Spotlight Assembly UK sent out invitations to 30,000 randomly chosen households, and 1,500 responded asking to be considered. The fact that 1,400 people were effectively rejected by CAUK, which sank its 520,000 budget into an assembly for 110 people, makes it, Vittles says, the worst example so far of depth of deliberation at the expense of breadth of participation. The narrow focus by the increasingly influential Sortition Foundation which was established in 2015, on the ancient Athenian practice of sortition (taking a random selection of citizens and sorting them into a representative sample) risks devaluing alternative legitimate methods used by researchers for decades, adds Vittles. The national climate assemblies in France, Scotland and Ireland didnt use postal sortition. In France, Harris did random phone sampling and, in Ireland, they took 60 random locations and sent interviewers door to door with quotas to reflect the population profile, to get a representative sample. A balanced approach While fundamentalists debate the finer methodological and philosophical points of citizens assemblies, research companies are, by and large, pragmatic. It is important to have standards, quality thresholds, a framework and principles to work to, if the process is to be respected, but there needs to be a balance, says Suzanne Hall, research director at Ipsos Mori. Rob Francis, head of central and local government at research and engagement consultancy Traverse, agrees: The methodology is very important, but it is also right to be able to challenge the criteria, rather than following them slavishly for the sake of orthodoxy. Cooke points out that one of the principles of deliberative methods is their effect on the decisions ultimately taken. Mackenzie believes the reason the government hasnt adopted any of the recommendations made by the citizens assembly on social care in 2018 is that it was commissioned by the 16 wrong people not the government, but MPs, who lack the power to take things forward. Even Climate Assembly UK is not guaranteed success according to these criteria so its not surprising that some agencies find working at a local council level, or for an individual government department, more rewarding. There is a lot of interesting work outside the big, high-profile, complex, public policy issues, says Cooke. She cites a recent project the agency did for the National Infrastructure Commission (NIC) on road congestion, which was designed to gauge peoples attitudes to different policy levers and help the NIC to understand whether deliberation was a good method to use in future. Success can be measured in a variety of ways. NatCen Social Research conducted what it claims to have been the UKs first online deliberative poll using video-conferencing software last year. Its aim was to determine peoples views on the future of Britain and governance post-Brexit, and whether and why those views changed as a result of the deliberations. It was largely an experiment to find out whether the conditions needed for effective deliberation could be replicated online. Ceri Davies, who leads NatCens programme of deliberative research on citizen engagement in policy-making and democratic innovation, explains that the agency essentially replicated online what it typically does in a physical location. It recruited 320 people through a random probability survey among its own panel of 3,000 people, drawn from the British Social Attitudes survey. It ran a feasibility study on what software to use, decided on Zoom, and then did lots of development work to build technological capability in its own team and participants. It sent out a presession questionnaire to gauge opinions on the topics being discussed immigration, food policy and consumer rights and followed that with briefing materials on the topics. The event took place over two days on the hottest weekend in June.