The boomerang effect of mandatory sanitary messages to prevent obesity
- First Online:
- Cite this article as:
- Werle, C.O.C. & Cuny, C. Mark Lett (2012) 23: 883. doi:10.1007/s11002-012-9195-0
- 493 Downloads
A variety of prevention measures are being adopted to counter obesity. One of them is to include health sanitary messages on advertisements for food products. We tested the efficacy of this type of measure in an experimental study with 131 participants who were randomly exposed to an advertisement for a hedonic product containing or not a sanitary message. Implicit memory representations (priming protocol), explicit attitudes (questionnaire) and a behavioral measure of food choice (healthy versus unhealthy snack) were collected. Results showed that participants associated negative concepts more easily to the product when the advertisement was presented without the sanitary message, while there were no differences in the explicit attitudes. Moreover, the choice of a healthy snack doubled in the absence of sanitary message. Contrary to its objectives, the obesity prevention sanitary message fills in consumers’ need for justification leading to a greater acceptability of the advertised product and increased choice of an unhealthy snack.
KeywordsObesity prevention Health policy Sanitary messages effectiveness Food advertising Need for justification
Obesity is a growing problem and a variety of prevention measures involving banning food ads during children’s programs (Dhar and Baylis 2011) or directly promoting healthy behavior on television are being adopted in different countries. In the USA, for example, a national campaign named “5-A-DAY” encouraging people to eat five fruits and vegetables a day was launched with a mitigated success (Basch et al. 1994). In France, since 2007, companies are obliged to include health sanitary messages on advertisements for food products. These messages have a positive tone and present solutions for avoiding weight gain, such as: “Eat five fruit and vegetables per day” or “Avoid snacking between meals”. Because other countries (e.g., Brazil or Thailand) are also considering the adoption of sanitary messages for food ads as a mean to prevent obesity (Hawkes 2007), it is important to understand the efficacy of this kind of prevention measure.
Previous research proved the efficacy of sanitary messages in cigarette packs (Hammond et al. 2004; Kees et al. 2006; Sabanne et al. 2009). The inclusion of graphic visual warnings and warning statements diminished the perceived attractiveness of the package and enhanced negative affect (Kees et al. 2006). Visual and verbal warning labels in cigarette advertising also led to negative attitudes and lower smoking intentions (Sabanne et al. 2009). Due to the positive results of sanitary messages in tobacco control, similar results could be expected for obesity prevention.
The inclusion of sanitary messages on advertisements is based on the assumption that food advertisements favor obesity. Indeed, previous studies emphasized the influence of food advertisements on weight gain (Halford et al. 2004,2007; Harris et al. 2009; Lobstein and Dibb 2005; Zimmerman and Bell 2010). Thus, imposing the use of sanitary messages in food advertisements could potentially counter these negative effects, easily disseminating health messages to the population. The post-test of the messages adopted in France revealed high levels of recall and positive attitudes (INPES 2007). However, prevention campaigns do not always reach expected results (Wilson 2011), some of them are counterproductive and have effects that are opposite to what was initially intended.
The presence of sanitary messages can have indeed a different effect on consumers’ perception of hedonic products. The consumption of hedonic products is associated with a need for justification and feelings of guilt (Kivetz and Simonson 2002). Although hedonic consumption generates positive emotions, hedonic choices are difficult to justify and may have negative consequences in the long run, increasing guilt (Chitturi et al. 2007; Hoch and Loewenstein 1997; Okada 2005). The policy makers’ objective for including sanitary messages in advertisements for foods is to encourage people to have a balanced diet consuming healthy foods and reducing the appeal of the advertised hedonic products. Nevertheless, the presence of the health message can be a justification for the hedonic choice. The presentation of a justification for the choice of a hedonic product increases its consumption (Khan and Dhar 2006; Kivetz and Simonson 2002; Lee and Shavitt 2009). In the same sense, making a hedonic decision and claiming to do something good for others just after it reduces the feeling of guilt and finally encourages the hedonic choice (Strahilevitz and Myers 1998). The Compensatory Health Beliefs Model, for example, argues that the negative effects of an unhealthy behavior can be neutralized by engaging in another healthy behavior (Rabiau et al. 2006). Here, presenting a hedonic food simultaneously with the solution to face the consequences of its consumption (sanitary message) may have an unexpected opposite effect. Individuals may look at the advertising for a hedonic product with the sanitary message and think “it is ok to eat this food as long as I exercise regularly later”.
The objective of this article is to verify the effects of obesity prevention sanitary messages used in France on the explicit and implicit representations towards hedonic foods and on the participants’ eating decisions. Recent research demonstrated that food advertising promoting snacking and fun automatically triggered consumption of available food, increasing food intake outside of participants’ conscious awareness (Harris et al. 2009). Previous research demonstrated that implicit attitudes can relevantly predict eating behavior (choice of a fruit or a snack), and that implicit and explicit attitudes did not correlate significantly between each other (Karpinski and Hilton 2001; Perugini 2005). Therefore, we expect justification effects of sanitary messages to happen only at an implicit level (Cunningham and Zelazo 2007). In contrast, answering explicit questions requires in depth elaboration of both the advertisement and the sanitary message. In this context, the justification effect of the sanitary message may be weaker, due to participants’ need for coherence.
Given the lack of research on the effects of obesity prevention sanitary messages in food advertisements on food decisions, we also wanted to investigate the effects of these messages on eating behavior. Following the conceptualization presented above, we expected that choice of an unhealthy food would be higher for those exposed to a hedonic food advertisement with the sanitary message than for those exposed to the advertisement without the sanitary message. We further wanted to test potential differences in calorie estimations (a proxy for how healthy a product is considered to be; Chandon and Wansink 2007) for the advertised product after exposure to the advertisement with or without the sanitary message. If the sanitary message acts as a justification for hedonic consumption, its presence should lead to reduced calorie estimates in comparison to when it is absent.
We conducted a survey to identify a sanitary message with the highest level of familiarity and a hedonic product that was both liked and familiar to the target population. In a separate pretest, 92 participants from the same population used in the primary study first indicated their spontaneous recall of prevention messages included in food advertisements. They were then asked to rate their familiarity with 17 sanitary messages for obesity prevention and tobacco control, on a seven-point scale (1 = “not familiar at all,” 7 = “very familiar”). The target messages were the four sanitary messages that are actually used to prevent obesity in France. Next, participants rated their familiarity and prior attitudes towards different hedonic snacking products, such as Mars chocolate bar, Big Mac from McDonald’s, or the French Burger from Quick (a popular fast-food restaurant in France).
A single health message was chosen for the main study: “For your health, eat at least five fruits and vegetables per day”. This was the best-recalled message among the four mandatory sanitary messages imposed by the French government (INPES 2007). Our pre-test results showed that this message had the highest level of spontaneous recall (65.2 %) and familiarity (M = 6.91; SD = 0.484) among the target population. Three paired-samples t tests were conducted to compare familiarity scores among the target and the three other sanitary messages, all revealing significant differences (“Avoid foods that are too fatty, too sweet or salty”: M = 6.71, SD = 0.764, t(91) = 3.023, p = 0.003; “Avoid snacking between meals”: M = 6.05, SD = 1.354, t(91) = 5.903, p = 0.000; “Adopt a regular physical activity”: M = 6.45, SD = 1.020, t(91) = 3.858, p = 0.000).
The hedonic food advertisement used was a real advertisement for McDonald’s showing the picture of a Big Mac. Our pre-test results indicate that this product is highly familiar (M = 6.09; SD = 1.545) and highly appreciated (attitude towards the product: M = 4.91; SD = 1.838) by the target population. Big Mac scored higher than all the other snacking products pre-tested both in familiarity and in product attitudes (all p’s < 0.05).
To implicitly measure the memory representations associated with the product, we developed a visuo-semantic priming protocol. The objective of this priming is to measure participants’ reaction time to process a stimulus (target) after the presentation of another stimulus (prime), without making any explicit link between these two phases (processing of the prime, processing of the target). If the prime is related (semantically or perceptively) to the target, then one should observe a facilitation effect, i.e., shorter reaction times to process the target than if the target’s presentation follows the processing of an unrelated prime. In the present study, participants were instructed to determine as quickly as possible if character strings were real French words or not (lexical decision task, originally developed by Meyer and Schvaneveldt 1971) after having been exposed to the advertisement. The presentation of primes and targets was controlled through computer using E-prime 2.0 (Psychology Software Tools, Inc., Pittsburgh, PA, USA). The test comprised a total of 96 sequences including 12 test and 84 distractive sequences designed to prevent participants from understanding the real objective of the study. Each sequence was comprised of a fixation point presented at the center of the screen for 1,000 ms, followed by the presentation of the prime (Big Mac advertisement or distractive image), and finally by the presentation of the target until the participant produced a response (if the participant did not respond after 3,000 ms, the computer terminated the trial and recorded an error). Due to the perceptual complexity of the prime, the Stimulus Onset Asynchrony between prime and target was 1,000 ms. The total duration of this test was approximately 5 min. The 12 test sequences were constituted by the presentation of the Big Mac advertisement (prime) followed by the presentation of a real French word (target). The words used as semantic concepts targets were six positive concepts (health, nutrition, wellness, pleasure, taste, and credibility) and six negative concepts (obesity, constraint, self-indulgence, guilt, weakness, and confusion). Half of the participants were exposed to the Big Mac advertisement alone as a prime in the test sequences and the other half were exposed to the Big Mac advertisement including the sanitary message.
Participants were then asked to fill in a questionnaire. Before answering, they were presented the Big Mac picture with or without the sanitary message for 5 s. Participants answered questions about their attitudes towards the product (Homer 1990; Voss et al. 2003) and general questions, such as gender, age, weight and height, and the time elapsed since their last meal. After finishing the questionnaire, participants received a McDonald’s coupon as a thank you gift for their participation in the study. They had the choice between two coupons for a McDonald dessert: a sundae (unhealthy option) or a bag of fruit (healthy option).
One-hundred thirty-one undergraduate students in a business school in France participated in the main experiment in exchange for course credit. Participants were 67.7 % female, their mean age was 20.03 years old (SD = 0.757), and 92.3 % of them had a normal body mass index (BMI).
For the analysis of the implicit measures results, the response time of the six positive and six negative concepts were grouped in two variables (Cronbach’s alphas were 0.71 and 0.65, respectively). Longer reaction times represent weaker associations between the prime (advertisement with or without the sanitary message) and the concept tested. We conducted separate analyses of variance using Bonferroni correction to control for inflated error on reaction times for the negative and positive semantic concepts, with the presence or absence of sanitary message as the independent variable. For the negative concepts, there was a significant effect of condition on reaction times, F(1,130) = 7.89, p < 0.01. Participants associated negative concepts more easily to the product when the advertisement was presented without the sanitary message (M = 681 ms; SD = 153) than when it was presented with a health sanitary message (M = 769 ms; SD = 201). The sanitary message diminishes negative associations with the advertised product. There were no significant differences across groups in reaction times for positive concepts, F(1, 130) = 1.28, p > 0.20.
To analyze the effects of the sanitary message factor on the explicit attitudes towards the advertised product, we first performed a factorial analysis of the scale of product attitudes (25 items). Results indicated three dimensions and three variables for the utilitarian, hedonic and guilt-related dimensions of the attitude towards the product were created. These variables were included as dependent variables in separate analyses of variance with the presence versus the absence of the sanitary message as the between-subjects factor. Results indicated no effects of the sanitary messages on the explicit attitudes towards the product (all p’s > 0.05).
To test for the effects of the sanitary message on food choice, we performed a logistic regression analysis. We entered the sanitary message as the independent variable (without SM coded 0 and with SM coded 1), the choice as the dependent variable (unhealthy choice coded 0 and healthy choice coded 1), and BMI, gender and time elapsed since last meal (an indicator of participants’ level of hunger) were included as covariates. Results indicated a main effect for the sanitary message (B = −0.897; Wald = 4.547; p = 0.033). This effect indicates that seeing the advertisement with the sanitary message leads to an increase in the probability of the unhealthy choice. The effects of the covariates were not statistically significant (all p’s > 0.05). An analysis of the frequencies per condition gives an indication of the size of this effect: the choice of a healthy food (bag of fruits) doubles in the absence of sanitary message (18 % with versus 35 % without; chi² = 4.617; p = 0.032).
We also investigated differences in the estimated number of calories in the advertised product (i.e., Big Mac sandwich). An ANOVA revealed a main effect of the sanitary message: participants exposed to the advertisement with the sanitary message made lower calorie estimations of the product (M = 503.03; SD = 311.792) than those exposed to the advertisement without the message (M = 646.32; SD = 420.387; F(1,123) = 4.330; p = 0.04).
Our results indicate a boomerang effect of the sanitary message. In the presence of the sanitary message, the choice of a healthy food diminishes significantly (18 % with versus 35 % without). Furthermore, the presence of the sanitary message conducted participants to make lower calorie estimations of the hedonic product in comparison to the situation when there was no sanitary message. These results indicate that the sanitary message influences participants’ eating decisions and their calorie estimations in the opposite sense: the sanitary message increases the likelihood of choosing an unhealthy dessert and decreases calorie estimations of the advertised product. These counter-intuitive effects probably take place out of the awareness of participants, since there were no differences in the explicit attitudes towards the product across conditions.
The presence of the sanitary message can automatically activate representations associated to a justification for the product consumption, leading to a less negative perception of the hedonic product. Our research shows that the presence of a health sanitary message on an advertisement for a hedonic product made the advertised product less negative at an implicit level. These results highlight an unconscious automatic process that could lead to the association of health messages to a justification for an indulgent behavior, leading therefore to opposite effects than those expected. Due to the positive tone and to the solutions for weight gain presented in the message, it represents a justification for hedonic consumption. The advertisement presents a temptation and simultaneously reminds consumers about the possibility to engage in a compensatory health behavior (Rabiau et al. 2006), favoring thus hedonic food consumption.
One limitation of the current study is that we cannot rule out the possibility that the reactions towards the presence of the sanitary message (relative to the absence) were dependent both on the level of familiarity with the sanitary message and with the target product and on previous attitudes towards the product. However, our pre-test indicated that the level of familiarity with the sanitary message was high and homogeneous among the target population. Pre-test results also revealed high levels of familiarity and positive attitudes towards the tested product. Therefore, we are assuming that familiarity for both the message and the product were overall high among target participants. That being said, a follow-up study measuring message and product familiarity before manipulation might be necessary to statistically control for the potential confounding role of these variables.
When designing prevention campaigns using sanitary messages to promote healthy consumption, public policy makers should consider the combined effect of advertisements and sanitary messages. Indeed, the sanitary messages could act as subtle cues that are priming related representations in memory networks (attitudes, beliefs, or evaluations) and impact future behaviors (Berger and Fitzsimons 2008). Consumers are not exposed to the sanitary message alone and this combined effect can have negative consequences in terms of obesity prevention. Therefore, the pre-test of sanitary messages should include several advertisements to test for their efficacy in both hedonic and utilitarian contexts.
Our findings also indicate that public policy makers should take into account potential justification effects associated with sanitary messages included in food advertisements. They characterize what Wilson (2011) conceptualizes as bloodletting policies, which have counterproductive effects. The sanitary messages actually used in France are very positive and give recommendations on how to have a healthier lifestyle. But sanitary messages can also be designed using a negative framing and highlighting risks associated with the consumption of unhealthy foods. Previous research showed that focusing on discouraging unhealthful behaviors is more effective to enhance health intentions than promoting healthful behaviors (for a meta-analysis, see Keller and Lehmann 2008). This kind of sanitary message should not act as a justification for hedonic consumption and could be considered in amended prevention strategies. Future research could also test the efficacy of this kind of approach.
Future research should also investigate the efficacy of story editing approaches applied to obesity prevention. Story-editing aims to change individuals’ personal interpretations of themselves and the social world in ways that lead to more desirable behaviors (Wilson 2011, p.18). This kind of intervention proved to be efficient in different domains (e.g., to prevent teenage pregnancies, alcohol and drugs’ consumption, or children abuse, and to enhance personal happiness), and several applications to healthy eating behavior could be experimentally tested. The basic idea is that just telling people what to do (e.g., eat five a day) is not enough. It would be better, for example, to create interventions that help them to recognize themselves as healthy eaters or as people that are highly conscious of what they eat.
Another interesting way to change eating habits that could be investigated in future research is to introduce a contextual change, bringing habits back into intentional control (Neal et al. 2006). Naturally changing contexts may be an opportunity to introduce interventions targeting at changing habits (Verplanken and Wood 2006). Obesity prevention strategies targeting kids could be introduced when kids are changing from middle to high school, for example.
Finally, future research should investigate the mechanism behind these justification effects, especially looking into the role of guilt. The presence of the sanitary message could reduce the guilt associated with the hedonic product due to a process of justification. Future research should specifically verify the levels of guilt activated by the presence of the sanitary message.
The objective of this study was to investigate explicit and implicit effects of the mandatory obesity prevention messages included in food advertisements. The results indicate a boomerang effect of the sanitary message diminishing negative implicit beliefs towards the advertised product and favoring the choice of a hedonic product in a future task. Thus, policy makers should investigate and control both explicit and implicit effects of the sanitary messages in different contexts in order to choose the most relevant and effective ones. These messages may have to be adapted depending on the type of advertised product (hedonic or utilitarian), and the use of negative framing may be relevant to avoid boomerang effects in the case of hedonic products. Future research should investigate the effects of sanitary messages that directly discourage the consumption of hedonic foods (e.g., “For your health, avoid foods that are too fatty, too sweet or too salty”) because this kind of format may prevent the justification effects found in the present research.
We thank Jennifer S. Coelho and Frédéric Basso for their helpful comments on an earlier version of this manuscript.