Digital Libraries currently face the challenge of integrating many different types of research information (e.g. publications, primary data, expert's profiles, institutional profiles, project information etc.), for which to date no general model for knowledge organization and retrieval exists. This causes the problem of structural and semantic heterogeneity due to the wide range of metadata standards, indexing vocabularies and indexing approaches used for different types of information. The research presented focuses on integrating reference data for publications and survey data in the social sciences, but also applies the problems existing in other domains. We present a model for the integrated retrieval of factual and textual data which combines the traditional content indexing methods for publications with the newer, but rarely used ontology-based approaches which seem to be better suited for representing complex information like that contained in survey data. The benefits of our model are (1) easy re-use of available knowledge organisation systems and (2) reduced efforts for domain modelling with ontologies.