The increase in qualitative research in family medicine raises a demand for critical discussions about design, methods and conclusions. This article shows how scientific claims for truthful findings and neutrality can be assessed. Established concepts such as validity, reliability, objectivity and generalization cannot be used in qualitative research. Alternative criteria for scientific rigour, initially introduced by Lincoln and Guba, are presented: credibility, dependability, confirmability and transferability. These criteria have been applied to a research project, a qualitative study with in-depth interviews with female patients suffering from chronic pain in the locomotor system. The interview data were analysed on the basis of grounded theory. The proposed indicators for scientific rigour were shown to be useful when applied to the research project. Several examples are given. Difficulties in the use of the alternative criteria are also discussed.