4 Policy Brief No. 44 Toda Peace Institute identify gaps in scholarship, help inform policy decisions, and translate academic knowledge for a broad audience of scholars, journalists, and interested citizens.” This research is further complicated by data availability. Again, this is why most studies use Twitter data — it is easy to get — but Twitter cannot necessarily stand in for other social platforms, and certainly not for non-social platforms. Plus, research fashions play a role. There is far less research (or concern) about the impact of Reddit or Amazon than Facebook, Google, or Twitter, even though more people currently visit Amazon or Reddit than Twitter (note that Amazon started demonstrating the ability to remotely delete books from users’ Kindles over ten years ago, and journalists at Vice have documented how Amazon actively experiments with which books they allow users to see). Finally, as we look at digital propaganda, false news, voter influence, and other methods of persuading people that use digital media, it is worth noting that there is a field of perspective change in social psychology, and one of its findings is that people’s perspectives change differently on different topics. Many factors affect decision-making, such as cognition effects (like whether someone is paying attention or not) and partisan effects (like whether a person is highly aligned with one side of the debate or not when they are exposed to new information). There can be strong differences between groups (for example, partisan attitudes may become more entrenched when exposed to different perspectives, while moderate partisan attitudes may not). As with all fields of human endeavour, these data may appear different and lead to different decisions when viewed at different scales. Altogether, these are significant limitations! Still, researchers have managed to learn some things about polarisation and other societal impacts of digital media. 2. Basic Research on Polarisation Suggests that Digital Media is not Driving Polarisation (But Polarisation Is Really High, and Getting Higher) The claim that polarisation is rising is uncontroversial; in fact, polarisation has been increasing for decades. With regards to polarization as a broad trend, the 2018 study “How Polarized are Citizens? Measuring Ideology from the Ground-Up” by Draca and Schwarz is a good place to start. The study documents polarisation in 17 countries across the European Union and North America. It notes that polarisation has increased more in the US — and is currently at a higher level in the US — than in any other countries studied. However, the fact that polarisation is rising does not necessarily mean that digital media is driving it. In research that takes a more granular look at the phenomenon, there is a range of approaches for measuring polarisation. In a 2017 study titled “Greater Internet use is not associated with faster growth in political polarisation among US demographic groups,” Boxell, Gentzkow, and Shapiro looked at eight proposed measures and gave references to previous discussions of the “appropriate definition of polarization and the extent of its increase.” Their study results are summarised by the title — i.e., they did not find any indication that polarisation correlates with internet use. Those authors also found that “demographic groups least likely to obtain information online