Social and behavioural

A lens onto fake news

Simon Knight considers epistemic cognition.

28 March 2017

Every day we face complex situations in which the information we need, and who we trust to provide that information, has a very real impact on our lives. How do we evaluate the competing claims of politicians on climate change policy, or Brexit; navigate medical information regarding vaccinating our children; or assess the relative merits of diet versus regular foods in adopting a healthy lifestyle?

In all of these everyday cases, we need to be able to make decisions about where to look for information, which information to select, and whether or not we should check the claims we found in one place against other sources. If we are trying to find out whether or not to take a medical supplement (such as a statin, or vitamin), we might come across a range of public-health, medical-research, and alternative-health sources, all varying in quality and claims. Whether we are experts or not, these sources must be navigated, making decisions about which sites to trust. Worse yet, the sources may well not agree with one another, even where we think they’re equally relatively trustworthy; so, to make a decision, these competing claims have to be integrated or synthesised to try and evaluate why they seem to disagree. Experts call the thinking that goes into these processes ‘epistemic cognition’.

Increasingly, there is a concern regarding the prominence of claims of dubious credibility that are made online along with people’s (particularly young people’s) abilities to make credibility judgements, with a recent spate of articles citing work from Wineburg’s group at Stanford University claiming that US high-school students cannot distinguish fake from real news. The general concern is twofold: first, that the internet has made it easier to spread information while simultaneously stripping that information of important contextual features (such as provenance), and second, that students ‘nowadays’ are particularly susceptible to faulty beliefs in their dealings with information in the world. So how do we evaluate claims found online? How do we figure out how these claims are connected? And how important is our own experience in doing this, rather than simply ‘taking it’ from sources?

A line of research has characterised these questions in terms of epistemic cognition. Could it be the key to addressing today’s challenges of navigating a multimedia landscape and the threat of fake news?

A lens onto information processing

A typical way that epistemic cognition research has explored these issues is by asking students to investigate a scientific or socio-scientific issue, such as climate change, vaccine acceptance, or the safety of mobile phones. Often, they’re asked to write a report and then rank the quality of the resources they used, which might include newspaper reports, textbooks, advocacy or lobbyist reports, and so on.

How does this all relate to fake news? Well, the suggestion is that people do not generally attend to source features such as the author, genre, publication date, venue, and so on, or corroborate claims across reliable sources. As a result, they are not able to bring together claims to form a coherent picture of an issue. This outcome is fundamental with respect to fake news, and results not only in mistrust around particular claims, but also in a general lack of trust in resources or a belief in false equivalence of resources.

Imagine, then, that you are a student, asked to research the efficacy of some food supplement; how might you proceed? Likely your first step would be to open a browser and head to Google (it’s highly likely to be Google you head to rather than one of the other search engines or specialist portals). Unfortunately this is your first pitfall as a student: as Hargittai et al.’s (2010) early work indicated, high-school students are likely to focus on search rank and superficially credible domain names (.gov., .edu, .org) rather than opening web pages and looking for the author’s credentials. In fact, even if search-interfaces are experimentally manipulated to swap ‘high’ and ‘low’ rank sites, students still trust the higher ranked sites more (Salmerón et al., 2012, 2013). That means that a search engine placing a poorer quality ‘alternative health’ website above a higher quality lay health advice site might lead to greater trust in the higher ranked site.

So what if, rather than sending students off ‘cold’, we prime them with some source articles that hold different perspectives, and ask them to investigate those views further? This is a task many of us would find familiar in our own professional and personal lives. Hsu and colleagues (2013) did just that, asking people to read two competing news articles about the risk of electromagnetic waves, do their own research, and then report which view they trusted more and why. Although the authors didn’t report which article was better, they did find that students who aligned more with justifications based on corroborating claims across sources, and using experts to support claims, had tended to engage in search strategies that indicated more sophisticated self-checking – or meta-cognitive – behaviours: for example, viewing more than one page of search results; navigating back and forth between search and pages; and conducting multiple queries based around the same language to refine the queries to find new information. So, when we look at how people approach evaluating and integrating complex sources, we need to be aware of the heuristics they might use – such as search rank and domain name – and the processes they engage in to find the information.

Do people use source features in evaluating sources?

In many cases though, we don’t need to find information ourselves; a friend or family member shares an interesting article on social media, or you hear about interesting research from a colleague, or a family member delves into a medical pamphlet looking for advice. In those cases it’s really important that people are capable of discriminating between a better and a worse source so they are not misled by poor information.

On that front, Kobayashi et al. (2014) provide more heartening results than recent coverage of ‘fake news’ might indicate. They asked participants to engage with two competing sources about the relationship between blood-type and personality. To test how participants used source features to form their opinion on the issue, one set of participants read a professionally authored introductory psychology textbook and an amateur ‘personal website’ with opposing explanations; a second set of participants read the same explanations but with the source credibility features (psychology textbook versus personal website) reversed. Participants tended to reconcile conflicts in information by adopting the view of the higher credibility source, although troublingly, they did not refer to features such as the credibility of the author in their justifications for why they held that perspective. This highlights the importance of explicit guidance on how to use source information to make judgements on sources of varying credibility.

To guide people on the effective use of sources and source-credibility features, psychologists need to understand a bit more about how people evaluate and justify their claims, and what that tells us about how they will behave when exposed to particular kinds of sources. One way of understanding this issue is by looking at how people use various justification strategies. How do people make use of the authoritativeness of a source and/or its methods; corroborate information from across multiple sources; and draw on personal opinion and direct experience? A body of work from Oslo University by Bråten, Strømsø, Ferguson, and Anmarkrud has investigated these questions. For example, they conducted studies asking students to read six conflicting documents on cell-phone radiation risks, finding that those who justified claims by checking across multiple sources had better comprehension of the sources, created better arguments, and were more likely to explicitly refer to the sources.

This checking behaviour matters. Imagine (as Kammerer, Amann, and Gerjets, 2015, instructed participants to) that a friend has asked you for your advice on a health issue that you don’t know much about. Would you try and find a single reliable source, would the sources have particular characteristics, and how would you deal with conflicting advice? If you are the kind of person who believes in justification by multiple sources, you might – like the participants in that study – be inclined to spend more time on ‘objective’ (institutional) websites and less time on subjective (forum) or commercial (health company) websites.

Happily, for psychologists interested in understanding how people think through these issues, both my own work on children’s small group dialogue (Knight & Mercer, 2014, 2016) and think-aloud data (reported by Mason et al., 2010; 2011) from web searching indicates that students do in fact engage – to varying degrees – in spontaneous reflection regarding both source credibility and selection issues, with those focusing solely on source selection performing worse. This body of research highlights the need for interventions to promote adaptive epistemic cognition.

Prompting adaptive epistemic cognition: Back to the Classroom

If you want to work out if a politician has said something that is factually inaccurate, you must navigate a range of sources that may differ in their claims. Sadly, you might simply fail to reconcile these different claims. But even if you do manage to reconcile the differences, you might do so by focusing on surface features such as domain name or appearance over relevant features for the task such as political leaning or verification methods. This problem might be expounded by only focusing on sources that appear in the top few results your search engine returned, and on your failure to explore and corroborate across sources, or use query refinement to explore the issue in more depth. A tendency to focus on one’s own opinion and experience compounds these issues, leading to less focus on authoritative sources and their corroboration.

Understanding how to address these issues is vital for modern day literacy practices. The recent Handbook of Epistemic Cognition offers some timely suggestions for how we might adopt this psychological research in educational settings. They flag the importance of learning contexts that open up explicit discussion around how we come to know things. Specifically they flag the importance of:

1. Encouraging students to consider their evidence, and its quality, against the particular claims or hypotheses being asserted;
2. Engaging students in identifying others’, and constructing their own, arguments;
3. Multiple perspectives (and the evidence on which they are founded) as a means to open discussion.

These are relatively simple heuristics for planning educational interventions. However, putting them into practical action requires considering some key questions: what information are we asking students to process? What perspectives on that information might they encounter? How prepared are they to deal with competing information? And are they equipped to process relevant source features on the issues they’re addressing?

Simple interventions – such as asking students how they would confirm their knowledge – do encourage them to engage with more sources, and to critically evaluate them (Porsch & Bromme, 2010). However, we need more research to understand how we can support those with poor practices to develop their skills.

Andreas Schleicher, the OECD education lead in charge of the international Pisa rankings, recently called for students to be taught how to spot fake news. Such tuition holds important outcomes if informed by psychology, including research in epistemic cognition. On encountering sources, students should (as Sam Wineburg would put it) ‘think like a historian’ and search laterally (about the source itself) rather than vertically (within the source itself). They should know which source features are important, and how to find multiple credible sources by engaging in sophisticated search practices. Perhaps most importantly, they should simply have a critical eye for evaluating the claims that they encounter, understanding the importance of evidence in making arguments.

Ultimately, psychology can offer insight into these issues, leading to practical implications for teaching and learning. Technologies have opened wonderful opportunities for learning, and technologies may also play a role in countering some of the worst impacts of the changing information landscape. Psychology, though, can guide both how we think about how the internet has changed our multimedia landscape, and how we support students in their interactions with this landscape.

- Simon Knight is a lecturer in the Connected Intelligence Centre at the University of Technology Sydney. He teaches on the centre’s Masters in Data Science and Innovation, and undergraduate module on quantitative literacy and argumentation. Simon researches epistemic cognition in information seeking and processing contexts, and is interested in developing social accounts of epistemic cognition (see Knight and Littleton, 2017), and research methods based on analysis of digital trace (learning analytics).


The International Society for the Learning Sciences created a number of video resources for an excellent introduction to epistemic cognition.

Bråten, I., Anmarkrud, Ø., Brandmo, C., & Strømsø, H. I. (2014). Developing and testing a model of direct and indirect relationships between individual differences, processing, and multiple-text comprehension. Learning and Instruction, 30, 9–24.
Bråten, I., Ferguson, L. E., Strømsø, H. I., & Anmarkrud, Ø. (2013). Justification beliefs and multiple-documents comprehension. European Journal of Psychology of Education, 28(3), 879–902.
Bråten, I., Ferguson, L. E., Strømsø, H. I., & Anmarkrud, Ø. (2014). Students working with multiple conflicting documents on a scientific issue: Relations between epistemic cognition while reading and sourcing and argumentation in essays. British Journal of Educational Psychology, 84(1), 58–85.
Hargittai, E. (2010). Digital Na(t)ives? Variation in Internet Skills and Uses among Members of the “Net Generation.” Sociological Inquiry, 80, 92–113.
Hsu, C.-Y., Tsai, M.-J., Hou, H.-T., & Tsai, C.-C. (2013). Epistemic Beliefs, Online Search Strategies, and Behavioral Patterns While Exploring Socioscientific Issues. Journal of Science Education and Technology, 1–10.
Kammerer, Y., Amann, D. G., & Gerjets, P. (2015). When adults without university education search the Internet for health information: The roles of Internet-specific epistemic beliefs and a source evaluation intervention. Computers in Human Behavior, 48, 297–309.
Knight, S., & Littleton, K. (2017). Socialising Epistemic Cognition. Educational Research Review, 21, 17-32.
Knight, S., & Mercer, N. (2016). Collaborative epistemic discourse in classroom information-seeking tasks. Technology, Pedagogy and Education, 1-18. 
Knight, S., & Mercer, N. (2015). The role of exploratory talk in classroom search engine tasks. Technology, Pedagogy and Education, 24(3), 303-319. 
Kobayashi, K. (2014). Students’ consideration of source information during the reading of multiple texts and its effect on intertextual conflict resolution. Instructional Science, 42(2), 183–205.
Mason, L., Ariasi, N., & Boldrin, A. (2011). Epistemic beliefs in action: Spontaneous reflections about knowledge and knowing during online information searching and their influence on learning. Learning and Instruction, 21(1), 137–151.
Mason, L., Boldrin, A., & Ariasi, N. (2010). Searching the Web to learn about a controversial topic: are students epistemically active? Instructional Science, 38(6), 607–633.
Porsch, T., & Bromme, R. (2010). Effects of epistemological sensitization on source choices. Instructional Science, 39(6), 805–819.
Salmerón, L., & Kammerer, Y. (2012). Selecting Pages from Google to Learn about a Controversial Topic: The Role of Epistemic Beliefs. Staging Knowledge and Experience: How to Take Advantage of Representational Technologies in Education and Training?, 181.
Salmerón, L., Kammerer, Y., & García-Carrión, P. (2013). Searching the Web for conflicting topics: Page and user factors. Computers in Human Behavior, 29(6), 2161–2171.