Health Education Research Advance Access published August 13, 2014
HEALTH EDUCATION RESEARCH
2014 Pages 1–11
Male smoker and non-smoker responses to television advertisements on the harms of secondhand smoke in China, India and Russia
1
World Lung Foundation, 61 Broadway, New York, NY 10006, USA, 2Centre for Behavioural Research in Cancer, Cancer Council Victoria, 615 St Kilda Road, Melbourne 3004, VIC, Australia and 3World Lung Foundation, Alexandria, NSW, Australia *Correspondence to: N. Murukutla. E-mail: [email protected] The International Anti-SHS Advertisement Rating Study Team comprises Cancer Council Victoria: Melanie Wakefield, Megan Bayly, Sarah Durkin; World Lung Foundation (WLF): Sandra Mullin, Nandita Murukutla, Trish Cotter; WLF, China: Yvette Chang, Winnie Chen; WLF, India: Shefali Gupta, Tahir Turk; WLF, Russia: Irina Morozova, Rebecca Perl. y
Received on September 2, 2013; accepted on July 15, 2014
Abstract Mass media campaigns can play an important role in strengthening support for smoke-free policies and reducing exposure to secondhand smoke (SHS). Identifying anti-SHS advertisements that are effective in diverse cultural contexts may allow for resource sharing in low- and middle-income countries. A convenience sample of 481 male cigarette smokers and non-smokers in three high tobacco burden and culturally dissimilar countries (India, China and Russia) viewed and rated five anti-SHS ads. Multivariate logistic regression analyses were conducted for ‘Message Acceptance’, ‘Negative Emotion’, ‘Perceived Effectiveness’ and ‘Behavioral Intentions’. Smokers and non-smokers in all countries consistently rated the strong graphic, health harm ads as the most effective, and the ‘informational’ ad as the least effective overall: the graphic ad ‘Baby Alive’ was at least 1.8 times more likely than the informational ad ‘Smokefree works’ to receive positive ratings on all four outcomes (all P < 0.001). Graphic, health harm messages about SHS exposure have the greatest universal appeal and are the most effective in motivating changes in behavioral
intentions. Similarity in reactions between smokers and non-smokers, and across countries, suggests that resource sharing and the use of a single graphic ad targeted at smokers and non-smokers would be cost-efficient strategies.
Introduction Secondhand smoke (SHS), or the smoke emitted by burning tobacco products, results in significant morbidity and mortality [1, 2]. Laws that ban smoking in indoor public places are an effective intervention in protecting people from SHS [1–4]. However, smoke-free laws are still non-existent, weak or are poorly enforced and SHS exposure continues to be high, particularly in low- and middle-income countries (LMICs) [5]. Mass media campaigns are a core, proven intervention for tobacco control [6, 7]. In addition to reducing tobacco uptake and prevalence, they can also serve a crucial role by creating awareness about SHS harms, changing social attitudes and practices around smoking and encouraging cessation [2]. Additionally, by changing social attitudes and practices, and creating public support for tobacco control, they can potentially help in the implementation of strong laws [7–11].
ß The Author 2014. Published by Oxford University Press. All rights reserved. For permissions, please email: [email protected]
doi:10.1093/her/cyu044
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Nandita Murukutla1*, Megan Bayly2, Sandra Mullin1, Trish Cotter3 and Melanie Wakefield2; for the International Anti-SHS Advertisement Rating Study Teamy
N. Murukutla et al.
Methods Three culturally dissimilar LMICs with high tobacco burdens, including SHS exposure, were selected. Given the high prevalence of smoking in men in developing countries relative to women (i.e. 43% men versus 2% women in China, 24% men versus 3% women in India, 60% men versus 22% women in Russia), men are a priority group in communication about SHS exposure in these countries [3, 4]. Our study therefore included male smokers and nonsmokers from China, India and Russia. 2 of 11
Design Using a standard ad rating methodology [12, 16], this study was conducted in 2010–2011 in one large city in each country: Changsha, China; New Delhi, India; Moscow, Russia. Participants rated the effectiveness of 10 ads: 5 that were shown in all countries and 5 that were selected by local stakeholders and differed across countries. This article focuses on the five ads shown in all countries.
Participants In each country, male cigarette smokers and nonsmokers aged 18–40 years were recruited. A local market research agency from each city completed recruitment using convenience-sampling methods that were appropriate for each country—faceto-face or telephone interviewing. Smokers of at least 5 cigarettes per day who had smoked for at least 1 year were targeted in recruitment. Eligible non-smokers were never-smokers or had smoked fewer than 100 cigarettes in their lifetime. Eligible smokers smoked at least five cigarettes daily. Participants were assigned to groups split by age (18–29/30–40 years), smoking status (smokers/nonsmokers) and ad order (order A/B). There were 8–12 participants in each group.
Advertisements All ads presented SHS messages with varying message style (Table I). These ads were selected from a resource of tobacco control ads (see http://www.worldlungfoundation.org/mmr) based on their effective performance in their country of origin [14, 17] and/or their suitability for adaptation. They were carefully translated (and back translated) and dubbed into local languages. On-screen text and end frames were replaced or deleted. The ads were counterbalanced in two orders, one being the reverse of the other. The country consistent ads were alternated with the country choice ads.
Ad rating questionnaire The non-smoker and smoker versions of the questionnaire were nearly identical. Both started
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
However, for mass media campaigns to be effective, they must use advertisements that resonate with and persuade the intended audience [8]. For campaigns targeted at active cigarette smoking, health messages that are graphic and negativeemotion arousing are most effective in encouraging smokers to quit [8]. These kinds of messages perform well not only in high-income countries (HICs) but also in LMICs [12], leading to the evolution of a cost-effective approach of taking readily available high-performing ads from HICs and minimally adapting them for effective use in LMICs for active smoking campaigns [13]. However, to date effective message styles for campaigns on ‘passive’ smoking and the health harms of SHS exposure have not been identified [14]. Furthermore, the adaptability of SHS ads from one country to another is yet to be determined. Effective mass media campaigns are critical, especially in LMICs with overburdened health systems, to reduce the burdens of SHS exposure and build support for effective smoke-free policies [15]. Thus, this study was designed to test the acceptability, relevance, perceived effectiveness and potential for behavior change of five television advertisements on SHS harms that had high potential for impact and adaptation. We sought to compare the performance of these ads between smokers and non-smokers both within and across LMICs with the aim of determining whether adaptation of existing SHS advertisements might be similarly effective for both smoker and non-smokers.
Male smoker and non-smoker responses to secondhand smoke television advertisements Table I. Descriptions of the SHS television ads Ad characteristics
Description
Cigarettes are eating you and your baby alive (Baby Alive)
Graphic and visceral ad focused on harms to children
Clinical
Strong graphic ad focused on harms to adults
Smoke-free works: Office
Informational style, focused on implementation of smoking ban in workplaces
Child
Narrative style, suggesting harm to the smoker and his family
Say no to SHS
Dramatic simulation, depicting toxic smoke transmitting from the smoker to non-smokers
Eerie music plays throughout. We see the silhouettes of a man and woman smoking cigarettes in front of a child. Their lungs are transposed on their silhouettes. A voiceover (VO) explains that cigarette smoke contains poisons that trigger severe health problems. Cut to images of real babies hooked to respirators. As the VO says that cigarette smoke can cause ear infections, asthma and pneumonia, we see shots of children with each of these conditions. Cut back to the child’s silhouette, then images of babies with low birth weight. The VO says that cigarette smoke is linked to low birth weight and doubles the risk of Sudden Infant Death Syndrome. Cut back to the silhouettes, then quick flashes of the sick children A non-smoking woman at a party is shown inhaling SHS. The camera follows the smoke down her throat and into her lungs, and simulated imagery of SHS being absorbed into the blood stream and damaging vessels is shown. VO: Secondhand smoke enters the throat. And it doesn’t just damage the lungs: Within minutes the smoke activates platelets, which thicken the blood and damages arteries. In time plaque accumulates in the damaged blood vessels, obstructing blood flow to the heart. This can result in stroke or heart attack. Text: A little secondhand smoke can do a lot of damage A VO announces that offices will be smoke-free beginning in March. In an office setting, we hear an employee light a cigarette and exhale. Zoom in on the chest of a coworker. A smoke-filled lung appears overlaid on the coworker’s shirt. Text states that secondhand smoke causes lung cancer. Zoom in on another office employee’s chest. An unhealthy heart and lungs appear overlaid on her shirt. Text states that secondhand smoke causes heart disease A man is seen smoking in different settings: at the dinner table, with a colleague, watching television with his daughter. An ill man is shown on the television, and the young daughter looks at her father with concern. He leaves the room and returns without his cigarette. VO: Everyone seeks happiness. But isn’t that a very high price to pay? Smoking can cost you and can also seriously affect the health and welfare of those around you. Text: Smoking kills. Passive smoking kills too. VO: Don’t have them pay the price for your smoking. Smoking will cost you The ad opens on a close-up of a lit cigarette. Pan out to see a man exhaling cigarette smoke. Pan out further to see his wife then also exhale cigarette smoke, and then his two young children also exhale smoke. The ad finishes with a close-up of daughter surrounded by a cloud of smoke. VO: The smoke from cigarettes contains 4000 hazardous chemicals, affecting innocent victims and slowly taking their lives. Secondhand smoke kills. Say no to secondhand smoke
Ads may be viewed at: http://67.199.72.89/mmr/english/ads_shs.html.
by assessing demographics, including gender, age, parental status, education level, residence with a smoker and rules around smoking indoors in their home. Smokers also reported number of cigarettes
smoked daily, intentions to quit in the next 12 months and any previous quit attempts. All ads were rated on an 11-item questionnaire using a 5-point scale: 1 ‘strongly disagree’, 2 3 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Ad name
N. Murukutla et al.
Procedure After completing the demographic questionnaire, participants were familiarized with the rating questionnaire by initially rating a ‘practice ad’ (an unrelated, generic ad, e.g. a shampoo ad). Thereafter, each SHS ad was viewed twice and privately rated before moving to the next.
Data analysis Frequency tables and 2 tests were produced to examine equivalency of the sample across countries and smoking status. Responses on the ad rating items were binary coded. Ratings of strongly agree and somewhat agree were coded as positive responses, while strongly and somewhat disagree and neither disagree or agree responses were coded as negative/neutral. Responses for the multiitem scales were summed and divided by the number of items in that scale. Average responses 4 of 11
of greater than 3.5 were coded as positive, and all other responses were coded as negative/neutral [12]. Multivariate logistic regression analyses were conducted using Stata/SE 12.1 for each outcome measure, for non-smokers and smokers combined. All models used robust standard errors to control for each participant having rated multiple ads. All models also controlled for age (18–29 or 30–40 years), education level (completed a tertiary degree or no tertiary degree), parental status (yes or no), whether smoking was allowed indoors in their home (yes or no), whether they lived with another smoker (yes or no), country (China, Russia, India), ad order (A or B) and all ads. Post-estimation Wald w2 tests were used to examine main effects of ad and smoking status. Two-way interactions were performed between ad and smoking status to examine whether responses to each ad were consistent between smokers and non-smokers, across the whole sample and within each country. Adjusted proportions of positive ad ratings were calculated for each smoking status group within each country from these models. Qualitative data from the structured discussions were consistent with the quantitative ratings and were primarily intended to guide local adaptations of the ads; hence, they are not discussed further here.
Results Sample characteristics A total of 481 participants were recruited (Table II). A significantly higher proportion of Indian smokers than non-smokers had completed a tertiary degree. Across all countries, significantly more smokers than non-smokers reported living with another smoker. Smoking inside the home was allowed more often for smokers than non-smokers in India and Russia. One-third of smokers reported smoking more than 15 cigarettes per day; however, this varied from 51% in China and Russia to 5% in India. While adherence to recruitment criteria was generally strong, 11.3% of Chinese smokers and 18.5% of Indian smokers reported smoking fewer than 5 cigarettes per day.
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
‘slightly disagree’, 3 ‘neither agree or disagree’, 4 ‘slightly agree’ and 5 ‘strongly agree’. The items were reduced to four subscales: ‘Message Acceptance’ (consisting of the item ‘ad was easy to understand’); ‘Negative Emotion’ (‘ad made me feel uncomfortable’); ‘Perceived Effectiveness’ (‘ad taught me something new’/‘made me stop and think’/‘was believable’/‘was relevant to me’/‘made me likely to talk to someone else about this ad’/ ‘made feel more concerned about smoking around other adults’ [smokers only]/‘made me feel more concerned about smoking around children’ [smokers only]/‘made me feel more concerned about others smoking around me’ [non-smokers only]/‘made me feel concerned about others smoking around children’ [non-smokers only]); ‘Behavioral Intention’ (‘makes me more likely to discourage smoking inside my home’ [nonsmokers]/‘makes me more likely to avoid places where people smoke’ [non-smokers]/‘more likely to avoid smoking indoors’ [smokers]/‘more likely to try to quit’ [smokers]). All scales showed strong internal reliability within each smoking status group (Cronbach’s a > 0.72).
Male smoker and non-smoker responses to secondhand smoke television advertisements Table II. Demographic characteristics of the sample China
Russia
Total sample
Nonsmokers
Smokers
Nonsmokers
Smokers
Nonsmokers
Smokers
Nonsmokers
Smokers
80
80
80
81
80
80
240
241
50 50 28 35 39 38
50 50 19 41 48 80***
51 49 36 40 15 9
49 51 53* 42 37** 28**
50 50 58 30 26 19
50 50 54 33 53** 50***
50 50 40 35 27 22
50 50 42 39 46*** 53***
– – – –
49 51 20 60
– – – –
95 5 35 51
– – – –
49 51 38 73
– – – –
64 36 31 61
50 50
50 50
50 50
51 49
50 50
50 50
50 50
50 50
Significant difference between smokers and non-smokers within country or total sample at *P < 0.05; **P < 0.01; ***P < 0.001.
Ad ratings Main effects Across countries, there was a significant main effect of ad on all outcomes (Table III): Message Acceptance (2 ¼ 21.57, P < 0.001); Negative Emotion (2 ¼ 117.36, P < 0.001); Perceived Effectiveness (2 ¼ 144.02, P < 0.001) and Behavioral Intention (2 ¼ 40.74, P < 0.001). On all outcomes, ‘Baby Alive’ had the highest odds of a positive rating relative to the referent ad ‘Smokefree Works’. Smoke-free Works had at least marginally lower odds of positive ratings than all other ads for all outcomes. Significant main effects of country were found on all outcomes: Message Acceptance (2 ¼ 14.82, P < 0.001); Negative Emotion (2 ¼ 13.50, P ¼ 0.001); Perceived Effectiveness (2 ¼ 43.76, P < 0.001) and Behavioral Intention (2 ¼ 113.87, P < 0.001). Participants from India had the highest odds of positive ratings on all measures. Participants from Russia had the lowest odds of positive ratings on Perceived Effectiveness and Behavioral
Intention, while China had the lowest odds for Message Acceptance and Negative Emotion. A main effect of smoking status was found only for Behavioral Intention, where smokers were 61% less likely to rate ads positively than non-smokers (odds ratio [OR] 0.39, 95% confidence interval [CI]: 0.27–0.56, P < 0.001). Participants aged 18–29 years had significantly lower odds than older participants of rating ads positively on Negative Emotion (OR 0.60, 95% CI: 0.42–0.87, P ¼ 0.007). Parents had significantly higher odds of ratings ads positively on Behavioral Intention than those who did not have children (OR 1.84, 95% CI: 1.18–2.87, P ¼ 0.007), while those who lived in homes where smoking was permitted indoors had lower odds of positive ratings on this measure than those who reported smoking was not allowed inside their home (OR 0.60, 95% CI: 0.41–0.87, P ¼ 0.008). No effects of ad order, education or residing with a smoker were found for any outcome. These models were repeated including only smokers to examine potential effects of smoker-only variables. 5 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
N Age group (%) 18–29 years 30–40 years Completed a tertiary degree (% yes) Parent (% yes) Smoking allowed inside home (% yes) Live with a smoker (% yes) Cigarettes smoked daily (%) 1–15 15 or more Thinking about quitting (% yes) Previously tried to quit (% yes) Ad order (%) Order A Order B
India
N. Murukutla et al. Table III. Multivariate logistic regression analysis of positive ad ratings for measures of Message Acceptance, Negative Emotion, Perceived Effectiveness and Behavioral Intentiona Negative Emotion
Perceived Effectiveness
Behavioral Intention
OR
(95% CI)
OR
(95% CI)
OR
(95% CI)
OR
(95% CI)
1.00 4.54 1.44 1.61 1.61
(2.35–8.78)*** (0.96–2.17)t (1.00–2.60)* (1.00–2.60)t
1.00 3.44 2.38 1.50 2.30
(2.68–4.42)*** (1.93–2.93)*** (1.23–1.83)*** (1.86–2.85)***
1.00 5.87 2.83 1.80 2.08
(4.29–8.03)*** (2.22–3.59)*** (1.43–2.27)*** (1.64–2.64)***
1.00 1.83 1.70 1.62 1.62
(1.47–2.28)*** (1.40–2.06)*** (1.31–2.00)*** (1.29–2.02)***
1.00 3.41 1.28
(1.82–6.38)*** (0.79–2.07)
1.00 1.90 1.13
(1.32–2.74)** (0.82–1.56)
1.00 2.00 0.45
(1.28–3.13)** (0.31–0.67)***
1.00 1.59 0.19
(1.01–2.51)* (0.12–0.28)***
1.00 1.06
(0.65–1.73)
1.00 1.01
(0.76–1.32)
1.00 0.79
(0.55–1.13)
1.00 0.39
(0.27–0.56)***
1.00 0.71
(0.40–1.27)
1.00 0.60
(0.42–0.87)**
1.00 0.81
(0.52–1.26)
1.00 0.98
(0.64–1.52)
1.00 0.89
(0.49–1.62)
1.00 1.29
(0.88–1.87)
1.00 1.25
(0.79–1.97)
1.00 1.84
(1.18–2.87)**
1.00 1.18
(0.72–1.93)
1.00 0.85
(0.63–1.13)
1.00 0.85
(0.59–1.23)
1.00 0.60
(0.41–0.87)**
1.00 1.13
(0.73–1.75)
1.00 0.93
(0.71–1.20)
1.00 1.19
(0.85–1.66)
1.00 0.95
(0.67–1.34)
1.00 0.84
(0.54–1.31)
1.00 0.82
(0.62–1.08)
1.00 0.90
(0.63–1.28)
1.00 0.89
(0.62–1.27)
1.00 1.20
(0.76–1.90)
1.00 1.00
(0.74–1.35)
1.00 1.06
(0.72–1.55)
1.00 0.79
(0.53–1.18)
OR: odds ratio; 95% CI: 95% confidence interval. a Behavioral intention is combined for smokers and non-smokers to allow for comparison of smoking status effects. This model was repeated for smokers and non-smokers separately, and no important differences to the overall model were observed. Significantly different to reference group at *P < 0.05; **P < 0.01; ***P < 0.001; tdenotes a trend at P ¼ 0.050–0.099.
Participants who smoked more than 15 cigarettes per day were significantly less likely to give positive ratings on Negative Emotion than those who smoked 15 or fewer cigarettes (OR 0.63, 95% CI: 0.40–0.99, P ¼ 0.045). Smokers who were considering quitting in the next 12 months were more likely to give positive Behavioral Intention ratings compared with those who were not (OR 2.25, 95% 6 of 11
CI: 0.12–4.13, P ¼ 0.009). The significant main effects of ad and country were still observed in these additional analyses.
Ad ratings, by smoking status and country Table IV shows the adjusted proportions of positive ad ratings by smoking status and country for all
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Ad Smoke-free Works (ref) Baby Alive Clinical Child Say no to SHS Country China (ref) India Russia Smoking status Non-smoker (ref) Smoker Age (years) 30–40 (ref) 18–29 Parent No (ref) Yes Smoking in home No (ref) Yes Ad order Order B (ref) Order A Tertiary degree No (ref) Yes Live with a smoker No (ref) Yes
Message Acceptance
Male smoker and non-smoker responses to secondhand smoke television advertisements Table IV. Adjusted percentages of positive ad ratings for Message Acceptance, Negative Emotion, Perceived Effectiveness and Behavioral Intention, by smoking status and country (including Wald tests for interactions between ad and smoking status)a India
Russia
Total sample
Non-smokers Smokers Adjusted %
Non-smokers Smokers Non-smokers Smokers Non-smokers Smokers Adjusted % Adjusted % Adjusted %
Message Acceptance Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 9.23, P ¼ 0.026 96 100b 93 89 91 88 84 99* 88 94
2 ¼ 3.00, P ¼ 0.223 100b 97 98 94 95 97 99 100b 95 96
2 ¼ 5.54, P ¼ 0.237 97 97 95 92 91 85 90 94 94 99
2 ¼ 17.01, P ¼ 0.002 99 97 95 92t 92 90 91 98** 92 96
Negative Emotion Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 8.00, P ¼ 0.092 57 59 68 64 33 40 43 55 63 51
2 ¼ 11.40, P ¼ 0.022 64 78t 71 67 58 56 70 71 73 63
2 ¼ 10.24, P ¼ 0.037 85 82 63 44* 35 35 32 44 62 61
2 ¼ 21.40, P < 0.001 68 74 67 59t 41 45 47 57* 65 60
Perceived Effectiveness Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 9.68, P ¼ 0.043 99 87* 89 88 68 62 80 85 73 80
2 ¼ 1.36, P ¼ 0.851 99 94 95 85t 87 80 91 81t 94 88
2 ¼ 3.92, P ¼ 0.417 81 86 73 68 50 46 57 64 66 71
2 ¼ 6.08, P ¼ 0.193 93 89 85 81t 68 63t 76 77 77 80
Behavioral Intention Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 2.80, P ¼ 0.591 80 81 80 73 62 62 73 78 76 74
2 ¼ 12.83, P ¼ 0.012 99 69** 95 74** 91 64*** 94 81* 97 74**
2 ¼ 1.16, P ¼ 0.884 54 28** 55 27** 45 23** 48 24** 49 28*
2 ¼ 5.39, P ¼ 0.250 77 61*** 76 59*** 65 51*** 71 63** 74 60***
Bold indicates the highest rated ad within a smoking status group for a given outcome, while italics indicates the lowest rated ad within a smoking status group for a given outcome. a Adjusted for age, education level, parental status, ad order and whether smoking was allowed indoors in home or lived with another smoker. All models used robust standard errors to control for each participant having rated multiple ads. b No variation in responses for that ad in that smoking status group. Significant difference between smokers and non-smokers within country or total sample at *P < 0.05; **P < 0.01; ***P < 0.001; t denotes a trend at P ¼ 0.050–0.099.
outcomes, and indicates where significant differences between smokers and non-smokers occurred for each ad. Overall, the pattern of findings by smoking status and across countries was highly similar; the few differences observed are summarized below. Message Acceptance. A significant interaction between ad and smoking status was found across the whole sample for Message Acceptance (2 ¼ 17.01, P ¼ 0.002), but only for China when examined by
country (2 ¼ 9.23, P ¼ 0.026). Within China, ‘Child’ was rated significantly higher by smokers compared with non-smokers (P < 0.010), while all other ads were rated similarly between smoking status groups. Baby Alive received the highest proportion of positive ratings by non-smokers across all countries. Child was rated highest by Chinese and Indian smokers, but received the lowest proportion of positive ratings for Chinese and Russian nonsmokers. Russian smokers rated Say no to SHS
7 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
China
N. Murukutla et al.
8 of 11
non-smokers ratings of each ad. In Russia, smokers gave significantly lower ratings for every ad compared with non-smokers (all at least P < 0.01). A similar effect was seen in India; however, there was also a significant interaction between smoking status and ad (2 ¼ 12.83, P ¼ 0.012), where the magnitude of difference in ad ratings between smokers and non-smokers for the ad Child was smaller than the other four ads. Smoke-free Works had the lowest proportion of positive ratings among every subgroup. Baby Alive and Clinical were generally highest among non-smokers, while Chinese smokers gave Baby Alive higher proportions of positive ratings, as did Russian smokers (equal with Say no to SHS). Smokers in India rated Child highest for Behavioral Intention.
Discussion This study found strong consistency across country, smoking status and outcomes in ad performance. Consistent with the literature on anti-smoking ads, SHS ads that were graphic and hard hitting had the greatest universal appeal, were most likely to create negative emotions and were most likely to motivate behavior change to reduce SHS exposure: Baby Alive was most often the top rated ad; Clinical, the other strong graphic ad, also showed consistently high ratings. While the information-style ad, Smoke-free Works, was comprehensible, it was the least likely to trigger negative emotion and reported intentions to change behavior around SHS exposure. The general consistency of findings across varied cultural contexts underscores adaptation of ads as a cost-effective strategy. In LMICs, where resources are limited, the considerable costs of creating and producing new ads are often a barrier to the implementation of effective tobacco control campaigns. Adaptation of ads that have performed well in one context to another context is a cost- and timeefficient approach: with minimal costs and production time, peripheral and executional elements of the ads may be substituted to make them culturally appropriate. However, these findings should equally
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
highest, while Indian non-smokers gave this ad and Smoke-free Works equally low ratings on Message Acceptance. Negative Emotion. Significant interactions between ad and smoking status were found for India (2 ¼ 11.40, P ¼ 0.022), Russia (2 ¼ 10.24, P ¼ 0.037), and the total sample (2 ¼ 21.40, P < 0.001) on Negative Emotion. All ads were rated similarly by Indian participants other than Baby Alive, where smokers tended to rate this ad higher than non-smokers (P ¼ 0.057). Within Russia, ‘Clinical’ was the only ad to show a significant difference in ad ratings, where non-smokers rated this ad significantly higher than smokers (P ¼ 0.014). While a weak trend toward an interaction was found for China (2 ¼ 8.00, P ¼ 0.092), there were no differences between smokers’ and non-smokers’ ratings of each ad. Clinical was highest in China, while Baby Alive was rated highest in Russia and for Indian smokers. Indian non-smokers gave Say no to SHS the highest proportion of positive ratings. Smoke-free Works received the lowest proportion of positive ratings across all countries and subgroups, except Russian non-smokers, where it was second lowest behind Child. Potential Effectiveness. China was the only country to show a significant interaction between smoking status and ad for Potential Effectiveness (2 ¼ 9.68, P ¼ 0.046). Chinese non-smokers rated Baby Alive highest, significantly higher than smokers did (P ¼ 0.033), while smokers rated Clinical highest (nominally higher than Baby Alive). Within India, non-smokers tended to give higher proportions of positive ratings to all ads compared with smokers, and this difference tended toward significance for Clinical (P ¼ 0.056) and Child (P ¼ 0.080). All ads were rated similarly by Russian smokers and non-smokers. Baby Alive received the highest proportion of positive ratings by all subgroups other than Chinese smokers, and Smoke-free Works was lowest for all subgroups. Behavioral Intention. For Behavioral Intention, no interaction between smoking status and ad was found for China, Russia or the total sample. In China, there were no differences in smokers and
Male smoker and non-smoker responses to secondhand smoke television advertisements the ads as increasing Behavioral Intention. Parents, who are likely to be more focused on health concerns, particularly for their children, would be expected to react more favorably to messages that protect their health from environmental toxins. Conversely, younger participants experienced less Negative Emotion than did older participants; and, those who lived in homes where smoking was permitted indoors were less likely than those who lived in smoke-free homes to see the potential in the ads for behavior change. Younger adults may be less concerned about their own health than older adults, and those who live in homes where smoking is allowed are likely to see smoking as more normative than those living in smoke-free homes—therefore, their reactions to the ads would be less favorable. The finding of no subgroup differences based on education is encouraging as it suggests that the ads are comprehensible and impactful even among those who are less educated. The study was limited by the use of five ads, which were not complete adaptations. It is possible that other examples of the types of ads selected in this study, and completely adapted versions of the ads, would have yielded different results. However, given the cost of production, only minimal adaptation was possible for the study by removing potentially distracting end tags, with the remainder of funds being reserved for the adaptation of most effective ads identified from the testing exercise. This article was also purposively limited to a convenience sample of men, for reasons already stated. Additionally, previous research in LMICs has shown that while women tend to rate anti-smoking advertisements higher than men overall, for strong graphic ads tend to be rated similarly highly by both genders [18]. However, the reactions of women and measurement of population-level effects are important for the future. Two other sampling criteria also limit the generalizability of the current findings: first, the study only included smokers of cigarettes and not smokers of indigenous smoking products (e.g. bidis in India). This was done to maintain equivalency across countries where cigarettes are the only commonly used form. Second, this study recruited the ‘heavier’ smokers (i.e. those who 9 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
inform strategies for HICs too, enabling the use of readily available campaign materials. There were very few significant differences in ad ratings between smokers and non-smokers. Across all ads, non-smokers were more positive in their ratings than smokers. This pattern of findings, demonstrated across the entire sample and within each country, is as expected: non-smokers, when engaged, typically tend to be advocates for tobacco control laws and policies and it is therefore consistent that they would respond more favorably to messages that encourage their protection from the harmful effects of SHS. There were no other meaningful differences and the graphic ads performed equivalently across smokers and non-smokers: they increased smokers’ intent to reduce exposing others to SHS expose and non-smokers’ intentions to avoid being exposed; importantly, the graphic SHS-focused ads also increased smokers’ motivations to quit. There are a number of important theoretical and programmatic implications from these findings. First, the general similarity in responses among smokers and non-smokers suggests that the mechanisms for behavior change may be similar for nonsmokers and smokers. Second, the finding that graphic SHS ads also encourage smokers to quit, although this may not be the primary goal of SHS ads, suggests that the same message can be used to influence both smokers and non-smokers, thus saving costs that would be associated with the development of different campaigns for smokers and non-smokers. Finally, the consistently higher ratings observed among non-smokers, and the emotional engagement observed among smokers and nonsmokers, suggest an avenue for future research. Specifically, it would be valuable for future research to examine whether the increased awareness and more negative attitudes toward SHS exposure created by campaigns can be used to create public support for the implementation of strong smoke-free laws. There were few other systematic subgroup differences, and these were largely consistent with expected levels of engagement in a message. Parents were more likely than those without children to rate
N. Murukutla et al.
Conclusions This study highlights the superior value of graphic, health harm-focused ads compared with information-style ads in generating negative emotion and persuading behavior change around SHS exposure. It finds that there are more commonalities than differences among countries in reactions to SHS ads. Graphic health harm focused ads have universal appeal and resonate with varied audiences, enabling the transfer of information and materials across HICs and LMICs for expediency and efficiency. It finds generally consistent reactions to the ads among smokers and non-smokers, suggesting that a single campaign with graphic, visceral ads may be used to target both groups cost effectively. Importantly, the ads may be used to not only encourage smokers and non-smokers to avoid being exposed to SHS, but may have a secondary benefit in increasing motivation to quit among smokers. Graphic, hard-hitting SHS messages delivered through mass media channels of communication are approaches that can be 10 of 11
used in many LMICs to change social norms and behaviors to support for smoke-free laws and policies.
Funding The study was supported by a grant from the Bloomberg Philanthropies to World Lung Foundation. However, Bloomberg Philanthropies was not involved in any aspect of the study or in the writing of this paper. References 1. Been JV, Nurmatov U, Cox B et al. Effect of smoke-free legislation on perinatal and child health: a systematic review and meta-analysis. The Lancet 2014; 383: 1549–60. 2. U.S. Department of Health and Human Services. The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2006. 3. IARC. IARC Handbooks of Cancer Prevention, Tobacco Control, Vol. 13: Evaluating the Effectiveness of Smokefree Policies. Lyon, France: International Agency for Research on Cancer, 2009. 4. Kaleta D, Koziel A, Miskiewicz P. MPOWER-strategy for fighting the global tobacco epidemic. Med Pr 2009; 60: 145–9. 5. King BA, Mirza SA, Babb SD GATS Collaborating Group. A cross-country comparison of secondhand smoke exposure among adults: findings from the Global Adult Tobacco Survey (GATS). Tob Control 2013; 22: e5. 6. National Cancer Institute (NCI). The Role of the Media in Promoting and Reducing Tobacco Use. Tobacco Control Monograph No. 19. Bethesda, MD: United States Department of Health and Human Services, National Institutes of Health, National Cancer Institute. NIH Pub. No. 07-6242, 2008. 7. Wakefield M, Loken B, Hornik R. Use of mass media campaigns to change health behaviour. Lancet 2010; 376: 1261–71. 8. Durkin S, Brennan E, Wakefield M. Mass media campaigns to promote smoking cessation among adults: an integrative review. Tob Control 2012; 21: 127–38. 9. Lane CH, Carter MI. The role of evidence-based media advocacy in the promotion of tobacco control policies. Salud Publica de Me´xico 2012; 54: 281–8. 10. Yanovitzky I, Stryker J. Mass media, social norms, and health promotion efforts: a longitudinal study of media effects on youth binge drinking. Commun Res 2001; 28: 208–39.
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
smoked 5 or more cigarettes a day), thereby potentially restricting the generalizability of findings to lighter smokers. That said, heavier smokers were the target audience since they are more likely to contribute toward SHS. As the more addicted smokers, they were also expected to provide a ‘tougher test’ of the advertisements’ effectiveness. Finally, reflective of the different behaviors of smokers and non-smokers, some of the items comprising the measure of behavioral intention scale were specific to each smoking status group, and therefore cannot offer an exact comparison. However, sensitivity analyses of the Behavioral Intention model for smokers and non-smokers conducted separately indicated that there were no important differences from the reported combined model. Furthermore, the other outcome that contained alternately worded questions for smokers and non-smokers, Perceived Effectiveness, had highly congruous items and strong internal reliability. Ultimately, perfect equivalency in the smoker and non-smoker ad rating items was neither necessary nor achievable.
Male smoker and non-smoker responses to secondhand smoke television advertisements 15. Kalkhoran S, Glantz SA. Smoke-free policies: cleaning the air with money to spare. The Lancet 2014; 383: 1526–8. 16. Wakefield MA, Durkin S, Murphy M, Cotter T. Pre-testing Anti-smoking Commercials: Process for the Conduct of Market Research. Alexandria, NSW: Cancer Institute NSW and the Cancer Council Victoria, 2007. 17. Perl S. Using Hard-hitting Media to Motivate Smokers in NYC: 2006 and 2007. Paper presented at the New York State Tobacco Control Conference, 6 May 2008, Albany, New York, 2008. 18. Durkin S, Bayly M, Cotter T et al. Potential effectiveness of anti-smoking advertisement types in ten low and middle income countries: do demographics, smoking characteristics and cultural differences matter? Social Science Medicine 2013; 98: 204–13.
11 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
11. Alday J, Murukutla N, Cedillo C et al. Smoke-free Sa˜o Paulo: a campaign evaluation and the case for sustained mass media investment. Salud Publica Mex 2010; 52(Suppl. 2): S216–25. 12. Wakefield M, Bayly M, Durkin S et al. Smokers’ responses to television advertisements about the serious harms of tobacco use: pre-testing results from 10 low- to middle-income countries. Tob Control 2013; 22: 24–31. 13. Cotter T, Perez D, Dunlop S et al. The case for recycling and adapting anti-tobacco mass media campaigns. Tob Control 2010; 19: 514–7. 14. Kosir M, Gutierrez K. Lessons Learned Globally: Secondhand Smoke Mass Media Campaigns. Saint Paul, MN: Global Dialogue for Effective Stop Smoking Campaigns, 2009.
HEALTH EDUCATION RESEARCH
2014 Pages 1–11
Male smoker and non-smoker responses to television advertisements on the harms of secondhand smoke in China, India and Russia
1
World Lung Foundation, 61 Broadway, New York, NY 10006, USA, 2Centre for Behavioural Research in Cancer, Cancer Council Victoria, 615 St Kilda Road, Melbourne 3004, VIC, Australia and 3World Lung Foundation, Alexandria, NSW, Australia *Correspondence to: N. Murukutla. E-mail: [email protected] The International Anti-SHS Advertisement Rating Study Team comprises Cancer Council Victoria: Melanie Wakefield, Megan Bayly, Sarah Durkin; World Lung Foundation (WLF): Sandra Mullin, Nandita Murukutla, Trish Cotter; WLF, China: Yvette Chang, Winnie Chen; WLF, India: Shefali Gupta, Tahir Turk; WLF, Russia: Irina Morozova, Rebecca Perl. y
Received on September 2, 2013; accepted on July 15, 2014
Abstract Mass media campaigns can play an important role in strengthening support for smoke-free policies and reducing exposure to secondhand smoke (SHS). Identifying anti-SHS advertisements that are effective in diverse cultural contexts may allow for resource sharing in low- and middle-income countries. A convenience sample of 481 male cigarette smokers and non-smokers in three high tobacco burden and culturally dissimilar countries (India, China and Russia) viewed and rated five anti-SHS ads. Multivariate logistic regression analyses were conducted for ‘Message Acceptance’, ‘Negative Emotion’, ‘Perceived Effectiveness’ and ‘Behavioral Intentions’. Smokers and non-smokers in all countries consistently rated the strong graphic, health harm ads as the most effective, and the ‘informational’ ad as the least effective overall: the graphic ad ‘Baby Alive’ was at least 1.8 times more likely than the informational ad ‘Smokefree works’ to receive positive ratings on all four outcomes (all P < 0.001). Graphic, health harm messages about SHS exposure have the greatest universal appeal and are the most effective in motivating changes in behavioral
intentions. Similarity in reactions between smokers and non-smokers, and across countries, suggests that resource sharing and the use of a single graphic ad targeted at smokers and non-smokers would be cost-efficient strategies.
Introduction Secondhand smoke (SHS), or the smoke emitted by burning tobacco products, results in significant morbidity and mortality [1, 2]. Laws that ban smoking in indoor public places are an effective intervention in protecting people from SHS [1–4]. However, smoke-free laws are still non-existent, weak or are poorly enforced and SHS exposure continues to be high, particularly in low- and middle-income countries (LMICs) [5]. Mass media campaigns are a core, proven intervention for tobacco control [6, 7]. In addition to reducing tobacco uptake and prevalence, they can also serve a crucial role by creating awareness about SHS harms, changing social attitudes and practices around smoking and encouraging cessation [2]. Additionally, by changing social attitudes and practices, and creating public support for tobacco control, they can potentially help in the implementation of strong laws [7–11].
ß The Author 2014. Published by Oxford University Press. All rights reserved. For permissions, please email: [email protected]
doi:10.1093/her/cyu044
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Nandita Murukutla1*, Megan Bayly2, Sandra Mullin1, Trish Cotter3 and Melanie Wakefield2; for the International Anti-SHS Advertisement Rating Study Teamy
N. Murukutla et al.
Methods Three culturally dissimilar LMICs with high tobacco burdens, including SHS exposure, were selected. Given the high prevalence of smoking in men in developing countries relative to women (i.e. 43% men versus 2% women in China, 24% men versus 3% women in India, 60% men versus 22% women in Russia), men are a priority group in communication about SHS exposure in these countries [3, 4]. Our study therefore included male smokers and nonsmokers from China, India and Russia. 2 of 11
Design Using a standard ad rating methodology [12, 16], this study was conducted in 2010–2011 in one large city in each country: Changsha, China; New Delhi, India; Moscow, Russia. Participants rated the effectiveness of 10 ads: 5 that were shown in all countries and 5 that were selected by local stakeholders and differed across countries. This article focuses on the five ads shown in all countries.
Participants In each country, male cigarette smokers and nonsmokers aged 18–40 years were recruited. A local market research agency from each city completed recruitment using convenience-sampling methods that were appropriate for each country—faceto-face or telephone interviewing. Smokers of at least 5 cigarettes per day who had smoked for at least 1 year were targeted in recruitment. Eligible non-smokers were never-smokers or had smoked fewer than 100 cigarettes in their lifetime. Eligible smokers smoked at least five cigarettes daily. Participants were assigned to groups split by age (18–29/30–40 years), smoking status (smokers/nonsmokers) and ad order (order A/B). There were 8–12 participants in each group.
Advertisements All ads presented SHS messages with varying message style (Table I). These ads were selected from a resource of tobacco control ads (see http://www.worldlungfoundation.org/mmr) based on their effective performance in their country of origin [14, 17] and/or their suitability for adaptation. They were carefully translated (and back translated) and dubbed into local languages. On-screen text and end frames were replaced or deleted. The ads were counterbalanced in two orders, one being the reverse of the other. The country consistent ads were alternated with the country choice ads.
Ad rating questionnaire The non-smoker and smoker versions of the questionnaire were nearly identical. Both started
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
However, for mass media campaigns to be effective, they must use advertisements that resonate with and persuade the intended audience [8]. For campaigns targeted at active cigarette smoking, health messages that are graphic and negativeemotion arousing are most effective in encouraging smokers to quit [8]. These kinds of messages perform well not only in high-income countries (HICs) but also in LMICs [12], leading to the evolution of a cost-effective approach of taking readily available high-performing ads from HICs and minimally adapting them for effective use in LMICs for active smoking campaigns [13]. However, to date effective message styles for campaigns on ‘passive’ smoking and the health harms of SHS exposure have not been identified [14]. Furthermore, the adaptability of SHS ads from one country to another is yet to be determined. Effective mass media campaigns are critical, especially in LMICs with overburdened health systems, to reduce the burdens of SHS exposure and build support for effective smoke-free policies [15]. Thus, this study was designed to test the acceptability, relevance, perceived effectiveness and potential for behavior change of five television advertisements on SHS harms that had high potential for impact and adaptation. We sought to compare the performance of these ads between smokers and non-smokers both within and across LMICs with the aim of determining whether adaptation of existing SHS advertisements might be similarly effective for both smoker and non-smokers.
Male smoker and non-smoker responses to secondhand smoke television advertisements Table I. Descriptions of the SHS television ads Ad characteristics
Description
Cigarettes are eating you and your baby alive (Baby Alive)
Graphic and visceral ad focused on harms to children
Clinical
Strong graphic ad focused on harms to adults
Smoke-free works: Office
Informational style, focused on implementation of smoking ban in workplaces
Child
Narrative style, suggesting harm to the smoker and his family
Say no to SHS
Dramatic simulation, depicting toxic smoke transmitting from the smoker to non-smokers
Eerie music plays throughout. We see the silhouettes of a man and woman smoking cigarettes in front of a child. Their lungs are transposed on their silhouettes. A voiceover (VO) explains that cigarette smoke contains poisons that trigger severe health problems. Cut to images of real babies hooked to respirators. As the VO says that cigarette smoke can cause ear infections, asthma and pneumonia, we see shots of children with each of these conditions. Cut back to the child’s silhouette, then images of babies with low birth weight. The VO says that cigarette smoke is linked to low birth weight and doubles the risk of Sudden Infant Death Syndrome. Cut back to the silhouettes, then quick flashes of the sick children A non-smoking woman at a party is shown inhaling SHS. The camera follows the smoke down her throat and into her lungs, and simulated imagery of SHS being absorbed into the blood stream and damaging vessels is shown. VO: Secondhand smoke enters the throat. And it doesn’t just damage the lungs: Within minutes the smoke activates platelets, which thicken the blood and damages arteries. In time plaque accumulates in the damaged blood vessels, obstructing blood flow to the heart. This can result in stroke or heart attack. Text: A little secondhand smoke can do a lot of damage A VO announces that offices will be smoke-free beginning in March. In an office setting, we hear an employee light a cigarette and exhale. Zoom in on the chest of a coworker. A smoke-filled lung appears overlaid on the coworker’s shirt. Text states that secondhand smoke causes lung cancer. Zoom in on another office employee’s chest. An unhealthy heart and lungs appear overlaid on her shirt. Text states that secondhand smoke causes heart disease A man is seen smoking in different settings: at the dinner table, with a colleague, watching television with his daughter. An ill man is shown on the television, and the young daughter looks at her father with concern. He leaves the room and returns without his cigarette. VO: Everyone seeks happiness. But isn’t that a very high price to pay? Smoking can cost you and can also seriously affect the health and welfare of those around you. Text: Smoking kills. Passive smoking kills too. VO: Don’t have them pay the price for your smoking. Smoking will cost you The ad opens on a close-up of a lit cigarette. Pan out to see a man exhaling cigarette smoke. Pan out further to see his wife then also exhale cigarette smoke, and then his two young children also exhale smoke. The ad finishes with a close-up of daughter surrounded by a cloud of smoke. VO: The smoke from cigarettes contains 4000 hazardous chemicals, affecting innocent victims and slowly taking their lives. Secondhand smoke kills. Say no to secondhand smoke
Ads may be viewed at: http://67.199.72.89/mmr/english/ads_shs.html.
by assessing demographics, including gender, age, parental status, education level, residence with a smoker and rules around smoking indoors in their home. Smokers also reported number of cigarettes
smoked daily, intentions to quit in the next 12 months and any previous quit attempts. All ads were rated on an 11-item questionnaire using a 5-point scale: 1 ‘strongly disagree’, 2 3 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Ad name
N. Murukutla et al.
Procedure After completing the demographic questionnaire, participants were familiarized with the rating questionnaire by initially rating a ‘practice ad’ (an unrelated, generic ad, e.g. a shampoo ad). Thereafter, each SHS ad was viewed twice and privately rated before moving to the next.
Data analysis Frequency tables and 2 tests were produced to examine equivalency of the sample across countries and smoking status. Responses on the ad rating items were binary coded. Ratings of strongly agree and somewhat agree were coded as positive responses, while strongly and somewhat disagree and neither disagree or agree responses were coded as negative/neutral. Responses for the multiitem scales were summed and divided by the number of items in that scale. Average responses 4 of 11
of greater than 3.5 were coded as positive, and all other responses were coded as negative/neutral [12]. Multivariate logistic regression analyses were conducted using Stata/SE 12.1 for each outcome measure, for non-smokers and smokers combined. All models used robust standard errors to control for each participant having rated multiple ads. All models also controlled for age (18–29 or 30–40 years), education level (completed a tertiary degree or no tertiary degree), parental status (yes or no), whether smoking was allowed indoors in their home (yes or no), whether they lived with another smoker (yes or no), country (China, Russia, India), ad order (A or B) and all ads. Post-estimation Wald w2 tests were used to examine main effects of ad and smoking status. Two-way interactions were performed between ad and smoking status to examine whether responses to each ad were consistent between smokers and non-smokers, across the whole sample and within each country. Adjusted proportions of positive ad ratings were calculated for each smoking status group within each country from these models. Qualitative data from the structured discussions were consistent with the quantitative ratings and were primarily intended to guide local adaptations of the ads; hence, they are not discussed further here.
Results Sample characteristics A total of 481 participants were recruited (Table II). A significantly higher proportion of Indian smokers than non-smokers had completed a tertiary degree. Across all countries, significantly more smokers than non-smokers reported living with another smoker. Smoking inside the home was allowed more often for smokers than non-smokers in India and Russia. One-third of smokers reported smoking more than 15 cigarettes per day; however, this varied from 51% in China and Russia to 5% in India. While adherence to recruitment criteria was generally strong, 11.3% of Chinese smokers and 18.5% of Indian smokers reported smoking fewer than 5 cigarettes per day.
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
‘slightly disagree’, 3 ‘neither agree or disagree’, 4 ‘slightly agree’ and 5 ‘strongly agree’. The items were reduced to four subscales: ‘Message Acceptance’ (consisting of the item ‘ad was easy to understand’); ‘Negative Emotion’ (‘ad made me feel uncomfortable’); ‘Perceived Effectiveness’ (‘ad taught me something new’/‘made me stop and think’/‘was believable’/‘was relevant to me’/‘made me likely to talk to someone else about this ad’/ ‘made feel more concerned about smoking around other adults’ [smokers only]/‘made me feel more concerned about smoking around children’ [smokers only]/‘made me feel more concerned about others smoking around me’ [non-smokers only]/‘made me feel concerned about others smoking around children’ [non-smokers only]); ‘Behavioral Intention’ (‘makes me more likely to discourage smoking inside my home’ [nonsmokers]/‘makes me more likely to avoid places where people smoke’ [non-smokers]/‘more likely to avoid smoking indoors’ [smokers]/‘more likely to try to quit’ [smokers]). All scales showed strong internal reliability within each smoking status group (Cronbach’s a > 0.72).
Male smoker and non-smoker responses to secondhand smoke television advertisements Table II. Demographic characteristics of the sample China
Russia
Total sample
Nonsmokers
Smokers
Nonsmokers
Smokers
Nonsmokers
Smokers
Nonsmokers
Smokers
80
80
80
81
80
80
240
241
50 50 28 35 39 38
50 50 19 41 48 80***
51 49 36 40 15 9
49 51 53* 42 37** 28**
50 50 58 30 26 19
50 50 54 33 53** 50***
50 50 40 35 27 22
50 50 42 39 46*** 53***
– – – –
49 51 20 60
– – – –
95 5 35 51
– – – –
49 51 38 73
– – – –
64 36 31 61
50 50
50 50
50 50
51 49
50 50
50 50
50 50
50 50
Significant difference between smokers and non-smokers within country or total sample at *P < 0.05; **P < 0.01; ***P < 0.001.
Ad ratings Main effects Across countries, there was a significant main effect of ad on all outcomes (Table III): Message Acceptance (2 ¼ 21.57, P < 0.001); Negative Emotion (2 ¼ 117.36, P < 0.001); Perceived Effectiveness (2 ¼ 144.02, P < 0.001) and Behavioral Intention (2 ¼ 40.74, P < 0.001). On all outcomes, ‘Baby Alive’ had the highest odds of a positive rating relative to the referent ad ‘Smokefree Works’. Smoke-free Works had at least marginally lower odds of positive ratings than all other ads for all outcomes. Significant main effects of country were found on all outcomes: Message Acceptance (2 ¼ 14.82, P < 0.001); Negative Emotion (2 ¼ 13.50, P ¼ 0.001); Perceived Effectiveness (2 ¼ 43.76, P < 0.001) and Behavioral Intention (2 ¼ 113.87, P < 0.001). Participants from India had the highest odds of positive ratings on all measures. Participants from Russia had the lowest odds of positive ratings on Perceived Effectiveness and Behavioral
Intention, while China had the lowest odds for Message Acceptance and Negative Emotion. A main effect of smoking status was found only for Behavioral Intention, where smokers were 61% less likely to rate ads positively than non-smokers (odds ratio [OR] 0.39, 95% confidence interval [CI]: 0.27–0.56, P < 0.001). Participants aged 18–29 years had significantly lower odds than older participants of rating ads positively on Negative Emotion (OR 0.60, 95% CI: 0.42–0.87, P ¼ 0.007). Parents had significantly higher odds of ratings ads positively on Behavioral Intention than those who did not have children (OR 1.84, 95% CI: 1.18–2.87, P ¼ 0.007), while those who lived in homes where smoking was permitted indoors had lower odds of positive ratings on this measure than those who reported smoking was not allowed inside their home (OR 0.60, 95% CI: 0.41–0.87, P ¼ 0.008). No effects of ad order, education or residing with a smoker were found for any outcome. These models were repeated including only smokers to examine potential effects of smoker-only variables. 5 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
N Age group (%) 18–29 years 30–40 years Completed a tertiary degree (% yes) Parent (% yes) Smoking allowed inside home (% yes) Live with a smoker (% yes) Cigarettes smoked daily (%) 1–15 15 or more Thinking about quitting (% yes) Previously tried to quit (% yes) Ad order (%) Order A Order B
India
N. Murukutla et al. Table III. Multivariate logistic regression analysis of positive ad ratings for measures of Message Acceptance, Negative Emotion, Perceived Effectiveness and Behavioral Intentiona Negative Emotion
Perceived Effectiveness
Behavioral Intention
OR
(95% CI)
OR
(95% CI)
OR
(95% CI)
OR
(95% CI)
1.00 4.54 1.44 1.61 1.61
(2.35–8.78)*** (0.96–2.17)t (1.00–2.60)* (1.00–2.60)t
1.00 3.44 2.38 1.50 2.30
(2.68–4.42)*** (1.93–2.93)*** (1.23–1.83)*** (1.86–2.85)***
1.00 5.87 2.83 1.80 2.08
(4.29–8.03)*** (2.22–3.59)*** (1.43–2.27)*** (1.64–2.64)***
1.00 1.83 1.70 1.62 1.62
(1.47–2.28)*** (1.40–2.06)*** (1.31–2.00)*** (1.29–2.02)***
1.00 3.41 1.28
(1.82–6.38)*** (0.79–2.07)
1.00 1.90 1.13
(1.32–2.74)** (0.82–1.56)
1.00 2.00 0.45
(1.28–3.13)** (0.31–0.67)***
1.00 1.59 0.19
(1.01–2.51)* (0.12–0.28)***
1.00 1.06
(0.65–1.73)
1.00 1.01
(0.76–1.32)
1.00 0.79
(0.55–1.13)
1.00 0.39
(0.27–0.56)***
1.00 0.71
(0.40–1.27)
1.00 0.60
(0.42–0.87)**
1.00 0.81
(0.52–1.26)
1.00 0.98
(0.64–1.52)
1.00 0.89
(0.49–1.62)
1.00 1.29
(0.88–1.87)
1.00 1.25
(0.79–1.97)
1.00 1.84
(1.18–2.87)**
1.00 1.18
(0.72–1.93)
1.00 0.85
(0.63–1.13)
1.00 0.85
(0.59–1.23)
1.00 0.60
(0.41–0.87)**
1.00 1.13
(0.73–1.75)
1.00 0.93
(0.71–1.20)
1.00 1.19
(0.85–1.66)
1.00 0.95
(0.67–1.34)
1.00 0.84
(0.54–1.31)
1.00 0.82
(0.62–1.08)
1.00 0.90
(0.63–1.28)
1.00 0.89
(0.62–1.27)
1.00 1.20
(0.76–1.90)
1.00 1.00
(0.74–1.35)
1.00 1.06
(0.72–1.55)
1.00 0.79
(0.53–1.18)
OR: odds ratio; 95% CI: 95% confidence interval. a Behavioral intention is combined for smokers and non-smokers to allow for comparison of smoking status effects. This model was repeated for smokers and non-smokers separately, and no important differences to the overall model were observed. Significantly different to reference group at *P < 0.05; **P < 0.01; ***P < 0.001; tdenotes a trend at P ¼ 0.050–0.099.
Participants who smoked more than 15 cigarettes per day were significantly less likely to give positive ratings on Negative Emotion than those who smoked 15 or fewer cigarettes (OR 0.63, 95% CI: 0.40–0.99, P ¼ 0.045). Smokers who were considering quitting in the next 12 months were more likely to give positive Behavioral Intention ratings compared with those who were not (OR 2.25, 95% 6 of 11
CI: 0.12–4.13, P ¼ 0.009). The significant main effects of ad and country were still observed in these additional analyses.
Ad ratings, by smoking status and country Table IV shows the adjusted proportions of positive ad ratings by smoking status and country for all
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
Ad Smoke-free Works (ref) Baby Alive Clinical Child Say no to SHS Country China (ref) India Russia Smoking status Non-smoker (ref) Smoker Age (years) 30–40 (ref) 18–29 Parent No (ref) Yes Smoking in home No (ref) Yes Ad order Order B (ref) Order A Tertiary degree No (ref) Yes Live with a smoker No (ref) Yes
Message Acceptance
Male smoker and non-smoker responses to secondhand smoke television advertisements Table IV. Adjusted percentages of positive ad ratings for Message Acceptance, Negative Emotion, Perceived Effectiveness and Behavioral Intention, by smoking status and country (including Wald tests for interactions between ad and smoking status)a India
Russia
Total sample
Non-smokers Smokers Adjusted %
Non-smokers Smokers Non-smokers Smokers Non-smokers Smokers Adjusted % Adjusted % Adjusted %
Message Acceptance Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 9.23, P ¼ 0.026 96 100b 93 89 91 88 84 99* 88 94
2 ¼ 3.00, P ¼ 0.223 100b 97 98 94 95 97 99 100b 95 96
2 ¼ 5.54, P ¼ 0.237 97 97 95 92 91 85 90 94 94 99
2 ¼ 17.01, P ¼ 0.002 99 97 95 92t 92 90 91 98** 92 96
Negative Emotion Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 8.00, P ¼ 0.092 57 59 68 64 33 40 43 55 63 51
2 ¼ 11.40, P ¼ 0.022 64 78t 71 67 58 56 70 71 73 63
2 ¼ 10.24, P ¼ 0.037 85 82 63 44* 35 35 32 44 62 61
2 ¼ 21.40, P < 0.001 68 74 67 59t 41 45 47 57* 65 60
Perceived Effectiveness Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 9.68, P ¼ 0.043 99 87* 89 88 68 62 80 85 73 80
2 ¼ 1.36, P ¼ 0.851 99 94 95 85t 87 80 91 81t 94 88
2 ¼ 3.92, P ¼ 0.417 81 86 73 68 50 46 57 64 66 71
2 ¼ 6.08, P ¼ 0.193 93 89 85 81t 68 63t 76 77 77 80
Behavioral Intention Baby Alive Clinical Smoke-free Works Child Say no to SHS
2 ¼ 2.80, P ¼ 0.591 80 81 80 73 62 62 73 78 76 74
2 ¼ 12.83, P ¼ 0.012 99 69** 95 74** 91 64*** 94 81* 97 74**
2 ¼ 1.16, P ¼ 0.884 54 28** 55 27** 45 23** 48 24** 49 28*
2 ¼ 5.39, P ¼ 0.250 77 61*** 76 59*** 65 51*** 71 63** 74 60***
Bold indicates the highest rated ad within a smoking status group for a given outcome, while italics indicates the lowest rated ad within a smoking status group for a given outcome. a Adjusted for age, education level, parental status, ad order and whether smoking was allowed indoors in home or lived with another smoker. All models used robust standard errors to control for each participant having rated multiple ads. b No variation in responses for that ad in that smoking status group. Significant difference between smokers and non-smokers within country or total sample at *P < 0.05; **P < 0.01; ***P < 0.001; t denotes a trend at P ¼ 0.050–0.099.
outcomes, and indicates where significant differences between smokers and non-smokers occurred for each ad. Overall, the pattern of findings by smoking status and across countries was highly similar; the few differences observed are summarized below. Message Acceptance. A significant interaction between ad and smoking status was found across the whole sample for Message Acceptance (2 ¼ 17.01, P ¼ 0.002), but only for China when examined by
country (2 ¼ 9.23, P ¼ 0.026). Within China, ‘Child’ was rated significantly higher by smokers compared with non-smokers (P < 0.010), while all other ads were rated similarly between smoking status groups. Baby Alive received the highest proportion of positive ratings by non-smokers across all countries. Child was rated highest by Chinese and Indian smokers, but received the lowest proportion of positive ratings for Chinese and Russian nonsmokers. Russian smokers rated Say no to SHS
7 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
China
N. Murukutla et al.
8 of 11
non-smokers ratings of each ad. In Russia, smokers gave significantly lower ratings for every ad compared with non-smokers (all at least P < 0.01). A similar effect was seen in India; however, there was also a significant interaction between smoking status and ad (2 ¼ 12.83, P ¼ 0.012), where the magnitude of difference in ad ratings between smokers and non-smokers for the ad Child was smaller than the other four ads. Smoke-free Works had the lowest proportion of positive ratings among every subgroup. Baby Alive and Clinical were generally highest among non-smokers, while Chinese smokers gave Baby Alive higher proportions of positive ratings, as did Russian smokers (equal with Say no to SHS). Smokers in India rated Child highest for Behavioral Intention.
Discussion This study found strong consistency across country, smoking status and outcomes in ad performance. Consistent with the literature on anti-smoking ads, SHS ads that were graphic and hard hitting had the greatest universal appeal, were most likely to create negative emotions and were most likely to motivate behavior change to reduce SHS exposure: Baby Alive was most often the top rated ad; Clinical, the other strong graphic ad, also showed consistently high ratings. While the information-style ad, Smoke-free Works, was comprehensible, it was the least likely to trigger negative emotion and reported intentions to change behavior around SHS exposure. The general consistency of findings across varied cultural contexts underscores adaptation of ads as a cost-effective strategy. In LMICs, where resources are limited, the considerable costs of creating and producing new ads are often a barrier to the implementation of effective tobacco control campaigns. Adaptation of ads that have performed well in one context to another context is a cost- and timeefficient approach: with minimal costs and production time, peripheral and executional elements of the ads may be substituted to make them culturally appropriate. However, these findings should equally
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
highest, while Indian non-smokers gave this ad and Smoke-free Works equally low ratings on Message Acceptance. Negative Emotion. Significant interactions between ad and smoking status were found for India (2 ¼ 11.40, P ¼ 0.022), Russia (2 ¼ 10.24, P ¼ 0.037), and the total sample (2 ¼ 21.40, P < 0.001) on Negative Emotion. All ads were rated similarly by Indian participants other than Baby Alive, where smokers tended to rate this ad higher than non-smokers (P ¼ 0.057). Within Russia, ‘Clinical’ was the only ad to show a significant difference in ad ratings, where non-smokers rated this ad significantly higher than smokers (P ¼ 0.014). While a weak trend toward an interaction was found for China (2 ¼ 8.00, P ¼ 0.092), there were no differences between smokers’ and non-smokers’ ratings of each ad. Clinical was highest in China, while Baby Alive was rated highest in Russia and for Indian smokers. Indian non-smokers gave Say no to SHS the highest proportion of positive ratings. Smoke-free Works received the lowest proportion of positive ratings across all countries and subgroups, except Russian non-smokers, where it was second lowest behind Child. Potential Effectiveness. China was the only country to show a significant interaction between smoking status and ad for Potential Effectiveness (2 ¼ 9.68, P ¼ 0.046). Chinese non-smokers rated Baby Alive highest, significantly higher than smokers did (P ¼ 0.033), while smokers rated Clinical highest (nominally higher than Baby Alive). Within India, non-smokers tended to give higher proportions of positive ratings to all ads compared with smokers, and this difference tended toward significance for Clinical (P ¼ 0.056) and Child (P ¼ 0.080). All ads were rated similarly by Russian smokers and non-smokers. Baby Alive received the highest proportion of positive ratings by all subgroups other than Chinese smokers, and Smoke-free Works was lowest for all subgroups. Behavioral Intention. For Behavioral Intention, no interaction between smoking status and ad was found for China, Russia or the total sample. In China, there were no differences in smokers and
Male smoker and non-smoker responses to secondhand smoke television advertisements the ads as increasing Behavioral Intention. Parents, who are likely to be more focused on health concerns, particularly for their children, would be expected to react more favorably to messages that protect their health from environmental toxins. Conversely, younger participants experienced less Negative Emotion than did older participants; and, those who lived in homes where smoking was permitted indoors were less likely than those who lived in smoke-free homes to see the potential in the ads for behavior change. Younger adults may be less concerned about their own health than older adults, and those who live in homes where smoking is allowed are likely to see smoking as more normative than those living in smoke-free homes—therefore, their reactions to the ads would be less favorable. The finding of no subgroup differences based on education is encouraging as it suggests that the ads are comprehensible and impactful even among those who are less educated. The study was limited by the use of five ads, which were not complete adaptations. It is possible that other examples of the types of ads selected in this study, and completely adapted versions of the ads, would have yielded different results. However, given the cost of production, only minimal adaptation was possible for the study by removing potentially distracting end tags, with the remainder of funds being reserved for the adaptation of most effective ads identified from the testing exercise. This article was also purposively limited to a convenience sample of men, for reasons already stated. Additionally, previous research in LMICs has shown that while women tend to rate anti-smoking advertisements higher than men overall, for strong graphic ads tend to be rated similarly highly by both genders [18]. However, the reactions of women and measurement of population-level effects are important for the future. Two other sampling criteria also limit the generalizability of the current findings: first, the study only included smokers of cigarettes and not smokers of indigenous smoking products (e.g. bidis in India). This was done to maintain equivalency across countries where cigarettes are the only commonly used form. Second, this study recruited the ‘heavier’ smokers (i.e. those who 9 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
inform strategies for HICs too, enabling the use of readily available campaign materials. There were very few significant differences in ad ratings between smokers and non-smokers. Across all ads, non-smokers were more positive in their ratings than smokers. This pattern of findings, demonstrated across the entire sample and within each country, is as expected: non-smokers, when engaged, typically tend to be advocates for tobacco control laws and policies and it is therefore consistent that they would respond more favorably to messages that encourage their protection from the harmful effects of SHS. There were no other meaningful differences and the graphic ads performed equivalently across smokers and non-smokers: they increased smokers’ intent to reduce exposing others to SHS expose and non-smokers’ intentions to avoid being exposed; importantly, the graphic SHS-focused ads also increased smokers’ motivations to quit. There are a number of important theoretical and programmatic implications from these findings. First, the general similarity in responses among smokers and non-smokers suggests that the mechanisms for behavior change may be similar for nonsmokers and smokers. Second, the finding that graphic SHS ads also encourage smokers to quit, although this may not be the primary goal of SHS ads, suggests that the same message can be used to influence both smokers and non-smokers, thus saving costs that would be associated with the development of different campaigns for smokers and non-smokers. Finally, the consistently higher ratings observed among non-smokers, and the emotional engagement observed among smokers and nonsmokers, suggest an avenue for future research. Specifically, it would be valuable for future research to examine whether the increased awareness and more negative attitudes toward SHS exposure created by campaigns can be used to create public support for the implementation of strong smoke-free laws. There were few other systematic subgroup differences, and these were largely consistent with expected levels of engagement in a message. Parents were more likely than those without children to rate
N. Murukutla et al.
Conclusions This study highlights the superior value of graphic, health harm-focused ads compared with information-style ads in generating negative emotion and persuading behavior change around SHS exposure. It finds that there are more commonalities than differences among countries in reactions to SHS ads. Graphic health harm focused ads have universal appeal and resonate with varied audiences, enabling the transfer of information and materials across HICs and LMICs for expediency and efficiency. It finds generally consistent reactions to the ads among smokers and non-smokers, suggesting that a single campaign with graphic, visceral ads may be used to target both groups cost effectively. Importantly, the ads may be used to not only encourage smokers and non-smokers to avoid being exposed to SHS, but may have a secondary benefit in increasing motivation to quit among smokers. Graphic, hard-hitting SHS messages delivered through mass media channels of communication are approaches that can be 10 of 11
used in many LMICs to change social norms and behaviors to support for smoke-free laws and policies.
Funding The study was supported by a grant from the Bloomberg Philanthropies to World Lung Foundation. However, Bloomberg Philanthropies was not involved in any aspect of the study or in the writing of this paper. References 1. Been JV, Nurmatov U, Cox B et al. Effect of smoke-free legislation on perinatal and child health: a systematic review and meta-analysis. The Lancet 2014; 383: 1549–60. 2. U.S. Department of Health and Human Services. The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2006. 3. IARC. IARC Handbooks of Cancer Prevention, Tobacco Control, Vol. 13: Evaluating the Effectiveness of Smokefree Policies. Lyon, France: International Agency for Research on Cancer, 2009. 4. Kaleta D, Koziel A, Miskiewicz P. MPOWER-strategy for fighting the global tobacco epidemic. Med Pr 2009; 60: 145–9. 5. King BA, Mirza SA, Babb SD GATS Collaborating Group. A cross-country comparison of secondhand smoke exposure among adults: findings from the Global Adult Tobacco Survey (GATS). Tob Control 2013; 22: e5. 6. National Cancer Institute (NCI). The Role of the Media in Promoting and Reducing Tobacco Use. Tobacco Control Monograph No. 19. Bethesda, MD: United States Department of Health and Human Services, National Institutes of Health, National Cancer Institute. NIH Pub. No. 07-6242, 2008. 7. Wakefield M, Loken B, Hornik R. Use of mass media campaigns to change health behaviour. Lancet 2010; 376: 1261–71. 8. Durkin S, Brennan E, Wakefield M. Mass media campaigns to promote smoking cessation among adults: an integrative review. Tob Control 2012; 21: 127–38. 9. Lane CH, Carter MI. The role of evidence-based media advocacy in the promotion of tobacco control policies. Salud Publica de Me´xico 2012; 54: 281–8. 10. Yanovitzky I, Stryker J. Mass media, social norms, and health promotion efforts: a longitudinal study of media effects on youth binge drinking. Commun Res 2001; 28: 208–39.
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
smoked 5 or more cigarettes a day), thereby potentially restricting the generalizability of findings to lighter smokers. That said, heavier smokers were the target audience since they are more likely to contribute toward SHS. As the more addicted smokers, they were also expected to provide a ‘tougher test’ of the advertisements’ effectiveness. Finally, reflective of the different behaviors of smokers and non-smokers, some of the items comprising the measure of behavioral intention scale were specific to each smoking status group, and therefore cannot offer an exact comparison. However, sensitivity analyses of the Behavioral Intention model for smokers and non-smokers conducted separately indicated that there were no important differences from the reported combined model. Furthermore, the other outcome that contained alternately worded questions for smokers and non-smokers, Perceived Effectiveness, had highly congruous items and strong internal reliability. Ultimately, perfect equivalency in the smoker and non-smoker ad rating items was neither necessary nor achievable.
Male smoker and non-smoker responses to secondhand smoke television advertisements 15. Kalkhoran S, Glantz SA. Smoke-free policies: cleaning the air with money to spare. The Lancet 2014; 383: 1526–8. 16. Wakefield MA, Durkin S, Murphy M, Cotter T. Pre-testing Anti-smoking Commercials: Process for the Conduct of Market Research. Alexandria, NSW: Cancer Institute NSW and the Cancer Council Victoria, 2007. 17. Perl S. Using Hard-hitting Media to Motivate Smokers in NYC: 2006 and 2007. Paper presented at the New York State Tobacco Control Conference, 6 May 2008, Albany, New York, 2008. 18. Durkin S, Bayly M, Cotter T et al. Potential effectiveness of anti-smoking advertisement types in ten low and middle income countries: do demographics, smoking characteristics and cultural differences matter? Social Science Medicine 2013; 98: 204–13.
11 of 11
Downloaded from http://her.oxfordjournals.org/ at Cleveland State University on August 19, 2014
11. Alday J, Murukutla N, Cedillo C et al. Smoke-free Sa˜o Paulo: a campaign evaluation and the case for sustained mass media investment. Salud Publica Mex 2010; 52(Suppl. 2): S216–25. 12. Wakefield M, Bayly M, Durkin S et al. Smokers’ responses to television advertisements about the serious harms of tobacco use: pre-testing results from 10 low- to middle-income countries. Tob Control 2013; 22: 24–31. 13. Cotter T, Perez D, Dunlop S et al. The case for recycling and adapting anti-tobacco mass media campaigns. Tob Control 2010; 19: 514–7. 14. Kosir M, Gutierrez K. Lessons Learned Globally: Secondhand Smoke Mass Media Campaigns. Saint Paul, MN: Global Dialogue for Effective Stop Smoking Campaigns, 2009.
