A citizen's assembly is a group of "random" yet representative individuals former with the aim to deliberate on a specific issue of national interest, listen to the view of a broad range of "relevant" parties on the issue, and produce a conclusive answer to it. Some suggest a citizen's assembly is the only/best way to solve the Brexit deadlock (e.g. here).
Yet, there could be several reasons why such an assembly might fail. For instance, the selection of individuals could end up not being random. A prime reason for this might be self-selection. Although the demographics of the group might be representative, some pre-selected individuals might choose not to participate, because of lack of interest, because the issue is (considered) not relevant to them, or because of distance, or lack of material resources. Another reason could be that those who are called to present arguments to the assembly might not represent all the views (the linked article speaks of a liberal bias). These and other problems might indicate that the conclusion reached by the group might not truly represent that of the population.
So, the question I have is, do citizens' assemblies work? By work I mean, do they "truly" represent society? Is there any empirical evidence on this?
PS: someone might validly say that at least they represent it better than parliament/MPs.