Social Climate Survey of Tobacco Control: A mixed-mode approach

The Social Climate Survey of Tobacco Control (SCS-TC) measures tobacco control indicators using data from random samples of the US population. From 2000-2008, we have administered this annual cross-sectional survey to a representative sample of US adults using Random Digit Dialing (RDD) methods. However, data from the NHIS highlight the growing problem of wireless substitution for survey researchers. The purpose of this study is 1) to compare two alternative sampling design approaches to RDD survey administration (N=1,532) in terms of cost, response rate, demographics, estimates on SCS-TC indicators, and burden on the survey administrators; and 2) to examine weighting strategies to create a valid mixed-mode data set from the RDD data set and the two alternative samples. The two alternative approaches are 1) an Internet Panel sample, N=1,731 (by Knowledge Networks); and 2) a cell-phone sample frame (N=1,509). Analyses reveal that the Knowledge Networks panel had the highest percentage of people with lower levels of education and the cell phone sample included the highest percentage of young adults. Age and education are correlated with many tobacco control indicators, and thus estimates from the cell phone sample and the Knowledge Networks panel were typically less favorable (from the perspective of tobacco control) than estimates from the RDD frame. This study will help guide tobacco and other public health researchers in selecting survey methods that will provide more representative samples of US adults, and therefore more valid survey results.