UNDER EMBARGO UNTIL OCTOBER 9, 2012, 12:01 AM ET

Promotion of Smoking Cessation with Emotional and/or Graphic Antismoking Advertising

Matthew C. Farrelly, PhD, Jennifer C. Duke, PhD, Kevin C. Davis, MA, James M. Nonnemaker, PhD, Kian Kamyab, BA, Jeffrey G. Willett, PhD, Harlan R. Juster, PhD
Background: Antismoking campaigns can be effective in promoting cessation, but less is known about the dose of advertising related to behavioral change among adult smokers, which types of messages are most effective, and effects on populations disproportionately affected by tobacco use. Purpose: To assess the impact of emotional and/or graphic antismoking TV advertisements on quit attempts in the past 12 months among adult smokers in New York State. Methods: Individual-level data come from the 2003 through 2010 New York Adult Tobacco Surveys. The influence of exposure to antismoking advertisements overall, emotional and/or graphic advertisements, and other types of advertisements on reported attempts to stop smoking was examined. Exposure was measured by self-reported confrmed recall and market-level gross rating points. Analyses conducted in Spring 2012 included 8780 smokers and were stratifed by desire to quit, income, and education. Results: Both measures of exposure to antismoking advertisements are positively associated with an increased odds of making a quit attempt among all smokers, among smokers who want to quit, and among smokers in different household income brackets ( $30,000 and $30,000) and education levels (high-school degree or less education and at least some college education). Exposure to emotional and/or graphic advertisements is positively associated with making quit attempts among smokers overall and by desire to quit, income, and education. Exposure to advertisements without strong negative emotions or graphic images had no effect. Conclusions: Strongly emotional and graphic antismoking advertisements are effective in increasing population-level quit attempts among adult smokers.
(Am J Prev Med 2012;xx(x):xxx) 2012 American Journal of Preventive Medicine

Background
large body of evidence demonstrates that televised mass media campaigns to encourage smoking cessation can contribute to reductions in adult smoking prevalence.1 4 Although the growing evidence base indicates that antismoking campaigns can promote cessation, less is known about the dose of adverFrom the Public Health and Environment Division (Farrelly, Davis, Nonnemaker, Kamyab), RTI International, Research Triangle Park, North Carolina; Public Health and Environment Division (Duke), RTI International, Boulder, Colorado; Kansas Health Foundation (Willett), Wichita, Kansas; and New York State Department of Health (Juster), Albany, New York Address correspondence to: Matthew C. Farrelly, PhD, RTI International, 3040 Cornwallis Road, P.O. Box 12194, Research Triangle Park NC 27709. E-mail: mcf@rti.org. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2012.07.023

tising related to behavioral change among adult smokers, the types of messages that are most effective, and the degree to which sustained televised mass media campaigns effectively influence important populations disproportionately affected by tobacco use.2,5,6 The impact of exposure on behavior change at the population level and the extent to which the media dose and message type influence various groups of smokers are crucial research questions for funding agencies given limited resources and the high cost of airing televised mass media campaigns. Although substantial evidence indicates that antismoking campaigns can increase call volume to telephone quitlines,4,715 these studies do not provide data on the population-level impact of media on smoking cessation because only a small fraction of smokers will ever call
Am J Prev Med 2012;xx(x):xxx 1

2012 American Journal of Preventive Medicine Published by Elsevier Inc.

2
16

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

a quitline. Only a few population-based studies2,6,17,18 of large-scale, sustained antismoking campaigns have found that market-level exposure to antismoking advertising is associated with smoking cessation or quit attempts. Hyland and colleagues17 found that the dose of advertising was associated with the likelihood of quitting, but this result was limited to respondents who felt the amount of anti-tobacco information in their community had increased substantially over the study period. Wakefeld and colleagues2 found that the dose of antismoking advertisements airing in a respondents media market in the previous 3 months, but not more distal time frames, was associated with making a quit attempt during the previous 3 months. Research indicates that highly emotional or graphic content is more likely to be attended to and recalled19,20 and is perceived to be more effective than less hardhitting content.21,22 Wakefeld and colleagues2 concluded that graphic advertisements are effective in promoting cessation from a study of media campaigns featuring almost exclusively graphic and/or emotional advertising. One empirical study to date examined the influence of different types of antismoking advertisements on population-level smoking cessation and found that advertisements that were highly emotional or included personal testimonials, but not other types, were associated with cessation.6 Although research suggests that threat or fear appeals are effective in eliciting behavioral change,23,24 some remain skeptical about their application to smoking cessation.25,26 Specifcally, Hastings and colleagues25 note that much of the support for fear-appeal effcacy is based on laboratory studies that may not be generalized to real-world antismoking campaigns in which audiences may become habituated to fear messages, leaving any benefcial short-term effects to fade. Aside from the effective dose and type of messages, understanding the population-level impact of antismoking campaigns among smokers of low SES is imperative given the large disparities in smoking prevalence and quit rates by SES in the U.S.27 Conflicting evidence remains on the ability of mass media campaigns to promote smoking cessation among low-SES populations relative to more advantaged populations.28,29 In their systematic review, Niederdeppe et al.28 note that SES disparities may occur because of differences in meaningful exposure to a campaign, motivational response to its messages, or opportunity to sustain cessation. One cohort study6 suggests that high exposure to campaigns that elicit negative emotions promotes increased cessation rates in lower SES populations. Efforts to understand the heterogeneity of antismoking campaign effects and effective message types among low-SES

populations are crucial to reduce disparities in smoking rates by SES. The New York Tobacco Control Program (New York TCP) is one of the largest statewide tobacco control programs in the U.S. From 2003 to 2009, New York TCP invested approximately $75 million in paid advertising to educate residents about the health risks of smoking and the dangers of secondhand smoke exposure.30 The current paper examines the association between exposure to the states antismoking TV advertisements and making a quit attempt among smokers overall and by desire to quit, income, and education. Making a quit attempt is an important marker of progress in the dynamic process of ultimately achieving successful cessation.31 The present study examines the impact of antismoking advertisements overall and by style of adthose that portray reasons for quitting smoking using a highly emotional or graphic style compared to advertisements without these features.

Methods
Data
The New York Adult Tobacco Survey (NY-ATS) is a crosssectional, random-digit-dial telephone survey representative of adults aged 18 years in New York State. Quarterly data from June 2003 through 2010 were examined from 8780 current smokers. The NY-ATS includes measures of cigarette and other tobacco product use, smoking cessation, exposure to secondhand smoke, and related attitudes, beliefs, and intentions; self-reported recall of antismoking advertisements; and sociodemographic characteristics. Data were weighted to reflect the state population of adults, adjusting for different probabilities of selection and survey nonresponse. Protocols were approved by the IRBs of RTI International and the New York State Department of Health.

Key Measures
The key outcome was past-year quit attempts (During the past 12 months, have you stopped smoking for one day or longer because you were trying to quit smoking?; yes [1]/no [0] indicator) among current smokers, defned as having smoked at least 100 lifetime cigarettes and smoking every or some days at time of interview. The following covariates were included in the overall analyses: age (18 24, 2539, 40 64, and 65 years); educational attainment (less than high-school diploma or equivalent, high school graduate or equivalent, some college, college graduate or more); race/ethnicity (non-Hispanic white, non-Hispanic African-American, Hispanic, non-Hispanic other race); gender (female, male); income ( $30,000, $30,000 $59,999, $60,000 $89,999, $90,000); and cigarettes per day (19, 10 19, 20). In addition, smokers interest in quitting was measured by responses to the question How much do you want to stop smoking?: a lot, somewhat, a little, or not at all.

Advertising Exposure Data

Exposure was measured with self-reported confrmed recall of advertisements and potential exposure based on TV gross rating points (GRPs). Similar to previous studies,2,3234 the primary meawww.ajpmonline.org

Farrelly et al / Am J Prev Med 2012;xx(x):xxx sure of interest is the annual number of GRPs for each of the ten media markets in the state. GRPs are based on the percentage of the population potentially exposed to advertisements (reach) and the average number of times they may have been seen (frequency) over a time period. Annual GRPs were calculated by summing the current quarterly GRPs at the time of interview and the three previous quarters. Past-year GRP variables were divided by 5000 such that an OR represents the change in odds for an increase of 5000 GRPs. Smokers saw an average of three emotional and/or graphic and three comparison advertisements per month across the study period. Although state-purchased advertisements airing in New York between March 2008 and September 2008 as part of the Legacy Foundations EX national campaign are included, some nationally placed EX advertisements airing over this time frame were unavailable for inclusion. Self-reported recall, which has been shown to be highly correlated with media exposure,35 was measured with a series of items that briefly describe each ad airing in the quarter. Among those who indicated recognition, respondents were asked for additional details and coded as having confrmed ad recall if they provided accurate details. Recall of one or more ads for each survey wave was classifed as having self-reported recall. Following previous work,11,21,36 all advertisements were coded to indicate whether they included strong negative emotions (e.g., sadness, fear) and/or intense or graphic images. The advertisements include those that describe the dangers of smoking, promote cessation, and/or highlight the dangers of exposure to secondhand smoke. Advertisements containing strong negative emotions were often personal testimonials about negative health consequences for smokers or their loved ones. Examples of graphic images include diseased lungs, clogged arteries, and tracheotomies. Advertisements were reviewed and coded by at least two coders in multiple rounds to resolve differing characterizations. According to Landis and Kochs k-scale, an indicator of strong negative emotions or graphic images resulted in almost perfect rater agreement (0.811.00).37 All advertisements not coded as either strong negative emotion or graphic were included in analyses as comparison advertisements. This category predominantly includes advertisements that offer encouragement and/or advice on quitting (e.g., Quitting Takes Practice, Time to Quit, Quit to Live) and those that highlight the dangers of exposure to secondhand smoke (e.g., Baby seat, Its like theyre smoking). Although coders did not code specifcally for an effcacy message within each advertisement that might complement the fear appeals,23 all advertisements informed smokers about the states quitline. Of the 142 study advertisements, 98 (69%) were comparison and 44 (31%) were emotional and/or graphic.
12000 10000 8000 6000 30 4000 2000 0 20 10 0 70 60 50 40

3
Current smokers making a quit attempt (%)

Population weighted GRPs

200 200 200 200 200 200 200 201 3 9 4 5 7 0 6 8

Graphic and/or emotional advertisements Comparison advertisements Quit attempts

Figure 1. Annual advertising levels and proportions of smokers making a quit attempt, 20032010
Note: Data are from the New York Adult Tobacco Survey. GRP, gross rating point

was signifcant individually or together and did not influence the magnitude of the advertising effects. Regressions were stratifed by desire to quit (a lot, somewhat, a little/not at all); income ( $30,000 vs $30,000); and education (less than a high-school degree vs higher education levels). All regressions were weighted and accounted for the survey design using Stata, version 12, in 2012. GRP information is available for the full study period (Quarter 3 of 2003 through Quarter 4 of 2010), whereas complete confrmed recall data are available from Quarter 2 of 2004 through Quarter 4 of 2010. Thus, analyses include 8780 smokers for GRP models and 5936 smokers for confrmed recall models.

Results
The prevalence of making a quit attempt increased steadily from 2003 to 2007, paralleling the increase in GRPs, and then leveled off as GRPs began to decline (Figure 1). The level and proportion of emotional and/or graphic advertising exposure changed over time. Table 1 displays data from two regression analyses on the association between making a quit attempt in the past year and two measures of exposure to antismoking advertisements: self-reported recall and past-year GRPs. Current smokers who recall recently seeing at least one advertisement have an increased odds of making a quit attempt in the past year of 31% (p 0.01). For every increase of 5000 GRPs annually, the odds of making a quit attempt increase by 21% (p 0.01). Across these model specifcations, the relationship between making a quit attempt and the various potential confounders is generally consistent. The quantity of cigarettes smoked per day is

Logistic Regressions
To illustrate the overall trend in quit attempts over time and the relationship with overall GRPs and GRPs by advertising type, the statewide prevalence of making a quit attempt and the populationweighted average annual level of GRPs over time were plotted (Figure 1). Quit attempts were regressed on the confrmed recall and past-year GRPs, controlling for the potential confounders noted above. Quit attempts were then regressed on GRPs for emotional/graphic and comparison advertisements. Initial models included controls for market-level cigarette prices and an annual secular trend variable but were deleted from fnal models as neither Month 2012

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

Table 1. Advertising exposures inuence on quit attempts, New York Adult Tobacco inversely correlated with Survey, 20032010 the odds of making a quit attempt for both models in Conrmed recall Past-year GRPs Table 1 (p 0.01). Age is signifcant in the GRP modCovariates OR (95% CI) p-value OR (95% CI) p-value el; the odds of making a quit Conrmed recall 1.31 (1.08, 1.57) 0.005 attempt are lower among Past-year GRPs (X 5000) 1.21 (1.11, 1.31) 0.000 smokers aged 65 years Cigarettes smoked per day 0.000 0.000 compared with smokers aged 1824 years. Educa19 1.00 1.00 tion, race/ethnicity, and 1019 0.68 (0.54, 0.84) 0.67 (0.55, 0.80) gender are not predictors of 20 0.42 (0.34, 0.53) 0.44 (0.37, 0.53) quit attempts in the two Age (years) 0.120 0.026 models. Income is margin1824 1.00 1.00 ally signifcant in the confrmed recall model. 2539 0.76 (0.55, 1.06) 0.83 (0.64, 1.09) Table 2 illustrates the 4064 0.78 (0.58, 1.06) 0.78 (0.61, 1.00) differential effects of 65 0.62 (0.42, 0.92) 0.62 (0.45, 0.85) emotional and/or graphic Education 0.972 0.973 advertisements and comLess than high-school diploma 1.00 1.00 pares a set of advertisements using confrmed reHigh school or GED 1.03 (0.77, 1.38) 1.03 (0.81, 1.31) call and annual GRPs. Both Some college 1.05 (0.78, 1.42) 1.03 (0.80, 1.32) models indicate that greater College graduate 1.00 (0.71, 1.39) 0.99 (0.75, 1.30) exposure to emotional Race/ethnicity 0.791 0.507 and/or graphic advertiseWhite (non-Hispanic) 1.00 1.00 ments is positively associated with making a quit atBlack (non-Hispanic) 1.13 (0.88, 1.45) 1.14 (0.93, 1.41) tempt in the past 12 months, Hispanic 1.07 (0.78, 1.46) 1.16 (0.89, 1.51) whereas the comparison adOther 1.09 (0.66, 1.79) 1.01 (0.69, 1.49) vertisements are not associGender 0.759 0.831 ated with quit attempts. Female 1.00 1.00 Recalling at least one emotional or graphic advertiseMale 0.97 (0.81, 1.16) 1.02 (0.88, 1.18) ment recently is associated Income ($) 0.061 0.176 with a 29% increase in the 30,000 1.00 1.00 odds of making a quit at30,00059,999 0.91 (0.73, 1.14) 0.91 (0.76, 1.09) tempt (p 0.05), whereas each additional 5000 GRPs 60,00089,999 0.89 (0.67, 1.19) 0.87 (0.68, 1.10) of exposure to emotional 90,000 0.60 (0.43, 0.84) 0.69 (0.52, 0.93) and/or graphic advertiseMissing income 0.91 (0.67, 1.22) 0.93 (0.73, 1.19) ments in the past year is asConstant 2.31 (1.54, 3.47) 0.000 1.72 (1.22, 2.42) 0.002 sociated with a 38% increase in the odds of making a quit N 5936 8780 attempt (p 0.01). The relaNote: Regressions that use conrmed awareness as the key covariate cover the period April 2004 2010, tionship between making a whereas regressions that use GRPs as the key covariate cover the period 20032010. For conrmed recall, elasticity 0.06; for past-year GRPs, elasticity 0.04. quit attempt and the various GED, General Educational Development test; GRPs, gross rating points confounders is similar to modelspresentedinTable1. illustrated in Figure 2, and a parallel set of models using Finally, the association between making a quit attempt and measures of exposure was examined separately by deconfrmed recall are summarized herein. Based on annual sire to quit, income, and education. The ORs from models GRPs, greater exposure to advertising is associated with an examining annual GRPs overall and by message type are increased odds of making a past-year quit attempt among
www.ajpmonline.org

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

smokers who want to quit a lot or somewhat, smokers with incomes $30,000 and $30,000, and smokers with a highschool degree or less and those with more education. The association between confrmed recall and making a quit attempt differed by sociodemographic group, as only smokers who want to quit somewhat, smokers with incomes of $30,000, and smokers with some college education or beyond were more likely to make a quit attempt if they reported recall of advertisements. The stratifed analyses by message type indicate that emotional and/or graphic advertisements are associated with quit attempts for multiple subgroups of smokers (Figure 2). Recall of emotional and/or graphic advertisements was associated with making a quit attempt for smokers with incomes $30,000 and those with a high-school degree or less (p 0.05), but not for any other groups (data not shown). Exposure to the comparison advertisements, as measured by pastyear GRPs and confrmed recall, was not associated with quitting for any group of smokers.

Table 2. Advertising exposures inuence on quit attempts by advertising type, New York Adult Tobacco Survey, 20032010
Conrmed recall Covariates Conrmed recallemotional and/or graphic advertisements Conrmed recallcomparison advertisements Annual GRPsemotional and/or graphic advertisements (X 5000) Annual GRPscomparison advertisements (X 5000) Cigarettes smoked per day 19 1019 20 Age (years) 1824 2539 4064 65 Education Less than high-school diploma High school or GED Some college College graduate Race/ethnicity White (non-Hispanic) Black (non-Hispanic) Hispanic Other Gender Female Male Income ($) 30,000 30,00059,999 60,00089,999 90,000 Missing income Constant N 1.00 0.89 (0.71, 1.12) 0.91 (0.67, 1.22) 0.60 (0.43, 0.85) 0.94 (0.69, 1.27) 2.45 (1.62, 3.70) 5801 0.000 1.00 0.98 (0.82, 1.17) 0.073 1.00 0.91 (0.76, 1.09) 0.86 (0.68, 1.09) 0.68 (0.51, 0.92) 0.92 (0.72, 1.17) 1.79 (1.27, 2.52) 8780 0.001 1.00 1.11 (0.86, 1.43) 1.07 (0.78, 1.47) 1.11 (0.67, 1.83) 0.830 1.00 1.01 (0.88, 1.17) 0.148 1.00 1.04 (0.78, 1.40) 1.05 (0.78, 1.42) 0.98 (0.70, 1.37) 0.851 1.00 1.13 (0.91, 1.39) 1.13 (0.87, 1.48) 0.99 (0.68, 1.45) 0.846 1.00 0.75 (0.53, 1.05) 0.76 (0.56, 1.04) 0.59 (0.40, 0.88) 0.945 1.00 1.02 (0.80, 1.30) 1.02 (0.79, 1.31) 0.97 (0.74, 1.28) 0.609 1.00 0.65 (0.52, 0.83) 0.41 (0.33, 0.52) 0.076 1.00 0.83 (0.64, 1.08) 0.78 (0.61, 0.99) 0.61 (0.44, 0.84) 0.967 0.000 1.00 0.67 (0.56, 0.80) 0.45 (0.37, 0.54) 0.019 OR (95% CI) 1.29 (1.05, 1.58) p-value 0.016 Past-year GRPs OR (95% CI) p-value

1.00 (0.78, 1.28)

0.992

1.38 (1.20, 1.58)

0.000

1.01 (0.87, 1.17)

0.864

0.000

Discussion
Exposure to New York TCPs antismoking TV advertisements, measured as self-reported recall or as a market-level dose of advertising, is associated with increased odds
Month 2012

Note: Regressions that use conrmed awareness as the key covariate cover the period April 2004 2010, whereas regressions that use GRPs as the key covariate cover the period 20032010. For conrmed recall, elasticity 0.06 (emotional and/or graphic advertisements) and 0.00 (comparison advertisements). For annual GRPs, elasticity 0.08 (emotional and/or graphic advertisements) and 0.00 (comparison advertisements). GED, General Educational Development test; GRPs, gross rating points

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

of making a quit attempt OR (95% CI) in the past year. Although Some college (n=4612) 1.20 (1.08, 1.34) Wakefeld and colHigh school (n=4168) 1.23 (1.09, 1.39) leagues2 found that the effect of discrete periods 1.17 (1.05, 1.31) $30,000 (n=4436) of antismoking advertising dissipate after 3 1.24 (1.07, 1.43) < $30,000 (n=3279) months, these results sug1.18 (0.98, 1.42) A little/Not at all (n=2454) gest that sustained periods of advertising afSomewhat (n=2594) 1.23 (1.06, 1.44) fected quit attempts over A lot (n=2454) 1.15 (1.00, 1.32) 1 year. However, measures of quit attempts Some college (n=4612) 1.23 (1.03, 1.48) over shorter time frames were unavailable to ex1.54 (1.26, 1.89) High school (n=4168) amine the possibility of 1.24 (1.03, 1.49) $30,000 (n=4436) diminishing GRP effects over the year. 1.40 (1.10, 1.79) < $30,000 (n=3279) Although it is likely 1.40 (1.00, 1.96) A little/Not at all (n=2454) that smokers quickly forget campaigns after they 1.36 (1.04, 1.77) Somewhat (n=2594) go off the aira notion supported by marked de1.31 (1.03, 1.68) A lot (n=2454) clines in quitline call vol1.15 (0.94, 1.42) Some college (n=4612) ume after a campaign endsit is also possible 0.92 (0.75, 1.14) High school (n=4168) that sustained exposure 1.09 (0.89, 1.34) $30,000 (n=4436) to advertising shifts attitudes and beliefs that sub< $30,000 (n=3279) 1.05 (0.83, 1.34) sequently lead to behavior change. Studies A little/Not at all (n=2454) 0.94 (0.66, 1.34) examining the influence Somewhat (n=2594) 1.09 (0.83, 1.43) of the truth antismoking campaign have found 0.96 (0.76, 1.21) A lot (n=2454) that cumulative exposure 0 1 2 to the campaign was associated with decreased Figure 2. Inuence of advertising exposure on quit attempts in the past 12 months, smoking prevalence32 2004 2010 33 and smoking initiation. Note: Data are from the New York Adult Tobacco Survey. GRP, gross rating point The present study shows that antismoking advertisements that have strong negative emotions or graphic images are more the goal is to motivate smokers to make quit attempts, effective in motivating quit attempts by adult smokers antismoking advertisements should be designed to elicit a than other comparison advertisements. This pattern is strongly negative emotional reaction. State and federal illustrated in Figure 1 where the rate of population quit public health agencies implementing antismoking camattempts is more related to the level of emotional and/or paigns should air hard-hitting campaigns, despite critigraphic advertising than the overall level of advertising. cisms that they go too far.39 This result supports the growing evidence that televised New Yorks advertising was equally effective regardless messages using emotional personal testimonials about of desire to quit, income, or education level. Similar to the consequences of smoking or graphic images to depict fndings by Durkin and colleagues,6 smokers exposed to the negative health consequences of smoking are effec1,6,11,38 tive. either highly emotional or graphic advertisements drove The implications are fairly straightforward; if
Past year GRPs Past year GRPs - graphic and/or emotional advertisements Past year GRPs - comparison advertisements Desire to quit Income Education Desire to quit Income Education Desire to quit Income Education

www.ajpmonline.org

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

the overall fndings. The emotional and/or graphic advertisements were effective with low-income and loweducation smokers, whereas the comparison advertisements were not. These results suggest that well-funded, hard-hitting antismoking campaigns can promote cessation among populations with historically high smoking rates. The New York TCP also provides a telephone quitline, offers free nicotine replacement therapy, and supports community-based mobilization. These efforts combined with a high cigarette excise tax and comprehensive smokefree laws likely increase the effectiveness of the antismoking campaigns.28 The emotional and graphic advertisements were effective also among those with varying levels of interest in quitting. The fndings suggest that concerns about the use of fear appeals25 may be misplaced given the growing body of evidence for their effectiveness in real-world settings. The current study has several limitations. First, the analysis relies on a repeated cross-sectional design. A longitudinal study similar to others in this area2,6,17 would provide stronger conclusions about the effectiveness of advertising. Second, although the study makes use of variation in exposure to advertising across ten media markets over a 7-year period, the results are based on the experience of a single state, limiting our ability to easily account for concomitant tobacco control activities or rule out alternative explanations. However, the similarity of results using self-reported recall and advertising dose suggests that the current results are robust. Increasing the rate of quit attempts at the population level is a critical component of reductions in overall smoking prevalence.40 This study strengthens existing evidence that antismoking campaigns can influence smoking cessation. In addition, there is growing evidence that advertisements that feature strong, negative emotions and/or graphic images are most effective in eliciting behavior change overall and among disadvantaged populations.
Support for this work came from the New York State Department of Health. No fnancial disclosures were reported by the authors of this paper.

3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13. 14.

15.

16. 17.

18.

19. 20.

21.

References
1. Durkin S, Brennan E, Wakefeld M. Mass media campaign to promote smoking cessation among adults: an integrative review. Tob Control 2012;21(2):12738. 2. Wakefeld MA, Spittal MJ, Yong HH, Durkin SJ, Borland R. Effects of mass media campaign exposure intensity and durability on quit at-

22.

23. 24.

tempts in a population-based cohort study. Health Educ Res 2011;26(6):988 97. Wakefeld MA, Loken B, Hornik RC. Use of mass media campaigns to change health behaviour. Lancet 2010;376(9748):126171. National Cancer Institute. The role of the media in promoting and reducing tobacco use. In: Tobacco Control Monograph No 19. NIH Pub No 07-6242. Bethesda MD: NIH, 2008. Niederdeppe J, Kuang X, Crock B, Skelton A. Media campaigns to promote smoking cessation among socioeconomically disadvantaged populations: what do we know, what do we need to learn, and what should we do now? Soc Sci Med 2008;67:134355. Durkin SJ, Biener L, Wakefeld MA. Effects of different types of antismoking ads on reducing disparities in smoking cessation among socioeconomic subgroups. Am J Public Health 2009;99:221723. Carroll T, Rock B. Generating Quitline calls during Australias National Tobacco Campaign: effective of television advertisement execution and programme placement. Tob Control 2003;12(S2):ii40 ii44. Cowling DW, Modayil MV, Stevens C. Assessing the relationship between ad volume and awareness of a tobacco education media campaign. Tob Control 2010;19(S1):i37i42. Erbas B, Bui Q, Huggins R. Investigating the relation between placement of Quit antismoking advertisements and number of telephone calls to Quitline: a semiparametric modelling approach. J Epidemiol Community Health 2006;60:180 2. Farrelly MC, Hussin A, Bauer UE. Effectiveness and cost effectiveness of television, radio and print advertisements in promoting the New York smokers quitline. Tob Control 2007;16(S1):i21i23. Farrelly MC, Davis KC, Nonnemaker JM, Kamyab K, Jackson C. Promoting calls to a quitline: quantifying the influence of message theme, strong negative emotions and graphic images in television advertisements. Tob Control 2011;20(4):279 84. Miller CL, Wakefeld MA, Roberts L. Uptake and effectiveness of the Australian Telephone Quitline service in the context of a mass media campaign. Tob Control 2003;12(S2):ii53ii58. Mosbaek CH, Austin DF, Stark MJ. The association between advertising and calls to a tobacco quitline. Tob Control 2007;16(S1):i24 i29. Pierce JP, Anderson DM, Romano RM. Promoting smoking cessation in the U.S.: effect of public service announcements on the Cancer Information Service telephone line. J Natl Cancer Inst 1992;84:677 83. Wilson N, Grigg M, Graham L. The effectiveness of television advertising campaigns on generating calls to a national Quitline by Maori. Tob Control 2005;14:284 6. North American Quitline Consortium. 2008/2009 annual report: moving quitlines forward. Phoenix AZ: NAQC; 2009. Hyland A, Wakefeld MA, Higbee C, Szczypka G, Cummings KM. Anti-tobacco television advertising and indicators of smoking cessation in adults: a cohort study. Health Educ Res 2006;21(2):296 302. Emery S, Yoonsang K, Choi Y, Szcaypka G, Wakefeld M, Chaloupka F. The effects of smoking-related television advertising on smoking and intentions to quit among adults in the U.S.: 1999 2007. Am J Public Health 2012;102(4):7517. Lang A, Newhagen J. Negative video as structure: emotion, attention, capacity, and memory. J Broadcast Electron 1996;40:460 77. Nabi RL. A cognitive-functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Commun Theory 1999;9:292320. Beiner L, McCallum-Keeler G, Nyman AL. Adults response to Massachusetts anti-tobacco television advertisements: impact of viewer and advertisement characteristics. Tob Control 2000;9(4):4017. Davis KC, Nonnemaker JM, Farrelly MC, Niederdeppe J. Exploring differences in smokers perceptions of the effectiveness of cessation media messages. Tob Control 2011;20(1):26 33. Witte K. Putting the fear back into fear appeals: the extended parallel process model. Commun Monogr 1992;59:329 49. Witte K, Allen M. Meta-analysis of fear appeals: implications for effective public health campaigns. Health Educ Behav 2000;27:591 615.

Month 2012

Farrelly et al / Am J Prev Med 2012;xx(x):xxx

33. Farrelly MC, Nonnemaker JM, Davis KC, Hussin A. The influence of the national truth campaign on smoking initiation. Am J Prev Med 2009;36(5):379 84. 34. Davis KC, Nonnemaker JM, Farrelly MC. Association between national smoking prevention campaigns and perceived smoking prevalence among youth in the U.S. J Adolesc Health 2007;41(5):430 6. 35. Southwell BG, Barmada CH, Hornik RC, Maklan DM. Can we measure encoded exposure? Validation evidence from a national campaign. J Health Commun 2002;7(5):44553. 36. Niederdeppe J, Davis KC, Farrelly MC, Yarsevich J. Stylistic features, need for sensation, and confrmed recall of national smoking prevention advertisements. J Commun 2007;57(2):27292. 37. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33(1):159 74. 38. Wakefeld M, Balch GI, Ruel E, et al. Youth responses to anti-smoking advertisements from tobacco-control agencies, tobacco companies, and pharmaceutical companies. J Appl Soc Psychol 2005;35(9): 1894 911. 39. Inbar M. Do new antismoking ads go too far? today.msnbc.msn.com/ id/30008941/ns/today-today_health/t/do-new-anti-smoking-ads-gotoo-far/#.Tw366CMpjKg. 40. Zhu SH, Lee M, Zhuang YL, Gamst A, Wolfson T. Interventions to increase smoking cessation at the population level: how much progress has been made in the last two decades? Tob Control 2012;21(2):110 8.

25. Hastings G, Stead M, Webb J. Fear appeals in social marketing: strategic and ethical reasons for concern. Psychol Market 2004;21:961 86. 26. Hastings G, MacFadyen L. Controversies in tobacco control: the limitations of fear messages. Tob Control 2002;11:73e5. 27. Mariolis P, Rock VJ, Asman K, Merritt R, Malarcher A, Husten C. Tobacco use among adults: U.S., 2005. MMWR Morb Mortal Wkly Rep 2006;55:1145 8. 28. Niederdeppe J, Fiore MC, Baker TB, Smith SS. Smoking-cessation media campaigns and their effectiveness among socioeconomically advantaged and disadvantaged populations. Am J Public Health 2008;98(5):916 24. 29. Lemelle A, Reed W, Taylor S, eds. Handbook of African American health: social and behavioral interventions. Dordrecht: Springer Science & Business Media, LLC 2011. 30. Davis KC, Farrelly MC, Duke J, Kelly L, Willett J. The implementation of health communication interventions for tobacco control and their potential impact on cessation outcomes: evidence from New York State, 2003 2009. Prev Chronic Dis 2012. 31. Nonnemaker J, Hersey J, Homsi G, et al. Self-reported exposure to policy and environmental influences on smoking cessation and relapse: a 2-year longitudinal population-based study. Int J Environ Res Public Health 2011;8(9):3591 608. 32. Farrelly MC, Davis KC, Haviland ML, Messeri P, Healton CG. Evidence of a dose-response relationship between truth antismoking ads and youth smoking prevalence. Am J Public Health 2005;95(3):42531.

www.ajpmonline.org