Validity of Survey Response Propensity Indicators: A Behavior Genetics Approach1. (13th December 2012)

• Record Type: Journal Article
• Authors: Littvay, Levente; Popa, Sebastian Adrian; Fazekas, Zoltán
• Is Part Of: Social science quarterly. Volume 94:Number 2(2013)
• Page Start: 569
• Page End: 589
• Publication Date: 2012-12-13
• Subjects: Political science -- Periodicals; Social sciences -- Periodicals; 300
• DOI: 10.1111/j.1540-6237.2012.00938.x
• Languages: English
• ISSNs: 0038-4941

Abstract

<abstract abstract-type="main"> <title> <x xml:space="preserve">Abstract</x> </title> <sec id="ssqu938-sec-0010" sec-type="section"> <title>Objectives</title> <p>This study explains how behavior genetic analysis using a twin design can help us assess the validity of our measures.</p> </sec> <sec id="ssqu938-sec-0020" sec-type="section"> <title>Methods</title> <p>We test multiple indicators of response propensity, a measure used by survey researchers to better understand the similarities and differences between survey respondents and nonrespondents. The response propensity indicators evaluated include response to follow‐up surveys and subsequent waves of a panel and the completion of a sensitive recontact information sheet to aid subsequent recontact efforts.</p> </sec> <sec id="ssqu938-sec-0030" sec-type="section"> <title>Results</title> <p>A classical and the newly proposed method of validation all point to insufficient validity of our response propensity measures. Construct validation using data from the National Survey of Midlife Development in the United States exhibited little correlation between indicators. Genetic analysis suggests that the success of subsequent data‐collection efforts is predominantly driven by additive genetic effects, while nonresponse to inquiries for recontact information is influenced predominantly by familial environmental predictors.</p> </sec> <sec id="ssqu938-sec-0040" sec-type="section"> <title>Conclusion</title> <p>Our results indicate that different underlying constructs drive the response propensity indicators, suggesting that nonresponse is, at minimum, multidimensional.</p> </sec> </abstract>