Trust and a sense of citizenship are the pillars of thriving democracies. They are essential to the willingness of citizens to make individual sacrifices in pursuit of collective endeavors that drive a society’s success. Unfortunately, that trust and sense of citizenship have been at historical lows worldwide recently and have been declining even more rapidly in Latin America and the Caribbean, in part perhaps, because of social media. What could be a catalyst for transparency and reform in the region, a platform for interpersonal bonding, communication and political news, has, the evidence suggests, become instead a platform where people have been empowered like never before to spread misinformation and insult those with opposing views. Uncivil discourse and polarization have flourished, and trust and the sense of citizenship have declined.
There are few things more important than trust in the political realm. Its decline threatens democracies and economies. Trust in government is fundamental to people’s desire to be politically engaged, to vote, to support existing policies and back institutional reforms that foster long-term, sustainable, and inclusive growth. Mistrust, by contrast, leads to political disengagement. It fosters a preference among politicians to promise quick fixes and impatience among the general public with reforms yielding higher but longer-term benefits.
The Gaps in Previous Research on Political Trust
Little research, however, has been done to examine the relationship between social media exposure, social media engagement, trust and trustworthiness. Is social media really guilty of helping destroy political and social trust as many observers and critics allege? Or is it merely a by-product of existing trends, including disillusionment with governments’ failures in handling economic crises, inequality, and the COVID-19 pandemic, among others? Is social media a catalyst for mistrust? Or is it simply incidental to it?
We decided to find out using a rigorous survey experiment to test whether social media exposure and engagement affect trust and trustworthiness. To do so, we used a variant of the well-known trust game, which usually involves trusting others to invest money that delivers shared benefits. Our goal was to fill the gap in the research about the relationship between these factors and determine whether social media might be really having the baleful influence on trust that so many observers regret.
Designing an Experiment to Test the Impact of Social Media
In our experiment, we take 4,800 respondents in Brazil and Mexico and have them alternate as principals (voters) and agents (politicians) in which the object is to collect as many votes as possible for a preferred candidate among two fictional candidates in an election. For each vote they collect for their candidate, respondents receive raffle tickets to win prizes that they can only win if their candidate wins the election.
Voters can win extra votes by entrusting their votes to politicians to cast them, but pay a heavy cost if those politicians prove untrustworthy and deliver their votes to a different candidate. Critically, after an initial round of the game, a subset of the participants are subjected to negative and positive tweets from incumbent and opposition politicians. Those tweets involve the handling of the COVID-19 crisis, with negative messages assigning blame to opposition parties for sowing conflict and weakening the government’s response and the positive ones expressing the benefit of interparty cooperation. The incumbent politicians are Eduardo Bolsonaro in the case of Brazil, and Marti Bartres in the case of Mexico, while the opposition ones are Fernando Hadad in Brazil and Francisco Calderon in Mexico.
An initial phase of the experiment shows that trust considerably declined between the first and second rounds of the game as the impact of the dissonant messaging from opposition politicians took effect. Indeed in the second round, respondents consistently reduced the number of votes entrusted to other players and retained for themselves a larger share. Interestingly, there was no significant decline in trustworthiness, the willingness to faithfully cast votes for another person, even if that person favored a different candidate.
The Level of Engagement with Social Media Makes a Difference
The experiment also tested the impact of engagement with social media by questioning respondents who responded to tweets by retweeting, liking or replying. The goal here was to see whether this more direct involvement with social media between the first and second rounds would also affect the respondents’ trust behavior. In this phase of the experiment, respondents in the treatment group who engaged with the tweets before the second round were compared to a control group that engaged with them only after the second round.
Here too, the results were revealing, showing a significant increase in distrust among those in the treatment group who actively responded to the tweets as compared to the control group or to those in the treatment group that didn’t engage. This shows how engagement, a crucial feature of the social media era, magnifies the declines in trust. Social media was expected to result in greater transparency, higher accountability and hence, more trust in the political realm. Our results show that this is not the case. While exposure to social media did not affect trustworthiness in our experiment, it significantly affected trust behavior, beyond mere attitudes. Moreover, those effects were magnified when users were engaged and sharing messages with others. That is worrisome indeed given the correlation between high trust, economic growth, social progress and stability.