Thread by @societyforepi

Whose ideas get funded and who gets to research is a very important question.

@hwitteman & colleagues ask why gender gaps in funding exist and conclude that women are scored less favorably as scientists even when their ideas score well!

#SERjournalclub
thelancet.com/journals/lance…

This paper makes creative use of the roll-out of a new grant funding mechanism in Canada to try to determine the causes of gender gaps in funding.

This is a good example of a “natural experiment” which we don’t see used very often in epidemiology.

The authors compared grant success rates (measured as “approved” or “not approved” during the peer review component) over time & between the two new post-2014 grant programs to see whether grant mechanism could explain the observed differences in success by gender.

One nice thing about this paper is that it takes a potentially vague causal question “does gender impact grant success?” and turns it into a well-defined causal question: “do reviewers rate female PIs themselves lower or just rate the ideas of female PIs lower (or both)?”

A well-defined causal question helps move us towards action. In their analysis, the authors found that the reduced success of women’s grant applications (at least in Canada) can be explained by peer-reviewers down-rating the women themselves, and not their ideas.

The obvious next step seems to be to change the way grants are evaluated to remove consideration of the submitting researcher, or at least change the way that this is incorporated into the final score.

Specifying well-defined causal questions has been a controversial idea in social epidemiology.

But, hot take, designing or studies without them can imply that problems with diversity & inclusion are the *fault* of the excluded or marginalized group.

What say you, #epitwitter?