To determine the effect of demographic variables on attitudes toward disability. Participants: Eighty-three female American and 89 female Taiwanese college students. Outcome Measure: A conjoint measurement of 16 stimulus cards (representing people with varying disability labels, severity of disability, age, gender, and education). Procedures: The participants were asked to sort the stimulus cards according to their personal preferences for working with people with disabilities. Results: Younger and higher educated women with milder disabilities were preferred by both Taiwanese and American students. Preference formation is affected by both disability-related variables and other demographic variables (e.g., educational levels). Conclusion: The use of conjoint analysis to examine multiple attributes of persons with disabilities may have higher external validity than single-attribute-design studies.