Title text: I understand large sample sizes are key to a low standard error of the mean, but the entire sophomore class?
College (undergraduate university, for international readers) is often a young adult's first experience with independent living. As a result, it is a time to "find yourself", so the aphorism goes. This often includes a certain amount of sexual experimentation, including same-sex experimentation.
This comic applies a scientific interpretation to a colloquial expression (a favorite on xkcd). Usually, "lesbian experimentation" refers to a woman engaging in one-night stands with other women to satisfy her curiosity about the experience and discover her own sexual orientation. Megan, however, does the "experimenting" according to the scientific method: hypothesize, experiment, draw conclusions, repeat. She formulates a hypothesis about her sexual orientation (apparently something along the lines of "I am a lesbian"), which she proceeds to test by experiments (sexual encounters) with female partners as the experimental group and male partners as the control group. She used large sample groups and multiple experiments (the reference to "control group B" implies at least two separate experiments). The title text implies that Megan "experimented" with the entire sophomore class (both males and females) before dating Cueball.