Following the first of Professor Lupia's posts on the problems of deliberative polls, the two masterminds of the deliberative polling method, Professor James Fishkin of Stanford and Professor Robert Luskin of the University of Texas, explain a bit more about the results and methodology of their approach:

It is true that studies of "deliberation" have produced some variety of results - unsurprisingly, given the considerable variety in definitions and operationalizations underlying them.

Much depends on whether a deliberation is part and parcel of a decision making process (as in many committee meetings) or a prelude to one (as in an election), whether it is consensus-seeking or not, to what extent it is balanced, etc.

We have tried to be clear about the kind of deliberation we are trying to capture, which is a prelude to decision-making, non-consensus-seeking, and balanced, among other characteristics.

Jury deliberations, the bedrock of the literature suggesting that deliberation homogenizes opinions within groups, polarizes them across groups, and allows the privileged disproportionate influence, are very different. The participants are generally less diverse, there is much less assurance of balance, the process is decision-making, and it is designed to achieve consensus if possible.

All we should claim is that the kind of deliberation we aim to capture (a) is relevant for what we might hope for in a mass public that learned, thought, and talked more about the issues before expressing its views in polls and casting its ballots and (b) seems to have effects of the sort we describe. Jury deliberations are interesting and important in their own right but far less relevant for the study of mass politics.

(The full, unedited reply can be found here - JCM)