akam| brk.b | goog | irm | lyv | net | nlsn | pm | ssd | t | tfsl | v | wfc | xom