That's a good link -- and to me it highlights something different: selection bias. Not of the people in the experiment, but of the people designing the experiment.
Instead of looking at it as "this person's a feminist, they're going to be biased to feminist results," look at it as "people who think to ask questions in this way tend to get this set of results, repeatedly. This will likely lead to them accepting the associated ideology." So instead of the studies proving the pre-conceived notions of the experimenters, what we could be seeing is the experiments selecting the appropriate experimenters. Since someone is unlikely to widely vary their methodology from one study to the next, they are likely to replicate the same "bias" purely because they are the same person going about things the same way.
To really break this cycle, you need to add some randomness from some outside force, such that a single person or group of people does not control the entire methodology of the study. Even if they are using methods to avoid bias, they are likely to always use the same methods, and so always get "affirming" results. In this, the single reviewer was correct, even though his assumptions of WHY he was correct are likely way off.
And yes, this line of thought completely affirms your comment about male vs female being incredibly stupid. If there's selection bias based on methodology, you're going to find men and women coming down on both sides -- there might be some clustering based on social norms of men vs. women, but that's a really fuzzy boundary at the best of times.