Assessing the relationship between interviewer effects and NSDUH data quality

To assess the impact of interviewer experience, two cohort analyses were conducted using 2002 and 2007 NSDUH data. RESULTS: The analyses of the relationships between several interviewer characteristics and various survey outcomes, including response rates and respondent self-reports on substance use...

Full description

Bibliographic Details
Main Authors: Wang, Kevin, Kott, Phillip (Author), Moore, Andrew (Author)
Corporate Authors: United States Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality (U.S.), RTI International
Format: eBook
Language:English
Published: Rockville, Maryland Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality 2013, October 24, 2013
Series:CBHSQ methodology report
Subjects:
Online Access:
Collection: National Center for Biotechnology Information - Collection details see MPG.ReNa
Description
Summary:To assess the impact of interviewer experience, two cohort analyses were conducted using 2002 and 2007 NSDUH data. RESULTS: The analyses of the relationships between several interviewer characteristics and various survey outcomes, including response rates and respondent self-reports on substance use and mental health indicators, have for the most part provided additional evidence in support of prior work in this area. Interviewer experience remains positively correlated with response rates and negatively associated with self-reported substance use. However, evidence suggests that the association between experience and substance use reporting has declined over time.
Not much evidence was found that other interviewer characteristics (e.g., demographics) and characteristics of their work (e.g., hours worked, miles traveled) were related to substance use and mental health outcomes in ways that suggest that changes in the distribution of these characteristics among interviewers would affect resulting survey estimates. CONCLUSION: The results of this study can be used to produce predictions of self-reported prevalence rates if a change in the sample design can be expressed in terms of the change in the distribution of a characteristic of interviewers. At the most basic level, one can assume that a design change, such as a reduction in the overall NSDUH sample size, will result in a change in the distribution of interviewer experience
BACKGROUND: For interviewer-administered surveys, interviewers can have a major effect on various measures of survey data quality, including unit response rates, item missing rates, estimates of uncertainty (variances), and the accuracy of survey estimates. As applicable to the National Survey on Drug Use and Health (NSDUH), this report summarizes research on the relationships between interviewer characteristics and survey participation, data quality, and survey outcomes. METHOD: This report used NSDUH data before 2002 to run logistic regression models that examine the relationship between interviewer characteristics and the likelihood of obtaining screener contact, screener cooperation, and interview cooperation. To assess the effects of various interviewer characteristics on the reporting of substance use and mental health items, single regression models used pooled 2002 to 2007 NSDUH data.
Physical Description:1 PDF file (various pagings) illustrations