Effects of Current and Enhanced Tobacco Corrective Messages on Smokers’ Intention to Quit Smoking and Intention to Purchase Cigarettes

Abstract

Introduction

A federal court has ordered tobacco companies to issue corrective messages to address tobacco-related misperceptions. This study examined the effects of viewing current versus two enhanced versions of tobacco corrective messages on smokers’ intention to quit smoking and intention to purchase cigarettes.

Methods

US adult smokers (N = 803) were randomly assigned to view (1) two current tobacco corrective messages (Current), (2) two corrective messages that include an industry deception statement (Industry Deception), or (3) two corrective messages with an industry deception statement and testimonials of people harmed by smoking (Industry Deception + Testimonial). Outcomes were pretest–posttest change in intentions to quit smoking and posttest intention to purchase cigarette measures.

Results

Intention to quit smoking increased significantly after viewing the Current corrective messages versus baseline. In addition, viewing the Industry Deception + Testimonial messages increased intention to quit smoking compared with the Current corrective condition and the Industry Deception condition. Hispanic smokers had increased intention to quit smoking and decreased intention to purchase cigarettes to a greater degree than non-Hispanic smokers in response to Industry Deception + Testimonial messages. There was no significant difference in intention to purchase cigarettes across conditions.

Conclusions

Enhancing the current corrective statements by including an industry deception statement and testimonials may strengthen effects and contribute to remedying the effects of tobacco misinformation.

Implications

Previous research has found that draft or proposed versions of tobacco industry corrective messages are effective in correcting beliefs and knowledge. However, studies have not examined how the current court-ordered corrective messages could change intention to quit smoking and intention to purchase cigarettes nor whether enhanced messages could perform better. Study findings suggest that the current corrective messages can increase smokers’ intention to quit smoking beyond their baseline intention. More importantly, enhancing corrective messages by including an industry deception statement and testimonial was found to be more effective than current corrective messages. Findings can inform future iterations of tobacco correctives and strategies to reverse the effects of tobacco misinformation.

Introduction

In 2006, Federal Judge Kessler ruled that the tobacco industry had intentionally misled the public about the harms and safety of their tobacco products, and ordered the industry to issue corrective statements meant to correct any tobacco-related misperceptions consumers might have held.^1,2 The correctives were to address five tobacco-related topics: health effects, secondhand smoke effects, manipulation of cigarette design, light and low-tar cigarettes, and addiction. The correctives were slated to be disseminated in 2013¹, but tobacco industry appeals against the ruling delayed the distribution until November 2017.³ The corrective messages have since appeared in major newspapers (until March 2018) and television (until November 2018) and are now appearing on tobacco company web sites and cigarette pack onserts.⁴

In addition to the setback in implementation of the correctives, the original court-ordered language for the correctives was changed due to appeals by the tobacco industry.⁵ Specifically, the original language included a statement acknowledging the deliberate deceptive practices of tobacco industry with the statement “A federal court has ruled that R.J. Reynolds Tobacco, Philip Morris USA, Altria and Lorillard deliberately deceived the American public about the health effects of smoking, and has ordered those companies to make this statement. Here is the truth.”¹ Unfortunately, the preamble describing tobacco industry’s culpability has been removed from the current correctives which only state “A federal court has ordered R.J. Reynolds Tobacco, Philip Morris USA, Altria and Lorillard to make this statement about the health effects of smoking” and potentially weakening the impact compared with the original correctives.^3,6–8

Although previous studies have examined how exposure to draft versions of corrective messages affect knowledge and beliefs, to our knowledge, the effects of viewing the final corrective messages that are being disseminated on smoking-related outcomes have not been evaluated. For example, Kollath-Cattano et al.¹ examined how adult smokers rate the original court-ordered corrective messages (that included the tobacco industry deception statement) on characteristics such as novelty, relevance, and elicitation of quit motivation. However, the study focused only on the comparison across the five topics and did not compare how cognitions would change pre–post or compared to a control condition. Smith et al.⁹ found that previously proposed versions of corrective messages were successful in changing beliefs and knowledge related to the five topics (eg, addiction, light and low-tar cigarettes). Similarly, Tangari et al.¹⁰ also found evidence for previously proposed versions of corrective messages increasing beliefs that smoking is harmful and that the industry manipulated the design of cigarettes to sustain addiction. However, these studies did not examine the effect of current correctives and behavioral outcomes such as intention to quit smoking.

This study examines the effect of viewing current corrective messages and enhanced versions of these messages on smokers’ intention to quit smoking and intention to purchase cigarettes. Although belief and knowledge correction is an initial goal of correctives, the more distal but important outcome of intention to quit smoking could be left unchanged.¹¹ Ultimately, the corrective messages should encourage tobacco cessation to partially undo or stop the damage that the tobacco industry has perpetuated by misleading smokers into smoking cigarettes. Therefore, we first examined whether the current correctives can influence smokers’ intention to quit smoking compared to baseline. Second, the study tested whether two forms of enhanced correctives, corrective messages with industry deception statements (that were excluded from the current corrective) and correctives with both industry deception statements and testimonials of people harmed by smoking could be more effective than the current correctives in affecting intention to quit smoking and intention to purchase cigarettes. This step was carried out to examine potential methods that would further increase the effectiveness of corrective messages.⁷ We are interested in the industry deception statement out of concerns that the exclusion of the statement would lead to weakened effects as prior research shows that emphasizing tobacco industry culpability in campaigns led to reduction in smoking.^12,13 In addition, prior research on corrective statements for industry products that falsely advertised health claims (e.g., Listerine prevents colds; Doan is more effective than other brands for back pain relief) shows that fact-based, simple correctives are ineffective or not effective enough.^14–17 Therefore, testimonials were included in enhanced corrective messages based on robust prior research that found testimonials to be an effective means of encouraging tobacco cessation.^18–22 Third, we explored whether the effect of enhanced corrective messages (vs. current corrective messages) would vary across different subgroups. We expect these findings to inform current and future strategies to remedy the effects of tobacco misinformation.

Methods

Participants (N = 803) were recruited from Survey Sampling International’s online panel to participate in the study in April 2018. Survey Sampling International is an opt-in panel comprising US adults. Survey Sampling International targets panelists that may be eligible for the study based on demographic information provided upon enrollment and invites them to participate in surveys. The inclusion criteria for this study included US adults, aged 18 years and older, who had smoked 100 cigarettes in their lifetime, and who currently smoke some days or every day in the past 30 days. Of 1182 panelists who began the survey, 920 were eligible and 803 completed the survey. Participants were informed that we were conducting a study on messages about smoking. Eligible participants first answered a series of questions about their smoking behavior (eg, nicotine dependence, readiness to quit) including intention to quit smoking (pretest) and were then randomly assigned to view (1) two current correctives (Current), (2) two enhanced correctives with industry deception statement (Industry Deception), or (3) two enhanced correctives with industry deception statement and testimonial (Industry Deception + Testimonial). Participants viewed each message one at a time on their computer screen for at least 10 seconds. The actual messages they viewed are described in Supplementary Table 1. After viewing corrective messages, participants answered questions about their intention to quit (posttest) and intention to purchase cigarettes. At the end of the survey, participants answered questions about demographic characteristics, such as age, race, and gender. Participants received points upon completing the survey that could be converted to cash per Survey Sampling International’s incentive policies. The university’s IRB committee reviewed and approved the study protocol. Study data were analyzed in 2018 using Stata v. 15.²³

Corrective Messages

Supplementary Table 1 shows the corrective messages about health effects and secondhand smoke for each condition. The current corrective messages focus on five topics: health effects, secondhand smoke effects, manipulation of cigarette design, addiction, light and low-tar cigarettes.¹ This study used two of these topics—health effects and secondhand smoke effects—when constructing corrective messages for each condition. The rationale for selecting the health effects and secondhand smoke topics was based on prior research showing high relevance and motivation to quit ratings for these two topics, and the availability of topic-relevant images and testimonials for the enhanced corrective messages.¹

Participants assigned to the Current condition viewed two (health effects and secondhand smoke effects) corrective messages identical in formatting and content to the actual messages that appeared on newspapers as stipulated by the Department of Justice.³ These messages have a statement about the court order for correctives and several topic-relevant (eg, health effects) facts in bullet points.

Enhanced correctives with industry deception statement (Industry Deception) retained the formatting and content of the current corrective messages but added a statement similar to the previous court-ordered language¹ that acknowledged deceptive practices in more specific terms to increase tobacco company culpability. The Industry Deception correctives included a statement that said: “Tobacco companies deliberately deceived the American public that there are no harmful effects of smoking” (health effects), and the phrase “Here is the truth” before topic-relevant facts. For the secondhand smoke effects topic, we included the statement that “Tobacco companies deliberately deceived the American public that there are no harmful effects of secondhand smoke” and the phrase “Here is the truth” before topic-relevant facts.

Enhanced correctives with industry deception statement and testimonial (Industry Deception + Testimonial) included the industry deception statement aforementioned and added topic-relevant images and testimonials from the Centers for Disease Control and Prevention (CDC) Tips From Former Smokers campaign materials (https://www.cdc.gov/tobacco/campaign/tips/index.html). However, only topic-relevant facts that coincided with the included testimonials were retained to keep the length of messages similar across conditions. For example, Terrie’s story about oral cancer was used to highlight health effects from tobacco smoking. To increase consistency and to keep the length of messages similar across conditions, only the bullet point about smoking causing cancer was retained.

Outcome Variables

Intention to quit smoking (pre and post) was measured with three items adapted from prior surveys. Participants rated their willingness to do the following in the next 30 days: (1) try to quit smoking, (2) reduce the number of cigarettes smoked per day, (3) quit smoking completely on a scale from 1 (definitely will not) to 4 (definitely will).¹⁸ Responses to the three items were averaged so that higher scores indicated greater intention to quit smoking (pretest, full data: M = 2.75, SD = .81; Cronbach’s alpha = .83).

Intention to purchase cigarettes (post only) was measured using the Juster scale adapted for tobacco-related behaviors.^24–28 Participants answered one item about how probable they were to purchase cigarettes in the next 30 days on an 11-point scale. Answer options ranged from 0 (no chance, almost no chance) to 10 (certain, practically certain) (M = 7.64, SD = 2.63).

Demographic and Smoking Behavior Variables

Participants answered questions about their demographic characteristics such as age, income, sex (male, female), education (high school or below, some college, college and higher), ethnicity (Hispanic, non-Hispanic), race (white, black, other). Smoking behavior questions included the number of days they smoked in the past 30 days, number of cigarettes smoked daily, nicotine dependence (Fagerström test²⁹) and readiness to quit (Contemplation Ladder³⁰).

Statistical Analysis

First, a paired t test comparing intention to quit smoking pre- and post-exposure to the Current correctives was conducted to examine whether smokers’ intention to quit changed in the Current correctives condition. Second, to examine whether enhanced correctives (Industry Deception or Industry Deception + Testimonial) can affect smokers’ intention to quit (pre–post difference; model 1) and intention to purchase (model 3) to a greater degree than Current correctives, outcome variables were regressed on a categorical variable indicating conditions where the referent category was the Current corrective condition. In addition, we compared the outcomes between the Industry Deception and Industry Deception + Testimonial conditions to analyze whether the addition of testimonials improved effectiveness beyond the industry deception statement. The regression models for intention to quit smoking included pretest intention to quit smoking as a covariate. Third, interaction terms between the condition variable and subgroup variables (gender, race, ethnicity, education, income, and smoking status [daily vs. non-daily]) were included in separate models (interaction with ethnicity presented in model 2 and model 4) to analyze differences across these groups in the effect of enhanced correctives (vs. current correctives). Finally, we conducted follow-up analyses for each subitem within the intention to quit smoking scale to examine which aspect of quitting was affected most by the study conditions.

Results

Demographic characteristics of the sample and F-statistics and chi-square values for comparisons across conditions are described in Table 1. We conducted randomization checks using one-way analysis of variance and chi-square tests of demographic characteristics across conditions which showed that participants’ mean age, nicotine dependence, readiness to quit, number of cigarettes smoked daily, and proportions of females, Hispanic origin, daily smokers, race, education, and income did not significantly differ across conditions.

Table 1.

Sample Characteristics by Condition

	Total N = 803	Current n = 286	Industry Deception n = 265	Industry Deception + Testimonial n = 252
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	F
Age	40.4 (12.6)	40.1 (12.4)	40.5 (12.9)	40.8 (12.6)	0.18
Nicotine dependence	5.1 (2.4)	5.1 (2.3)	5.0 (2.3)	5.0 (2.6)	0.05
Readiness to quit	6.5 (2.7)	6.5 (2.8)	6.5 (2.5)	6.5 (2.8)	0.03
Number of cigarettes smoked per day	15.19 (19.33)	15.16 (19.32)	15.55 (18.36)	14.83 (20.38)	0.09
	%	%	%	%	χ2
Sex (female)	55.7	57.5	55.9	53.6	0.86
Ethnicity					0.13
Hispanic	8.0	8.4	7.6	7.9
Race					1.14
White	82.2	80.8	82.6	83.3
Black	7.5	8.7	6.8	6.8
Other	10.3	10.5	10.6	9.9
Education					3.93
High school and below	24.3	25.2	25.0	22.6
Some college	32.4	28.3	34.5	34.9
College and higher	43.3	46.5	40.5	42.5
Income					1.56
$0–$34999	25.9	25.5	28.0	24.2
$35000–$74999	35.5	34.3	35.6	36.9
$75000 or higher	38.5	40.2	36.4	38.9
Daily smoker	69.7	69.2	73.6	66.3	3.33

Nicotine dependence was measured with the Fagerström test²⁹ and ranged from 0–10. Readiness to quit was measured with the Contemplation Ladder³⁰ and ranged from 0–10.

First, we analyzed whether viewing the Current correctives changes smokers’ intention to quit compared to their baseline. Participants’ intention to quit smoking (M = 2.90, SD = 0.83) increased significantly after being exposed to the Current correctives compared to their baseline intention to quit smoking (M = 2.81, SD = 0.81); t(284) = −3.14, p < .01, Cohen’s d = .11. We also compared baseline and post-exposure means for each item in our intention scale and found that all items significantly increased from pre- to post-exposure. Specifically, exposure to the Current correctives increased willingness to “try to quit smoking” (t(284) = −2.35, p < .05), “reduce the number of cigarettes smoked per day” (t(284) = −2.68, p < .01), and “quit smoking completely” (t(283) = −2.21, p < .05).

Next, we analyzed whether viewing the enhanced correctives could increase intention to quit smoking and decrease intention to purchase cigarettes compared to the current correctives. Table 2 presents the regression models that compared outcomes across the Current corrective condition versus the enhanced corrective conditions (Industry Deception, Industry Deception + Testimonial). Intention to quit smoking did not significantly differ between participants who viewed the Current correctives and those who viewed the Industry Deception correctives (Table 2, model 1). However, participants exposed to the Industry Deception + Testimonial correctives had significantly greater intention to quit smoking compared to those exposed to the Current correctives (B = 0.09, p < .05; Table 2, model 1). Follow-up analysis indicated that Industry Deception + Testimonial correctives significantly increased the intention scale subitem, “reduce the number of cigarettes smoked per day” (B = 0.11, p < .05). In addition, those who viewed the Industry Deception + Testimonial correctives had significantly higher intention to quit smoking than those who viewed the Industry Deception corrective messages (B = 0.08, p < .05). Intention to purchase cigarettes did not differ between participants across conditions (Table 2, model 3).

Table 2.

Effect of Industry Deception Corrective and Industry Deception + Testimonial Corrective on Intention to Quit and Intention to Purchase Cigarettes

	Post-intention to quit N = 801				Intention to purchase cigarettes N = 803
	Model 1		Model 2		Model 3		Model 4
	B	[95% CI]	B	[95% CI]	B	[95% CI]	B	[95% CI]
Condition (reference category is Current corrective)
Industry Deception corrective	0.00	[−0.07 to 0.08]	0.01	[−0.07 to 0.09]	0.29	[−0.15 to 0.73]	0.28	[−0.18 to 0.74]
Industry Deception + Testimonial corrective	0.09*	[0.01 to 0.16]	0.06	[−0.02 to 0.14]	−0.13	[−0.57 to 0.32]	0.02	[−0.45 to 0.48]
Pre-intention to quit	0.82***	[0.78 to 0.85]	0.82***	[0.78 to 0.86]	(Not included)
Ethnicity (reference category is non-Hispanic)	(Not included)				(Not included)
Hispanic			0.02	[−0.17 to 0.21]			−0.14	[−1.24 to 0.96]
Industry Deception corrective × Hispanic	(Not included)		−0.05	[−0.34 to 0.23]	(Not included)		0.11	[−1.51 to 1.74]
Industry Deception + Testimonial corrective × Hispanic	(Not included)		0.31*	[0.04 to 0.59]	(Not included)		−1.83*	[−3.45 to −0.20]
Constant	0.60	[0.48 to 0.73]	0.60	[0.48 to 0.72]	7.59	[7.28 to 7.89]	7.60	[7.28 to 7.92]
R  2	0.68		0.68		0.004		0.02

Listwise deletion was used for missing values. CI = confidence interval.

*p < .05, ***p < .001.

Third, we analyzed whether the effects of enhanced correctives compared to the current correctives differ by subgroups. There were no differences in the effect of enhanced correctives (vs. current correctives) for subgroups based on age, sex, education, race, income, and smoking status (daily vs. non-daily). The interaction between ethnicity (Hispanic vs. non-Hispanic) and the Industry Deception + Testimonial corrective condition was significant for intention to quit (B = 0.31, p < .05; Table 2, model 2) and intention to purchase cigarettes (B = −1.83, p < .05; Table 2, model 4). Follow-up analysis indicated that the interaction between ethnicity (Hispanic vs. non-Hispanic) and the Industry Deception + Testimonial corrective condition was significant for the intention subitem, “reduce the number of cigarettes smoked per day” (B = 0.36, p < .05). Hispanic participants who viewed the Industry Deception + Testimonial corrective messages (vs. Current correctives) had higher intention to quit and lower intention to purchase cigarettes compared to non-Hispanics who viewed the Industry Deception + Testimonial corrective messages (vs. Current correctives). Figure 1 illustrates the predicted marginal means of intention to quit and intention to purchase cigarettes across conditions for Hispanic and non-Hispanic participants.

Figure 1.

Predicted intention to quit smoking and intention to purchase cigarettes by study condition and ethnicity. Predicted marginal means and standard errors are based on the regression model estimates in Table 2.

Discussion

This study is, to the best of our knowledge, the first study to examine how viewing current corrective messages can change intention and the effects of enhanced versions (vs. current correctives) on outcomes of intention. The study found that viewing the Current corrective messages increased smokers’ intention to quit smoking compared to baseline. In addition, we found that the Industry Deception + Testimonial correctives were more effective than Current correctives in increasing intention to quit smoking, but not for reducing the intention to purchase cigarettes. The effect of corrective condition did not differ by age, sex, education, race, and income. However, ethnicity (Hispanic vs. non-Hispanic) was a significant moderator such that Hispanic smokers had higher intention to quit smoking and lower intention to purchase cigarettes after viewing the Industry Deception + Testimonial correctives (vs. Current correctives) compared to non-Hispanic smokers.

This study’s finding that viewing current correctives increased smokers’ intention to quit smoking beyond their baseline should, however, be interpreted with caution. This result may have manifested from repeated testing and may reflect an instrument effect. Therefore, more future studies are needed to assess the impact of the current correctives. For example, future research could use a control condition where participants see neutral messages unrelated to tobacco to avoid this concern to inference.

More importantly, the study results suggest that while corrective messages that included the industry deception statement (similar to the original court-ordered correctives) alone may not improve effects on intention to quit smoking beyond the current correctives, the addition of testimonials could further improve outcomes. These results are consistent with previous research that found that graphic warning labels with images and testimonials are more effective in increasing intention to quit smoking compared to text warning labels.^18,19,31 The effect of corrective messages could be enhanced with the addition of industry deception statement and testimonials of people harmed by smoking. Although this may not be feasible for completed print and television advertisements, messages appearing for a longer period on cigarette package onserts and tobacco industry web sites could be bolstered with an industry deception statement and testimonials similar to those used by the current study to increase effectiveness. Of note, the enhanced versions of corrective messages did not have any effects on intention to purchase cigarettes compared to the current corrective messages. This may be because almost 70% of the samples were daily smokers with moderately high levels of nicotine dependence (Table 1). The brief exposure to enhanced versions of messages may not have been enough to reduce already established levels of intention to purchase cigarettes in the next 30 days.

The significant moderation effect of Hispanics is in line with previous research that found that Hispanic smokers perceived Tips From Former Smokers campaign advertisements (that were used in the enhanced corrective messages) to be more effective than non-Hispanic whites and that the relationship between perceived effectiveness and quit attempts was stronger for Hispanics compared to non-Hispanic whites.²⁰ In general, Hispanics have higher receptivity to antismoking messages³² and this may be because Hispanic smokers are more ready to quit as can be seen in their more frequent quit attempts compared to non-Hispanic whites.³³ However, given the small number of Hispanic persons per condition in this study, more research with a larger Hispanic population is needed.

The study has some limitations that could be addressed in future research. First, the study used only two messages that focused on health effects and secondhand smoke effects of five topics required by the court. Although exposure was brief, viewing both current and enhanced corrective messages increased intention to quit smoking. However, effects may be stronger with repeated exposure. Future research could test whether viewing all five topic corrective messages has stronger effects on intention to quit smoking and intention to purchase cigarettes. Future studies could also identify which topics are the most effective in influencing intention. Second, the study population was not representative of the population of US smokers and lacked racial diversity with more than 80% being white. Future studies could test the effect of current and enhanced messages on a nationally representative sample to examine whether the effects of current and enhanced correctives are generalizable to the US population. Third, the study was not able to disentangle the effects of an image of a person versus the person’s textual testimonial. Although there is some research in the context of graphic warning labels that an image may be more influential in increasing quit attempts than testimonial statements, future research that includes a condition without pictures is needed to examine this notion.³⁴

Despite these limitations, the study results are significant as they demonstrate that enhanced corrective messages that include an industry deception statement and testimonial can enhance effects on intention to quit smoking compared to the current correctives. Given the potential for limited exposure^35,36 to current correctives, enhanced corrective messages may be an alternative for current and future efforts to remedy the effects of tobacco industry misinformation.

Funding

There is no funding to report for this study.

Declaration of Interests

None declared.