Health Communication Effectiveness of Cigarette Warning Labels ...

4 downloads 13873 Views 138KB Size Report
Mar 14, 2014 - impact of graphics, message framing (gain vs. loss), and temporal framing ..... digitally signing the consent form, participants were directed.
This article was downloaded by: [University Of Maryland] On: 18 July 2015, At: 07:05 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London, SW1P 1WG

Health Communication Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hhth20

Effectiveness of Cigarette Warning Labels: Examining the Impact of Graphics, Message Framing, and Temporal Framing a

b

a

Xiaoli Nan , Xiaoquan Zhao , Bo Yang & Irina Iles a

a

Department of Communication, University of Maryland

b

Department of Communication, George Mason University Published online: 14 Mar 2014.

Click for updates To cite this article: Xiaoli Nan, Xiaoquan Zhao, Bo Yang & Irina Iles (2015) Effectiveness of Cigarette Warning Labels: Examining the Impact of Graphics, Message Framing, and Temporal Framing, Health Communication, 30:1, 81-89, DOI: 10.1080/10410236.2013.841531 To link to this article: http://dx.doi.org/10.1080/10410236.2013.841531

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Health Communication, 30: 81–89, 2015 Copyright © Taylor & Francis Group, LLC ISSN: 1041-0236 print / 1532-7027 online DOI: 10.1080/10410236.2013.841531

Effectiveness of Cigarette Warning Labels: Examining the Impact of Graphics, Message Framing, and Temporal Framing Xiaoli Nan Department of Communication University of Maryland

Downloaded by [University Of Maryland] at 07:05 18 July 2015

Xiaoquan Zhao Department of Communication George Mason University

Bo Yang and Irina Iles Department of Communication University of Maryland

This study examines the effectiveness of cigarette warning labels, with a specific focus on the impact of graphics, message framing (gain vs. loss), and temporal framing (present-oriented vs. future-oriented) among nonsmokers in the United States. A controlled experiment (N = 253) revealed that graphic warning labels were perceived as more effective, stronger in argument strength, and were generally liked more compared to text-only labels. In addition, loss-framed labels, compared to their gain-framed counterparts, were rated higher in perceived effectiveness, argument strength, and liking. No significant difference was observed between the present- and future-oriented frames on any of the dependent variables. Implications of the findings for antismoking communication efforts are discussed.

Thought of as “the foremost preventable cause of premature death” (Kees, Burton, Andrews, & Kozup, 2010, p. 265), smoking is associated with many health problems including lung cancer, respiratory difficulty, impotence, eye problems, gum disease, and early aging (U.S. Department of Health and Human Services [DHHS], 2004). About 6 million people die from tobacco use each year, and half of tobacco users eventually die from chronic tobacco-related health problems (World Health Organization [WHO], 2011). In the United States, every year smoking accounts for approximately 20% of deaths and incurs an economic cost of $190 billion (Centers for Disease Control and Prevention [CDC], 2009). These daunting figures urge health communication professionals and scholars to develop more effective communication approaches that can reduce the prevalence of cigarette use. One of the communication strategies that Correspondence should be addressed to Xiaoli Nan, PhD, Associate Professor, Department of Communication, University of Maryland, 2105A Skinner Building, College Park, MD 20742-7635. E-mail: [email protected]

have been experimented with is the use of warning labels on cigarette packages. Cigarette warning labels have a long history of being used to inform the public of smoking-related risks. Compared to traditional mass media channels such as TV or radio, cigarette warning labels are at a lower cost and have an “unparalleled reach among smokers” (Hammond et al., 2007, p. 202), which makes warning labels an appealing tool for spreading anti-smoking messages. To improve the impact of cigarette warning labels, Canada launched an initiative to replace the traditional text-only warning labels with graphic warning labels in 2000. In 2003, the World Health Assembly adopted a significant health treaty, the WHO Framework Convention on Tobacco Control (FCTC), which recommended application of vivid pictorial images on cigarette packages. As of 2011, more than 169 countries in the world had signed the treaty (Hammond, 2011). In the United States, the Food and Drug Administration (FDA) announced in 2011 its new rule to require all cigarette brands in America to place graphic warning labels on the

Downloaded by [University Of Maryland] at 07:05 18 July 2015

82

NAN ET AL.

back and front of cigarette packages starting from September 2012, marking a significant step in the U.S. government’s antismoking efforts (U.S. DHHS, FDA, 2011). This policy was eventually rejected by the U.S. Court of Appeals for the District of Columbia in August 2012, citing lack of evidence showing the effectiveness of graphic labels (Reynolds et al. v. FDA, 2012). Despite the court decision, previous studies have shown that graphic warning labels hold strong promise in curbing cigarette consumption (e.g., Hammond, Fong, McDonald, & Brown, 2004; Peters et al., 2007; Strasser, Tang, Romer, Jepson, & Cappella, 2012; White, Webster, & Wakefield, 2008). Given the relatively novel nature of graphic warning labels and the small number of studies conducted in the United States, however, the current research seeks to further assess the effectiveness of graphic versus text-only warning labels among a group of American young adults. Adding graphics to text-only labels is not the only way to potentially improve label effectiveness. Other label features warrant systematic investigation as well. So far, limited research attention has been paid to message framing— whether the label focuses on the positive outcomes of not smoking (gain frame) or the negative consequences of smoking (loss frame)—and temporal framing—whether the label highlights the short-term consequences of the target behavior (present-oriented) or the long-term consequences of the target behavior (future-oriented). The current study, then, also aims to examine the role of message and temporal framing in the effectiveness of both text-only and graphic warning labels. Both smoking cessation and prevention are important public health goals. In the present study, we focus on smoking prevention and examine how nonsmokers respond to cigarette warning labels—in terms of perceived effectiveness, argument strength, and general liking, as well as subsequent attitudes and intentions about smoking—as a function of the use of graphics, message framing, and temporal framing.

GRAPHIC VERSUS TEXT-ONLY WARNING LABELS Since their implementation, graphic warning labels have received plenty of research attention. Longitudinal studies (e.g., Borland et al., 2009; Hammond et al., 2004; White et al., 2008) found graphic warning labels to be an effective means of antismoking communication. Experimental investigations point to graphic warning labels as more effective than text-only ones, with their advantages ranging from greater emotional impact and memorability to higher perceived effectiveness in deterring smoking (e.g., O’Hegarty et al., 2006; Peters et al., 2007; Strasser et al., 2012). Because a great majority of graphic warning labels feature terrifying images of smoking-related health consequences, such as blackened teeth or a ravaged lung, studies have shown that graphic warning labels elicit greater

negative affect toward the labels or smoking (e.g., Kees, Burton, Andrew, & Kozup, 2006; Peters et al., 2007). Kees et al. (2006), for instance, conducted three studies among both adult and student smokers and consistently found that smokers reported more negative affect such as fear and anxiety toward graphic warning labels than text-only ones. Similar findings were reported in Kees et al. (2010). In another study, Peters et al. (2007) demonstrated that graphic warning labels produced more negative affective reactions to smoking cues and to the smoker image among both smokers and nonsmokers. Perhaps due to their vividness, graphic warning labels tend to generate greater attention to and recall of the warning statements (e.g., Peterson, Thomsen, Lindsay, & John, 2010; Strasser et al., 2012). Strasser et al. (2012) reported that viewers of graphic warning labels accurately recalled 83% of the warning information, compared to 50% among viewers of text-only warning labels. Peterson et al. (2010) showed that graphic warning labels generated higher levels of visual attention and were more likely to be accurately recalled than the traditional warning messages. Graphic warning labels are perceived as more effective than text-only ones by smokers and nonsmokers alike. Vardavas, Connolly, Karamanolis, and Kafatos (2009) showed that nonsmoking adolescents rated the graphic labels as more effective in preventing them from smoking in comparison to the text-only warnings. Another study found that young current and former smokers thought cigarette warning labels with text plus graphics were substantially more deterrent than text-only labels (O’Hegarty et al., 2006). Not only are graphic warning labels perceived as more effective, they also tend to reduce smoking intentions to a greater extent than text-only ones (Kees et al., 2006; Kees et al., 2010; Sabbane, Lowrey, & Chebat, 2009). Kees et al. (2006), for example, showed that the addition of visual warnings to the text-only statements increased smokers’ intentions of quitting smoking. In another study, Sabbane et al. (2009) found graphic warning labels to be more effective in curbing smoking intentions among a group of Canadian adolescent nonsmokers. Surprisingly, however, they also found that graphic warning labels had a slight boomerang effect on smoking intentions among adolescent nonsmokers in the United States. The authors attributed the unexpected finding to the novelty of graphic warning labels in the United States. The majority of studies on graphic warning labels were conducted outside of the United States. As Sabbane et al. (2009) suggest, it is not clear whether Americans might respond to graphic warning labels similarly to responses of people from other countries where familiarity with this type of messages is higher. To further evaluate the effectiveness of graphic warning labels versus text-only in terms of both message evaluation (perceived effectiveness, perceived argument strength, and liking) and behavioral antecedents (attitudes and intentions) among a group of American young nonsmokers, we propose the following hypotheses:

EFFECTIVENESS OF CIGARETTE WARNING LABELS

H1: Graphic warning labels, compared to text-only ones, will lead to (a) greater perceived effectiveness; (b) greater perceived argument strength; (c) greater general liking of the labels; (d) less favorable attitudes toward smoking; and (e) less smoking intentions.

Downloaded by [University Of Maryland] at 07:05 18 July 2015

MESSAGE FRAMING A large body of literature has been dedicated to evaluating the impact of message framing on persuasion. In the context of health communication, message framing refers to the type of outcome associated with adopting or failing to adopt a particular health behavior (Rothman & Salovey, 1997). Loss-framed messages attempt to persuade by pointing at the negative consequences or costs incurred by not adopting the recommended behavior. On the other hand, gain-framed messages emphasize the benefits or the positive outcomes that are accrued through adopting the behavior (Rothman & Salovey, 1997). Prospect theory constitutes one of the main theoretical bases for research on message framing (Kahneman & Tversky, 1979). According to the theory, individuals tend to be risk averse when facing potential gains, and relatively risk seeking when confronted with potential losses. In addition, individuals assess gain or loss information in relation to a reference point (i.e., their current state). In the case of health behaviors, if the perceived risk of performing a suggested behavior when compared to one’s present level of health is low, individuals are more likely to prefer messages framed in terms of relative benefits. Conversely, when the risk associated with that behavior is high, emphasizing the relative disadvantages of performing the behavior will elicit more responsiveness (Salovey & Williams-Piehota, 2004). Because not smoking is a low-risk behavior that prevents the occurrence of future illnesses, it is expected that gain-framed antismoking messages will be more effective than loss-framed ones. In support of this hypothesis, Schneider et al. (2001) demonstrated that gain-framed messages were more effective in reducing smoking behaviors among young smokers. Toll et al. (2008) found an advantage for gain-framed messages in encouraging continuous abstinence among a group of adult smokers who had completed a smoking cessation treatment. Both investigations used relatively complex antismoking messages involving audio and visual materials. Studies on cigarette warning labels (Bansal-Travers, Hammond, Smith, & Cummings, 2011; Goodall & Appiah, 2008), however, appear to support loss framing as the more effective message type. Goodall and Appiah (2008), for example, showed that adolescent smokers and nonsmokers held significantly more favorable attitudes toward loss-framed graphic labels. They also found that adolescent smokers were less likely to intend to smoke after being exposed to loss-framed labels. Bansal-Travers et al. (2011) reported that loss-framed warning labels were rated

83

by adult smokers and nonsmokers as more likely to attract attention, encourage thoughts about health risks, motivate quitting, and be most effective. The reason for the discrepancy in findings concerning message-framing effects in the two message contexts is unclear. Perhaps in the limited space provided by a cigarette package, it is more difficult to execute a gain frame than a loss frame since smoking is predominantly associated with negative health consequences in public awareness. Through more complex messages, however, a gain frame may be delivered with sufficient clarity and effectiveness. Because our study focuses on cigarette warning labels, we propose the following hypotheses in line with the current literature concerning this message format: H2: Loss-framed labels, compared to gain-framed labels, will lead to (a) greater perceived effectiveness; (b) greater perceived argument strength; (c) greater general liking of the labels; (d) less favorable attitudes toward smoking; and (e) less smoking intentions.

TEMPORAL FRAMING The time reference employed in a message could impact message perceptions and persuasiveness. Referring to health costs or benefits as occurring now or later in the future might affect the way risk is perceived and, consequently, individuals’ compliance with a health message. Early literature on temporal distance maintains that the projection of negative events in the future lowers risk perceptions in the present (Lewin, 1951). Construal Level Theory (CLT; Trope & Liberman, 2000) suggests that the more distant an object or event is from the individual, the more abstract it will be perceived as being. As such, distant risks may be perceived as less threatening than proximal risks. Chandran and Menon (2004) analyzed how a specific type of temporal framing (a day or a year) might affect risk perceptions and intentions to perform a recommended behavior. According to their findings, the day frame (e.g., “every day, a great many people die of cancer”) was perceived to be more proximal and concrete and, in turn, was associated with higher self-risk estimates and concern about health hazards compared to the year frame (e.g., “every year, a great many people die of cancer”). Regardless of the valence of the outcome (i.e., loss-framed vs. gain-framed outcomes), present-oriented events elicited “a greater sense of immediacy in risk judgments and behavioral intention” (Chandran & Menon, 2004, p. 384). Their conclusions, along with the research already discussed here, support the idea that present-oriented framing (as opposed to future-oriented framing) intensifies the proximity of health risks and, therefore, may elicit more desirable responses to a health message. The role of temporal framing in the communication of smoking cessation or prevention, however, has not received much scrutiny. Crawford, Balch, and Mermeistein (2002) reported that members of a focus group believed future

84

NAN ET AL.

cigarette warning labels should emphasize the immediate effects of smoking. Yet there is scant empirical evidence that either supports or refutes this claim. Given that the role of temporal framing in the effectiveness of cigarette warning labels has not been adequately studied, we propose the following research question: RQ1: Are present-oriented labels more effective than futureoriented labels in terms of (a) perceived effectiveness; (b) perceived argument strength; (c) general liking of the labels; (d) attitudes toward smoking; and (e) smoking intentions?

METHOD

Downloaded by [University Of Maryland] at 07:05 18 July 2015

Participants In total, 290 undergraduate students (mean age = 19.62, SD = 1.77) recruited from communication classes in a large northeastern university in the United States participated in the study for exchange of course credit. We asked participants to indicate whether they currently smoke cigarettes. Those who reported “not at all” were classified as nonsmokers and constituted the working sample for this study (n = 253, mean age = 19.47, SD = 1.77). Of the 253 nonsmokers, 68.4% were females, 68.8% Whites, 12.6% Blacks, 11.1% Asians, and 4.3% Hispanics.1 Design and Procedure This study, which was conducted as a Web-based experiment, followed a 2 (graphics: present vs. absent) × 2 (message framing: gain vs. loss) × 2 (temporal framing: present-oriented vs. future-oriented) factorial design. After digitally signing the consent form, participants were directed toward the study website and were randomly assigned to one of the eight experimental conditions. Participants first responded to a preliminary survey that included a number of questions about demographic information and smoking status. They were then presented with three digitally modified cigarette warning labels placed on mock cigarette packages, displayed one at a time and in a rotated order (i.e., the three labels were displayed in six different orders with each order seen by one-sixth of the participants). To identify appropriate health problems for use in the three labels, we sampled common themes covered in existing graphic warning labels used around the world. We consulted a useful website (http://www.tobaccolabels.ca/healthwarningimages), which archived a large collection of graphic cigarette warning 1 A power analysis showed that, given the design of our study, a sample of 272 would be needed to detect a small effect size, Cohen’s f 2 = .02, given α = .05 and power = .80. The actual sample size of this study (N = 253) was slightly lower than this desired sample size and had a power of .77 for the same f 2 and α.

labels, organized by countries as well as by common themes. We then selected three health problems (i.e., eye problem, skin aging, and impotence) that appeared most relevant for college students. After seeing all three warning labels, participants responded to a battery of questions assessing several dependent variables, including perceived effectiveness, perceived argument strength, general liking of the labels, attitudes toward smoking, and smoking intentions.

Manipulations We created the experimental warning labels using Adobe Photoshop, a digital graphic editing software package. While working on designing the labels, we tried to make the labels look similar to the FDA’s proposed cigarette warning labels in terms of size and position of the graphics as well as the color, size, and font of the texts. We adopted the same mock cigarette packaging that the FDA used for presenting the sample graphic warning labels.2 Graphics. The graphic warning labels were identical to the text-only labels except that the former had graphics on them while the latter had only texts. All other aspects of the labels were the same. Message framing. On the gain-framed warning labels, the texts emphasized the reduced health risks as a result of less smoking (i.e., reducing the damage to the back of the eye; reducing the damage to the skin; increasing blood flow to prevent impotence). The graphic images on the gainframed labels portrayed a healthy eye, a woman with good skin, and a happy young couple, respectively. On the lossframed warning labels, the texts stated the negative health consequences associated with smoking (i.e., causing damage to the back of the eye; causing damage to the skin; reducing blood flow to cause impotence). The graphic images appearing on the loss-framed labels were an eye under surgery, a woman with wrinkled skin, and an unhappy young couple, respectively. Temporal framing. The present-oriented warning labels focused on the short-term effects of smoking (e.g., “Just one cigarette may cause immediate damage to the back of the eye”) or not smoking (e.g., “One less cigarette may immediately reduce damage to the back of the eye”). The future-oriented labels highlighted the long-term consequences of smoking (e.g., “Smoking causes long-term damage to the back of the eye”) or not smoking (e.g., “Quitting smoking reduces long-term damage to the back of the eye”).

2 The graphic warning labels used in the experiment can be obtained from the first author.

EFFECTIVENESS OF CIGARETTE WARNING LABELS

Key Measures

Downloaded by [University Of Maryland] at 07:05 18 July 2015

Perceived effectiveness. Participants were asked to what extent they thought that the three warning labels they just saw made them feel like staying away from smoking, how effective they thought the three warning labels were in stopping nonsmokers from smoking, and how effective they thought the three warning labels were in motivating smokers to quit smoking. The questions were answered on a 7-point scale (1 not at all, 4 moderately, and 7 very much) and averaged into an overall perceived effectiveness measure (M = 3.56, SD = 1.66, α = .81). Perceived argument strength. Perceived argument strength of the warning labels was measured using nine items adapted from a previously published scale (Zhao, Strasser, Cappella, Lerman, & Fishbein, 2011). Examples of items used included “The warning labels grasped my attention immediately,” “The warning labels put thoughts in my mind about not wanting to smoke,” and “The warning labels made me feel like the whole society is against smoking now.” Participants rated these statements on a 7-point scale (1 strongly disagree to 7 strongly agree). Ratings were averaged across the 9 items to generate an index for perceived argument strength (M = 3.95, SD = 1.40, α = .89). General liking. General liking was measured by asking participants how much they liked or disliked the three warning labels they just saw, a question that was adapted from Legacy Media Tracking Survey (LMTS), a periodical national tobacco survey sponsored by the American Legacy Foundation (see Nan & Zhao, 2010). Answers were indicated on a 7-point scale (1 strongly dislike to 7 strongly like) (M = 3.68, SD = 1.84). Attitudes toward smoking. Attitudes toward smoking were assessed by a question asking “What is your attitude now towards smoking?” Answers were indicated on three 7-point Likert-type scales ranging from 1 (negative/bad/unfavorable) to 7 (positive/good/favorable). The items were averaged to produce the summary measure (M = 1.55, SD = 1.04, α = .96). Similar attitudes measures have been used in previous studies on substance use (e.g., Cho & Boster, 2008; Goodall & Appiah, 2008). Smoking intentions. Smoking intentions were assessed by asking participants to indicate their agreement or disagreement with the statement “I will not start smoking in the future” on a 7-point scale (1 strongly disagree to 7 strongly agree) (M = 6.66, SD = 1.10).

RESULTS Manipulation Checks Manipulation checks were performed on message framing and temporal framing using a series of multivariate analyses

85

of variance (MANOVAs). A multivariate main effect of message framing was found on the two items assessing perceived message focus on costs vs. benefits (F[2, 248] = 22.82, p < .001, Wilks’ λ = .85). Participants in the gain frame condition perceived greater message focus on the benefits of quitting (M = 4.86) than those in the loss frame condition (M = 3.48, F[1, 249] = 30.08, p < .001, d = .70). On the other hand, participants in the loss frame condition reported greater message focus on the costs of smoking (M = 4. 73) than those in the gain frame condition (Mgain = 4.19, F[1, 249] = 3.48, p = .06, d = .20). A multivariate main effect of temporal framing was also found on the two items tapping perceived message focus on long- versus short-term effects of smoking or quitting (F [2, 248] = 63.08, p < .001, Wilks’ λ = .66). Participants in the present-oriented condition perceived greater message focus on short-term effects (M = 4.68) than those in the future-oriented condition (M = 3.01, F[1, 249] = 55.24, p < .001, d = .94). At the same time, participants in the future-oriented condition perceived greater message focus on long-term effects (M = 5.91) than those in the present-oriented condition (M = 4.19, F[1, 249] = 64.96, p < .001, d = 1.03). These results suggest that both message framing and temporal framing manipulations were successful. We did not perform a check on the graphic manipulation because it was straightforward, marked by the presence or absence of graphics in the warning labels.

Hypotheses and Research Question To test H1 and H2, and to answer RQ1, a MANOVA was conducted on perceived effectiveness, perceived argument strength, general liking of the labels, attitudes toward smoking, and smoking intentions. Correlations of these dependent variables are presented in Table 1. Graphics, message framing, and temporal framing were entered as fixed factors. The multivariate main effect of graphics was significant (F[5, 237] = 5.1, p < .001, Wilks’ λ = .90), as was that of message framing (F[5, 237] = 4.16, p = .001, Wilks’ λ = .92). The multivariate main effect of temporal framing was not significant (F[5, 237] = .72, p = .61, Wilks’ λ = .98). No interaction in the model was significant on the multivariate level (all p > .25).

TABLE 1 Correlations of Dependent Variables

Effectiveness Argument strength Liking Attitude Intention ∗p

Effectiveness

Argument Strength

Liking

Attitude

.73∗∗∗ .64∗∗∗ −.11 .01

0.76∗∗∗ −.09 .02

−.06 .08

−.14∗

< .05; ∗∗∗ p < .001.

86

NAN ET AL. TABLE 2 Means (SDs) Related to the Hypotheses and Research Question a) Effectiveness H1 H2 RQ1

Graphics Text Gain Loss Present Future

(1.68)a

3.84 3.21 (1.69)b 3.95 (1.62)a 3.23 (1.72)b 3.46 (1.69) 3.67 (1.73)

b) Argument Strength (1.25)a

4.27 3.64 (1.54)b 4.35 (1.30)a 3.68 (1.45)b 3.80 (1.40) 4.16 (1.40)

c) Liking (1.78)a

4.04 3.22 (1.83)b 4.04 (1.77)a 3.37 (1.87)b 3.50 (1.82) 3.85 (1.86)

d) Attitude

e) Intention

(.85)a

6.62 (1.10) 6.71 (1.11) 6.75 (.91) 6.57 (1.25) 6.69 (1.02) 6.63 (1.17)

1.44 .85 (.55)b 1.38 (.85) 1.20 (.64) 1.31 (.79) 1.34 (.72)

Downloaded by [University Of Maryland] at 07:05 18 July 2015

Note. For each dependent variable, means with different superscripts in each panel were significantly different at p < .05. Means without superscripts were not significantly different.

On the univariate level, the main effect of graphics was significant on perceived effectiveness (F[1, 241] = 8.46, p = .004, d = .35), perceived argument strength (F[1, 241] = 13.73, p < .001, d = .46), general liking (F[1, 241] = 12.50, p < .001, d = .46), and smoking attitudes (F[1, 241] = 6.59, p = .01, d = .35), but not on intention (F[1, 241] = .61, p = .48, d = .09). Means and standard deviations associated with these tests are presented in Table 2. Consistent with H1a through H1c, nonsmokers in the graphics condition reported higher perceived effectiveness, higher perceived argument strength, and greater liking than those in the text-only condition. H1a through H1c were supported. Unexpectedly, nonsmokers in the graphics condition reported more positive smoking attitudes than those in the text-only condition. H1d was not supported. Graphic versus text-only conditions produced similar smoking intentions. H1e was not supported. For message framing, its main effect was significant on perceived effectiveness (F[1, 241] = 9.96, p = .002, d = .41), perceived argument strength (F[1, 241] = 13.26, p < .001, d = .46), and general liking (F[1, 241] = 6.86, p = .009, d = .35), but not significant on smoking attitudes (F[1, 241] = 1.87, p = .17, d = .20) or smoking intentions (F[1, 241] = 1.83, p = .18, d = .20). As shown in Table 2, participants in the loss frame condition reported greater perceived effectiveness, greater perceived argument strength, and greater liking than those in the gain frame condition. H2a through H2c were supported. Smoking attitudes were similar between the two conditions, as were smoking intentions. H2d and H2e were not supported. RQ1 is concerned with the effect of temporal framing. There was no significant difference between the present- and future-oriented frames on any of the dependent variables (see Table 2). Although no significant effect of the independent variables on attitudes and intentions was observed in the preceding analysis, we still explored the possibility of indirect effects of the independent variables on attitudes and intentions through the mediation of message evaluation variables. This analysis is allowable because recent research on mediation analysis suggests that a significant total effect is not

always a necessary condition for mediation to occur (Shrout & Bolger, 2002). Using the SPSS macro Process (Hayes, 2013), we performed a series of multiple mediation tests (1000 bootstrap resamples) with graphics, gain versus loss framing, and temporal framing as the independent variables, smoking attitudes and intentions as the dependent variables, and perceived effectiveness, perceived argument strength, and liking as mediators. None of the indirect effects emerged as significant, however (all 90% bootstrap confidence intervals included 0). There was no evidence of message evaluation mediating the effect of the experimental manipulations on smoking attitudes and intentions.

DISCUSSION Although the attempt to mandate the placement of graphics on cigarette warning labels in the United States has met with much resistance and controversy, evidence accumulated to date generally points to graphic warning labels as a more effective means of communicating smoking risks compared to text-only ones. The purpose of the current research was to further evaluate the effectiveness of graphic versus textonly labels among a group of young American nonsmokers. We also sought to expand current literature by assessing the impact of message framing (gain vs. loss) and temporal framing (present-oriented vs. future-oriented) on the effectiveness of the labels. Consistent with our expectation, nonsmokers in our study who were presented with the graphic warning labels reported greater perceived effectiveness than those exposed to the text-only ones, a finding that is aligned with the results of previous studies (e.g., O’Hegarty et al., 2006; Vardavas et al., 2009). We also expected that the graphic warning labels would be perceived as greater in argument strength and that nonsmokers would like the graphic warning labels more. Our findings supported these hypotheses. Overall, it appears that nonsmokers’ evaluations of graphic (vs. text-only) warning labels, whether they are more cognitively derived (e.g., perceived effectiveness, perceived argument strength) or more affectively based (e.g., liking), are generally favorable.

Downloaded by [University Of Maryland] at 07:05 18 July 2015

EFFECTIVENESS OF CIGARETTE WARNING LABELS

Our findings concerning the relative effects of graphic versus text-only labels on smoking attitudes and intentions are less conclusive and somewhat surprising. First, label format (graphic vs. text-only) had no impact on smoking intentions. Second, contrary to our expectation, nonsmokers exposed to the graphic warning labels reported more favorable attitudes toward smoking than those presented with the text-only ones. How to account for this surprising finding is not clear. Normally, one would expect message responses to be consistent with the targeted attitude and behavior—this is indeed a key assumption in the cognitive response tradition in persuasion research (Greenwald, 1968; Petty & Cacioppo, 1986). But in this study, we observed opposite effects between evaluative responses to the message and smoking attitudes. If these findings reflect true effects, one might suspect that different mechanisms were at work in nonsmokers’ message evaluations and attitudinal considerations. Perhaps in message appraisal, nonsmokers approached the warning labels from a more disinterested perspective—assessing their merits as information applicable to others, particularly smokers, rather than themselves, even though the questions were supposed to elicit their personal reactions. When they were asked to express their attitudes directly, however, the disinterested perspective was replaced with a more personally involved one. It was here that some form of resistance started to kick in. For example, the blunt graphic warning labels might be perceived by some nonsmokers as an unjustified threat to their autonomy. As a consequence, psychological reactance might arise, which could in turn lead to attitude shifts incongruent with the message-advocated position. While our finding concerning the attitudinal influence of graphic versus text-only labels is somewhat surprising, Sabbane et al. (2009), as mentioned previously, appear to have obtained a similar pattern of results with respect to smoking intentions. In their study, a group of American adolescent nonsmokers reported higher intentions to smoke after seeing graphic (vs. text-only) warning labels. Sabbane et al. mentioned the novelty of graphic warning labels among the American participants as a possible contributing factor to the boomerang effect as the same graphic labels reduced smoking intentions among a group of Canadian adolescent nonsmokers, who presumably were much more familiar with these labels. It seems plausible that familiarity with graphic labels might have reduced psychological reactance to the abrasive warning labels among the Canadian participants. This possibility is consistent with the speculative explanation we ventured earlier. But again, we should be cautious about the level of consistency between the two studies, because no effect of graphics on intention was observed in this study. More research is clearly needed to assess both the robustness of the current findings and the verity of our proposed account. An important component of our study was to examine the influence of message framing and temporal framing on the effectiveness of cigarette warning labels. With respect to

87

message framing, the results showed an advantage for the loss-framed labels in terms of message evaluation measures (i.e., perceived effectiveness, perceived argument strength, general liking). However, we found the loss-framed labels to be no more effective than the gain-framed ones in eliciting desirable attitudinal or behavioral responses toward smoking. Again, these findings may appear to suggest different mechanisms at work in nonsmokers’ message evaluations and target behavior evaluations or intentions. Nonsmokers tend to evaluate loss-framed labels more favorably than gain-framed labels, perhaps basing the assessment on their naive theories about what might impact smokers the most or how persuasion works, but their personal attitudinal or behavioral responses toward smoking are not substantially influenced by message framing. The latter finding is consistent with O’Keefe and Jensen’s (2006) meta-analysis, which found limited influence of message framing on persuasive outcomes such as attitudes and intentions. Overall, graphics and message framing influenced message evaluations as predicted but exerted limited (and to some extent unexpected) influence on attitudes and intentions toward smoking. Furthermore, as shown in the mediation analyses, graphics and message framing both had no indirect effect on attitudes or behavioral intentions through any of the message evaluation measures. These findings together raise interesting questions about assessing message effectiveness in the context of cigarette warning labels. To the extent that attitudes and intentions are strong predictors of behaviors, our study suggests that adding graphics or using loss frames may not make much difference in enhancing the effectiveness of the labels. However, stated attitudes and behavioral intentions toward smoking tend to be highly skewed toward the antismoking extremes (perhaps due to a social desirability bias), as is the case in our study; it may be premature to write off the utility of using graphics or loss-frames in cigarette warning labels, especially considering that the message evaluation measurements tend to support the use of graphics and loss-frames. In any event, more research is needed to evaluate the effectiveness of different types of cigarette warning labels on actual smoking behaviors (initiation or cessation). Our data show that temporal framing had little influence on message evaluations, attitudes, or intentions toward smoking. This set of findings appears to be inconsistent with relevant theory (Lewin, 1951; Trope & Liberman, 2000) and previous formative research (Crawford et al., 2002). One possibility is that for nonsmokers the difference between present and future consequences is simply not that striking since neither is really personally relevant. The difference may matter more to smokers, who should be more concerned about the consequences of smoking and hence more sensitive to how the consequences are projected. We also did not find any interactions among the three manipulated variables (graphics, message framing, temporal framing), indicating that the effects we observed operated independently.

Downloaded by [University Of Maryland] at 07:05 18 July 2015

88

NAN ET AL.

Our findings provide some useful insights into how nonsmokers respond to cigarette warning labels, but need to be interpreted with several constraints in mind. First, due to the experimental setting of the study, we purposefully made participants focus on and think about the warning labels, which might have inflated the effects of the warning labels. In the natural environment, people, especially nonsmokers, might simply ignore the warning information, or react to it less strongly due to low involvement. Second, in the current study, we only assessed three types of warning labels that dealt with eye, skin, and impotence problems. As our sample was composed of college students, who might have a different level of concerns about these problems than the general population, we need to be cautious when generalizing the findings to other individuals. There might also be gender differences in responses to these health issues that we did not account for. For instance, female participants might find aging skin to be more personally relevant, whereas male participants might be more concerned about impotence. To the extent that these differences could potentially affect the effects of graphics and framing, they need to be incorporated in future theoretical frameworks and empirical testing. Third, because we examined only nonsmokers, some key outcome measures such as smoking intentions and attitudes had only limited variance in this study. They were also highly skewed because a large majority of the participants were clustered toward the antismoking extremes of the measurement scales. These distributional problems could have limited our ability to detect meaningful influence of the warning labels on behavioral antecedents. Finally, we did not examine smokers in this study. We intend to conduct another study to assess smokers’ reactions toward cigarette warning labels using a similar design. Findings from this other study should add to the current evidence to provide a more complete understanding of the utility of warning labels as a broad scale intervention method. Despite the limitations, our study offers some implications for antismoking communication efforts using cigarette warning labels. First, although nonsmokers rated graphic warning labels more favorably than text-only ones, there was no evidence that adding graphics would reduce positive attitudes toward smoking or intentions to smoke. Because we observed that adding graphics actually slightly enhanced attitudes toward smoking, we are concerned about the use of graphic warning labels in attempts to influence nonsmokers in the United States. We urge that more empirical work be conducted to assess the intended and unintended effects of graphic warning labels among nonsmokers and smokers alike. Second, our study suggests that loss-framed labels are rated more favorably by nonsmokers than gain-framed labels, but gain versus loss framing does not make a difference in smoking attitudes and intentions. As such, we recommend the use of loss frames for cigarette warning labels as far as nonsmokers are concerned, but caution

against unrealistic expectations of enhanced impact on attitudes or intentions toward smoking. Finally, whether the labels focus on short-term or long-term consequences of smoking (or not smoking) does not seem to make a difference for nonsmokers. Given the lack of previous research in this area, however, it is important to gather additional evidence before dismissing the relevance of temporal framing in warning labels for nonsmokers. REFERENCES Bansal-Travers, M., Hammond, D., Smith, P., & Cummings, K. M. (2011). The impact of cigarette pack design, descriptors, and warning labels on risk perception in the U.S. American Journal of Preventive Medicine, 40, 674–682. doi:10.1016/j.amepre.2011.01.021. 674. Borland, R., Yong, H., Wilson, N., Fong, G. T., Hammond, D., Cummings, K. M., . . . McNeill, A. (2009). How reactions to cigarette packet health warnings influence quitting: Findings from the ITC four-country survey. Addiction, 104, 669–675. doi:10.1111/j.1360-0443.2009.02508.x. Centers for Disease Control and Prevention. (2009). Smoking and tobacco use: Fact sheet. Retrieved from http://www.cdc.gov/tobacco/ data_statistics/fact_sheets/fast_facts/index.htm Chandran, S., & Menon, G. (2004). When a day means more than a year: Effects of temporal framing on judgments of health risk. Journal of Consumer Research, 31, 375–389. doi:10.1086/422116. Cho, H., & Boster, F.J. (2008). Effects of gain vs. loss frame anti-drug ads on adolescents. Journal of Communication, 58, 428–446. Crawford, M. A., Balch, G. I., & Mermeistein, R. (2002). Responses to tobacco control policies among youth. Tobacco Control, 11, 14–19. doi:10.1136/tc.11.1.14. Goodall, C., & Appiah, O. (2008). Adolescents’ perceptions of Canadian cigarette package warning labels: Investigating the effects of message framing. Health Communication, 23, 117–127. doi:10.1080/10410230801967825. Greenwald, A. G. (1968). Cognitive learning, cognitive response to persuasion, and attitude change. In A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 147–170). San Diego, CA: Academic Press. Hammond, D. (2011). Health warning messages on tobacco products: A review. Tobacco Control, 20, 327–337. doi:10.1136/tc.2010.037630. Hammond, D., Fong, G. T., Borland, R., Cummings, K. M., McNeill, A., & Driezen, P. (2007). Text and graphic warnings on cigarette packages: Findings from the international tobacco control four country study. American Journal of Preventive Medicine, 32, 202–209. doi:10.1016/j.amepre.2006.11.01. Hammond, D., Fong G. T., McDonald, P. W., & Brown, K. S. (2004). Graphic Canadian cigarette warnings labels and adverse outcomes: Evidence from Canadian smokers. American Journal of Public Health, 94, 1442–1445. doi:10.2105/AJPH.94.8.1442. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–291. doi:10.2307/1914185. Kees, J., Burton, S., Andrews, J. C., & Kozup, J. (2006). Tests of graphic visuals and cigarette package warning combinations: Implications for the framework convention on tobacco control. Journal of Public Policy & Marketing, 25, 212–223. doi:10.1509/jppm.25.2.212. Kees, J., Burton, S., Andrews, J. C., & Kozup, J. (2010). Understanding how graphic pictorial warnings work on cigarette packaging. Journal of Public Policy & Marketing, 29, 265–276. doi:10.1509/jppm.29.2.265. Lewin, K. (1951). Field theory in social science. New York, NY: Harper.

Downloaded by [University Of Maryland] at 07:05 18 July 2015

EFFECTIVENESS OF CIGARETTE WARNING LABELS Nan, X., & Zhao, X. (2010). The influence of liking for anti-smoking PSAs on adolescents’ smoking-related behavioral intentions. Health Communication, 25, 459–469. O’Hegarty, M., Pederson, L. L., Nelson, D. E., Mowery, P., Gable, J. M., & Wortley, P. (2006). Reactions of young adult smokers to warning labels on cigarette packages. American Journal of Preventive Medicine, 30, 467– 473. doi:10.1016/j.amepre.2006.01.018. O’Keefe, D. J., & Jensen, J. D. (2006). The advantages of compliance or the disadvantages of noncompliance? A meta-analytic review of the relative persuasive effectiveness of gain-framed and loss-framed messages. In C. S. Beck (Ed.), Communication yearbook 30 (pp. 1–43). Mahwah, NJ: Lawrence Erlbaum Associates. O’Keefe, D. J., & Jensen, J. D. (2007). The relative persuasiveness of gainframed and loss-framed messages for encouraging disease prevention behaviors: A meta-analytic review. Journal of Health Communication, 12, 623–644. Peters, E., Romer, D., Slovic, P., Jamieson, K. H., Wharfield, L., Mertz, C. K., & Carpenter, S. M. (2007). The impact and acceptability of Canadian-style cigarette warning labels among U.S. smokers and nonsmokers. Nicotine & Tobacco Research, 9, 473–481. doi:10.1080/1462220070123963. Peterson, E. B., Thomsen, S., Lindsay, G., & John, K. (2010). Adolescents’ attention to traditional and graphic tobacco warning labels: An eye-tracking approach. Journal of Drug Education, 40, 227–244. doi:10.2190/DE.40.3.b Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York, NY: Springer-Verlag. Reynolds et al. v. FDA. (2012). No. 11-5332, U.S. Court of Appeals for the District of Columbia Circuit, August 24. Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate healthy behavior: The role of message framing. Psychological Bulletin, 121, 3–19. doi:10.1037/0033-2909.121.1.3. Sabbane, L.I., Lowrey, T. M., & Chebat, J.C. (2009). The effectiveness of cigarette warning label threats on nonsmoking adolescents. Journal of Consumer Affairs, 43, 332–345. Salovey, P., & Williams-Piehota, P. (2004). Field experiments in social psychology: Message framing and the promotion of health protective behaviors. American Behavioral Scientist, 47, 488–505. Schneider, T. R., Salovey, P., Pallonen, U., Mundorf, N., Smith, N. F., & Steward, W. T. (2001). Visual and auditory message framing

89

effects on tobacco smoking. Journal of Applied Social Psychology, 31, 667–682. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422–445. doi:10.1037//1082-989X.7.4.422 Strasser, A. A., Tang, K. Z., Romer, D., Jepson, C., & Cappella, J. N. (2012). Graphic warning labels in cigarette advertisements: Recall and viewing patterns. American Journal of Preventive Medicine, 43, 41–47. doi:10.1016/j.amepre.2012.02.026. Toll, B. A., O’Malley, S. S., Katulak, N. A., Wu, R., Dubin, J. A., Latimer, A., . . . Salovey, P. (2007). Comparing gain- and loss-framed messages for smoking cessation with sustained-release bupropion: A randomized controlled trial. Psychology of Addictive Behaviors, 21, 534–544. doi:10.1037/0893-164X.21.4.534. Trope, Y., & Liberman, N. (2000). Temporal construal and time-dependent changes in preference. Journal of Personality and Social Psychology, 79, 876–889. U.S. Department of Health and Human Services, Food and Drug Administration. (2011). Required warnings for cigarette packages and advertisements; final rule (Docket no. FDA-2010-N-0568). Retrieved from http://www.regulations.gov/#!documentDetail;D=FDA2010-N-0568-0251 U.S. Department of Health and Human Services, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. (2004). The health consequences of smoking: A report of the surgeon general. Washington, DC: U.S. Government Printing Office. Vardavas, C.I., Connolly, G., Karamanolis, K., & Kafatos, A. (2009). Adolescents’ perceived effectiveness of the proposed European graphic tobacco warning labels. European Journal of Public Health, 19, 212–217. White, V., Webster, B., Wakefield, M. (2008). Do graphic health warning labels have an impact on adolescents’ smoking related beliefs and behaviours? Addiction, 103, 1562–1571. doi:10.1111/j.1360-0443.2008.02294.x World Health Organization. (2011). WHO report on the global tobacco epidemic, 2011: Warning about the dangers of tobacco. Retrieved from http:/ /whqlibdoc.who.int/publications/2011/9789240687813_eng.pdf Zhao, X., Strasser, A., Cappella, J. N., Lerman, C., & Fishbein, M. (2011). A measure of perceived argument strength: Reliability and validity. Communication Methods and Measures, 5, 48–75. doi:10.1080/19312458.2010.547822