First, I am surprised that a serious study would settle for self-assessment, the least accurate of all possible polls (for example, it's hard to tell if the liberal-identified participants exaggerated their bias due to their pangs of liberal guilt or understated it, as usually happens). Second, I am even more surprised by the high degree of identification with such a broad movement, like it's all Berkeley out there. Surely many of them disagree with at least some of the ideas proclaimed by other "liberals". I guess they are not up to date on their Paul Graham studies:

there is a step beyond thinking of yourself as x but tolerating y: not even to consider yourself an x. The more labels you have for yourself, the dumber they make you.