We present a new approach for characterizing the potential of scientific studies to reduce conflict among stakeholders in an analytic-deliberative environmental decision-making process. The approach computes a normalized metric, the Expected Consensus Index of New Research (ECINR), for identifying where additional scientific research will best support improved decisions and resolve possible conflicts over preferred management actions. The ECINR reflects the expected change in agreement among parties over preferred management actions with the implementation and consideration of new scientific studies. We demonstrate the ECINR method based on a preliminary application to coral reef protection and restoration in the Guanica Bay Watershed, Puerto Rico, focusing on assessing and managing anthropogenic stressors, including sedimentation and pollution from land-based sources such as sewage, agriculture, and development. Structured elicitations of values and beliefs conducted at a coral reef decision support workshop held at La Parguera, Puerto Rico, are used to develop information for illustrating the methodology. The ECINR analysis was focused on a final study group of seven stakeholders, consisting of resource managers and scientists, who were not in agreement on the efficacy and respective benefits of reducing loadings from three sources: sewage, agriculture, and development. The scenario assumed that loadings would be reduced incrementally from each source through a series of management steps, which would be ranked in order of maximizing anticipated benefits. An examination of whether beliefs exhibited greater confidence and coherence between stakeholders when informed by plausible study results followed. The results suggest that new scientific research would be generally likely to bring people who initially disagreed to agree. Seventy-five percent of the hypothetical research results were projected to result in more agreement among the stakeholders. However, there can be situations where prior beliefs may be too different from the study results to shift perspectives enough to result in more agreement. Furthermore, in a few cases, hypothetical research results were projected to lead to more conflict among stakeholders. Priority research, according to the seven stakeholders, would be to quantify loadings from agriculture and sewage, and not loadings from development, since it is predicted to make little difference in the outcome. Assuming the stakeholders are conflict-averse, they would likely opt for research on sewage loadings as the highest priority. Though preliminary, these results suggest that ECINR can provide useful insights into the social implications of a research program.