A comparison of computer-assisted and paper-and-pencil self-administered questionnaires in a survey on smoking, alcohol, and drug use

This study compares estimates of self-reported substance use and psychological well-being obtained in computer-assisted and paper-and-pencil self-administered surveys. We examine the extent to which mode effects on survey responses are moderated by respondent characteristics and respondent attitudes, including attitudes toward computers, attitudes about confidentiality and privacy, and their general mistrust of others. Respondents age 12-34 were selected through a multistage area probability sample of urban and suburban areas nationally and were randomly assigned to interview mode. There were few main effects of mode on self-reported substance use and well-being. Significant mode-by-age interaction terms revealed that adolescents were more sensitive to mode of administration than older respondents. Adolescents reported significantly higher levels of alcohol use, illicit drug use, and psychological distress in the computer mode than on paper self-administered questionnaires. Significant mode-by-mistrust interactions were also found. Respondents with higher levels of mistrust in others were less likely to report substance use in the computer mode than in the paper-and-pencil format.