Skip to main content

Statistical Methods Whitepaper 

In FY 2015/2015, as part of its commitment to reduce fraud related to the Supplemental Nutrition Assistance Program (SNAP), the United States Food and Drug Administration (USDA) undertook a series of steps to address SNAP system integrity. The Consolidated Appropriations Act of 2014 provided funding to augment integrity efforts going forward. 

In 2015, the Mississippi Department of Human Services (MDHS) submitted an application for funding (under grant opportunity CDFA 10.546) to develop and install an Early Detection and Fraud Investigation Management System and to develop statistical methods for automatically detecting possibly fraudulent activity and prioritizing investigations. The system allows tracking of investigations throughout the lifecycle from tip through to procedural hearings, and provides case-specific workflows and report templates. The statistical models are designed to be used to analyze transaction data to detect suspicious activity and automatically analyze benefit redemption patterns of individuals referred for investigation. These models will also automatically conduct cost-benefit prioritization for potential investigations by analyzing the likelihood of fraud and its potential value. The project proposal was approved, and MDHS has worked with the National Strategic Planning and Analysis Research Center at Mississippi State University to undertake the research and system development tasks necessary to fulfill the goals of the project. This whitepaper contains a description of the statistical work developed as part of the grant.