Best of #econtwitter - Job market paper RCT

This paper is doubly navel-gazing (non-derogatory!) in that it is (1) about the profession itself + (2) about twitter itself. Combined with the fact that it intervenes on ~the job market~… the perfect recipe for lots of attention:

“Twitter is real life” indeed. The tweets were usually things like this:

What did the paper find?

First, did these quote tweets increase twitter views and likes? Yes:

Did this increase interviews or flyouts? Here’s the data:

^additional interviews, flyouts, and offers are all indistinguishable from zero. The point estimates are roughly positive +1 interview, flyout, and offer (these are marginally significant if you add in the right controls) — which would be huge

Finally, “women in the treatment group received 0.9 more job offers than women in the control group (p = 0.024)”

Altogether, if you want to be maximally sassy about it,

That said,

Experiments are hard

Here are a few points, thank you @paulgp:

The timing is also a little confusing. The authors write, “The lack of a statistically significant treatment effect at the interview stage could be due to the timing of our intervention”

but it seems odd to claim to affect offers without affecting interviews, given how these things work

Ethics

^the randomistas make other good points in the replies. Here’s a more coherent concern:

^for what it’s worth, most of the tweets were so dry and leaden that I think a lot of people figured it was an experiment (there was discussion of this at the time)

^“RTs are not endorsement” IRL,,,

Anyway, Tyler Cowen has the best synthesis:

Here’s one of the authors for the last word:

Fin

^🔥