Recruitment for any citizens panel (or citizens assembly, citizens jury) normally aims to be strictly random. This avoids the kind of self-selection seen at public meetings where a small number of ‘the usual suspects’ – the more articulate, confident and vocal – may dominate a meeting and exclude others from having a say.
Random selection seeks to get round this issue by recruiting a small but diverse sample of the local public. The approach can be more or less ‘scientific’ depending on time and resources. Nevertheless, however you do this there is always an element of bias. For example, using the electoral roll excludes those who don’t bother to register or vote. Polling agencies who use online panels exclude those who are not online. And whichever method you use people must want to be part of a the process – which means there will always be an element of self-selection.
We are following a 2-step process:
The first stage is a short online form
This asks a set of basic questions such as name, address, age and full postcode (so we can check you actually live in the area). We also include a couple of questions on your perceptions on homelessness. These are not trick questions and there is no right answer; we are seeking a range of viewpoints even if these later on.
We also include a couple of questions on personal experience of struggling to pay the rent while putting food on the table as we want to ensure that the citizens panel is composed of one or more people with direct experience of these issues.
Using Google Forms, the responses automatically populate a linked spreadsheet. This creates a pool of applicants from which 15 people will be randomly chosen. Ideally we need a pool of around 30 or 40 people.
The second stage is the random selection of the citizens panel itself.
A copy of the spreadsheet is made and names are replaced by a unique number. Email addresses and postcodes are also removed.
The anonymised spreadsheet (minus names and email addresses) will then be submitted to the Steering Group and a person present will be asked to call out a random set of unique numbers in order to arrive at a 15 member panel. This helps ensure that the final selection is scrupulously fair and random.
Stratified random selection
However even here things can go wrong; suppose we have a pool of 40 people who apply online. If it turns out that thirty of them are older, retired males, then however the panel is randomly chosen it will be overwhelmingly male. That would be a mistake and doesnt reflect the demographic profile of Penzance where women form just over half the population. Click here to see the profile
To get round this, we intend to segment the pool of applicants into male and female. Depending on how large the pool of applicants are, we might further stratify into young/old, home owner/non-home owner and so on. Random selection will then be applied to each of these segments. This is called ‘stratified random selection’ and allows for a more diverse panel that better reflects the make-up of the local population.
To help understand this approach, take a look at the pie chart below:
For more on stratified random sortition, see the additional note at the end.
One final note
If you apply for the Citizens Panel but are not chosen, please do not be disappointed. We really appreciate the fact that you have volunteered and we know that everyone has something of value to say. If you want to contribute online, please click the UPDATES button -see bottom of page. You can be sure that every comment will be read and weighed. We will also alert you to the final report and recommendations by the panel when this is published.
Additional notes – The Sortition Foundation approach
An alternative and more rigorous approach (and therefore more expensive) is adopted by the Sortition Foundation. This is outside our budget but we are sharing these notes with you in case you or your group wishes to experiment with citizens assemblies and have a larger budget to play with.
The more rigorous approach is adopted by the Sortition Foundation costs several hundred pounds – peanuts in terms of local government expenditure but a hefty additional expense for small voluntary organisations. They offer a Community service in which they send out letters to a random selection of around a thousand people in a particular location. From that thousand they normally expect to get a rough response of 30 to 40 people. That forms the initial pool of applicants from which they then conduct a stratified random sortition approach to reflect age and gender. To include social background, educational level and occupation, you would likely have to send out several thousand letters to build a pool of 300 or so applicants. Clearly, you need a much bigger budget for this approach – which we don’t have.