Wednesday, September 05, 2018

Polarisation May Be Caused By Opposing Views Rather Than 'Echo Chambers'

Polarisation has been an issue attracting much attention recently. The RSA has devoted a whole podcast series to the topic. One of the potential explanations put forward for increasing polarisation is the presence of 'echo chambers' on social media. It was therefore interesting to read recent research which found the reverse to be the case, namely that exposure to opposing views on social media can increase political polarisation.

The latest research by Christopher Ball et al was based on a large field experiment to test three hypotheses:
  1. that contact with opposing groups and exposure to their views can increase deliberation and compromise
  2. that exposure to those with opposing political views may create a backfire effect and exacerbate political polarisation as users double down on their own beliefs
  3. whether this backfire effect is more likely to occur among conservatives than liberals, based on recent studies indicating conservatives hold values that prioritise certainty and tradition, whereas liberals value change and diversity


The research found no evidence that exposing Twitter users to opposing views reduces political polarisation but the study did reveal backfire effects.

The study found that Republicans who followed a liberal Twitter bot for a month became substantially more conservative. Democrats also became slightly more liberal after following a conservative Twitter bot, although these effects were not statistically significant for Democrats.

These findings suggest that trying to use Twitter to change people's perspectives by exposing them to opposing political views could be ineffective and worse, actually counterproductive.

Methodology and limitations

The study recruited self-identified Republicans and Democrats who visited Twitter at least three times a week . They were offered $11 to complete a survey which 1,652 people completed. They were then offered the opportunity to follow a Twitter bot. 64.9% of Democrats and 57.2% of Republicans accepted and were paid $6 each week for correctly answering questions about the content of messages retweeted by the Twitter bot. They were offered a further $12 to repeat the survey at the end of the study. The participants followed either a liberal or conservative Twitter bot that retweeted messages randomly sampled from those with opposing political ideologies.

Participants completed a survey in mid October 2017 at the start of the study and again one and a half months later. These surveys were designed to measure any change in political ideology during the study period.

The authors are at pains to point out their findings are not evidence that exposure to opposing political views will increase polarisation in all settings. Twitter users are not representative of all Americans and there may be different effects in other countries and cultures. Users were also paid to read messages that they might normally ignore. The authors say "the most important limitation of our study is that we were unable to identify the precise mechanism that created the backfire effect among Republican respondents."

You can read the full study here.