Few of the depressing statistics on women in science pinpoint hard evidence for bias against individuals (see, for example, Nature 495, 22–24; 2013). So I conducted a small experiment of my own.
We know that successful grant applications are important drivers of promotion and tenure. In my first year as a researcher in 2005, I submitted 16 grant applications under my full name, which is not gender neutral. Just one received funding.
In my second year, I applied using only my initials and my last name: my success rate went up fivefold. This was only an n = 1 experiment, but I didn't care to repeat it.
In my third year, the university adopted an electronic grant-application system, which, unbeknown to me, automatically entered my full first name. In this blinded study, my success rate went down fivefold again, coincident with changing back to a female name on the application.
The following year, I asked the university to modify my entry to use only initials for my first names, thereby frustrating the automated system. My success rate went back up fivefold.
I am the same applicant. The replicates are low, but the outcome apparently differed only when it was obvious to the reviewers that I was female.