The conservation of biodiversity is a major goal in nature conservation, but measuring the total biodiversity of a site or a region is not possible; thus there is a great demand for indicators to represent biodiversity. To be able to make use of indicators, criteria must first be established for their selection, and the degree to which the indicators meet the criteria must be tested. However, the purposes for which indicators are appliedand thus sometimes the criteria themselves-differ between ecological science and environmental policy. As transparency in choosing and testing suitable biodiversity indicators will optimize the results of an indicator, this article first aims to determine if there are common approaches in selecting biodiversity indicators in ecology and environmental policy. Second, we asked which criteria biodiversity indicators were scientifically tested against to determine their suitability. To answer these questions, we analyzed papers on biodiversity indicators referenced in the Web of Science. Our results demonstrate different patterns for selecting biodiversity indicators in the different fields of application. In ecology, the quality of indicators is mainly determined by a close relationship between indicator and indicandum (i.e., indicated phenomenon), while the relevance of an indicator for a given issue, e.g., reserve selection or an assessment of a certain impact, is of paramount importance for conservation policy. Surprisingly, few biodiversity indicators are empirically tested to determine if they meet the criteria by which they were purportedly chosen. We argue that this is due to the different conceptualizations of biodiversity indicators in science and environmental policy. Since the suitability of biodiversity indicators remains untested in many cases, our findings suggest room to make better use of indicators in ecology and environmental policy. As the results of ecological research are put to use to solve environmental problems, the selection of indicators for ecological research should correspond to a large extent with those used in environmental policy. Further, to assess the suitability of a biodiversity indicator, it should be tested against all of the criteria relevant for its selection.