Thread by @SlickRockWeb

Besides the acknowledged narrow scope of the new NYU #disinfo study one of the biggest issues IMHO with the study was the utilization of voter survey data. The issues targeted by the Kremlin was much more nuanced than voter survey data available 2 the authors.. AFTER THE FACT 1/4

This was no fault of the authors given they had to do a retrospective study. The authors did find subtle shifts (but not statistically significant) and if you looked at the questions that would have been most relevant to the Kremlin ... these pro-Kremlin shifts were largest. 2/4

Important nuanced questions that would not likely have been in most voter survey data collected prior to the 2016 election & appeared to be important in the targeting by the Kremlin would have been below (I'm sure there are others). Chinese tariffs & Obamacare... not so much 3/4

With 20/20 hindsight the perfect study would have been 2 look at some of the more nuanced questions & how the sentiment / attitudes shifted after exposure to Russian disinformation (Facebook + Twitter). And look at differences based on Sex (M/F) & based on race (white / POC) 4/4

Its important 2 note that this study ONLY looked at the effects of #Twitter on a subset of votes based on voter survey data they could access. The effects of the Kremlin hack & dump and Kremlin #disinfo on Facebook was not looked at. Heres our prior thread

Lastly with this study design, any shifts in voting behaviors towards or away from Trump were always going 2 be minimal given the fact that 4 this dataset the Twitter accounts with the highest concentration of exposure to Russian #disinfo was among strongly Republican #RussiaGate

My good friend @ushadrons made another good pt. One segment that was most definitely targeted by the Russian IRA trolls would not have had much if any presence in the YouGov voter survey data that the authors paired with 2 do this study. I'd be very curious of their age breakdown