In this article the authors review methodological issues that arise when interviews and self-report questionnaires are used with people with mental retardation and offer suggestions for overcoming some of the difficulties described. Examples are drawn from studies that use qualitative methodology, quantitative studies assessing different question types, and studies reporting on the development of instruments measuring psychiatric symptoms, self-concept, and quality of life. Specific problems that arise with respect to item content (e.g., quantitative judgments, generalizations), question phrasing (e.g., modifiers), response format (e.g., acquiescence, multiple-choice questions), and psychometric properties (factor structure and validity) are discussed. It is argued that because many self-report questionnaires include questions that have been found to be problematic in this population, more attention needs to be paid to establishing the validity of such measures and to clearly defining the population for which the instrument is designed. The importance of obtaining information from people with mental retardation 1 is increasingly stressed by researchers, clinicians, and those involved in service development. This is a result of the limitations of informant reports for many psychological phenomena (such as subjective states), as well as a change in the philosophy of services toward user consultation and a greater emphasis on people with mental retardation making decisions about their own lives. Structured or semistructured interviews are used in the formulation stage of many psychological therapies, in the assessment of abilities, IQ, and neuropsychology, as well as in forensic, research, and individual planning contexts. Verbal selfreport questionnaires are used in clinical assessment (e.g., to assess psychopathology, self-esteem, personality), research, and service evaluation. However, difficulties in conducting interviews and using self-report questionnaires with this population are widely reported. Such difficulties have been described in psychiatric and psychological assessments (e.g.