How do Citizen Preferences Shape Politician Behaviour?

The question of whether and how politicians attend to and act on citizen preferences is of central importance to both social science and the practical manifestation of public policy. This paper deals with two related questions: First, do politicians pay attention to public opinion? Second, when they are exposed to public opinion what actions do they take in response? To answer these questions, we are currently conducting the largest ever field experiment with Members of Parliament. In short, we are sharing accurate, objective public opinion with 338 Members of the Canadian House of Commons over a four year period and observing how they react to receiving this information. By randomizing the provision of public opinion data, we can make definitive causal statements about the link between public preferences and the behaviour of MPs. Our study relies on two sources of information about public opinion: • The Local Parliament Project – this is the largest ever public opinion study conducted in Canada. A representative sample of ~40,000 Canadians shared their opinions and the importance they place on approximately 40 political issues. This is supplemented by a study of more than 75,000 additional voters for a smaller number of policy items. The result of this is precise public opinion data in all of Canada’s federal electoral constituencies; • The 42nd Election Twitter Project – teaming up with the Social Media and Political Participation Lab at New York University, we have collected every election related tweet sent during the 42nd Canadian federal election (more than 5 million). The basic logic of our study is that information on public opinion will be shared with randomly-selected subsets of MPs. There are two broad questions our experiment allows us to cleanly answer. First, we intend to vary the presentation of information to MPs. Sometimes, MPs will be told the opinions of all constituents. Other times, opinions will be broken out by partisanship. Sometimes MPs will receive data framed in a positive fashion. Other times, it will be given an oppositional frame. By varying the presentation of data and observing the behavior of receiving MPs and their offices – behaviours such as clicking through on monitored links, asking for more information, writing or calling experts whose names are included in the briefs – we can begin to estimate what features of public opinion MPs are most likely to attend to. Second, by comparing the various representative actions that MPs take – making a statement in the House, including information on their policy position in householder leaflets, attending constituency meetings, proposing private members motions or bills, or voting on government bills – we can understand how public opinion influences their actual representation of voters. This phase of our project is launching in January, 2016 and thus we will only have preliminary data in time for the workshop. However, we will have these data for one dozen experiments by that point in time. Accordingly, we will be able to share our important insights into the links between public opinion and the actions of Members of Parliament. Are we excluding the PM and speaker? Who else? Let's just say 338 for now.