A replication of the study finding Trump rallies associated with increased hate crimes finds an even stronger association with Clinton rallies. These apparent effects go to near zero once you account for county population
Clinton rallies cause a bigger "increase" in hate crimes than Trump rallies do, per someone who replicated the original study
An important update on the claim that Trump rallies increased hate crimes by 226%: Harvard economist found that controlling for population size reduces the effect to zero.
File this as another entry in glaring errors in social research getting through institutional gatekeepers who know better and achieving wide acclaim in the media -- with no one noticing the obvious flaw -- because it confirms their preferred narratives
When a study about Trump and hate crimes sounds too good (or bad) to be true, maybe it is.
The paper reported on by everyone before peer-review process claimed Trump rallies correlate w/226% increase in hate incidents. The methodology & analysis was flawed. Same methodology applied to Hillary rallies show even greater increase of hate incidents.
As someone who declared skepticism of the "Trump rallies increased hate crimes two hundred percent!" story, I am pleased to see it is indeed a case of remarkably bad statistics
Where do (Trump) rallies happen? Counties with higher population. This makes the original paper's errors seem egregious and irresponsible.
for every media screw up we see, remember there is doing something pretty important almost every day. this piece especially
In fact, by sticking with the underlying, flawed methodology, they found that Hillary Clinton rallies produced an even larger hate crime spike than Trump rallies. (In actuality, they did not, because the methodology is flawed for both.)
"Trump rallies linked to 226% jump in hate crimes", claimed academic study, shared across media -hadn't gone through peer review -replication study finds no effect -applies dodgy method to Clinton rallies & finds even greater increase ht
This finding was debunked by Harvard PhD students. Study did not consider relevance of county size. Hillary Clinton rallies correlates with an even bigger hate spike, according to this flawed approach.
There appears to be A LOT of action around this post, which is being cited as if it were an actual study. I feel like the real study here involves watching political polarization around the dissemination of quasi-facts.
There is definitely a desire to publish clickbait based on people's fears. For some reason, people often would rather read things are getting worse than better. But the truth is even more important than getting some clicks
That 226% figure was absurd on its face, and never should have been reported.