This field experiment study will evaluate an existing commercially produced online sexual harassment training program in an educational setting and will examine the effects of three instructional strategies on knowledge and behavioral identifications in sexual harassment training for college students. College students completed a written knowledge-based pretest and then received sexual harassment training via one of three content-similar treatment groups: online instruction, face-to-face instruction, or by reading a pamphlet. After training, an equivalent posttest was taken to assess knowledge gain; after watching video vignettes, participants identified sexually harassing behaviors. Three weeks later a follow-up posttest was given to assess knowledge retention. Training did not produce an immediate gain in knowledge scores regardless of training condition. However, reading and face-to-face training conditions predicted the correct answer of case-related questions on the posttest; reading and online training conditions predicted knowledge retention three weeks after the training. From the videos, participants correctly identified 54% of verbal sexual harassment cues; participants over identified 19% of verbal cues as sexual harassment. Participants correctly identified 30% of nonverbal sexual harassment cues; participants over identified 16% of nonverbal cues as sexual harassment. Correct identification of nonverbal sexual harassment cues was higher for females, those with prior sexual harassment training, and those in the face-to-face condition. Implications for future training sessions and future research directions are also discussed.