I've concluded that I don't have enough data to be sure but wouldn't be surprised if it's a mixture of all three - testosterone levels correlating with assertiveness/aggressiveness seems sufficient to suggest biology may play a part, and I've definitely seen the latter two happen.

I don't, however, think it's contradictory to say that "women are less likely to interrupt back" does not automatically mean that the person interrupting them is therefore sexist, even though the results of the behaviour end up impacting a higher percentage of women than men - they may simply be oblivious to the consequences, even if from the sounds of it in this case that wasn't it.