Deliberative Polling

Both expose citizens to balanced information and deliberation, but they differ in scale and aim. A citizens' jury convenes a small panel of a dozen or two to produce reasoned recommendations; its legitimacy comes from being a deliberating microcosm, not from statistical representativeness. A Deliberative Poll uses a much larger random sample — typically hundreds — and measures opinion change with before-and-after surveys, so its results can be generalised to the population. Deliberative Polling answers 'what would the public think if informed?'; a citizens' jury answers 'what does this deliberating panel recommend?'.

Fishkin's central idea is that conventional polls measure actual public opinion, which is often uninformed and inattentive, whereas a Deliberative Poll measures what public opinion would be if people were well informed and had genuinely deliberated. Because it is impractical for the entire public to deliberate, the method uses a representative sample as a stand-in: their post-deliberation views estimate this counterfactual, considered opinion. The gap between the before and after surveys is the empirical signature of how information and deliberation would change the public's mind.

It can, which is why balance is built into the design. Briefing materials are prepared with input from advocates of different positions and reviewed for even-handedness, and plenary sessions feature competing experts whom participants question directly. Moderators are trained to keep small-group discussion balanced and inclusive. These safeguards reduce but do not eliminate the risk of bias, so credible Deliberative Polls document how materials and panels were assembled and subject the process to independent scrutiny, allowing observers to judge whether the deliberation was genuinely balanced.