University of Amsterdam Graduate School of Communication
Master’s Thesis
The persuasive effects of message framing
combined with self-efficacy messages on cigarette
packages among people in Europe
Author: Luana Herrerias Stade Student ID: 10602259
Master’s Programme Communication Science
Supervisor: Saar Mollen
Persuasive Communication Word Count: 7812
Pages: 26
TABLE OF CONTENT
INTRODUCTION 1
THEORETICAL BACKGROUND 3
Framing matters 3
The importance of self-efficacy 5
METHOD 10
Participants and design 10
Pre-test 11 Stimulus material 12 Procedure 13 Measures 14 Dependent variables 14 Manipulation check 15 Control variables 15 Post measurement 16 RESULTS 17 Randomization 17 Manipulation check 18 Main analysis 19 DISCUSSION 22 CONCLUSION 26
REFERENCES IV
APPENDIX IX
FIGURES AND TABLES
Figure 1. Extended Parallel Process Model – own illustration according to Witte (1992) 7
Abstract
Objectives: The goal of the study was to assess the effects of message framing on intention to quit smoking, attitude towards smoking, information seeking behavior and actual behavior. In addition, self-efficacy messages were tested for their moderation effect on the main effect (N=161).
Methods: A total of 161 participants were randomly assigned into one of four conditions in a 2 (Frame: gain-frame vs. loss-frame) x 2 (Self-efficacy: high-efficacy vs. low-efficacy) between-subjects experiment. Participants were invited to the online experiment via e-mail and also through social networks like Facebook.
Results: The analysis yielded that there was no significant main effect of message framing on intention to quit smoking, attitude towards smoking and information seeking behavior found. However, it was found that high self-efficacy messages combined with loss framed messages, opposed to low self-efficacy messages combined with loss framed messages were more effective in generating cognitive responses such as intention to quit smoking and information seeking behavior.
Conclusion: There was a moderation effect of self-efficacy on the main effect of message frame and intention to quit smoking and information seeking behavior found. Participants with high self-efficacy combined with loss framed messages scored higher on intentions to quit smoking than people with low self-efficacy and loss framed messages. In addition, a moderation effect of self-efficacy on the main effect of message framing on information seeking behavior was found. It is also suggested, that self-efficacy can be enhanced on health warning labels on cigarette packages.
Keywords: message framing, gain frame, loss frame, self-efficacy, high self-efficacy, low self-efficacy, threat, smoking cessation.
Introduction
In the Fifties, Marlboro, in their huge media campaigns used to promote an image of smoking linked to freedom to encourage people to smoke. Today, 50 years later, it is well established that at least four actors of the Marlboro campaign have died due to smoking related diseases, as for example lung cancer (Pearce, 2014). Society’s opinion on smoking has changed and people have become aware of the heath threatening consequences of smoking. Nevertheless, tobacco use is still one of the biggest public health threats
worldwide. The World Health Organization (WHO) announced, that tobacco kills nearly 6 million people each year, 5 million of those are direct tobacco users and more than 600 000 deaths are the result of non-smokers being exposed to secondhand smoke (WHO, 2014a). Compared to other parts of the world, Europe has the highest proportions of deaths caused by smoking (WHO, 2014b).
Health warning labels on cigarette packages are designed to inform people about the negative consequences of smoking cigarettes. A total of 140 countries have ratified the Framework Convention on Tobacco Control (FCTC), meaning that those countries use at least 30 percent of a cigarette package to display health warning labels. Those labels are either displayed as a graphic depiction of consequences of smoking or as a textual warning only (Hammond et al., 2007). Current warning labels on cigarette packages are usually framed as losses, emphasizing the negative consequences of continuing smoking, which are likely to enhance negative emotions such as fear (Cox & Cox, 2001).
Researchers differ in their opinions on how to maximize the effectiveness of cigarette warning labels. A study comparing the impact of warning labels on cigarette packages in four different countries (i.e., Canada, United States, UK, Australia) has found that graphic warning labels (i.e., threatening pictures) are more effective in generating cognitive responses, such as recall, compared to text-only warnings (Borland et al., 2009).
In addition, research conducted in Australia has found evidence that the
introduction of graphic health warning labels on cigarette packages is an effective step towards quitting smoking habits. Those pictures display negative outcomes of continuing smoking (Wakefield, Germain & Durkin, 2008). These studies claim to have found support for the effectiveness of messages that stress negative outcomes of smoking behavior.
Other researchers, however, have criticized these studies, as they lack both in a true experimental design and behavioral outcome measures. A causal relationship between threatening cigarette warning labels and actual smoking cessation can therefore not be proven (Ruiter & Kok, 2005; Ruiter & Kok, 2006). A recent meta-analysis researching the effects of threatening messages on actual health behavior, showed that threatening
messages, as currently used on cigarette packages, are not effective, especially if people have a low sense of self-efficacy (Peters, Ruiter & Kok, 2013). Self-efficacy refers to the belief that one is able to perform certain behavior in order to avoid the health threat that is stressed in a health warning. Research suggested that people with a high level of self-efficacy are more likely to be able to quit smoking whereas people with low self-self-efficacy are more likely to avoid the threat in other ways, because they do not feel able to perform the given behavior such as quitting smoking. Therefore, threatening health warning labels
can only be effective when combined with high self-efficacy beliefs (Wong & Capella, 2009), while current health warning labels only refer to threatening messages displayed as losses. A vast amount of studies can be found that analyzed the effects of health warning labels on attitude and intention to quit smoking (Peters, Ruiter & Kok, 2013; Hammond et al., 2007; Wakefield, Germain, & Durkin, 2008). However, the findings of previous research differ remarkably as both lack on a true experimental design and behavioral outcome measures.
Therefore, it is crucial to get more insight into how health warning labels should be formulated to evoke behavior changes such as smoking cessation. In addition, it is
necessary to analyze the impact of efficacy and whether it is possible to increase self-efficacy through messages on cigarette packages.
The research question central in this study is whether warning labels that stress positive outcomes associated with quitting smoking (i.e., gains) are more effective in generating cognitive and behavioral responses such as attitude towards smoking and intention to quit smoking and actual behavior, than messages that in contrast stress
negative consequences of continuing smoking, which currently are most frequently used as health warning labels. In addition, the crucial role of manipulating self-efficacy on
cigarette packages will be studied.
Theoretical Background
Framing matters
Prospect theory (Tversky & Kahneman, 1981) claims that people presented with messages framed as losses are more likely to become risk seeking in order to prevent those losses, whereas people presented with messages framed as gains, are more likely to avoid taking a risk to secure those gains (Kühberger, Schulte-Mecklenbeck, & Perner, 1999). Framing theory, which applies the premises of Prospect theory to health behavior, explains how health messages influence behavior, suggesting that using differential frames (gain frame vs. loss frame) of messages will result in different outcomes of behavior. There is a distinction to make between so called detection behavior and prevention behavior
(Rothman & Salovey, 1997).
According to Framing Theory a detection behavior, as for example getting checked for skin cancer, includes high risk for a negative outcome. Therefore, it is more likely to be performed when emphasizing potential losses that may be incurred if one does not perform the recommended behavior, since people become more risk seeking. A loss framed
message highlights the disadvantages of inaction. A gain framed message, on the other hand, is seen as being effective for prevention behavior since it highlights the advantages of performing the recommended behavior. People are more likely to make a decision that ensures certainty, because they become more risk-averse and try to secure those gains. Gain framed messages, therefore, should be more effective for behavior that associates little risk. In summary, health behaviors can differ in the amount that is associated with them (Rothman, Bartels, Wlaschin & Salovey, 2006). Prevention behaviors, such as quitting smoking are generally associated with little risk, as the decision to continue smoking is riskier.
A recent meta-analysis (Gallagher & Updegraff, 2012) assessed the persuasive impact of message framing on intention, attitude and behavior, by using 189 effect sizes out of 94 peer-reviewed studies, which compared gain framed messages to loss framed messages. The meta-analysis found that gain-framed messages, opposed to loss-framed messages are more effective for prevention behavior such as, skin cancer and physical activity, but also smoking cessation. Contrary, loss-framed messages were not found to be significantly more likely to promote detection behavior. Furthermore, the study
emphasized that it is not sufficient to only assess beliefs such as intentions to quit smoking and attitudes towards smoking but that actual behavior has to be taken into account when assessing the impact of framing in health communication (Gallagher & Updegraff, 2012).
Another critic on previous research that claimed the superior effects of loss framed messages, is the lack on both a true experimental design and behavioral outcome measures. Therefore, causality cannot be proven (Ruiter & Kok, 2005; Ruiter & Kok, 2006). Findings of previous research have been mixed in their results of effectiveness. It is necessary to prove causality and also measure whether people actually changed their behavior after
being exposed to different message frames. Based on the literature following is hypothesized:
H1: A gain-framed message stressing positive outcomes associated with quitting smoking, opposed to a loss-framed message stressing negative consequences of continuing smoking on cigarette packages is more effective in generating cognitive and behavioral responses such as intention to quit smoking, attitude towards smoking, information seeking behavior and actual smoking
cessation.
The importance of self-efficacy
Research has shown that framing can have different effects on people.
Nevertheless, different studies found that the main effect of message framing on different outcomes such as intention to quit smoking and behavior might be moderated by different factors, as for example self-efficacy (Covey, 2014). Self-efficacy refers to the belief that one is able (not able) to perform a certain behavior, such as for example smoking
cessation. Wong and Capella (2009) found that people with low self-efficacy, presented with threatening messages are more likely to avoid the threat because they are not able to perform the given behavior, such as quitting smoking. People with high-self efficacy, on the other hand, when presented with threatening messages are more likely to engage in the recommended behavior. Therefore, threatening health warning labels, emphasizing the
losses of inaction, can only be effective when combined with high self-efficacy beliefs (Wong & Capella, 2009).
Current health warning labels on cigarette packages are mostly framed as losses of inaction. Those labels display consequences of continuing smoking, such as ‘If you smoke, your lungs will get damaged’. A research that investigated the effects of framing on
consumers’ beliefs and attitude towards screening, found that loss framed messages, opposed to gain framed messages tend to elicit negative emotions and, especially fear (Cox
& Cox, 2001). In this regard the Extended Parallel Process Model (EPPM) provides an explanation for why emphasizing benefits can be more effective than highlighting the losses in health behavior in some cases. The model predicts how people react differently when presented with fear appeals. Fear appeals emphasize the negative consequences of a behavior if someone continues with risky behavior such as smoking (‘If you smoke, your lungs will get damaged’). The EPPM makes predictions about the cognitive (danger control) and emotional (fear control) processes that may occur when faced with fear appeals. The perceived threat refers to cognitions about a harm or danger that exist, which includes two underlying dimensions: severity and susceptibility
Severity refers to the beliefs about the magnitude of the threat. Susceptibility is the belief about the likelihood of experiencing the threat. When a person is presented with a fear appeal, such as for example a health warning label on a cigarette package, two different reactions can occur depending on the threat and the efficacy. Efficacy includes two underlying dimensions: response efficacy and self-efficacy.
Response efficacy refers to the beliefs about the effectiveness of a response in averting the threat, while self-efficacy refers to the belief whether a person is able (not able) to perform a certain behavior. According to the EPPM, after exposure to a
threatening message, a person will evaluate the threat first. If the threat experienced by a health warning label is low, the person will not process the message further and there will be no effect. If people feel threatened, they will try to reduce their fear. Subsequently, the person evaluates the efficacy (self-efficacy, response efficacy) of the recommended response to the threat. This will determine whether a person will control the danger of the threat or control his fear. If a person believes he or she can perform the recommended response (high self-efficacy, high response efficacy) it will result in danger control, since they will start to think about how to reduce the threat. Danger control is a cognitive process eliciting protection motivation. People will think about strategies to avert the threat.
Figure 1. Extended Parallel Process Model – own illustration according to Witte (1992) On the other hand, if a person thinks he or she is not able to act in the
recommended response (low self-efficacy, low response efficacy), the person will control the fear, leading to denial, defensive avoidance or reactance. This theory postulates that fear appeals can only be effective when efficacy is high (Witte, 1992; Witte & Allen, 2000).
Applying the EPPM to health warning labels on cigarette packages can cast light on how people process the warning messages. Health warnings on cigarette packages mostly communicate the negative consequences of smoking, which are likely to trigger negative emotions such as fear (Cox & Cox, 2001).
A vast amount of studies found evidence that many smokers react defensively to threatening health warning labels and therefore, emphasize the ineffectiveness of the current health warning labels on cigarette packages (Kessels, Ruiter, & Jansma, 2010; Maynard, Munafò, & Leonards, 2013). De Vries and Backbier (1994) confirmed the importance of self-efficacy in persuasion. According to their study it is a major determinant of people’s motivation to engage in healthy behavior.
A systematic review and meta-analysis (Covey, 2014) investigating a variety of dispositional factors on the effect of gain and loss framing in health messages found that the effectiveness of framing might depend on the disposition of a recipient. The systematic review of 47 reports published between 1990 and 2012 tested 23 different moderators of framing (gain frame vs. loss frame). The review emphasized the crucial role of taking recipients’ dispositional factors into account when using framing techniques. According to the meta-analysis consistent interactions were found for moderators such as self-efficacy beliefs, regulatory focus, and approach avoidance motivation (Covey, 2014). The review revealed that most studies found that loss framed messages were most persuasive with people with high self-efficacy (Riet, Ruiter, Werrij, & De Vries, 2008; Williams et al., 2001). Only one study did not find significant interactions of self-efficacy and framing (Sanchez, 2006). However, it is suggested that further research should investigate the effects in more in-depth in different contexts (Covey, 2014).
A revised meta-analysis (Peters, Ruiter & Kok, 2013), testing fear appeal theory, found that threat and efficacy indeed interact with each other in their effects on health behavior, meaning that the threat can only be effective in enhancing behavioral change if a persons’ efficacy is high. In line with those findings a study examining the impact of self-efficacy on the effects of framing found that people with high self-self-efficacy presented with loss framed messages resulted in a higher intention to perform skin self-examination than gain framed messages (Riet, Ruiter, Werrij & de Vries, 2008). The authors proved that loss frame messages could indeed be effective but only if a person has high self-efficacy. Additionally, a trend was found that people with low self-efficacy presented with fear appeals showed a negative effect. This means that people are more likely to engage in health defeating behavior and are endangered to engage in risky behavior (Peters, Ruiter & Kok, 2013). Applying this to the case of smoking, it might be a dangerous path to use loss framed messages without considering the crucial effects of self-efficacy.
According to the EPPM, threatening messages can only lead to a desired outcome, such as smoking cessation, when the efficacy of a person is high. That means, when the person perceives that he or she is able to perform the given behavior. Therefore, it is crucial that the concept of self-efficacy is considered when analyzing the effects of gain and loss framed messages on cigarette packages. In order to change certain health related behavior such as smoking, it seems to be necessary to access the effectiveness of gain and loss framed messages and the impact of self-efficacy.
This study aims to analyze whether enhancing smokers' self-efficacy on cigarette packages through messages could make a contribution to the effect of gain frame and loss frame messages on intention to quit smoking, attitude towards smoking, information seeking behavior and actual behavior. Knowing how people in Europe are affected by framing techniques can be a contribution to the current debates in many countries. Gain and loss frame messages might only be effective if certain conditions such as strong self-efficacy beliefs are present. Previous research failed in providing a true experimental design, therefore a causal relationship between threatening cigarette warning labels and quitting smoking was not proven. In addition, actual behavior was not included as
measurement (Ruiter & Kok, 2005; Ruiter & Kok, 2006). Furthermore, this study does not analyze a persons’ self-efficacy belief but adds knowledge to the effects of manipulating self-efficacy on cigarette packages. In order to close the existing gap in research the following hypothesis is formulated:
H2: A high self-efficacy message in combination with a loss framed message, opposed to a low self-efficacy message in combination with a loss framed messages is more effective in generating cognitive and behavioral responses such as attitude towards smoking, intention to quit smoking, information seeking behavior and actual smoking cessation. The difference of low
self-efficacy and high self-self-efficacy, however is not expected for gain framed messages.
Method
Participants and design
Participants were approached on the street and were invited by email to participate in the online experiment. In addition, the online experiment was shared on social networks like Facebook, which resulted in a convenience sample. A total of 193 participants were randomly assigned into one of four conditions in a 2 (Frame: gain-frame vs. loss-frame) x 2 (Self-efficacy: high-efficacy vs. low-efficacy) between-subjects experiment.
Other studies have been criticized, as they lack both on a true experimental design and behavioral outcome measures (Ruiter & Kok, 2005; Ruiter & Kok, 2006). Therefore, an online experiment was conducted, which can prove causality and is strong in internal validity. In addition, different outcomes, such as intention to quit smoking, attitude towards smoking, information seeking behavior and actual behavior were assessed in the online questionnaire.
The research was conducted among smokers residing in Europe, older than 18 years with sufficient English language skills. This was necessary since the study was conducted in English.
At the beginning of the questionnaire, participants were asked about their age, about the level of English, whether they reside in Europe and whether they are smokers. Non-smokers (n = 9), participants under 18 (n = 3) and people that did not reside in Europe (n = 6) were excluded from the experiment. Additionally, those participants that dropped out of the experiment earlier (n = 12) and were not fluent in English (n = 2) were excluded. A total of 161 participants (Mage = 25.61; SDage = 6.60; 74 men) remained ranging from
18 until 88 years. It can be concluded that the sample is well educated since the biggest group of participants consisted of people with University degrees (44.10 % Bachelor level, 28 % Master level). Participants declared that they are currently students (55.30 %) or working (43 %). Generally, it can be said that the sample included people that are not particularly heavy smokers (37.40 % indicated to smoke 0-5 cigarettes a day). Marlboro turned out to be the most liked brand (31.30 %), followed by Gauloises (17.80 %) and Lucky Strike (14.70 %). Participants were equally distributed into the four conditions, low efficacy vs. loss frame (n = 40), low efficacy vs. gain frame (n = 42), high self-efficacy vs. loss frame (n = 38) and high self-self-efficacy vs. gain frame (n = 41). The ethical committee approved the study design.
Pre-test
To select appropriate gain and loss framed messages (positive outcomes of quitting smoking vs. negative consequences of not quitting smoking) and self-efficacy messages (low self-efficacy vs. high self-efficacy), a pre-test was conducted. The aim of this pre-test was to detect one of the most negative and one of the most positive perceived message of a total of 11 messages framed to either emphasize losses associated with inaction, or gains associated with quitting smoking (see Appendix table 1). In addition, 18 self-efficacy messages were pre-tested in order to choose the most helpful message and the least helpful message for self-efficacy (see Appendix table 2).
A total of 31 people between 22 and 32 years old participated in the pre-test (Mage = 25.52; SDage = 6.52; 15 men) Participants were asked to rate the 22 outcomes of
(quitting) smoking on a scale running from 1 (very negative) to 5 (very positive). In addition, in order to find out whether the loss framed messages were indeed perceived as more threatening, than the gain framed messages, participants were asked to rate the consequences on a scale running from 1 (not threatening at all) to 5 (extremely
threatening). Lastly, participants were asked to rate 18 self-efficacy messages on a scale running from 1 (not helpful at all) to 5 (very helpful). Based on the findings the two most suitable messages for framing and self-efficacy were chosen.
Valence. There was a significant difference for the messages ‘If you quit smoking, your lungs will last longer’ (M = 4.19, SD = .80) and ‘If you smoke, your lungs will get damaged’ (M = 2.16, SD = 1.40); t (30) = 6.80, p = .000 found. Out of the 11 loss framed messages and 11 gain framed messages, those two messages showed the largest difference between each other in regards of positive and negative outcomes.
Threat. There was a significant difference for the messages ‘if you quit smoking, your lungs will last longer’ (M = 1. 81, SD = 1.25) and ‘if you smoke, your lungs will get damaged’ (M = 4.19, SD = .87); t (30) = -6.89, p = .000 found. Out of the 11 loss framed messages and 11 gain framed message, those two messages showed the largest difference to each other in regards to threat.
Self-efficacy. There was a significant difference for the messages ‘Craving for a cigarette? Have a gum instead’ (M = 3.42, SD = 1.31) and ‘Craving for a cigarette? Count to 10’ (M = 1.84, SD = .97); t (30) = 5.47, p = .000 found. Out of 18 self-efficacy
messages, those two messages showed the largest difference between each other in regard of helpfulness.
Stimulus material
Based on the results of the pre-test images of cigarette packages including messages were developed aiming to be as realistic as possible. Twelve cigarette brands (i.e., Marlboro Red, L&M Red, Lucky Strike Red, Pall Mall Red, Marlboro Gold, Camel Filter, Gauloises Blondes Blue, Kent Red, Johns Player Special Red, Camel Blue,
cover the most popular cigarette brands in Europe. Pictures of those packages were used in which the original messages were replaced by the ones that were selected as the strongest messages for the aim of the study (see Appendix).
Procedure
Participants were told that the aim of the study was to analyze smoking preferences and that therefore, they would be presented with 12 different pairs of cigarette brands and that they would be asked to indicate which ones they would prefer to smoke. The smoking preferences procedure was used to avoid demand awareness and to provide a natural exposure to the manipulation. After a brief explanation of the procedure of the study and providing their consent, participants were asked some general demographic and other background questions (i.e., age, gender, favorite cigarette brand, educational level). In addition, they were asked about their nicotine dependence. If people did not meet the requirements of the study, such as being under 18, not living in Europe and not being a smoker, they could not continue with the experiment and were immediately redirected to the end of the survey.
The online experiment included 12 different cigarette brands (Gauloises, Marlboro, Marlboro Gold, Marlboro Flavor Mix, Kent, Camel, Camel Blue, Lucky Strike, John Player Special, L&M, Kent iSwitch, Pall Mall). Each brand was presented twice through out every condition, once on the left screen and once on the right screen, in order to avoid associations with a certain brand. Participants were always first exposed to a self-efficacy message (low self-efficacy: ‘Craving for a cigarette? Count to 10’ / high self-efficacy: ‘Craving for a cigarette? Have a gum instead’) followed by a framed health message (loss frame: ‘If you smoke, your lungs will get damaged’ / gain frame: ‘If you quit smoking, your lungs will last longer’). This parallel displaying method was then repeated 5 times, so participants saw each self-efficacy message six times and each frame message six times.
After the smoking preferences task, the dependent variables intention to quit smoking and attitude towards smoking were assessed. In addition, participants were asked to what extent they would like to receive more information regarding smoking cessation. Finally, participants were asked if they would like to participate in a short follow-up study of five minutes seven days later. They were told that when taking part in both parts of the study they could win one out of ten vouchers of 10 euros. The follow up measurement was sent to the ones that agreed, seven days later.
Measures
Dependent variables
Attitude. The attitude towards quitting smoking was measured through five items. Participants were asked whether they thought quitting smoking is (1 = unhealthy -7 = healthy; 1 = bad - 7 = good; 1 = unwise - 7 = wise; 1 = harmful - 7 = beneficial; 1 = unpleasant - 7 = pleasant) (Latimer et al., 2012; Moorman & van den Putte, 2008). The scale including all five items yielded a Cronbach’s α=.86, (M = 5.45, SD = 1.59).
Intention. The intention to quit smoking was measured by asking five statements with which participants indicated their agreement (1 = fully disagree, 7 = fully agree). “I’m willing to quit smoking within the next three months”, "I intend to quit smoking within the next three months", "I plan to quit smoking within the next three months", "I will try to quit smoking within the next three months", "I expect to have quit smoking in three
months" (Gibbons, Gerrard, Ouellette, & Burzette, 1998; Latimer et al., 2012; Moorman & van den Putte, 2008). The scale including all five items yielded a Cronbach’s α =.97, (M = 3.62, SD = 1.85).
Information seeking behavior. To measure information seeking behavior,
participants were asked to indicate on a scale running from 1 (not at all) to 7 (very much) if they would like to be provided with information regarding quitting smoking.
Manipulation check
Message frame. In order to examine whether the manipulation of message frame was successful, participants were presented with the health warning messages they had seen before and were asked to indicate on a 7-point scale to what extent they thought the consequences were positive or negative (gain-frame, loss-frame).
Self-efficacy. To examine whether the self-efficacy manipulation was successful, participants were presented with the self-efficacy message they had seen before and were asked to indicate on a 7-point scale to what extent they thought the message was helpful or not helpful at all.
Control variables
Nicotine dependence. The level of addiction to nicotine was assessed using the Heaviness of Smoking Index (Heatherton, Kozlowski, Frecker, Rickert, & Robinson, 1989). This scale consists of two items: ‘On average, how many cigarettes do you smoke each day?’ (1 = 0-5, 2 = 6-10, 3 = 11-20, 4 = 21-30, 5 = 31 or more), (M = 2.03, SD = .98) and ‘How many minutes after you wake up in the morning do you light your first
cigarette?’ (1 = maximum of 5 minutes, 2 = 6-30 minutes, 3 = 31-60 minutes, 4 = 61-120 minutes, 5 = more than 120), (M = 2.40, SD = 1.23).
A correlation analysis showed that the two items of nicotine dependence were correlated with each other, r = .64, p = .000, meaning that people that smoke more cigarettes on average a day are more likely to light up their first cigarette a day, earlier, than those that smoke less cigarettes on average a day.
Post measurement
Previous research criticized several studies of framing effects, since they lack both on a true experimental design and behavioral outcome measures (Ruiter & Kok, 2005; Ruiter & Kok, 2006). In order to measure actual behavioral change as a result of exposure to the self-efficacy messages and framing messages it was necessary to conduct a post measurement test.
A total of 38 people participated in the follow up study. After removing two participants from the study because they did not finish it, a total of 36 participants remained (22%) (Mage = 25.39; SDage = 3.80; 20 men). First, participants were asked some general background questions (i.e., gender, age). In addition, they were asked whether they made an attempt to quit smoking in the past seven days (n = 11). In order to measure actual behavior as an outcome, participants were asked how many cigarettes they smoked daily on average in the last seven days. Only two people indicated of not having smoked at all. In addition, one person indicated of having smoked 20 cigarettes a day on average, which was the highest score. Most people indicated of having smoked on average between 2 and 5 cigarettes a day (16 %). Furthermore, they were asked about their
intention to quit smoking (M = 3.65, SD = 2.11) and their attitude towards smoking (M = 5.43, SD = 1.76). For this the same questions were used as in the main experiment. In addition, they were asked on a scale running from 1 (not at all) to 7 (very much) whether they want to receive more information regarding quitting smoking (M = 3.56, SD = 2.41). To measure actual information seeking behavior, participants were asked if they wanted to be redirected to a website with information regarding quitting smoking after finishing the study. Participants who indicated yes (n = 11) were redirected to a website from the
National Cancer Institute's (NCI) Tobacco Control Research Branch which provides tobacco- and other health-related information to help people to be healthier and quit
smoking (http://www.smokefree.gov). The sample size was too small and therefore not representative to analyze in terms of actual behavioral change. Therefore, this study focused on the main survey.
Results
Randomization
To examine whether randomization was successful and participants were equally distributed between the four conditions a factorial ANOVA was conducted with, age and nicotine dependence as the dependent variables and self-efficacy and message frame as independent variable. The analyses showed that there was no significant difference between groups that received a gain or loss framed message in terms of age found, F (1, 157) = 1.34, p = .248, η2= .00. Furthermore, there was no main effect of self-efficacy, F (1, 157) = 2.70, p = .604, η2= .00, and no interaction between message frame and self-efficacy found, F (1, 157) = 2.23, p = .137, η2= .01
Additionally, the results showed that there was no significant difference between groups that received a gain or loss framed message in terms of nicotine dependence found, F (1, 157) = .43, p = .514, η2= .00. There was no main effect of self-efficacy either, F (1, 157) = .008, p = .927, η2= .00, nor was there an interaction between message frame and self-efficacy found, F (1, 157) = 1.45, p = .230, η2= .00.
To assess whether the randomization into the four conditions was successful in terms of gender, a Chi-square test was conducted. The test yielded that the percentages of females and males in each group did not differ significantly from each other, χ2 (3, N = 161) = 1.97, p = .578. The low self-efficacy and loss frame condition consisted of 15 (9.30 %) males and 25 (15.50 %) females, the low self-efficacy and gain frame condition
frame condition had 20 (12.40 %) males and 18 (11.20 %) females and the high self-efficacy and gain frame condition included 20 (12.40 %) males and 21 (13 %) females.
A correlation analysis revealed that there is no correlation between nicotine dependence and the dependent variables intention to quit smoking, r = .007, p = .926 and attitude towards smoking, r = -.144, p = .069. Therefore, it was not deemed necessary to control for nicotine dependence.
Manipulation check
Message frame. In order to test whether the manipulation in the main experiment was successful a factorial ANOVA with message frame and self-efficacy as the
independent variable and perceived valence as the dependent variable was conducted. As expected a significant difference between loss frame and gain frame messages on
perceived valence was found, F (1, 157) = 73.51, p = .000, η2= .31, those in the gain framed condition rated the messages as significantly more positive (M = 5.05, SD = 1.86), compared to those in the loss framed condition (M = 2.68, SD = 1.66). The main effect of self-efficacy on perceived valence, F (1, 157) = 2.11, p = .147, η2= .01 and the interaction between message frame and self-efficacy, F (1, 157) = 1.65, p = .200, η2= .01 was non-significant.
Self-efficacy. To analyze whether the manipulation of self-efficacy on the cigarette packages was successful another factorial ANOVA was conducted. As expected a
significant difference between low self-efficacy and high self-efficacy messages was found, F (1, 157) = 11.28, p = .001, η2= .07, those in the low self-efficacy condition rated the messages as significantly less helpful (M = 2.41, SD = 1.80), than those in the high self-efficacy condition (M = 3.38, SD = 1. 84). The main effect of self-efficacy on helpfulness, F (1, 157) = .493, p = .484, η2= .00 and the interaction between message frame and self-efficacy, F (1, 157) = .364, p = .547, η2= .00 was non-significant.
Main analysis
Attitude. In order to analyze the first hypothesis, that message framing has a main effect on the attitude towards smoking, a factorial ANOVA was conducted. The test yielded a non-significant effect, F (1, 159) = 0.03, p = .853, η2= .00. Therefore the hypothesis (H1), that gain framed messages are more effective in generating responses such as attitude towards smoking, can be rejected. In order to analyze the second hypothesis, that self-efficacy has a moderating effect on the main effect of message framing and attitude towards smoking a factorial ANOVA was conducted. As mentioned before H1 was rejected, no main effect of message framing on attitude towards smoking was found. In addition no main effect of self-efficacy on attitude towards smoking was found, F (1, 157) = 2.90, p = .591, η2= .00. Furthermore, there was no interaction effect of self-efficacy on the main effect of message framing and attitude towards smoking found, F (1, 157) = 0.65, p = .420, η2= .00. Therefore the hypothesis (H2) that high self-efficacy messages combined with loss framed messages, opposed to low self-efficacy messages combined with loss framed messages result in stronger negative attitudes towards smoking, can be rejected.
Intention to quit smoking. In order to analyze the first hypothesis, that message framing has a main effect on intention to quit smoking, a factorial ANOVA was
conducted. The test yielded a non-significant main effect of message framing on intention to quit smoking, F (1, 159) = 2.08, p = .151, η2= .01. Therefore the hypothesis (H1) that gain framed messages are more effective in generating positive cognitive responses such as a higher intention to quit smoking has to be rejected. In order to test the second hypothesis, that self-efficacy has a moderating effect on the main effect of message framing and intention to quit smoking a factorial ANOVA was conducted, using intention to quit smoking as the dependent variable and message frame and self-efficacy as the independent variable. As mentioned above the H1 was rejected. There was also no main effect of
self-efficacy on intention to quit smoking found, F (1, 157) = 2.47, p = .118, η2= .02. However the interaction between message frame and self-efficacy was significant, F (1, 157) = 5.81, p = .017, η2= .04, meaning that self-efficacy can be seen as a significant moderator for the main effect of framing on intention to quit smoking. Those in the high self-efficacy
condition combined with loss framed messages showed higher intentions to quit smoking (M = 4.42, SD = 1.84) than those in the low self-efficacy and loss framed condition (M = 3.28, SD = 1.77). In order to analyze whether the difference between the low self-efficacy and loss framed condition is significantly different, than the high self-efficacy and loss framed condition, a t-test was conducted. The test revealed that the difference between those two groups is significant, t (76) = -2.79, p = .007. Those in the high self-efficacy and gain framed message condition (M = 3.31, SD = 1.88), however, did not show higher intention to quit smoking, than those in the low self-efficacy and gain framed message condition (M = 3.55, SD = 1.77). Therefore the hypothesis (H2) that loss frame messages in combination with high self-efficacy messages are more effective than low self-efficacy and loss framed messages on intention to quit smoking can be accepted.
Information seeking behavior
.
In order to test the first hypothesis, that message framing has a main effect on information seeking behavior, a factorial ANOVA was conducted. The test yielded a non-significant effect, F (1, 159) = .17, p = .682, η2= .00. Therefore, the hypothesis (H1) that gain frame messages, opposed to loss framed messages result in higher information seeking behavior can be rejected. In order to test thehypothesis (H2), that self-efficacy moderates the main effect of message framing and information seeking behavior another factorial ANOVA was conducted. As mentioned above, H1 was rejected, meaning no main effect of message frame on information seeking behavior found. The test yielded that there was a non-significant main effect for self-efficacy on information seeking behavior found, F (1, 157) = .59, p = .448, η2= .00. However, there was a significant interaction effect of self-efficacy on the main effect of
message frame and information seeking behavior found, F (1, 159) = 3.97, p = .048, η2= .03, meaning that self-efficacy can be seen as a moderator of the main effect of framing and information seeking behavior. Those in the high self-efficacy condition combined with loss framed messages showed more interest in receiving information about quitting
smoking (M = 3.74, SD = 2.29) than those in the low self-efficacy and loss framed
condition (M = 2.83, SD = 1.99). Those in the high self-efficacy and gain framed message condition (M = 2.93, SD = 1.96), however, did not show more interest in receiving
information about smoking cessation than those in the low self-efficacy and gain framed message condition (M = 3.33, SD = 2.12).
In order to analyze whether the difference between the low self-efficacy and loss framed condition differs significantly from the high self-efficacy and loss framed condition a t-test was conducted. The test revealed that there is only a marginal significant difference between the groups, t (76) = -1.88, p = .064. Therefore the hypothesis (H2) that
self-efficacy moderates the main effect of gain framed messages and loss framed messages on information seeking behavior can be accepted. For an overview of all means and standard deviations for H2 see table 1.
Table 1. Overview of means and standard deviations of interaction effects of self-efficacy on the main effect of message frame and attitude towards smoking, intention to quit smoking, information seeking behavior
Loss Gain
MlowSE SDlowSE MhighSE SDhighSE MlowSE MlowSE MhighSE SDhighSE
Attitude towards smoking 5.64 1.56 5.30 1.89 5.39 1.35 5.46 1.57 Intention to quit smoking 3.27 1.77 4.41 1.83 3.55 1.77 3.31 1.88 Information seeking behavior 2.83 1.99 3.74 2.28 3.33 2.12 2.93 1.96
In conclusion, hypothesis 1 has to be rejected since there was no main effect of framing on intention to quit smoking, attitude towards smoking and information seeking behavior found. Hypothesis 2 can be partly accepted since there was a moderation effect of self-efficacy on intention to quit smoking and information seeking behavior found.
However, there was no moderation effect of self-efficacy on attitude towards smoking found. The sample size of the post measurement (N = 36) was too small and therefore not representative to analyze in terms of actual behavioral change. Therefore the hypothesis (H1) that gain framed messages, opposed to loss framed messages are more effective in evoking behavioral change such as smoking cessation could not be tested in this study. In addition, the hypothesis (H2) that high self-efficacy messages combined with loss framed messages, opposed to low self-efficacy messages combined with loss framed messages are more effective in evoking behavioral change such as smoking cessation could not be analyzed either.
Discussion
The aim of the study was to analyze whether gain and loss framed messages and self-efficacy enhancing messages displayed on cigarette packages can affect smokers’ intention, attitudes, information seeking behavior and actual behavior in Europe
differently. It was hypothesized that gain framed messages stressing positive outcomes of quitting smoking, opposed to loss framed messages stressing negative outcomes of continuing smoking, result in more positive cognitive responses such as intention to quit smoking, the attitude towards smoking, information seeking behavior and actual behavior (H1).
Furthermore, it was hypothesized that high self-efficacy messages combined with loss framed messages, opposed to low self-efficacy messages combined with low framed
messages are more effective in evoking intention to quit smoking, attitude towards
smoking, information seeking behavior and actual behavior. However, this difference was not expected for gain framed messages. It was expected that people that were presented with a high self-efficacy messages and a loss framed messages will score higher on intention to quit smoking, score more negatively on attitude towards smoking, show more interest in information about smoking cessation and finally will engage in actual behavior such as smoking cessation, compared to those exposed to a low self-efficacy message and loss framed message (H2).
The results showed that there was no superior effect of gain framed messages, opposed to loss framed messages on intention to quit smoking, attitude towards smoking and information seeking behavior found, therefore H1 was rejected. These findings are not in line with theory on message framing (Rothman & Salovey, 1997) and with a meta-analysis that found a superior effect of gain framed messages, opposed to loss framed messages for prevention behavior such as smoking cessation (Gallagher &
Updegraff, 2012).
The effect of message framing on actual behavior could not be analyzed in this study due to a small sample size in the post measurement test. However, according to Gallagher and Updegraff (2012) actual behavioral change should be a core measurement when testing for framing effects. Additionally, behavior change is the most important outcome when talking about smoking cessation. It is suggested that attitudes and intentions do not necessarily lead to actual behavioral change. Someone might have the intention to quit smoking but whether this person will indeed quit smoking is not naturally given by higher intentions towards it (Webb & Sheeran, 2006). Therefore, further studies should aim to incorporate behavior as an outcome when testing for framing effects.
The analyses, testing the second hypothesis showed that there was a moderating effect of self-efficacy on the main effect of message framing and intention to quit smoking
found. In addition, a moderating effect of self-efficacy on the main effect of message framing and information seeking behavior was found. The results yielded that high self-efficacy messages combined with loss framed messages, opposed to low self-self-efficacy messages and loss framed messages, scored higher on intention to quit smoking and also resulted in more interest in receiving information about smoking cessation. However, there was no moderating effect of self-efficacy on the main effect of message framing and attitude towards smoking found. Furthermore, as mentioned above actual behavior was not accessed in this online experiment due to the small sample size of the post measurement test. Therefore, it could not be analyzed whether self-efficacy has a moderating effect on the main effect of message framing and actual behavior. Further research should examine this in more detail.
The second hypothesis could be partly accepted. It was shown that loss framed messages are more effective when combined with high self-efficacy messages, compared to low self-efficacy messages and loss framed messages. This is in line with previous research that underlined the importance of a recipients’ dispositional factor when talking about message framing and its effects (Covey, 2014; Peters, Ruiter & Kok, 2012; Wong & Capella, 2009). However, future research should investigate already existing self-efficacy beliefs of a person, combined with self-efficacy enhancing messages on cigarette packages. Testing how existing self-efficacy beliefs combined with manipulating self-efficacy on cigarette packages interact can add a more in-depth knowledge of how to maximize self-efficacy of a person.
Additionally, further studies should investigate in more in-depth how self-efficacy messages can influence people differently. Based on the pre-test, only one message for high self-efficacy and one message for low self-efficacy were chosen. The sample of this research consisted of people between 18 and 88 years old. However, it might be the case that someone with 18 is differently affected by a self-efficacy message than someone with
88 years old. Self-efficacy might increase over time and be enhanced through experience (Bandura, 1989). Indeed, research has emphasized the dynamics of self-efficacy and that it can change over time (Gwaltney, Shiffman, Balbanis & Paty, 2005). Therefore, it is
necessary to compare different age groups with different self-efficacy message in order to maximize their effects. Analyzing how different self-efficacy messages influence people could add more in-depth knowledge of how to evoke smoking cessation.
An important part of the online experiment was to test whether it is possible to enhance self-efficacy through health label warnings on cigarette packages. The results showed that this might be a solution for people with low self-efficacy beliefs.
The research adds evidence for the crucial role of self-efficacy in threatening health communication, formulated as losses of consequences of inaction. Research found that loss framed messages are more likely to evoke negative emotions such as fear, than gain
framed messages (Cox & Cox, 2001). Currently, health warning labels emphasize on losses of continuing smoking, which according to research are likely to induce fear. Applying the Extended Parallel Process Model (EPPM) on the effects of health warning labels on cigarette packages, it was stressed that current health warning labels might be ineffective due to the cognitive (danger control) and emotional (fear control) processes that occur when a person is confronted with a fear appeal. If a person thinks he or she can perform the recommended response (high self-efficacy) it will result in danger control, since he or she will start to think about how to reduce the threat. If a person believes he or she is not able to quit smoking (low self-efficacy), the person will control the fear, leading to denial, defensive avoidance or reactance (Witte & Allen, 2000; Witte, 1992). Applying the theory and the findings of this study to current health warning labels it is suggested that warning labels might be ineffective since the crucial part of self-efficacy messages is missing.
Conclusion
Overall, the findings suggested that current health warning labels on cigarette packages should be treated with caution. Warning labels stressing negative outcomes of continuing smoking should be reconsidered, as they are not effective when used without encouraging smokers’ self-efficacy. This study adds to the knowledge of how self-efficacy enhancing messages interact with message framing and what effect it has on intention to quit smoking and information seeking behavior. Public policies should be adjusted, so that not only loss framed messages are displayed on cigarette packages but also self-efficacy enhancing messages.
References
Bandura, A. (1987). Regulation of cognitive processes through perceived self-efficacy. Developmental Psychology 25(5), 729-735.
Borland, R., Wilson, N., Fong, G.T., Hammond, D., Cummings, K.M., Yong, H.H., Hosking, W., Hastings, G., Trasher, J., McNeill, A. (2009). Impact of graphic and text warnings on cigarette packs: findings from four countries over five years. Tobacco Control, 18, 358-364.
Covey, J. (2014). The role of dispositional factors in moderating message framing effects. Health Psychology, 33(1), 52-65.
Cox, D, & Cox, A. (2001). Communicating the consequences of early detection: The role of evidence and framing. Journal of Marketing, 65(3), 91-103.
De Vries, H.., & Backbier, E. (1994). Self-efficacy as an important determinant of quitting among pregnant women who smoke: The ø-Pattern. Preventive Medicine, 23, 167– 174.
Gallagher, K. M., & Updegraff, J. a. (2012). Health message framing effects on attitudes, intentions, and behavior: a meta-analytic review. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 43(1), 101–16.
Gibbons, F. X., Gerrard, M., Ouellette, J. a., & Burzette, R. (1998). Cognitive antecedents to adolescent health risk: Discriminating between behavioral intention and
behavioral willingness. Psychology & Health, 13(2), 319-339
Gwaltney, C. J., Shiffman, S., Balabanis, M. H., & Paty, J. A. (2005). Dynamic
self-efficacy and outcome expectancies: Prediction of smoking lapse and relapse. Journal
Hammond, D., Fong, G., Borland, R., Cummings, M., McNeill, A. & Driezen, P. (2007). Text and graphic warnings on cigarette packages. Findings from the international tobaccocontrol four country study, American Journal of Preventive Medicine, 32(3), 202-209.
Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., Rickert, W., & Robinson, J. (1989). Measuring the heaviness of smoking: using self-reported time to the first cigarette of the day and number of cigarettes smoked per day. British Journal of Addiction, 84(7), 791– 9.
Kessels, L. T. E., Ruiter, R. A. C., & Jansma, B. M. (2010). Increased attention but more efficient disengagement: neuroscientific evidence for defensive processing of
threatening health information. Health Psychology : Official Journal of the Division of Health Psychology, American Psychological Association, 29, 346–354.
Kühberger, a, Schulte-Mecklenbeck, M., & Perner, J. (1999). The Effects of Framing, Reflection, Probability, and Payoff on Risk Preference in Choice Tasks.
Organizational Behavior and Human Decision Processes, 78(3), 204–231
Latimer, A. E., Krishnan-Sarin, S., Cavallo, D. a, Duhig, A., Salovey, P., & O’Malley, S. a. (2012). Targeted smoking cessation messages for adolescents. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 50(1), 47–53.
Maynard, O. M., Munafò, M. R., & Leonards, U. (2013). Visual attention to health warnings on plain tobacco packaging in adolescent smokers and non-smokers. Addiction (Abingdon, England), 108(2), 413–9.
Moorman, M., & van den Putte, B. (2008). The influence of message framing, intention to quit smoking, and nicotine dependence on the persuasiveness of smoking cessation messages. Addictive Behaviors, 33(10), 1267–75.
Peters, G.-J. Y, Ruiter, R. a C., & Kok, G. (2013). Threatening communication: a critical re-analysis and a revised meta-analytic test of fear appeal theory. Health Psychology Review, 7(Suppl 1), S8–S31.
Pearce, M. (2014). At least four Marlboro men have died of smoking related diseases. Retrieved 12/07, 2014 from http://www.latimes.com/nation/nationnow/la-na-nn-marlboro-men-20140127-story.html
Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate healthy behavior: the role of message framing. Psychological Bulletin, 121(1), 3–19. Retrieved 10/07, 2014 from http://www.ncbi.nlm.nih.gov/pubmed/9000890
Rothman, A. J., Bartels, R., D., Wlaschin, J. & Salovey, P. (2006). The strategic use of gain-and loss-framed messages to promote heathly behaviour: How theory can inform practice. Journal of Communication, 56, 202-220.
Riet, J., Ruiter, R. A. C., Werrij, M., & de Vries, H. (2008). Self-efficacy moderates message framing effects: The case of skin-cancer detection. Psychology and Health, Vol. 25, Nr. 3, 339-349.
Ruiter, R. A. C., & Kok, G. (2005). Saying is not (always) doing: cigarette warning labels are useless. European Journal of Public Health, 15(3), 329.
Ruiter, R. A. C., & Kok, G. (2006). Response to Hammond et al . Showing leads to doing, but doing what? The need for experimental pilot-testing. European Journal of Public Health, 16(2), 225.
Sanchez, J. C. (2006). Effects of message framing to produce healthy behaviour: The role of autoefficacy and cognitive motivation. Interna- tional Journal of Clinical and Health Psychology, 6, 613–630.
Smokefree.gov. Retrieved 10/11, 2014 from http://smokefree.gov
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453-458.
Wakefield, M., Germain, D. & Durkin, S.J. (2008). How does increasingly plainer cigarette packaging influence adult smokers’ perceptions about brand image? An experimental study, TobControl, 17, 416-421.
Webb T. L. & Sheeran P., (2006). Does change behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132, 249–268.
Williams, T., Clarke, V., & Borland, R. (2001). Effects of message framing on breast-cancer-related beliefs and behaviors: The role of mediating factors. Journal of Applied Social Psychology, 31, 925–950.
Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59(4), 329-349.
Witte, K., & Allen, M. (2000). A Meta-Analysis of Fear Appeals: Implications for
Effective Public Health Campaigns. Health Education & Behavior, 27(5), 591–615.
World Health Organization, Tobacco Fact sheet N339 (2014a). Retrieved 10/07, 2014 from http://www.who.int/mediacentre/factsheets/fs339/en/
World Health Organization, regional Office for Europe (2014b). Data and statistics. Retrieved 10/07, 2014 from http://www.euro.who.int/en/health-topics/disease-prevention/tobacco/data-and-statistics
Wong, N. & Cappella, J.N. (2009). Antismoking threat and efficacy appeals: Effects on smoking cessation intentions for smokers with low and high readiness to quit. Journal of Applied Communication Research, Vol. 37, 1-20.
Appendix
Table 1. Pre-test. Means (SDs) of Framing and perceived threat
MFRAME SDFRAME MTHREAT SDTHREAT
If you quit smoking you will
be healthier 4.19 .654 1.97 1.378
If you smoke you will get ill
2.23 .845 3.55 1.121
If you quit smoking your
lungs will last longer 4.19* .792 1.81* 1.250
If you smoke your lungs will
get damaged 2.16* 1.369 4.19* .873
If you quit smoking this will
save you a lot of money 4.06 .929 1.87 1.384
If you smoke this will cost you
a lot of money 2.45 1.234 3.52 1.029
If you quit smoking you will
get healthy teeth 3.71 .864 1.90 1.326
If you smoke you will get
rotten teeth 1.97 .912 3.65 1.082
If you quit smoking you will
live longer 4.03 .795 1.74 1.237
If you smoke you will die
sooner 2.19 1.250 4.00 1.065
If you quit smoking you will
get a positive social image 3.26 .893 1.71 1.131
If you smoke you will get a
negative social image 2.42 .765 2.65 1.018
If you quit smoking you will advance the health of people around you
3.61 .761 1.81 1..250
If you smoke you will hurt the
health of people around you 2.32 .832 3.23 1.055
If you quit smoking you will
smell better 3.77 .717 1.87 1.204
If you smoke you will smell
Note. An asterisk indicates the health outcomes of smoking that were used in the experiment.
Table 2. Pre-test. Means (SDs) of self-efficacy messages.
MFRAME SDFRAME MTHREAT SDTHREAT
If you quit smoking your
stamina will improve 3.77 .805 2.00 1.366
If you smoke your stamina
will reduce 2.77 1.117 3.48 1.122
If you quit smoking this will make you attractive to your social environment
3.45 .888 1.87 1.118
If you smoke this will make you unattractive to your social environment
2.35 .755 2.74 1.237
If you quit smoking you will
advance your health 4.10 .651 2.00 1.342
If you smoke you will hurt
your health 2.71 1.243 3.77 .990
MSE SDSE Craving a cigarette? Take a walk instead 2.55 1.362 Craving a cigarette? Call a friend for help 2.45 1.312
Craving a cigarette? Call a friend instead 2.68 1.447
Craving a cigarette? Eat an apple instead 2.61 1.308
Craving a cigarette? Have a gum instead 3.42* 1.311
Craving a cigarette? Drink some water instead 2.65 1.253
Craving a cigarette? Visit your favorite webpage instead 2.29 1.371
Note. An asterisk indicates the health outcomes of smoking that were used in the experiment.
MSE SDSE
Craving a cigarette? Take a deep breath instead 2.26 1.210
Craving a cigarette? Text a friend instead 2.55 1.234
Craving a cigarette? Ignore it 2.06 1.263
Craving a cigarette? Count to 10 1.84* .969
Craving a cigarette? Eat a fruit instead 2.74 1.390
Craving a cigarette? Play your favorite song instead 2.58 1.148
Craving a cigarette? Watch a YouTube video instead 2.00 1.065
Craving a cigarette? Take a bike ride instead 2.61 1.256
Craving a cigarette? Call your doctor for help 2.00 1.033
Figure 1. Two cigarette packages (i.e., Gauloises Blondes Blue) featuring the self-efficacy messages used for the experiment.
Low self-efficacy message (i.e. ‘Craving for a cigarette? Count to 10’) and high self-efficacy message (i.e. ‘Craving for a cigarette? Have a gum instead’)
Figure 2. Two cigarette packages (i.e., Gauloises Blondes Blue) featuring the framing messages used for the experiment
Loss framed message (‘If you smoke your lungs will get damaged’) and gain framed message (‘If you quit smoking your lungs will last longer’)
