This article aims to understand decision making under uncertainty and risk, with a case study on Cape Cod, Massachusetts. Decision makers need to consider imperfect information on the cost and effectiveness of advanced nitrogen-removing on-site wastewater treatment systems as options to mitigate water quality degradation. Research included modeling nitrogen load reduction to impaired coastal waters from seven treatment system technologies and eliciting expert knowledge on their costs. Predictions of nitrogen load removal and cost for each technology incorporated variation in effectiveness and uncertainty in household water use, costs, and expert confidence in costs. The predictions were evaluated using the Pareto efficiency concept to reveal tradeoffs between cost and effectiveness. The stochastic dominance index was used to identify preferred technologies for risk-averse decision making, assuming no further learning is possible. Lastly, the predictions were combined into a cost-effectiveness metric to estimate the expected payoff of implementing the best treatment system in the face of uncertainty and the expected payoff of learning which treatment systems are most cost-effective over time. The expected value of perfect information was calculated as the difference between the expected payoffs. Three technologies revealed Pareto efficient tradeoffs between cost and effectiveness, whereas one technology was the preferred risk-averse option in the absence of future learning. There was a high expected value of perfect information, which could motivate adaptive management on Cape Cod. This research demonstrated decision analysis methods to guide future research and decision making toward meeting water quality objectives and reducing uncertainty.