Citizens’ Assemblies: Less Accurate Than Tossing a Coin? analyses previous data from jurisdictions that have held citizens’ assemblies, to assess their accuracy and relevance in reflecting and understanding public attitudes.

The report argues there are fundamental defects in how they work, including the composition of those who take part in them and the group dynamics that occur within them. Importantly, these flaws manifest even when the citizens’ assemblies are run with integrity, by groups that are not seeking to push an agenda.

New analysis presented in the report shows that, when a citizens’ assembly is followed by a popular vote, the citizens’ assembly is not only a poor predictor of the result, but is systematically biased towards progressive outcomes:

• In 58% of cases – more than half – the citizens’ assembly’s recommendation was rejected by the general public following a vote – less accurate than tossing a coin.
• The citizens’ assemblies were systematically biased towards more progressive options. They gave more progressive takes than the general public on 88% of cases – by an average of 25 percentage points more progressive.
• The literature suggests citizens’ assemblies are likely to perform worst on the most controversial and emotive issues where they are proposed – for example assisted suicide or abortion.

The report argues that contentious issues involving conflicts of rights are best dealt with by Parliament, which is the only forum with the legitimacy to do so.