The spread of information on social media shapes everyday decisions. As customers, we decide what to buy, as patients, we decide which doctors to consult, and as voters, we decide which choices to make at the ballot box partly based on the information we see on social media. Yet, the facts and opinions that we are presented with are a result of a complex web of factors. Driven by psychology, we make choices–which social media to sign up for and check regularly, who to follow or become friends with–that affect what we see. Driven by business interests, platforms employ filtering algorithms that gauge which content we are likely to respond positively to. This results in a steady mix of new content. Consciously or subconsciously, we then have to take the decision whether to ‘like’it, forward it to friends, or interact with it in some other way. Whether we actually take the advice we see to heart, and believe the facts we are presented, is yet another question.

Against this backdrop, it is essential to understand which content spreads on social media, and thus affects social media users’ decisions. The spread of content such as news stories on social media is also known as information diffusion (cf. Gruhl, Liben-Nowell, Guha, & Tomkins, 2004; Kim, Bae, & Hastak, 2018). In particular, understanding the patterns and mechanisms of information diffusion allows us to explain, predict and perhaps even counter the negative consequences of the spread of misinformation. One way in which researchers approach the topic of information diffusion is by studying real-life social media interactions. This is known as social media analytics, and it is emerging as a distinct research area that is grounded in a variety of disciplines such as computer science, information systems and psychology (Stieglitz et al. 2014). It is characterised by a distinct methodological approach, which is different, for example, from the laboratory experiments and survey research methods employed by psychologists, and it comes with its own advantages and disadvantages, opportunities and challenges. Vast amounts of unstructured data are difficult to analyse. Attempts to structure the data, to quantify text data, for example, may prove unreliable.