A bridge over troubled water: Interdisciplinarity, Novelty, and Impact

This paper analyses the relationship between novelty, interdisciplinarity and impact of scientific research. We discuss the limitations of the established indicators of novelty and interdisciplinarity (Variety, Disparity and Balance) and how they overlap. We correct the biases with normalized indicators that exploit a Configuration Null Model and we propose a protocol to verify the reliability of the indicators and standardize the measurement. The proposed methodology and new indicators are illustrated exploiting 231450 articles in Physics, their references and citations, from 8 Journals of the American Physical Society (APS) from 1985 to 2015. We show that Novelty à la Uzzi et al. (2013) is highly correlated with Disparity and Novelty à la Wang et al. (2017) is deeply affected by the properties of the articles’ citation network. We show the importance of taking into account the specific definition of the knowledge space and considering separately the different dimensions of interdisciplinarity. Using the new unbiased indicators we show that Variety has a positive effect on the articles’ Impact while Disparity and Balance have a negative effect. Previous results tend to underestimate the negative Impact of Disparity and overestimate the positive effects of Novelty.