Biomedical data collections are typically compiled on the basis of assessment instruments and associated terminologies and their data structure explained by means of data dictionaries. The Information Artifact Ontology (IAO) is an attempt to give a realism-based account of the essence of information entities and how components of such entities relate to each other and to that what they are information about. Changes in the taxonomy and the definitions of the IAO, most importantly the addition of the terms ‘representational artifact’ and ‘representational unit’, are proposed to make the IAO a useful tool to clarify formally the distinctions and commonalities between data collections and associated artifacts that are compiled independently from each other, yet cover the same domain.