Totally you should run experimental design by Twitter before doing it, else everyone will critique it if they don't like the results! Happy to RT the design to get input.
-
-
Replying to @DrAnneCarpenter @Erysiphales and
Two options I can think of: 1) Relabel everyone as Person A, B, C and ask "who is more emotional" (or angry/aggressive) 2) Swap identities randomly (so sometimes men are me etc) and ask "is the last comment more emotional than the others in the thread?" (or angry/aggressive)
1 reply 0 retweets 1 like -
Replying to @DrAnneCarpenter @Erysiphales and
I suspect (2) might be better - with 1 I think I'd be labeled 'worst' (even if minutely so) regardless of whether ppl know I'm female or not. Not really getting at the Q. (p.s. in (1) you'd also run a parallel test where people DO keep their identities)
1 reply 0 retweets 1 like -
Replying to @DrAnneCarpenter @Erysiphales and
Hmm. What about presenting a bunch of the replies, and for each you get to choose from a list of adjectives to describe the reply. I think any ranking yours might end up 'worst' simply because it's the most pointed. Whilst not emotional, it *elicits* an emotional response...
1 reply 0 retweets 1 like -
Replying to @DrDeCaluwe @DrAnneCarpenter and
Just so you know, collecting data on MTurk if you have never done it before is a much bigger time commitments than a single evening. You need JS, PHP, and SQL usually and at least.
2 replies 0 retweets 2 likes -
Replying to @o_guest @DrDeCaluwe and
Wut? I wasn’t thinking this would be an actual published study. If you just throw some screenshots of the tweets together, it should take less than a day
1 reply 0 retweets 0 likes -
Replying to @Psycholosopher @DrDeCaluwe and
The languages I mentioned are just used to collect the data.
1 reply 0 retweets 1 like -
Replying to @o_guest @Psycholosopher and
I'm sure nobody means to but to assume what we do to (I'm lucky enough to not have to do it myself) to collect data in psych is easy is kind of offensive.
2 replies 0 retweets 2 likes -
Replying to @o_guest @DrDeCaluwe and
I am a psychologist. It really is easy to collect data. There are free survey platforms that can handle this. No programming needed
1 reply 0 retweets 0 likes -
Replying to @Psycholosopher @DrDeCaluwe and
The screenshots won't make good stimuli since they are going to induce context collapse IMHO.
1 reply 0 retweets 0 likes
I guess it all depends what point what you are testing. But if you want to show somebody (a single person, not a population level inference) is sexist you need more than a few Tweet screenshots to chuck to some Mturkers.
-
-
Replying to @o_guest @Psycholosopher and
Also the fact that you want to run a test to check if somebody is sexist makes me think you have a strange prior for how likely somebody (a white dude in science) is sexist? You can use Occam's razor here instead of run some essentially pointless experiment.
2 replies 0 retweets 3 likes -
Replying to @o_guest @Psycholosopher and
Another issue is wiping the names off might not be enough to hide gender. As gender might be predictable from content of tweet: http://u.cs.biu.ac.il/~koppel/papers/male-female-text-final.pdf …
0 replies 0 retweets 0 likes
End of conversation
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.