Their main results are on the page numbered 20 (page 21 of the PDF) – their data shows political polarisation has grown (over the period 1980-2020) in six OECD countries and declined in six OECD countries. The six countries where it is has grown (in order from greatest to smallest growth) are the US, Switzerland, France, Denmark, Canada, and New Zealand. The six countries where it has declined (in order from greatest to smallest decline) are (West) Germany, Sweden, Norway, Britain, Australia and Japan.
They don't have any direct measures of social media use in their source data; the closest things they have are Internet penetration and consumption of online news, but they found no statistically significant correlation between those and polarisation. The only clearly statistically significant correlation they could find was a positive correlation between polarisation of societal elites and that of the general population (p=0.011). They also found a positive correlation between increasing racial diversity and political polarisation which was of borderline statistical significance (p=0.052).
Thanks, I'd also like to point-out that 40 years is a long time in relation to the more recent FB algo changes that reordered timeline in order to increase engagement.
Their main results are on the page numbered 20 (page 21 of the PDF) – their data shows political polarisation has grown (over the period 1980-2020) in six OECD countries and declined in six OECD countries. The six countries where it is has grown (in order from greatest to smallest growth) are the US, Switzerland, France, Denmark, Canada, and New Zealand. The six countries where it has declined (in order from greatest to smallest decline) are (West) Germany, Sweden, Norway, Britain, Australia and Japan.
They don't have any direct measures of social media use in their source data; the closest things they have are Internet penetration and consumption of online news, but they found no statistically significant correlation between those and polarisation. The only clearly statistically significant correlation they could find was a positive correlation between polarisation of societal elites and that of the general population (p=0.011). They also found a positive correlation between increasing racial diversity and political polarisation which was of borderline statistical significance (p=0.052).