The authors recognize this in their FIRST paper: http://faculty.washington.edu/kstarbi/Stewart_Starbird_Drawing_the_Lines_of_Contention-final.pdf … "It is not necessarily representative of the broader *LM discourse. Specifically, our sample has likely biased our analysis towards discursive acts that connect *LM narratives with violence."
-
-
Replying to @schock @katestarbird
To emphasize this point: look at Figure 1 in the first paper (p 96:6), it clearly shows that the vast majority of tweets in the set are from the Dallas and Baton Rouge shootings of police.
#BlackLivesMatter
.1 reply 1 retweet 9 likes -
Replying to @schock @katestarbird
So methodologically, IF YOU KNEW THAT SAMPLING THIS WAY WOULD LIMIT YOU TO THE (RIGHT WING) FRAME OF BLM AS VIOLENT COP KILLERS, WHY DID YOU SAMPLE THIS WAY... And then make claims about the broader discursive battle?
#BlackLivesMatter
2 replies 4 retweets 48 likes -
Replying to @schock @katestarbird
The extremely limiting approach to gathering the corpus of tweets is why this study's key finding should be: "When police officers are shot, russian bots try to convince people Black Lives Matter is involved."
#BlackLivesMatter
6 replies 22 retweets 80 likes -
Replying to @schock
We should chat. I agree to a large extent here w/ the limitations you note here. We are still analyzing this data and working on the correct framing. And I think it will be important for other research to look at these efforts over time and across other data.
1 reply 5 retweets 30 likes -
Replying to @katestarbird @schock
Our evidence shows trolls trying to infiltrate these conversations. In most cases, they fail. But for
@Crystal1Johnson (left) &@TEN_GOP (right), they were quite effective. We have evidence these accounts were integrated into the following networks on their respective sides.5 replies 14 retweets 18 likes -
Replying to @katestarbird @schock and
This is a tiny piece of a much larger phenomenon. Our claims are not that the
#BlackLivesMatter
movement was at all compromised by these efforts, but that the online conversations around BLM (and people’s perceptions of BLM) are being influenced by trolls.2 replies 11 retweets 28 likes -
Replying to @katestarbird @schock and
Thank you for your comments. We will take a deep look at both the specifics and the concerns they represent (which I share) and take those into account as we move this currently preliminary work towards future publication.
3 replies 4 retweets 32 likes -
Replying to @katestarbird @schock and
This objection misses the point of much of this fakery. These accounts are screenshotted as “evidence” of what BLM represents to groups already suspicious or hostile. They aren’t influential within their respective movements usually—expect when they are like the “TEN GOP” one.
6 replies 2 retweets 20 likes -
Replying to @zeynep @katestarbird and
I agree with this point, but it's the original paper that misses it.
1 reply 1 retweet 4 likes
Much of our data analytics is text-based and siloed “Internet vs. cable” while the real story is in the whole ecology and is mostly playing out visually, not verbally.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.