The Citizens’ Assembly on Brexit – a group of citizens that will consider options for Brexit – meets for the first time today. In this post Alan Renwick, Rebecca McKee, Will Jennings and Aleksei Opacic explain the process by which members were selected to be representative of the UK electorate, both demographically and in terms of how they voted at the 2016 EU referendum.
Over two weekends this month in Manchester, the Citizens’ Assembly on Brexit – a group of around 45 citizens – will meet to learn about, deliberate on, and make recommendations relating to the options for Brexit. As we set out in a previous post, the Assembly follows a well-established model for fostering quality public deliberation around major policy decisions. One key feature of this model is the process through which the members are selected. They are not elected or allowed to self-select. Rather, they must be identified through a rigorous process of random selection designed to ensure that – so far as possible for a group of this size – they reflect the diversity of the wider population. This post sets out how we have done that for the Citizens’ Assembly on Brexit and what we can say so far about the results.
Our goal has been to secure a broadly representative sample of the population. The relevant population in this case is the UK electorate: the people who are entitled to participate in public decision-making in the UK. It is their views that need to be heard by policy-makers involved in deciding the UK’s approach to Brexit.
Achieving a representative sample is far from straightforward. It relies on careful planning and design, as well as reflection, long before an assembly takes place. The first step is to decide what it actually means to be representative of the population at large. This is normally determined through stratification: you decide in advance what proportions of certain groups must be included, or set minimum levels for people from each group. Past citizens’ assemblies in places such as Ireland and Canada have used various stratification criteria, including age, ethnic background, geography, social class, and employment status. Each characteristic can be treated more or less strictly, depending on the context of the assembly. All official citizens’ assemblies have required equal numbers of men and women. In Canada, which has specific issues with ethnic inequalities, assemblies have required minimum numbers of people from the First Nations.
The stratification design is only the first step. Next comes the process of finding people to fill the stratification quotas. We know very clearly that one potential method – allowing people to self-select into the assembly, doesn’t work. Those who opt in are usually very different from those who don’t, most notably in terms of their levels of political interest and participation: they are more likely to have an interest in politics and to be already engaged in political activities. Self-selection may also attract people with particularly strong views about the topic of debate. Additionally, they are more likely to be people with plenty time to attend: those without caring or childcare responsibilities, those who are older and retired, and also people who can afford to take the time off work. To avoid the major issues of self-selection into a citizens’ assembly, some method of random selection is required.
Here there is a choice of methods depending on the resources available. Assemblies in Ireland, Canada, and the Netherlands have used methods making it possible for any member of the electorate to be selected into the assembly. Such methods – which are described in detail in a recent Constitution Unit report – are the gold standard for official assemblies. But they are beyond the financial means of an unofficial body such as the Citizens’ Assembly on Brexit, and they also do not lend themselves to detailed research into the recruitment dynamics.
Instead, we have built on methods trialled in 2015 by Democracy Matters: a project led by Matthew Flinders at the University of Sheffield, in which many of the Brexit assembly team were also involved, which ran two local citizens’ assemblies. This method is selection through an existing representative online panel. Democracy Matters used YouGov’s panels in the areas concerned. For the Citizens’ Assembly on Brexit, we have worked with ICM. Starting with an existing panel cuts costs considerably. It also allows us to track very closely who is signing up or dropping out at each stage of the process.
Recruitment design for the Citizens’ Assembly on Brexit
For the Citizens’ Assembly on Brexit, we adopted six stratification criteria: age, gender, region, ethnic background, social class, and, whether people voted to Leave or Remain, or didn’t vote at all, in the 2016 referendum. Considering the nature of the Brexit topic, this last criterion was crucial for the legitimacy of the Assembly and its findings and for the quality of the deliberations. We calculated the quotas from ONS mid-year population estimates and the National Readership survey – which is used to calculate social class – as well as the official referendum results.
We initially contacted potential members through an ICM survey, reaching a representative sample of 5,000 people across the UK. We asked respondents a variety of questions about their political views and interests. We also told them a little about the assembly and asked if they would be interested in attending.
Having thus identified potential Assembly members, we used our stratification criteria to work out whom to approach for a conversation about signing up. We initially contacted 53 people (our target number of 45, plus an extra eight people to allow for possible no-shows on the day), replacing those who were unable to attend when needed. We contacted potential members in the first instance by e-mail, asking them to fill out a Doodle poll indicating when we could phone them. We called those who did not respond after a couple of days. We tried calling three times and left messages; after that, if they still had not responded, we replaced them on the list with another person meeting the same stratification criteria.
Recruiting members is not the only challenge: it is also necessary to retain them, so that they do come to the first meeting and remain throughout the subsequent process. One lesson from previous citizens’ assemblies is the importance of good communications with members before the first meeting. Being both approachable and professional is vital, and multiple channels of communication are needed, depending on people’s preferences. We have a team of three people who contacted potential Assembly members initially and have maintained contact subsequently. They have been able to build up relationships with many members by name.
Another lesson is the importance of considering the barriers to attendance that people may face. In addition to their travel expenses and hotel accommodation, members will be gifted £200 per weekend. This is designed as an incentive: if we did not offer it, the membership would be strongly skewed towards those who are unusually interested in politics. But it also helps people with childcare and other responsibilities to meet the costs of making alternative arrangements. Because this is provided as a gift, members who are receiving out-of-work benefits will not face sanctions. Finally, we have ensured that our meeting venue and materials meet members’ particular requirements, including those with mobility issues, hearing impairments or dyslexia.
Recruitment in practice
Of the 5,000 people contacted by ICM, 1,179 said they were willing and able to attend both weekends of the Assembly. This is an exceptionally high response rate: Martin Boon, Director of ICM has described himself as ‘stunned’ by the level of public interest. It reflects the great ongoing interest in Brexit from voters (and, indeed, non-voters) on all side of the debate.
At the time of writing, 51 Assembly members are signed up to attend. A few people who initially accepted our invitation have, for various reasons, had to pull out. But the vast majority are highly engaged with the project and say they are really looking forward to the first weekend. One commented that she ‘would like to be involved in one of the most important decisions we are making as a country’. We have exchanged e-mails and phone calls, and all members have received a briefing pack in the post.
The members are, as planned, broadly representative of the UK electorate on all of our stratification criteria. Importantly, the proportions of Leave and Remain voters reflect the proportions in the population as a whole, and there is representation from those who did not vote in the referendum.
In addition to the members of the Assembly, we have identified a control group of people who responded positively to the original ICM survey, but whom we have not invited to participate in the Assembly. We have constituted the control group through what is called a ‘matched pair design’, ensuring that each Assembly member has a pair in the control group who matches them on our six stratification criteria. Follow-up surveys with the Assembly members and the control group will provide a powerful way of ascertaining the degree to which any changes in the Assembly members’ views are due to their participation in the Assembly or other external factors.
The Citizen’s Assembly on Brexit will give a small but broadly representative sample of people the chance to come together and deliberate about the most important issue facing the UK today. It will give decision-makers in Whitehall and Westminster deep insights into the reasoned and informed conclusions of a group of regular UK citizens.
About the authors
Dr Alan Renwick is the Director of the Citizens’ Assembly on Brexit, and the Deputy Director of the Constitution Unit.
Dr Rebecca McKee is Research Associate for the Citizens’ Assembly on Brexit, based at the Constitution Unit.
Aleksei Opacic is part of the Member Recruitment and Liaison Team for the Citizens’ Assembly on Brexit, and a Research Volunteer at the Constitution Unit.