Within the Advocacy Coalition Framework (ACF), policy-oriented learning is understood as a change in policy beliefs. Additional work has noted that belief reinforcement, not just belief change, is also a potential policy learning outcome. Yet, little work has attempted to reconcile how learning could involve both belief change and belief reinforcement. In this article, I propose a policy-oriented learning model where policy beliefs – deep core, policy core, or secondary aspects – are understood as having a distribution with a central tendency (that is, the belief) as well as variance (that is, certainty associated with the belief). With policy beliefs considered as distributions, learning can be understood as changes in beliefs (that is, a change in the central tendency) as well as changes in certainty (that is, variance), and conversely, a decrease in belief uncertainty would constitute belief reinforcement. Using data from a deliberative forum that brought together various stakeholders including experts, natural resource managers, and the public to discuss environmental issues impacting coastal communities, I explore policy-oriented learning as changes in concern regarding several key issues before and after the forum. Additionally, I examine the association between concern following the forum and self-reported learning. I find support for the proposed policy-oriented learning model as shown by significant changes in average concern as well as average variance among participants across several of the issues discussed. In this way, the article makes a theoretical contribution to the ACF literature by testing the use of distributions to assess policy learning.