With just a couple of days to go and opinion polls showing a strongly polarised electorate in Scotland, the only certainty is that the result of the independence referendum will be extremely tight. In such a context, delivering convincing messages can persuade a crucial handful of voters to take a stance and lead one side or another to a victory. This claim is supported by an experimental study that Céline Colombo and I carried out at BLUE Lab at the University of Edinburgh.
Convincing arguments from both sides reduced indecision about how to vote, but also increased the likelihood of a voter supporting Scottish independence. This increase was particularly apparent among those who described themselves as more politically engaged
Given that the study was conducted in May, these findings might have predicted the rise in Yes-vote intentions over the last few weeks. The question, however, is why such an asymmetrical effect occurred, if we took extreme care to provide all the participants with a balanced set of equally strong arguments from both sides?
The interpretation lies in the higher level of uncertainty related to the choice of independence compared to keeping the status quo. Since provision of information reduces indecision, reading convincing arguments has a stronger effect on the Yes side, because it contributes to reducing the uncertainties of independence to an ‘acceptable’ level. In a sense, providing information means the Yes option becomes less a ‘leap of faith’ and a more considered choice.
The study involved around 300 eligible voters who had to fill in a survey, in addition to reading different texts under different conditions. In one condition, voters were presented with a list of 16 texts – eight in favour of independence and eight against – but could select and read only eight. In another group everyone had to read the same balanced set of eight arguments supporting both sides of the debate. The texts provided were created after a detailed content analysis of newspapers, official documents and policy papers, and were presented in the most ‘neutral’ way, without mentioning media sources or the names of politicians or political parties. All these adjustments aimed to induce the subjects to focus as much as possible on the substance of the information provided.
The media and deliberative democracy
The preliminary results of the study, which have just been published by the European Union Democracy Observatory in a working paper, confirm that referendum campaigners’ messages do affect voters’ opinions. This is good news for academic research, which traditionally has struggled to demonstrate the effects of media exposure on political and social behaviour, but it also implies that media have a strong responsibility in the referendum debate. Recent positions such as Rupert Murdoch’s support for independence, The Scotsman favouring the union, or the allegedly non-neutral role of the BBC can all play an important role in shifting the electorate towards one side or another.
This research contributes to ongoing debates on deliberative democracy and provides further implications for the use of deliberative settings as a means to reach more considered decisions. In line with previous research in political psychology, our findings confirm that voters tend not only to select evidence in line with their prior opinions, but also to see evidence consistent with their prior opinions as stronger. These mechanisms, however, apply especially when voters have already made up their minds. On the contrary, as the level of indecision increases, voters tend to be more even-handed both in the selection and the evaluation of arguments from both sides.
In practical terms, this suggests that referendum campaigners should not waste time trying to convince supporters from the other side of the fence, but, instead, should focus on those who still have doubts about how to vote. Since, according to the polls, a relevant share of the electorate is still undecided, both camps still have a chance to attract new voters and they should keep on campaigning intensely until the very last moment. More generally, our findings highlight media’s responsibility to provide information as balanced and as accurate as possible. In a context in which each single vote can make the difference, instead of taking part in the campaign, media should limit their role to giving citizens the widest range of arguments necessary to cast a considered vote.