Background: An important issue in alcohol and drug use research is the degree to which study participants cooperate with survey interviewers and provide accurate information. We examine the year-by-year trends in the perceived cooperativeness of participants in a large national survey focused on alcohol and drug use in the United States between 2002 and 2015.
Methods: We examine fourteen years of cross-sectional data (2002-2015) from the National Survey on Drug Use and Health (NSDUH) using the NSDUH's Restricted Data Analysis System. The main variable of interest was field interviewer reported participant cooperativeness (i.e., "How cooperative was the respondent?"). We present estimates of proportional rates of cooperation and examine the degree to which the estimated proportions for cooperativeness vary from 2002 to 2003 estimates based on non-overlapping 95% confidence intervals as a proxy for secular trend tests.
Results: The proportion of respondents classified as "very cooperative" was consistently elevated in all survey years, increasing significantly from 95.6% (CI = 95.4-95.8) in 2002-2003 to 96.7% (CI = 96.5-96.8) in 2014-2015. Elevated levels of cooperation were observed for participants reporting no/any past-year alcohol, marijuana, cocaine, and heroin use. Rates were elevated-96 to 98% in 2014-2015-among respondents of all sociodemographic backgrounds (i.e., age, gender, race/ethnicity, income, nativity). Only a fraction of participants were classified as "not very cooperative" (0.2-0.4%) or "openly hostile" (0.1%).
Conclusions: Cooperativeness with NSDUH survey research has been very high since the early 2000s with perceived participant cooperativeness increasing in recent years and consistently low rates of non-cooperativeness across all years.
Keywords: Bias; Cooperation; Substance use; Survey research; Trends.
Copyright © 2019 Elsevier B.V. All rights reserved.