, 1991). However, this increased arousal did not appear to produce differential effects of AS on attitudinal or intention measures. Finally, although the IM measures reasonably predicted quit intentions, the inclusion namely of psychophysiological measures produced no increase in explained variance. Research has demonstrated the utility of physiological responses in understanding message processing and recall (Lang, Chung, Lee, Shin, & Schwartz 2005; Palmgreen et al., 2001). Physiological responses may be less informative in characterizing effective messages in an older, smoking population such as ours, than in younger subjects viewing illicit drug content media (Lang, Chung, Lee, & Zhao, 2005; Palmgreen et al., 2001). Further research is needed. Limitations A few considerations should be noted.

First, participants in the present study were exposed to a single set of four PSAs within each condition. Although we observed some interesting effects of MSV and AS on physiological and self-report measures, the single exposure appears to have been insufficient to produce persuasive effects, similar to what has been observed elsewhere (Worden & Flynn, 2002). Yet, sustained low-level (monthly) exposure to antitobacco messages has been shown to effectively reduce smoking prevalence and increase antitobacco sentiment (Emery et al., 2005). Laboratory studies with repeated sessions might better capture the effect of repeated, low-frequency PSA exposures akin to an antitobacco campaign.

Comprehensive media campaigns, such as those launched in California, have been shown to effectively dissuade smoking initiation and increase quit attempts and, therefore, are an essential component of attempts to address tobacco dependence (Pierce et al., 2002). Second, moderating effects of sensation seeking were observed only for self-efficacy and were not robust and consistent across multiple outcomes. Sensation seeking may have weaker persuasive effects among adults with established behaviors, such as smoking, compared with effects on the initiation of new behaviors among adolescents (Ling & Glantz, 2002a). Also the persuasive effects of the smoking-themed PSAs used in the present study may differ AV-951 from those of the anti-drug PSAs used in most prior research. It would be interesting to test the PSA feature manipulations in the present study in a population of adolescents to identify those features that are most persuasive with respect to not engaging in smoking behavior. Summary The present study represents the first comprehensive theory-based experimental investigation of the effects of different features of antitobacco PSAs. As such, it provides a framework for future research to identify the critical elements of effective antitobacco PSAs.