The findings of artist research seem to depend to an extraordinary extent on the operationalisation of the object and the methods of data collection and analysis. This article considers the usability of census data in the follow-up of the living and working conditions of artists for arts policy purposes. In Finland, census data are no longer gathered by questionnaires but are derived direct from official registers, which causes particular problems in the case of artists who in many respects differ from the standard salary earner. The data tend to reflect current legislation and administrative practices rather than people's subjective conceptions. Some census variables prove to be useless for research purposes. The census data nonetheless offer unique opportunities for horizontal and vertical comparisons. To serve the needs of artist researchers the data have to be partially de- and reconstructed; e.g., the identification of artists may be improved by reclassifying the original uncoded occupational titles coming from the various registers. The production of occupation data is so laborious that the census cannot meet the urgent needs of policymakers. Methodological triangulation, supplementing census data with targeted questionnaires and interviews, is suggested for the monitoring of the status of artists.