Implications of Coarse Data Allocation Methods for Flood Mitigation Analysis

• Activity Number: 421
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): James Howard*+
• Companies: UMBC / Kore Federal
• Keywords: data science ; coarse data ; statistical allocation ; missing data

Abstract

Efforts to perform fine-grained analysis are often hampered by data provided by government agencies that does not reflect appropriate granularity. Coarse-grained government data may reflect the data collection methods, strategies, or may reflect the reality of what the data represents, and cannot be made more granular. For example, the United States Flood Mitigation Assistance (FMA) grant program makes grants to both state and local governments. By employing data on the FMA program, this analysis examines allocation strategies for coarse data. Between 1996 and 2011, approximately one-quarter of FMA grants were given to state governments with the remaining three-quarters given to local governments. Performing a local-level analysis of the impacts of these grants requires an allocation method that fairly reflects the local impact of statewide grants. This note considers several allocation strategies and how these strategies affect the implementation and interpretation of statistical models for public policymaking.