An Investigation of Decennial Census Effects on Estimates of Substance Use and Mental Illness from the National Survey on Drug Use and Health (NSDUH)

• Activity Number: 87
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Neeraja Sathe*+ and Patrick Chen and Art Hughes and Jonaki Bose and Lanting Dai and Misty Foster
• Companies: RTI International and RTI International and Center for Behavioral Health Statistics and Quality, SAMHSA and Center for Behavioral Health Statistics and Quality, SAMHSA and RTI International and RTI International
• Keywords: census effect ; mental illness estimates ; substance use estimates ; NSDUH ; poststratified weights ; population control totals

Abstract

The National Survey on Drug Use and Health (NSDUH), an annual survey of the U.S. civilian, noninstitutionalized population aged 12 or older, is a major source of substance use and mental illness data. The 2010 estimates were produced using weights poststratified to 2010 population control totals (intercensal estimates) derived from the 2000 decennial census; however, the 2011 estimates were produced using weights poststratified to 2011 population control totals derived from the 2010 decennial census. This study was done to determine whether the change in the source of the control totals had an effect on the level of change observed between the 2010 and the 2011 estimates. To examine this "census effect," 2010 estimates were also produced using weights poststratified to 2010 population control totals derived from the 2010 decennial census, resulting in two sets of weights for use on the 2010 data. NSDUH estimates were compared using both sets of 2010 estimates along with the 2011 estimates. Substance use estimates were more affected by the census effect than were mental illness estimates, and they were more notable for estimated totals compared with rates.