Just days ago, presidential candidate Bernie Sanders repeated a claim from a highly dubious Washington Post study, that “in 2016, counties hosting a Trump rally saw a 226% spike in hate crimes.” Ilhan Omar has repeated the claim, as have the NAACP, MSNBC’s Joy Reid and Joe Scarborough, among many others.
Politifact gave a generous ruling to the claim in grading it “half-true” because of some caveats they found in the study, such as the fact that “the data can be subject to ‘statistical noise’ [unexplained variability] and jurisdictional differences in hate crime definitions and police aggressiveness, and that cause and effect are hard to pinpoint.”
A bigger problem isn’t that the definitions of hate crimes differ from jurisdictions, its that the Post doesn’t use a database of hate crimes at all in their study. In fact, that Politifact even thought jurisdictional issues would be relevant here calls into question how heavily they scrutinized the study. As the popular Facebook fact checker “Meme Policeman” noticed when he began diving into the numbers, the list of hate crimes that the Post used in their analysis wasn’t a list of hate crimes as traditionally recorded:
The Post examined the counties which hosted 275 Trump campaign rallies in 2016 and used the Anti-Defamation League’s (ADL) “HEAT map” to compile data about hate crimes in the subsequent months. They list literally hundreds of instances where (generally far right) activist groups distributed fliers or other propaganda material as hate crimes. Some of these had slogans which were fairly benign (“America First,” “Defend America,” “Life, Liberty and the Pursuit of Victory,” “Better dead than red”) while others were outright racist or anti-Semitic (“Join the KKK and fight for race and nation,” “European roots, American greatness”).
Their definition of hate crimes included anti-Semitic comments made on Snapchat, and trash talk between 13-year olds at a soccer game. While acts of bigotry, the ADL’s definition of hate crimes is far more liberal than the FBI’s (which includes physical assault, threats, harassment, and vandalism). As I pointed out in a prior article regarding the ADL’s statistical looseness, the ADL had previously made headlines with a study claiming that anti-Semitic hate crimes rose 57% in 2017. The entire rise was attributable to a single individual who was phoning in bomb threats to Synagogues. Without him, anti-Semitic violence actually had decreased by 47% according to the ADL’s own statistics. That’s an enormous amount of variability.
Furthermore, Sanders tried to give the impression that counties which held Trump rallies then saw 226% explosions in hate crimes, but the Post didn’t actually do a “before and after” comparison in their study. Rather, they compared counties that held Trump rallies with “comparable” counties, without ever telling us what a comparable county is. A county of the same size that never held a Trump rally? A county with the same income demographics? The same racial demographics? We’re not told.
Luckily, the Meme Policeman actually did an exhaustive analysis of the FBI’s statistics, and it contradicts the Posts findings:
For this analysis, I took the list of Trump’s presidential campaign rallies. These began in 2015, but only the ones during 2016 up until the election were used, as that’s what the meme references. There were 275 total rallies, but some cities were visited multiple times, giving us 223 total cities/counties to analyze. The hate crimes data was taken from the FBI database from 2015 and 2016 in order to compare hate crimes for each year for every area.
Results: Of the 223 cities and towns Trump had rallies in, over 1/3 (86) reported zero hate crimes for all of 2016. Over half (127) reported one or less. Considering the relatively low frequency of these crimes, statistics for a given area can be wildly misleading. For instance, Tampa, which Trump visited more than any other city, had hate crimes rise by a whopping 300%! But this is because it went from 1 incident in 2015 to 4 in 2016. Meanwhile, hate crimes in Las Vegas, where Trump visited 4 times, plummeted by 40% (48 in 2015 to 29 in 2016). Thus, beware anytime you see hate crimes stats for a local area or even state, as these can fluctuate dramatically.
There were 1,433 hate crimes in all cities and counties that Trump had campaign rallies in during 2015, and that increased 1% to 1,450 in 2016. Population in those cities and counties rose more than 1%, representing a decline in hate crimes committed per-capita.
Or to put this in other terms, hate crimes increased 4.6% in 2016 but only 1% in areas where Trump held rallies.