Problem/Objective: In 2004 and 2005, the New York Adult Tobacco Survey (ATS) observed a significant decline in smoking prevalence estimates. During this same period, prevalence rates observed in the New York Behavioral Risk Factors Surveillance System (BRFSS) remained steady. Two features that differentiate the ATS survey protocol from the BRFSS protocol, and have some potential to influence smokers' willingness to participate, are the introductory scripts and respondent incentives. We designed an experiment to see whether and how introductory scripts and incentive offers affected participation among current smokers in one quarter of the ATS.

Methods: This experiment involved two alternative versions of the introductory scripts, the original ATS script that highlighted the survey focus on tobacco use and an alternative script that closely followed the BRFSS script and indicated a broader focus on health behaviors. The two incentive conditions tested were no incentive initially offered and an initial offer of $20. Telephone numbers sampled for the first quarter of the ATS in 2006 were randomly assigned to one of the four possible conditions represented by these alternatives.

Results: The results indicated that the initial offer of an incentive had a significant positive impact on smokers' participation. Smokers' participation was also slightly higher in both conditions with the alternative script compared to the non-incentive condition with the original script, but these differences were not statistically significant.

Conclusions: Smokers' participation in ATS surveys appears to be conditioned by incentive offers and possibly the presentation of the survey, but other factors like sponsorship could also play a role.