Individual admissions officers have biases, but many of these are washed out when one considers them in the aggregate (one doesn't know ahead of time which admissions officers will be reading one's application), though Gwern points out evidence of systemic bias on at least one dimension.
I would guess a lot of important biases don't wash out in this way. For one thing, some biases are pretty much universal. For another, it is likely that most admissions officers, particularly in elite schools, generally belong to the elite liberal sub-culture. So if you do extracurriculars which go against that subculture, one can expect to be at a disadvantage, all things being equal.
I agree with this. But applicants are often from the elite liberal sub-culture, too, and if one restricts consideration amongst the activities that they would plausible engage in, there will be less systematic bias.
[Edit: The post below gives the impression that our conversations with admissions officers are our only reasons for believing the claims. We've also consulted with other sources such as How to Be a High School Superstar: A Revolutionary Plan to Get into College by Standing Out (Without Burning Out) which corroborate the admissions officers' remarks]
We spoke with admissions officers at Harvard, Yale, University of Chicago, Columbia, Stanford, MIT, Duke, University of Pennsylvania, Dartmouth, Williams, Johns Hopkins, Swarthmore, Brown, Northwestern and Caltech, about how they evaluate student participation in extracurricular activities, for 15 colleges total. Some things that we found based on college's statements are below.
Kawoomba suggests that colleges' statements on the first point below can't be taken at face value. What do you think?
Cross-posted from the Cognito Mentoring blog
See also High school extracurricular activities: factors to consider and College statements about extracurricular activities