Business Law paper

Marketing and Public Policy Conference Proceedings/2017 47

For further information contact: Janet Hoek, Professor of Marketing, University of Otago ([email protected]).

Estimating the Effects of Novel On-Pack Warnings on Young Adult Smokers and Susceptible Non-Smokers Philip Gendall, University of Otago Christine Eckert, University of Technology Sydney Janet Hoek, University of Otago Jordan Louviere, University of South Australia

Keywords: standardized tobacco packaging, on-pack warnings, temporal construal, choice modeling

EXTENDED ABSTRACT

Research Question Young adult smokers often view smoking as a temporary phe- nomenon, believe they will quit before they become addicted, and dismiss the health risks of smoking as irrelevant. Their propensity to make decisions based on emotion rather than rea- son suggests warnings that call on them to forgo physiological pleasure and potential peer approval in the hope of living lon- ger, may lack compulsion.

Focussing on more proximal social effects could shape how young adults assess the gains and losses of smoking, and more effectively deter initiation and promote cessation. Warnings showing smoking’s effects on others, particularly children, ba- bies, and companion animals, may also create tensions between the choices young adults wish to assert for themselves and the risks they will impose on others. We developed and tested nov- el warnings that illustrated smoking’s effects on others, chal- lenged temporal perspectives, foregrounded adverse social out- comes, exposed industry practices, and presented new health themes, and then tested these with young adult smokers and susceptible non-smokers.

Method and Data We conducted a behavioural likelihood experiment and a choice modelling experiment in which 474 smokers and 476 susceptible non-smokers aged between 16 and 30 evaluated 12

novel warnings, which we categorised into five groups: graphic health effects, smoking’s effect on others, loss of physical or social attractiveness, tobacco industry manipulation, and the economic consequences.

We developed two sets of warnings (for smokers and non-smok- ers) comprising ten images, of which seven were identical. Each respondent saw ten warning images (nine test images and the status quo) featuring one of three message themes (factual, so- cial or question). Smokers chose the pack they would be most/ least likely to buy while susceptible non-smokers chose the pack most likely to encourage someone like them to turn down a cigarette. We expected smokers to choose the least dissuasive images and susceptible non-smokers to choose the most dissua- sive images. Respondents also used the Juster Scale to estimate how the warning labels could affect their behaviour. Data were analysed using independent sample t-tests and by estimating sequential best worst scale adjusted latent class models.

Summary of Findings The Juster Scale estimates showed warnings featuring the ef- fect of smoking on others, graphic health outcomes, and loss of physical attractiveness, were between 25% and 50% more like- ly to prompt smokers to consider quitting, reduce their tobac- co consumption, or seek quit advice, than the other warnings. Warnings showing graphic health harms elicited the strongest responses from susceptible non-smokers.

Marketing and Public Policy Conference Proceedings/2017 48

The choice analyses used the images with the lowest proba- bility scores as references (a wallet, showing economic costs, for smokers; a perfume bottle, showing social effects, for sus- ceptible non-smokers). Three warnings were particularly dis- suasive for both smokers and non-smokers; these featured a dying smoker, a tracheotomy scar, and cigarettes being tested on dogs. Smokers were also significantly less likely to choose a pack showing a stillborn foetus or a baby exposed to second hand smoke. Susceptible non-smokers were significantly more responsive to two further warnings showing graphic health risks.

Smokers were around 20 times more likely to choose the status quo than a pack featuring a stillborn foetus warning. The mag- nitude of effects was smaller for susceptible non-smokers; their odds of turning down a cigarette package picturing a dying smoker were nearly five times higher than for the status quo.

Key Contributions Young adult smokers and susceptible non-smokers reacted

most strongly to warnings featuring graphic health outcomes, harm to others, and external effects of harm. These findings are consistent across the two methods employed, and add con- fidence to our conclusion that more diverse themes could en- hance the overall impact of on-pack warnings.

New warnings that performed strongly in our study could com- plement the standard approach: health warnings using infor- mational approaches. More diverse warnings that evoke both health and other risks may be particularly effective in reaching different segments within the wider smoker and susceptible non-smoker populations. In addition to graphic health warn- ings, policy makers should develop warnings that feature harm to innocent parties and tobacco industry practices could reso- nate with groups less responsive to health messages. An on-go- ing work programme is also required to identify, test and then evaluate novel warnings, so these reflect current evidence and concerns.

References are available on request.

Copyright of AMA Marketing & Public Policy Academic Conference Proceedings is the property of American Marketing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.