What is - really - in a dataset?

AIM - The article considers methodological problems of working with register data and shows how deficiencies in a quantitative dataset may constitute a tool for discovery if data shortcomings are used as input in a qualitative investigation of data genealogy. DATA - Based on a specific research case example the article demonstrates how qualitative moments are intrinsically embedded both in quantitative datasets and in the statistical processing. CONCLUSIONS - It is argued that statistical analyses of datasets cannot be made without considerations of institutional organisation and perceptions of persons implied in the production of data. Lastly it is suggested that prison-based drug misuse treatment research currently needs to face the paradoxical challenge of excessive statistical programmatic power, which has encouraged formulations of ‘causalist’ research questions. It is further suggested that a reorientation towards theoretical explanation as an addition to the statistical demonstration of factor associations is important if quantitative studies are to further open the black box of treatment.