Fig. 2. Predeliberative (panel a) and postdeliberative (panel b) weights elicited for valuation criteria, distributed by spatial scale: error bars show 95% confidence interval around mean.