3. Representative

How many members does a citizens’ assembly need?

The idea of a citizens’ assembly is to reflect, as closely as possible, the diversity of the wider population with the members of the assembly. As far as possible, people should be able to see “people like them” in the membership of the assembly. This typically means that citizens’ assemblies covering larger populations (e.g. national) require more members than those covering smaller populations (e.g. local). 

Citizens’ assemblies typically have between 50 and 200 members. 

The number of members is one of the distinguishing features of a citizens’ assembly, compared with other deliberative mini-publics, such as citizens’ juries (which have fewer) or deliberative polls (which have more).

Who takes part in a citizens' assembly?

Members of the general public take part in citizens' assemblies. They are recruited to be broadly reflective of the demographics of the place in question (town, city, local authority or nation). In practice this will mean that the final make-up of the assembly will reflect the gender, age, socio-economic status and ethnicity of the population in question. Often where this might lead to one member from a particular minority (say ethnicity, or in the case of a UK-wide assembly the devolved nations) that minority is over-recruited. 

Some assemblies also try to reflect the wider population’s attitudes to the policy the assembly will be discussing. For example, the 2017 Assembly on Brexit was recruited so that the members had voted 48% remain and 52% leave in the EU referendum. Other assemblies have recruited on the basis of attitudes to government or to mirror value modes. 

Assembly members will have no professional connection to the issue under discussion by the assembly, so they will not work in a government department working directly on the issue, or be a member of a campaign group actively advocating for a specific position on the issue, for example. 

The composition of the assembly's members will be determined by the demographics of the geography which it is intended to reflect. For example, a citizens' assembly deliberating on a UK wide issue would recruit participants who, as a whole, reflected the UK’s demographics. This would include the basic demographics above, but would also be likely to include geographic spread across the English regions and the devolved nations. 

Citizens’ assemblies called by a combined authority region, city or local authority are likely to mirror this type of make-up, but at the smaller geographic scale, ie by borough or ward.  

If the number of participants in the assembly is relatively small, it may be necessary to over-recruit certain demographics to ensure that more than one person ‘represents’ a specific demographic criteria. 

Sometimes participants are also recruited to ensure that the mini-public is reflective of wider attitudes such as the way participants voted in the 2016 EU referendum, or concern for the environment. The proportions of such attitudes in the wider population will already need to be known through data from elections or robust survey methodologies, for example. 

It is important to be able to show that who you have recruited measures up against existing data on any demographic criteria chosen as the basis for recruitment (e.g. census data)

How are public participants selected to take part in an assembly?

As noted above, assembly members are recruited to be broadly reflective of the demographics of the geography in question (town, city, local authority or nation and potentially attitudes to the policy under discussion). 

There are broadly three methods for recruiting participants:

• Sortition/ civic lottery  - which uses random selection to identify potential assembly members initially, followed by selecting participants at random from that pool to match the sampling criteria. In practice what this means is that 5,000 - 10,000 letters (depending on the size of the assembly) are sent to random postcodes in the sampling area. Recipients are asked to register to join the assembly by providing some basic demographic information. This creates the pool from which the final sample is randomly selected from. 
• In person - uses either telephone, or on street recruitment (or a mixture of both) to identify assembly members. Once the sampling criteria have been agreed, a recruitment company will approach people at random (either at their homes from the electoral register or in the street), complete a basic survey with them and if the person is willing to take part in the assembly and fits the sampling criteria then they are recruited. 
• From a panel - most polling companies maintain a panel of the general public who have volunteered to take part in polls, surveys and focus groups. Assembly members can be recruited directly from such a panel. 

There are pros and cons for all methods.  

The choices of methodology are likely to be informed by the question, the commissioning organisation and the geographical nature of the assembly and the stratification criteria.  In particular it is important to think about what scrutiny the choice of approach will come under.

It may be that multiple methods are used – for example, an initial use of a civic lottery but with supplementary in person recruitment to “top up” categories that are under-represented. This might include younger people, BAME participants or particular socio-economic groups, for example.