Prejudice against sexual and gender minorities (e.g., LGBT people) is quite prevalent and is harmful. We examined an existing—and often-used—contact intervention in pre-existing groups in an educational setting and assessed its effectiveness in reducing different forms of LGBT negativity. We focused particularly on modern LGBT negativity: a relatively subtle form of prejudice, involving ambivalence, denial, and/or the belief that there is too much attention for LGBT prejudice. We used a mixed design in which condition (experimental vs. control group) was the between-participants factor, which was randomized at the group level, and time (pretest vs. posttest vs. follow-up) was the within-participants factor (N = 117). Interventions were video recorded and the behavior of LGBT educators and participants was coded. Participants responded positively to the intervention, especially to the LGBT educator’s “coming-out story.” Exploratory analysis of the video data indicated that the perceived effectiveness of the intervention was higher in groups where participants were more engaged, although caution is necessary in interpreting this finding. The most important measure indicated that modern LGBT negativity decreased in the intervention groups directly after the intervention, but returned to baseline levels one week later. However, in the control condition, modern LGBT negativity had increased over time. Taken together, this suggests that an actual reduction in modern LGBT negativity was short-lived (i.e., the intervention effect disappeared within 7 days).