Insect injury/crop yield response regression relationships were developed for the bollworm and cotton fleahopper on select cultivars of cotton in the Lower Gulf Coast of Texas. Regression relationships were developed for different phenological stages of plant development for each cultivar. One of the following methods was used to accomplish this: (1) a range of insect densities was established by caging known numbers of insects on cotton plants in the field and then densities were related to yield changes; or (2) densities of insects in insecticide field trials were related to yield changes. An economic decision-aid was deterministic plant submodel, a stochastic insect submodel, and a heuristic submodel of management decisions represented by a rulebased expert system. Stochastic economic relations were included to transform model output into profit. The model is driven by a random weather generator. The goal is to compare three management strategies, representing an extension specialist, an extension publication and a static economic threshold. Multiple simulation provide distributions of profit under each strategy over a wide range of input conditions. These distributions are compared with stochastic dominance.