OK, it's pretty telling IMO that you've rejected an **actually** testable (in practice, not just in theory) hypothesis for ones that are either impossible or extremely cumbersome to test.
-
-
Replying to @NateSilver538 @BrendanNyhan
Is it possible that it's not a good one? Why would people actively respond *more* to trolling that nevertheless impacted them? How would you ever measure impacts of abusive behavior online? We know: people tend to go silent. There really was a reason we moved beyond Skinner.
5 replies 0 retweets 18 likes -
Replying to @zeynep @BrendanNyhan
OK, then give me some other hypothesis that's *actually testable* and that someone (you or me or a smart graduate student out there) could *actually test* (i.e.. the data is available publicly) in a reasonable length of time.
3 replies 0 retweets 7 likes -
Replying to @NateSilver538 @BrendanNyhan
You'd never have any complex research by that standard. Allow only existing data plus a grad student—preclude all other questions? Here's a reasonable ask: 1-an independent research team; 2-gets access to non-public social media data; 3-combine it with public data; 4-gets a year.
1 reply 0 retweets 15 likes -
A version of this is happening, but 2016 is excluded. Plus, we need a proper longitidunal panel for the future. My point in raising the objection is that there is genuinly doable research, but it's not easy. Framing the question right plus some resources plus, yeah, data access.
2 replies 0 retweets 14 likes -
Replying to @zeynep @BrendanNyhan
Given that low-complexity analysis doesn't support your hypothesis, it's just not a good set of facts for you that you also reject any attempt at *medium*-complexity analysis.
5 replies 0 retweets 6 likes -
Moreover, you still haven't even listed an actual, *high*-complexity hypothesis that could be tested (given some *reasonable* parameters, e.g. a smart graduate student with a year and cooperation from one of the major social media platforms to fulfill additional data requests).
2 replies 0 retweets 6 likes -
Replying to @NateSilver538 @BrendanNyhan
I don't understand why we're limited to one graduate student here when physicists are publishing papers with 5000+ authors BUT. Yes, a reasonable sized team plus past data from social media platforms plus a year or so. You'd have a good start.
2 replies 0 retweets 13 likes -
For example: prevalence of exposure to RU-originated memes to voting (voter records are public). Shift in narrative (sentiment analysis of a true random sample) as related to exposure (obviously confounds with algorithm plus self-selection but I've seen papers that get at this).
2 replies 1 retweet 11 likes -
Social media data may have natural experiments (when Facebook was testing what kind of news to promote: were there comparable groups we can test effects on?). Interaction between liking an RU-page and downstreat impact. Then impact of downgrading of content from liked pages. etc.
1 reply 0 retweets 6 likes
(Lessened likelihood of voting: among left side of the spectrum, obviously). etc. There is no shortage of stuff to look for.
-
-
Replying to @zeynep @BrendanNyhan
That's not where you and I have a difference of opinion. We're arguing about the scale and magnitude of the impacts. I think the magnitude was small, maybe very small. You've claimed it was medium or large. I'm asking you to propose a way to test whether it was small or large.
2 replies 0 retweets 4 likes -
I’m fascinated that anyone claims they can come up with such a test; where is the alternate political universe that we can use as a control?
1 reply 0 retweets 0 likes - Show replies
New conversation -
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.