Using Local Knowledge During Data Collection: Does It Make a Difference Who Applies It and When?

• Activity Number: 690
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Rachael Walsh* and James Christy and John Marshall
• Companies: U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau
• Keywords: reassignment ; paradata ; multilevel modeling ; random intercepts

Abstract

As part of a recent reorganization of its data collection activities, the Census Bureau created a new management structure. In addition to consolidating twelve Regional Offices to six, it created several new supervisory positions, like the Field Supervisor (FS) and the Survey Statistician - Field (SSF). The FS directly supervises interviewers within a geographic area, and the SSF manages multiple FSs within their collective area. At inception, only the SSF was granted rights to reassign cases to interviewers across FS areas. Initially, managers in some Regional Offices extended these rights directly to the FS. A uniform policy to reassign cases across FS areas was implemented in all Regional Offices in November 2014. This research assesses this change in management of case reassignment, specifically attempting to determine the effectiveness of the policy of allowing case reassignment by the FS early in the data collection process. After including covariates known to increase the level of effort necessary to resolve a case, as well as those affecting interview completion, multilevel models and logistic regression with random intercepts can use paradata to determine whether this modification of reassignment rights is changing data collection outcomes.