A Non-Response and Measurement Error Analysis for the National Survey of College Graduates

• Activity Number: 76 - Paradata for Adaptive Survey Designs and Other Applications
• Type: Contributed
• Date/Time: Sunday, July 29, 2018 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Kayla Varela* and Allison Zotti and Kevin Tolliver and Amanda Nagle
• Companies: U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau
• Keywords: non-response error; data linkage; adaptive design; measurement error; response rates

Abstract

Surveys often combat nonresponse by actively pursuing reluctant respondents in order to increase response rates. High response rates do not provide evidence of data quality alone, in particular with respect to nonresponse bias, (Groves and Heeringa 2006) and (Groves and Peytcheva 2008). While converting reluctant respondents may increase response, it may also increase measurement error if disinclined respondents do not provide information that is of as high of quality as eager respondents. Several studies have examined the possible tradeoff between data collection strategies aimed to convert reluctant respondents and measurement error, such as (Olson, 2006) and (Kreuter et al., 2010). The National Survey of College Graduates (NSCG) have implemented adaptive contact strategies to target under-represented households that are likely to respond and deprioritize households that are over-represented and unlikely to respond. This presentation discusses the research planned to analyze the possible tradeoff of non-response and measurement error with the contact strategies in NSCG using administrative data linkage.