This study examines the factors that determine the likelihood of submitting a potentially fraudulent prevented planting claim. A theoretical model is developed and the theoretical predictions are empirically verified by utilizing a binary choice model and crop insurance data from the southern United States. The empirical results show that insured producers with higher prevented planting coverage, lower dollar value of expected yield, and a history of submitting prevented planting claims are more likely to submit an anomalous prevented planting claim. The empirical model also suggests revenue insurance plans may be more vulnerable to prevented planting fraud than the traditional yield-based insurance plan. Results of this study can be valuable to compliance offices in their efforts to find "indicators" of fraudulent behavior in crop insurance, especially with regard to prevented planting.