Does information increase opinion quality? Differential information effects on the quality of opinions towards defence depending on issue attributes

The capacity of citizens to voice an opinion on foreign policy and military issues is a much-debated topic. The minimalist view – otherwise known as the “Almond-Lippmann consensus” – states that citizens lack both interest in it and knowledge about it; consequently, their opinions are unstructured, incoherent and unstable; and for this reason, public opinion should not be relevant for policy-making. Whereas research in the US has shown that public opinion on defence is more rational, stable and consistent than what the “Almond-Lippman consensus” suggested, in Europe, little attention has been paid to the way citizens form their opinions on those issues, monopolized by experts and the top of the executive, and for which a certain left-right consensus exists. As it has been demonstrated that the quality of opinions – that is, their stability and consistency – depends on how well informed the people are, this paper contributes to bridging this gap by focusing on the effects of providing information on opinions towards defence issues. To better understand the effect of (the lack of) information on people’s opinion on these matters, we conduct a series of four experiments in France, the UK and Germany. These three countries offer different combinations of military power, the professionalisation of their armed forces, the visibility of the military, their participation in multilateral operations, and the approval of missions by their national democratic institutions. Each experiment tests the effect of providing information on respondents’ answers, for a given topic. We compare the answers of respondents randomly assigned to the group that receives information to those of respondents randomly assigned to the group that doesn’t. We also compare the results between the experiments, that is between the different issues, and over time (two waves experiment). Each issue has been selected depending on its attributes, and more specifically how much debated it is in at least one of the three countries (based on debates in the Parliament) and how salient it is (measured in terms of visibility in the media): arms exportation, participation in NATO, the fight against terrorism abroad, and nuclear deterrence. We hypothesize that the more respondents know about a specific defence issue, the less the information we provide should influence their opinions, and the more stable their opinions should be over time. Reversely, respondents who know less should be more sensitive to information effects and express fewer stable opinions. However, for them, we expect that providing information increases the stability of their opinions on a given topic. Therefore, the effect of providing information should matter at three levels. Individually, the most educated and the most interested in defence issues should be less sensitive to information effects. At the issue level, information about the most salient ones should have less effect on respondents’ answers than information on the least salient issues. Finally, at the country level, issue salience varies, and therefore, information effects should also be less important among respondents of the country in which a given issue is the most debated and salient.