Sexual harassment is difficult to define. A study was undertaken to examine four variables related to sexual harassment: (1) sexual harassment when behaviors are unwelcome and sexual; (2) unwelcome nonsexual behavior called nonsexual aggravation; (3) welcome sexual behavior which includes organizationally dysfunctional relations and nondysfunctional ones; and (4) welcome nonsexual behavior including social support. Participants (N=308) were employees of a military organization including officers, enlisted soldiers, and civilian workers. Since women were considered the sex likely to be harassed, they were oversampled. Surveys measured sex, age, marital status, race, supervisory status, pay grade, percentage of males in work unit, group membership, incidence of harassment, and type of harassment. The results showed that incidence of sexual harassment was associated with age, marital status, and sex. Enlisted personnel were more likely to be harassed than were civilians or officers. Nonsexual aggravation was related to the incidence and severity of harassment. The most frequent harassment was sexist humor. Managers of organizations might monitor nonsexual aggravation as a less controversial indicator of sexual harassment than common measures. (ABL)