This paper describes the steps taken in generating and implementing a qualitative research design for Phase II of the Ethnocultural Communities Facing AIDS Study. Theoretically framed by the macro-level, sociocultural model of health behaviour developed by Kleinman, the methodological procedures are an adaptation of Scrimshaw's Rapid Assessment Procedures (RAP) and a participatory approach involving stakeholders from each ethnocultural community. Qualitative data-on behaviours conducive to HIV transmission in six ethnocultural communities in Canada-were elicited using a combination of key communicator interviews, focus groups, and participant observation techniques. Data were analyzed using systematic content analysis techniques. Inter-rater reliability checks and procedures of triangulation demonstrated the validity of evidence generated. A commitment to research partnership with community persons, and an accountability loop that provided assurances of how the data would be scientifically represented, were critical elements in the process of design construction.