A Method for Improving List Building: Cluster Profiling

• Activity Number: 186
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): William Cecere* and Denise Alexandra Abreu+
• Companies: National Agricultural Statistics Service and National Agricultural Statistics Service
• Keywords: hierarchical clustering ; cluster profiling ; k-means ; dual frame

Abstract

The USDA National Agricultural Statistics Service (NASS) conducts the quinquennial Census of Agriculture in years ending in 2 and 7. Also, NASS conducts an annual area frame based survey, the June Area Survey (JAS). The census has a dual frame: an independent list frame and the area frame from the JAS. The JAS is used to identify farming operations missed on the list frame. In 2007, a full census questionnaire was sent to all JAS records that were not found on the census mail list. Multiple clustering techniques were used to characterize farming operations missed during the census mail list building. Hierarchical methods (average linkage, centroid, and Ward's method) and non-hierarchical k-means clustering were used to identify groupings. Through cluster profiling, potential improvements to future list building efforts are discussed.