In this article, we model various forms of non-optimizing behavior in a newsvendor setting, including biases such as recency, reinforcement, demand chasing, and anchoring, as well as unsystematic decision errors. We assume that a newsvendor may evaluate decisions by examining both past outcomes and future expected payoffs. Our model is motivated by laboratory observations under several types of supply chain contracts. Ordering decisions are found to follow multi-modal distributions that are dependent on contract structures and incentives. We differ from previous research by using statistics to determine which behavioral factors are applicable to each decision maker. A great deal of heterogeneity was discovered, indicating the importance of calibrating a contract to the individual. Our analysis also shows that the profit performance and the effectiveness of co-ordinating contracts can be affected by non-optimizing behaviors significantly. We conclude that, in addition to the aggregate order quantities, the decision distributions should be considered in designing contracts.