Watershed ecosystems are characterized by multiple attributes which are difficult to collapse into single money-metric. Attribute ranking without prices and group deliberation can be used to process information about these complex systems. We develop this approach and use it with a stakeholder group to classify attributes in subwatersheds for restoration. We examine the relationship between individual valuation and valuation arrived from deliberation and information exchange by the group. While group consensus values tended toward an average of individual responses, significant differences existed between mean values and consensus values for some attributes, emphasizing the role of ecosystem attribute information and deliberation.