Published on EUDO Café – 29 August 2014
In less than two months the citizens of Scotland are going to decide on the future of their country in a clear-cut referendum. If Yes votes win the majority, Scotland will take the road for independence and might soon become the 29th member of the European Union. According to the polls on whatscotlandthinks.org, the No camp has steadily gathered more supporters. However, the gap between the two sides is fairly small, and reaching the crucial 50 percent still seems a feasible goal for the supporters of independence.
Clearly, those who will manage to bring the undecided voters to their camp will win the battle over independence. Depending on the polls, for the last three months the share of undecided voters has remained substantial, ranging from 10 to 28 percent. Thus, the question is how to convince these voters to choose either the “Yes-Scotland” or the “Better-together” side. What type of information should they be provided for? And what would be the effect of this information on their attitudes?
Motivated by these questions, Céline Colombo and I conducted a lab-experiment at the BLUE Lab of the University of Edinburgh between April and May 2014. The experiment, which was made possible thanks to the financial support of the EUI SPS Department and the EUI Presidency, involved around 300 subjects (almost entirely university students) for one week. Each subject had to fill in a survey in addition to reading different arguments both in favour and against independence under different conditions. The information provided came only from real sources – such as newspapers, official documents and policy papers – and was presented in a completely ‘naked’ format – i.e. without mentioning any politicians, political parties or organizations of any kind.
Part of the results will be published in a forthcoming EUDO working paper, but some preliminary findings can already be anticipated. First, provision of information clearly reduces indecision, especially when subjects have the possibility to select information. In this case, the percentage of undecided or not-willing-to-vote subjects decreases by half (from 27 to 14 percent) compared to a control group in which subjects don’t read any information material before stating their voting intentions.
Second, provision of information increases the likelihood to vote Yes by 11 to 15 percentage points, depending on the group. In this case, the effect is larger when subjects have to read a balanced set of arguments – i.e. four pro and four against independence – suggesting that being exposed to substantial information supporting independence reduces the indecision related to this option and makes citizens more willing to vote Yes. It is worth stressing that in all groups subjects were provided with a balanced set of texts supporting both sides of the referendum campaign and that all the arguments were rated almost as equally strong.
Finally, our results suggest that information increases the likelihood to vote Yes especially when it meets a ‘fertile soil’. In particular, those who were born in the UK (and especially in Scotland) and those who have been living in Scotland for more than two years all become more likely to support independence after reading the same set of arguments. Interestingly, information has a stronger effect also among those who dislike to take risks, since reading convincing arguments makes a Yes vote a less risky option and independence doesn’t appear as much as a ‘leap into the unknown’.
An interesting study, but I think it leaves a rather important question open: “is it more information or the illusion of more information that makes the difference?” Let me illustrate the issue. Suppose there is some subject matter about which a test subject is ignorant, but which is relevant in some manner to the test subject. Said subject matter has a clear and provable state of truth or falsity, but the test subject is ignorant of this. Now, I develop several convincing arguments for the truth and for the falsity of said subject matter. What would your study conclude when performing this experiment? Given that you have ground truth about the subject matter, what would this tell you about people’s belief systems?