Thread by @AnechoicMedia_

The analysis in question, written up in the NYT by notorious liar Max Fisher, is based on secret, unverifiable data from a researcher who won't share his methods or allow replication.

https://twitter.com/axios/status/1157995851693715456

Here's the NYT author @axios is citing here -- look at this blatant, unretracted lie he wrote in a previous gun column:

https://twitter.com/AnechoicMedia_/status/963980126877900800?s=20

@axios Here's an illustration of that bait-and-switch NYT writer Max Fisher used:

"Gun deaths" are correlated with guns because of higher suicide rates. Homicide alone does *not* correlate.

Fisher cited a handful of such "gun death" metrics in a paragraph explicitly about "murder".

@axios The research @axios is mentioning here, reported by that same Mr. Fisher, comes from Adam Lankford, who claims to have assembled a proprietary dataset containing an authoritative sample of every mass shooting in every country (absurd on its face). He refuses to share this data.

@axios Lankford's key claim is that mass shootings in America aren't a cultural oddity of ours, but rather the predictable result of our having more guns, which he says correlates with more mass shootings everywhere.

The only glimpse we get of his secret dataset is this scatterplot:

@axios Now, I don't find that plot informative at all, but just to be sure, I hand-traced every single data point in it to run my own statistical tests.

Contrary to Lankford's claims, the correlation doesn't hold when using simple statistical checks for outliers (rank order, logscale)

@axios Similarly, removing the US and Yemen from the correlation makes all results non-significant. This is highly suggestive of these countries having unique circumstances, rather than guns being predictive of shootings among all countries generally.

@axios Here's what the guns/shooting relationship looks like among all countries in both rank order and logscale plots. There's no significant correlation in either case -- if guns caused shootings, these plots would easily show it.

@axios @axios is citing a guns and shooting claim from a researcher who refused to answer questions about his methods or hand over data to other researchers.

In the only plot he actually shared, his key claim falls apart under basic statistical scrutiny.

@axios For what it's worth, a random reddit stats teacher says my analysis is correct:

@axios Here's a full explanation of the weak, suspicious claim Max Fisher and Adam Lankford put in the Times, which is being cited by @axios