Early detection and excision of thin lesions may be important in reducing mortality from melanoma. Periodic skin self-examination may be beneficial in identifying thin lesions. The purpose of this study was to evaluate factors associated with skin self-examination. The study population was comprised of 549 Caucasian residents of Connecticut 18 years of age or older who were selected as controls as part of a population-based case-control study on skin self-examination and melanoma conducted during 1987-1989. Personal interviews were conducted to obtain information on skin self-examination, demographics, history of cancer, phenotypic characteristics, sun exposure habits, and screening and health behaviors. Nevus counts were performed by trained nurse interviewers. Logistic regression was used to model the relationship between the variables of interest and skin self-examination. Female gender was identified a priori as a predictor of skin self-examination, and thus all analyses were stratified by gender. Age, education, and marital status were also identified a priori as important predictor variables and were selected for inclusion in the final models. Skin awareness was a strong factor associated with skin self-examination for both females and males. For females, previous benign biopsy or the presence of an abnormal mole was identified as important for future skin self-examination using our criteria. A family history of cancer, physician examination, and change in diet to reduce cancer risk increased the likelihood of skin self-examination in males but not females. In women, light hair color may increase the likelihood of performing skin self-examination. Older age and college or postgraduate education was associated with a decreased likelihood of performing skin self-examination in both males and females. Identifying factors associated with skin self-examination will enable health care providers to target individuals who may not be performing skin self-examination but who are at increased risk for developing melanoma.