A comparison of paper vs computer-assisted self interview for school alcohol, tobacco, and other drug surveys Denise Hallfors a, *, Shereen Khatapoush b , Charles Kadushin b , Kim Watson b , Leonard Saxe b a University of North Carolina at Chapel Hill, 401 Rosenau Hall CB 7400, Chapel Hill, NC 27599 7400, USA b Brandeis University, 415 South Street, Wattham, MA 02254,USA Received 1 October 1998; received in revised form 1 September 1999; accepted 1 October 1999 Abstract School surveys of alcohol, tobacco, and other drug (ATOD) use play an important role in prevention programs and policy development. Improving the quality of such data collection is critical. Most surveys are conducted with paper and pencil (PAP) instruments, but computer-assisted self interviews (CASI) are now being developed and there is some evidence that they elicit higher rates of positive responses to sensitive questions than traditional measures. The present study examines whether ATOD school surveys using CASI are feasible and improve the quality of data. Seventh, ninth, and eleventh grade students in two California communities were randomly assigned to PAP or to CASI n  2296). The ®ndings indicate that while CASI did not increase reported rates of substance use over PAP, it signi®cantly improved the speed of data processing and decreased the incidence of missing data. CASI was well accepted by students and school sta despite problems such as lack of computer resources. 7 2000 Elsevier Science Ltd. All rights reserved. Keywords: School surveys; Computer-assisted surveys; Survey evaluation; Parental consent; Substance abuse; Adolescent substance use 1. Introduction National, state, and local institutions routinely use school surveys to assess trends in alcohol, tobacco, and other drug (ATOD) use (Hallfors, Stirratt, Rinds- kopf & Saxe, 1997). Data from school surveys, most prominently the Monitoring the Future (MTF) studies (Johnston, O'Malley & Bachman, 1996), have played a key role in developing national policy (Oce of National Drug Control Policy, 1997). This study focuses on two key questions. First, given that ®ndings must rely on respondents to report accurately on beha- viors that are sensitive and illegal (Johnson et al., 1996), can the veracity of school survey responses be improved by implementing computer-assisted self inter- viewing (CASI)? Second, is CASI a feasible innovation for school districts to implement? School districts and communities are under growing pressure to assess substance use and evaluate whether prevention programs are working. Thus, for example, United States Department of Education (DOE) rules require school districts to select research-based pro- grams and assess their impact against measurable goals if they are to continue receiving Safe and Drug Free Schools funding (Notice of Final Principles of Eectiveness, 1998). This is a formidable expectation for school districts who report that their greatest tech- nical assistance need from state DOE agencies is evalu- ation (US Department of Education & US Department of Health and Human Services, 1987). As schools and communities try to reduce rates of sub- stance use among youth, better tools to assess accu- rately indicators of prevention program eectiveness Evaluation and Program Planning 23 (2000) 149±155 0149-7189/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S0149-7189(00)00011-2 www.elsevier.com/locate/evalprogplan * Corresponding author. Tel.: +1-919-966-6287; fax: +1-919-966- 0458. E-mail address: denise_hallfors@unc.edu (D. Hallfors).