Creating Indicators for the Output of Interdisciplinary Research

The United States Government and other entities such as the Organization for Economic Cooperation and Development (OECD) publish reports of science and engineering indicators.  The increasing attention paid to interdisciplinary research suggests the need to address this new area of interest by identifying and characterizing the interdisciplinary content within the total output of research in terms of indicators. The search for relevant measures of the interdisciplinarity of research output needed to create indicators. In order to select for detailed consideration several metrics of interdisciplinary research output that appear most promising for eventual use as indicators, the following criteria are suggested: 1) The indicators selected for inclusion must be based on existing data and not require the collection of new data; 2) The proposed metrics should be readily understandable by non-experts, i.e., educated readers who may have a notion of interdisciplinary research but little or no knowledge of the techniques for measuring it. 3) Metrics should be appropriate to a broad range of contexts in S&E rather than applicable to narrow areas alone. 4) Metrics should be sufficiently well-developed in the research literature to permit meaningful assessment. This paper focuses on selected literature of 20 articles relevant to the measurement of the output of interdisciplinarity research (IDR) - exclusive of questions of quality, impact, or scope, which are related but tangential to the question of how to measure output, and exclusive of interdisciplinary approaches to education or medicine which are not directly related to research. Five different approaches are considered and compared as they would contribute to indicators: 1) co-authorships; 2) citations; 3) entropy; 4) diversity; and 5) betweenness centrality. Recommendations are made for IDR indicators that could be created at the national or inter-governmental levels.