‘SORT BY RELEVANCE’ – WHOSE RELEVANCE? A CRITICAL EXAMINATION OF ALGORITHM-MEDIATED ACADEMIC LITERATURE SEARCHES

Abstract

The academic literature has arguably never been more accessible. Through the internet, academics - and the wider public - are presented with an increasing array of platforms which offer sources of academic literature and information sources. Large-scale platforms – such as Google Scholar – often utilise algorithms in order to manage how search query results are prioritised and presented to searchers. This ‘sorting by relevance’ introduces an opaque layer to how readers engage with the literature, with potentially important implications for academic rigour and equity. Academic publishing and citation practices often serve to preserve privilege; in turn, there is a risk that algorithms which draw upon data sources such as these will compound biases. There is a need to interrogate how such algorithms work, and academics’ assumptions about the mechanisms behind them. This paper starts a critical discussion of the issue of algorithm-mediated academic literature searches. First, it draws upon the existing literature – primarily related to Google Scholar. This is followed by a mapping of the use of algorithms across a sample of major academic databases. Algorithms are found to be used extensively, although how they operate varies and is often not clear. The paper concludes with next steps for this project.