The below comments have me wondering if “Steve” and “Jane” get more engagement on the forum that “Ahmed” and “Fatima”. I mean, this is an empirical question.
Thoughts?
The below comments have me wondering if “Steve” and “Jane” get more engagement on the forum that “Ahmed” and “Fatima”. I mean, this is an empirical question.
Thoughts?
I think a bigger problem than names is the enforcement of tone and style. Kaleem points this out in the screenshot.
In my experience the forum is too harsh on posts that contain minor grammatical errors or deviate from the typical EA tone, characterized by dispassionate, clinical language, and the use of rationalist in-group jargon.
This acts as an extra barrier to entry for anyone that does not fit the typical EA mold, which disproportionately includes people from non-native English countries.
yeah i feel this. i post very off the cuff stuff and people definitely don’t like it
To me it comes across as disrespectful of people’s time
posting off the cuff posts is disrespectful of peoples time?
I at least think it’s important to distinguish:
People who make minor errors because they don’t speak English as a first language or otherwise find it difficult to avoid minor errors; from
People who make minor errors because they write things quickly and off the cuff and can’t be bothered to put the effort into making the work tighter.
Whether or not 2 is a problem, I at least don’t want to blur these two things together and don’t think a greater sympathy for 1 should necessarily imply a greater tolerance of 2.
(“Off the cuff” can often mean low quality posts that are harder to engage with because not clearly written. These posts can decrease the time cost to the writer but increase it for the reader, especially if we focus on time cost per unit of value received. I don’t think it’s crazy to think there’s something in this that is disrespectful to the reader.)
Interesting question—I doubt this would be the case, but agree it can be at least partially answered empirically.
Perhaps the forum team could try and do some kind of name analysis vs karma to rule this out? This would be difficult, because you would have to somehow control for “quality of post/comment” which would be hard but possible. For example you could get a team of people to rate post quality blinded to the poster name, then compare the karma of similarly rated posts with different names.
Alternatively you could do a prospective study, perhaps putting up similar-ish types of posts or quicktakes written by the same person/people under randomly assigned pseudonyms. This could be a bit disengenuous and bad actor-ish though as you would be kind of experimenting with us all without our consent, but hey that’s what big tech does all the time.
I’m sure there are other ways of assessing this I am missing too.
What if you could only see who wrote a post after you’ve rated the comment?