Role: Principal investigator

Funding body: Society for Research in Higher Education (SRHE)

Award value: £10,000

Duration: February to December 2022

‘Sort by relevance’: Exploring assumptions about algorithm-mediated academic literature searches

Project Aims and Research Questions

The Internet has revolutionised access to information. It has arguably never been easier to access academic literature and research evidence, and a range of online scholarly databases and academic social networking sites provide platforms to search the literature. 

However, there are reasons to be cautious about the range of available sources. For example, different databases vary in terms of their coverage and the sources they include; most are run as commercial enterprises, which may bring hidden priorities for promoting particular content. As a result, relying on a particular database as a source may bring only a partial view of a research field. 

The use of algorithms – even if intended to aid the user, by providing what it calculates to be the most ‘relevant’ material – can further obscure exactly why particular literature has been included in search results. Following in the wake of its market dominance as a search engine, Google Scholar is an extremely popular way of searching the academic literature:

“Google Scholar aims to rank documents the way researchers do, weighing the full text of each document, where it was published, who it was written by, as well as how often and how recently it has been cited in other scholarly literature.” (Google Scholar, 2021).

This has two potential risks associated with it. First, that by drawing upon citations and favouring certain publications, it will serve to amplify the inequalities present in scholarly publishing. Second, it carries methodological risks when carrying out literature reviews, as it is not clear exactly why a particular article has deemed to be of high relevance in search results.

These issues were recently the focus of an article in the Times Higher Education (Matthews, 2021). However, the article also highlighted the lack of research on how academics use these platforms and navigate these risks in practice. The proposed study intends to be the first to fill this gap.

Through a mixed methods research design, the study will address the following research questions:

  • What are academics’ assumptions about how algorithm-mediated literature searches (such as Google Scholar) work?
  • Is there any evidence – through the choice of sources, or presentation of different search results– to suggest that there is a risk of ‘filter bubbles’ being formed?
  • Does the perceived quality of results vary by platform, and what are the practical implications of this?

 

References

Google Scholar (2021) About Google Scholar. https://scholar.google.com/intl/en/scholar/about.html

Matthews, D. (2021) Will a Facebook-style news feed aid discovery or destroy serendipity? Times Higher Education. https://www.timeshighereducation.com/features/will-facebook-style-news-feed-aid-discovery-or-destroy-serendipity

%d bloggers like this: