,
8 tweets,
2 min read
Read on Twitter
If the 84 applicants were ~40% women (a reasonable stab in the dark for this career stage) and women were equally likely as men to be funded, then the Fisher’s Exact test result for the observed scenario is p=.04.
I don’t know if this review process was fair or unfair. Small numbers (n=10) mean it’s easy to have unbalanced results just through chance. But I’d be curious to know more about what measures were put in place to ensure fairness.
We have robust evidence that women tend to do less well than men when the reviewers focus largely on the scientist (versus largely the science). I’d be curious to know how much weight “excellence” had here.
In the meantime, you know where we don’t see these kinds of imbalances in success rates? CIHR project grants, which have 100k bridge funds *in addition to* full grants with overall success rates higher than 10/84=12% *and* equalized success rates for ECIs.
To fund the best, most rigorously peer-reviewed science in the public interest, our tax dollars should be prioritized for the arms-length Tri-Council funding agencies (NSERC, CIHR, SSHRC) plus CFI. cc @KirstyDuncanMP
Re-threading the thread for easier reading
Some useful detail from @BrainCanada:
Re-running Fisher Exact with 35% (29/84) instead of 40% (34/84) gives p=.15.
