Descriptions of the steps in the development of the SHI include the development of the content, determination of the psychometric properties, and the development of severity weights. The SHI was found to meet some of the fundamental criteria of a good, standardized instrument; its content domain is comprehensive as judged by experts; its internal consistency reliability is excellent; and it has evidence of factorial validity. By asking about sexual harassment in behavioral terms, the SHI avoids "unacceptable disease bias" (Sackett, 1979). For epidemiological research, the SHI offers an additional enhancement, in that the behavior measured may be severity- weighted, thus allowing investigators to explore dose-response relations and threshold effects between sexual harassment and various health outcomes, or, conversely, dose-response relations and threshold effects between putative risk factors and sexual harassment. The SHI has approximate interval properties, and as such, may be used either as a continuous outcome variable or as an explanatory variable in multivariable regression modeling. 3 tables and 47 references