Background. Epidemiologic studies are dependent on simple and valid questions to assess the prevalence and type of urinary incontinence. Objective. To examine the reproducibility and validity of two standardized questions seeking to identify stress and urge incontinence among elderly women. Methods. A random sample of 421 women 75 years or older living in the center of Odense, Denmark, were invited to participate, and 223 (53%) accepted a first interview in their home. A sample of 154 women were selected for a second interview with the same questions [accepted by 144 (94%)]. The second visit also included a long open interview about incontinence and was followed by a prospective registration of leakages. Results. In the first interview 39% (95% confidence interval (CI) 34-47%) of the elderly women reported incontinence. The reliability of the questions was acceptable [kappa of 0.81 (95% CI 0.34-0.89)] with percent agreement of 90% (95% CI 84-95%) between first and second interviews for all incontinence (stress and/or urge). When the open interview was used as a "gold standard," the questions showed acceptable validity: a sensitivity of 0.91 (95% CI 0.86-0.96) and a specificity of 0.86 (95% CI 0.80-0.92) for all incontinence. Conclusion. Standardized questions about incontinence give reproducible answers and produce information that is comparable to a long open interview.