Increasing Alcohol Control Policy Support: The Mediating Role of Empathy and Emotions

Abstract

Background: Mass media messages have the potential to reduce alcohol-related harm through increasing support for alcohol-control policies. Objectives: In Study 1, we experimentally examined the incidental effects of alcohol mention in news articles about accidents and crime and found evidence that messages mentioning alcohol as a causal factor in the accident or crime can activate empathic and emotional responses, which in turn increase support for alcohol control policy. In Study 2, anti-drinking and driving public service announcements (PSAs) were edited to either portray consequences of drinking and driving for the driver only or for both the driver and innocent others. Results: The versions of the PSAs that showed consequences to others were more successful at eliciting alcohol-control policy support than the versions showing only consequences to the self. As hypothesized, empathy that was directed toward victims (but not drivers under the influence) and negative emotions were supported as mediators of this relationship. Although negative affect and feelings of empathy for the victims of drunk driving are not pleasant emotions to experience, they appear to play an important role in increasing support for public policies to increase safety without having a negative impact on viewer’s evaluations of the PSA. Conclusions: Results from the two studies provide evidence that empathy and emotions can play a mediating role between message characteristics (either intentional or incidental) and alcohol control policy support.

Keywords:

• Alcohol
• empathy
• emotions
• prosocial
• policy support
• health

Alcohol control public policy support

An effective way to reduce alcohol-related harms is through alcohol-control policies such as enforcing laws prohibiting distribution of alcohol to minors (Wagenaar & Perry, 1994). Media messages can influence the public to support these policies, but there remains much to learn about the types of message characteristics that are successful in inducing this support and the mechanism that explain why those message characteristics are successful. In the present research, we examine the role of the emotions of empathy and disgust in response to news coverage of alcohol-related harms, and experimentally compare messages that focus on harm to self alone vs including an emphasis on harm to victims.

Study one

The mention of alcohol as a causal factor in news articles regarding accidents and crimes increases support for enforcement of alcohol related policies already in effect (Slater et al., 2012). Anger mediates this effect of alcohol mention in such news stories on alcohol enforcement support (Goodall et al., 2013). Further, emotional reactions (e.g. anger and sadness) elicited by coverage of alcohol-related crimes and accidents increase policy support via concern (Solloway et al., 2013).

In addition, this study investigates the role of disgust—an emotion that psychologically operates differently than anger, but that we posit may produce similar effects on policy support. Disgust is an unpleasant emotion, but like anger, it can encourage increased policy support. This study also investigates the potential effect of priming empathy, a variable we anticipate may co-activate with anger and disgust.

Emotions

As noted by previous research (see Goodall et al., 2013; Nabi, 2003) negatively valanced emotions are likely to occur in response to media coverage about crimes and accidents. Indeed, previous work (e.g. Goodall et al., 2013) has supported the idea that anger can mediate the effect of alcohol mention in news stories on alcohol enforcement support; we look to replicate this finding.

H₁: Anger will mediate the relationship between alcohol mention and alcohol control policy support such that participants who read versions of news articles that mention alcohol as a causal factor will have stronger policy support (via anger) than participants who read versions of articles that do not mention alcohol as a causal factor.

Unlike the approach tendency of anger, disgust makes a person want to recoil from and reject a stimulus (see Newhagen, 1998). We propose that the avoidance reaction associated with disgust may involve rejection of the emotion-inducing behavior depicted in the stories (in this case, harm caused to victims by intoxicated perpetrators). This distinction from a flight-based avoidance reaction is important, and we argue it is possible that disgust may also mediate the relationship between coverage of alcohol related crimes/accidents and alcohol control policy support, particularly since numerous policies involve punitive action toward perpetrators or stricter enforcement of actions intended to address the problem.

H₂: Disgust will mediate the relationship between alcohol mention and alcohol control policy support such that participants who read versions of the news articles that mention alcohol as a causal factor will have stronger policy support (via disgust) than participants who read versions of the articles that do not mention alcohol as a causal factor.

Empathy

Anger and disgust reactions are partly motivated by injustice of the harm inflicted upon victims. As such, trait empathy will likely co-activate when anger and disgust are elicited. Empathy occurs when a person feels an emotional response congruent with the perceived welfare of another (Batson & Moran, 1999).

Previous work shows that empathy can be an important variable influencing persuasion as well as altruistic motivation (e.g. Batson et al., 1981). For example, Johnson et al. (2009) found that empathy mediated the relationship between stereotype primes and support for public policy regarding victims of a natural disaster, and Gault and Sabini (2000) found empathy to be a mediator between gender and policy support for distressed others.

H₃: Empathy will mediate the relationship between alcohol mention and alcohol control policy support. Participants who read versions of news articles that mention alcohol as a causal factor in crimes or accidents will experience more empathy, and thus more policy support than those who read versions that do not mention alcohol as a causal factor.

Methods

Design

The study employed a 3 × 2 multiple message experimental design in which Qualtrics was programmed to evenly distribute participants across conditions. The three-level factor was story topic. Participants read a series of three stories presented in random order about (a) a motor vehicle accident, (b) non-motor vehicle accident (e.g. drowning, firearm accident, electrocution, etc.), and (c) violent crime. Additionally, one story for each of the three topics came from a pool of 60 unique stories taken from a national sample of crime and accident news. These 60 stories were manipulated to either mention or not mention alcohol as a causal factor in the crime or accident (the between-subjects factor). These stories were used in prior research but with a new respondent population and a measurement instrument expanded to include measures of disgust and empathy in addition to the anger measures that were used in previous research (Slater, et al., 2012).

Procedures and measures

Participants were randomly assigned to either a condition mentioning the role of alcohol as a causal factor in the incidents or not mentioning alcohol as a causal factor. After reading each story, participants answered a series of questions about their emotional reactions to the content, completed a measure of policy support (see Goodall et al., 2013). Items were measured on an 11-point scale from 0 (Do not support at all) to 10 (Support very strongly), and included the following stems for items measuring support for enforcement of current laws: Stricter enforcement of laws in my community that forbid liquor stores, bars, or restaurants to sell alcohol to someone who is visibly intoxicated, Stricter enforcement of laws in my community that forbid liquor stores, bars, or restaurants to sell alcohol to youth under age 21, Stricter enforcement of laws that forbid having open alcoholic beverage containers in motor vehicles (α = .90). To measure support for new policies, stems included: Policies in my community that would limit the number of liquor stores in a neighborhood, Policies in my community that would limit the number of bars in a neighborhood, Laws in my community that would ban alcohol-related billboards, and State laws making bars and restaurants liable for lawsuits if someone becomes intoxicated there and causes someone to be killed or injured in a car crash or assault (α = .91). A 17-item measure of general empathy, adapted from the Interpersonal Reactivity Index (Davis, 1983) was measured on a 0 (Strongly Disagree) to 7 (Strongly Agree) scale (α = .79).

Participants

A total of 911 participants from Qualtrics online panels participated in the study, 50.30% female, 7.24% Hispanic, 84.28% White, 7.90% Black, 3.24% Asian. The mean age was 31.09 years old (SD = 14.25). For power analysis relevant to the indirect effects examined in his PROCESS macro, Hayes (2013) recommends Fritz and MacKinnon’s (2007) table, which estimates that for .80 power, 558 participants would be a minimum.

Results

The mention of alcohol as a causal factor had an indirect effect on both new and existing policy support through anger, replicating prior results (Goodall et al., 2013). For existing policies, the indirect effect was statistically significant as the confidence intervals (CI) did not contain zero lower-level confidence interval [LLCI] = .02; upper-level confidence interval [ULCI] = .05; β = .04. For new policies, the LLCI = .06, ULCI = .14, β = .10. Thus, H₁ was supported.

Although anger is an approach emotion and disgust is an avoidance emotion (e.g. Elliot et al., 2013), we expected that disgust would act similarly to anger because the avoidance response is driven by a sense of rejection for the object eliciting the emotion, for example, drunk driving (Newhagen, 1998). Further, both anger and disgust are negative emotions focused on the cause of the incident. The data support this hypothesis, as disgust mediated the effect of alcohol mention on policy support very similarly to anger. For existing policies, the indirect effect was statistically significant as the CI did not contain zero (LLCI = .04; ULCI = .08; β = .06). For new policies, the LLCI = .07, ULCI = .15, β = .11. Thus, H₂ was supported.

We also expected that empathy may play a mediational role between message characteristics and alcohol control policy support. Three models were examined to see how the variables (alcohol mention, emotions, and empathy) worked together to increase policy support. One model was a serial effect from alcohol mention to anger (or disgust) to empathy to policy support. The second was a serial effect from alcohol mention to empathy to anger (or disgust) to policy support. The third looked at anger, disgust, and empathy as distinct mediators. Although all three models appeared viable, the most parsimonious was the model looking at emotions and empathy as distinctly contributing to the effect of alcohol mention on policy support via parallel (see Table 1 and Figure 1). Thus, H₃ was supported.

Figure 1. Mediation model for empathy and anger producing largest total indirect effect.

Table 1. Total indirect effects of alcohol mention on policy support for the three models.

Model type	Effect	BootSE	BootLLCI	BootULCI
Parallel mediation	.04	.01	.02	.05
Serial: anger then empathy	.04	.01	.02	.06
Serial: empathy then anger	.04	.01	.02	.06

Discussion

This study provides initial evidence that for the purpose of influencing health policy support, messages can facilitate empathetic reactions, which can influence outcomes above and beyond the effect of emotions such as anger and disgust. However, it is unclear whether it is empathy for the drunk driver (e.g. a shared feeling of fear or guilt) or empathy for the victim (e.g. a shared feeling of sadness or loss) that drives this effect.

The messages in Study 1 either did or did not include alcohol mention as a causal factor in alcohol-related accidents and crime, which is an incidental message characteristic. Incidental effects are not designed to increase policy support, but we see here that they can. The question remains whether public service announcements (PSAs), which differ from news articles in format (print vs video) and intent (incidental or intentional) will have the same effect.

Study two

In this study, we focus on how anti-drinking and driving PSAs may influence public support for regulatory action as well as influence outcomes such as attitude change.

Individual verses collective interest as a motivator of responsible behavior

PSAs often show the consequences of engaging in risky behavior. This is particularly the case for anti-drinking and driving PSAs which can be found to show consequences to the driver (e.g. physical harm, arrest, losing a driver’s license, etc.) and/or consequences to other people (e.g. unknown others injured in accidents by drunk drivers, financial consequences for the family of a drunk driver, etc.).

Yet, it is unclear if the general public is more influenced by appeals to their own safety and well-being or by appeals for the safety and wellbeing of other people. Evidence suggests that this may vary based on the topic and situations involved in the appeal. For example, Hupfer (2006) found that for blood donation campaigns, an appeal to self-interest (making blood available for themselves if required) was more successful than a more altruistic appeal. However, Perea and Slater (1999) found that participants responded to a collectivist alcohol warning label more positively than the individually-oriented one. Here, we seek to examine whether messages concerned with risks to others influence policy support and its antecedents through empathy and evoked emotion.

Empathy and emotions as mediating mechanisms

Study 1 found evidence that empathy and negative emotions mediate the effect of news stories on support for alcohol control policy. We extend that research by using empathy and evoked emotions as a mechanism for explaining the effects of other-harm versus harm to self only as threats in drinking and driving PSAs. We also examine the concept and operations for empathy more closely, by examining the locus of empathy.

Emotions

Goodall et al. (2013) found evidence that emotions explain distinct responses to alcohol-related messages. To more thoroughly understand the effects of the PSAs, several dependent variables were used in addition to measures regarding support for various alcohol control policies. We included behavioral measures of willingness to advocate in support of alcohol control policies, alcohol evaluations, concern about drinking and driving, approval of drinking and driving, and evaluations of the PSAs.

Prior research has not distinguished between alcohol-related harms to the intoxicated individual or alcohol-related harms to innocent others. However, it seems plausible to infer that portraying the victim of someone else’s intoxicated behavior enhances anger, sadness, and disgust, which tend in turn to increase issue concern and support for alcohol control policies. If so, we should see that anti-drinking and driving PSAs portraying harm to others should be more effective.

H₁: Messages that depict consequences to both the self and others will indirectly increase a) positive message evaluation, b) negative attitudes about drinking and driving, c) concern about alcohol-related risks, d) support for alcohol control policy, and e) willingness to actively support such policies through negatively valenced emotions.

Empathy

Study 1 indicated that the news stories about alcohol-related crimes and accidents could prime increases in a trait empathy measure. In the present study, we adapt the trait empathy measure to assess empathy that is specifically directed toward either victims or intoxicated drivers. It is possible that while watching a PSA about drinking and driving, participants could feel empathy specifically directed toward people who drink and drive. Perhaps they have experience engaging in this behavior or understand how it would feel to be caught drinking and driving or to harm another person. This empathy with the intoxicated driver in the PSA could increase alcohol control policy support as viewers feel any kinds of fear, guilt, or shame that the drunk drivers may have experienced in the PSA. However, it seems more likely that policy support will be driven by empathy with the victim of drinking and driving. Previous work (e.g. Gault & Sabini, 2000; Johnson et al., 2009) did not flesh out whether effects were driven by empathy for the instigator, however, their data provide evidence that it is likely that empathy for the victims drives altruistic intensions. Thus, in addition to emotions, empathy may also play a role in mediating effects PSA characteristics on engagement with narratives. Therefore, we predict:

H₂: Messages that depict consequences to both the self and others will indirectly increase a) positive message evaluation, b) negative attitudes about drinking and driving, c) concern about alcohol-related risks, d) support for alcohol control policy, and e) willingness to actively support such policies, through state empathy with the victims of drunk driving accidents.

Methods

Participants

A total of 437 participants were recruited from a Qualtrics online panel. Participants were 62.93% female, an average of 47.36-year old (SD = 16.972), 14.25% were Hispanic and 72.85% were White. For power of .8, 462 participants are needed (Fritz & MacKinnon, 2007), making this study just slightly underpowered.

Design and stimulus

After consenting to participation, participants were randomly and evenly assigned to one of three conditions: (a) a condition featuring a PSA showing a drunk driver causing self-harm through an accident, (b) a condition featuring a PSA showing a drunk driver causing self-harm and harm to victims through an accident, and (c) a control condition in which participants viewed a Jim Beam or a Budweiser product advertisement. Since there were two videos in each condition (i.e. two videos that showed harm to the drunk driver only, two videos that showed hard to self and others, and two videos that advertised alcohol), we ran the data analysis with each video separately to verify that they were equal in terms of significance before collapsing the two videos into each condition. Indeed, in terms of significant impact on mediators as well as indirect efforts on the consequent variables, the videos did not differ within conditions and thus were collapsed.

To develop the stimuli, two anti-drinking and driving PSAs were manipulated so that there were two versions of each. Ads were chosen if they were not created in the United States (to reduce the chance of prior exposure) and did not contain spoken language as languages other than English or strong accents may be distracting for American participants. In addition, the ads were less than a minute long and were deemed to be possible to edit in such a way as there could be two versions (one version showing consequences to the driver only and one version that also showed consequences to innocent others) that both seemed credible and impactful.

Measures

The discrete emotions anger, sadness, fear, and disgust were measured on an 11-point scale (from “0, did not feel at all” to “10, felt extremely strongly”).

Empathy was measured both for the victims of drunk driving (α = .86) and for drunk drivers themselves (α = .83) with a modification of the trait empathy measure, the Interpersonal Reactivity Index (Davis, 1983). For example, “I can easily put myself in the place of characters” was modified to “I could easily put myself in the place of someone hurt by a driver under the influence of alcohol.”

Concern about alcohol related accidents was measured via four items asking whether participants rated motor vehicle accidents caused by a driver under the influence alcohol from “Just a minor concern” (0) to “One of your biggest concerns” (10) and differentiating between “in my state,” “happening to a family member or close friend,” “happening to me,” and “happening in the United States,” α = .82.

As in previous research (Goodall et al., 2013), alcohol control policy support was measured for both existing (α = .78) and new policies (α = .87).

Positive behavioral intentions were measured via items asking how willing participants would be to sign a petition for stronger enforcement of existing policies, or to donate money or volunteer to pass out a petition (α = .80).

Alcohol evaluations were measured on 11-point semantic differential scales for good-bad, beneficial-harmful, and like-dislike (α = .84).

Evaluations of the PSA were also measured via items asking how clear, well-done, and interesting the videos were, again on an 11-point scale (α = .90).

Results

Direct effects of the experimental manipulation

Our primary outcome is policy support, and we first tested to see if the PSAs illustrating harms to others increased such support. For new policy support, this turned out to be the case with those who saw the version of the video with harm to self only (m = 6.84, SD = 2.79) reporting lower support than those who saw the version with harm to both self and others (m = 7.42, SD = 3.16), t (435) = −2.04, p = .04, direct effect β = .42. Findings were similar for support of existing policies with the harm to self only group (m = 9.38, SD = 1.96) showing lower policy support than the harm to both self and others group (m = 9.82, SD = 1.71), t (435) = −2.48, p = .01, direct effect β = .25. However, main effects on concern about drinking and driving, approval of drinking and driving, behavioral intention to engage in active support of policy, and evaluations of alcohol were non-significant.

Control videos

Although our main interest was not in comparing the PSAs to the control videos (alcohol ads), it is worthwhile to note that the versions of PSAs that showed harm to self only (m = 5.11, SD = 2.59) and the control videos (m = 2.04, SD = 1.74) differed significantly on negative affect t (642) = 17.76, p < .01. The versions of the PSAs showing harm to both self and others (m = 6.85, SD = 2.78) also differed significantly from the control videos, p < .001. Regarding empathy for victims of drinking and driving, while the PSAs that showed harm to both self and others (m = 6.78, SD = 1.30) differed from the controls (m = 6.42, SD = 1.53); t (654) = 3.24, p = .001. However, the versions of the PSA which did not show harm to innocent others (m = 6.47, SD = 1.43) did not differ significantly than the alcohol ads for empathy for victims. Thus, ads that showed harm to the self only did no better at eliciting empathy for victims of drinking and driving than did alcohol advertisements.

Negative affect as a mediator

H₁ proposed that messages that depict consequences to both the self and others will indirectly affect alcohol control policy support through negatively valenced emotions. When Model 4 was run for Hayes’s PROCESS Macro (Hayes, 2013), the type of video influenced negative affect (β = 1.77, p < .01) which in turn influenced new alcohol control policy support (β = .26, p < .01). The indirect effect was significant as the CI did not include zero. The LLCI was .27 and the ULCI was .73, β = .47. The indirect effect of negative affect on existing policy support was also significant as it was on other dependent measures (see Table 2). H₁ was consistently supported across each outcome measure.

Table 2. Indirect effect of PSAs showing harm to others through negative affect.

Dependent measure	Lower-level confidence interval	Upper-level confidence interval	Beta
New policy support	.27	.73	.47
Existing policy support	.18	.49	.31
Behavioral intensions to support policies	.28	.74	.47
Evaluations of alcohol	.08	.31	.18
Concern for drinking and driving	.32	.68	.47
Approval of drinking and driving	–.32	–.01	.15
Evaluations of PSA	.24	.61	.39

Empathy as a mediator

H₂ argued that messages that depict consequences to both the self and others will indirectly affect alcohol control policy support, more so than messages depicting consequences to the self only, through state empathy with the victims of drunk driving accidents. This hypothesis was supported across each DV. The results for empathy for the victims of drinking and driving were nearly identical to the results of negative emotions, see Table 3. Indirect effects through empathy for the driver did not reach significance for any of the DVs. Therefore, it is clear that effects were due to empathy elicited for the victim, and not more general empathetic responses, see Figures 2 (support for new policies) and 3 (support for existing policies).

Figure 2. Indirect effects of video on support for new policies through affect and empathy.

Figure 3. Indirect effects of video on support for existing policies through affect and empathy.

Table 3. Indirect effect of PSAs showing harm to others through empathy.

Dependent measure	Lower-level confidence interval	Upper-level confidence interval	Beta
New policy support	.05	.32	.16
Existing policy support	.06	.39	.21
Behavioral intensions to support policies	.06	.47	.26
Evaluations of alcohol	.01	.12	.05
Concern for drinking and driving	.05	.35	.19
Approval of drinking and driving	–.28	–.04	–.13
Evaluations of PSA	.05	.42	.21

Evaluations of the PSAs

We assessed evaluative responses to the PSA as described earlier. The video showing harm to both self and others (m = 9.41, SD = 2.03) was rated more highly on how well-written, clear, and interesting it was than the video that showed harm to only the driver (m = 8.84, SD = 2.33), t (435) = −2.72, p = .01. The indirect effect of video type on new policy support via evaluation of the PSA was also significant (β = .14, LLCI = .04. ULCI = .30), as was that for existing policy support (β = .13, LLCI = .03, ULCI = .26). To the extent that this may be a result of responding to the effects of editing instead of responding to the presence of portrayals of victims, it is a threat to interpretation of findings.

Therefore, we also tested the impact of evaluation of the PSA in a model looking at empathy and emotions as simultaneous mediators, to assess whether or not positive evaluations would account for results and eliminate the effects of empathy and emotions in response to the experimental manipulation. That did not prove to be the case. Once empathy and emotions were added to the model, the effect of evaluation of the PSA disappeared for both new policy support (β = .06, LLCI = −.01, ULCI = .19) and existing policy support (β = .04, LLCI = −.01, ULCI = .12), and the effects of empathy and emotions survived. For the indirect effects on new policy support, the CI for negative affect are significant (LLCI = .15. ULCI = .59) as are the CI for empathy (LLCI = .02, ULCI = .23), but the video evaluation indirect effect is no longer significant (LLCI = −.01, ULCI = .19). Similarly, for existing alcohol policies, negative affect was a significant mediator (LLCI = .01 and ULCI = .24), as was empathy (LLCI = .05, ULCI = .35). However, the video evaluation was again non-significant (LLCI = −.01, ULCI = .12).

Discussion

This study provides evidence that negative affect and empathy are mechanisms explaining the impact of emphasizing victims of drinking and driving in the PSAs. Both negative affect and empathy for the victims of drinking and driving provided significant indirect paths for the effect of video type on support for new and existing alcohol control policies, alcohol evaluations, concern about drinking and driving, and approval of drinking and driving.

One of the strengths of this study is that we used multiple dependent measures, and each of these measures behaved very similarly, which make our findings more robust. Particularly, these findings indicate that these PSAs not only influence attitudes about alcohol and intoxicated driving but also influence public policy support. The impact of PSAs on alcohol control public policy support appears to be a function of feelings of anger, sadness, fear, and disgust associated with the impact of drunk driving on victims, and on empathy for such victims and their families. Additionally, as we had two videos in each condition, and the effects were equal in terms of significance for both, this increases our confidence that the effects were not due to the unique effects of a single stimulus and editing efforts.

These findings have practical implications for the creation of anti-drinking and driving PSAs. Clearly, our findings support use of PSAs that include consequences to others in addition to consequences to the person engaged in the behavior. Although negative affect and feelings of empathy for the victims of drunk driving are not pleasant emotions, they appear to play a strong role in increasing support for policies to increase safety without having a negative impact on viewer’s evaluations of the PSA.

From a theoretical perspective, our findings further support an emerging model regarding the potential of mediated messages to elicit support for altruistic prosocial behaviors that improve the well-being of others. By altruistic prosocial behaviors, we refer to behaviors that benefit others without any direct benefit (and indeed sometimes might involve inconvenience) to oneself. We suggest that mediated messages do so through two paths. One path is through priming existing tendencies toward empathetic responses, which in turn translate into increased willingness to support altruistic behavior. The second, and related, path is through the emotions of anger, sadness, and so forth that are evoked by the message.

A second theoretical finding of interest is the impact of these drinking and driving PSAs on public policy support. Critics of behavior change interventions, such as behavior change PSAs, have pointed out that such interventions focus on individual behavior and obscure a focus on public policies that may underlie public health issues (e.g. Wallack, 1993). These findings suggest that some kinds of public service-type messages may increase support for policy-level interventions.

From a substantive vantagepoint, then, this research suggests that creation of PSAs illustrating impact on victims, and encouraging news coverage that clearly communicates the causal role of alcohol in damaging the lives of others, not only impacts attitudes about alcohol risks but also can increase support for alcohol control policies. Public support for such policies increase the likelihood of their being instituted, which in turn are likely to reduce alcohol-related harms.

Limitations and future research

In this study, we did not test a video that only included harm to others. Such a video would help understand the relationship among these variables; however, it is very difficult, if not impossible, to manipulate a video to be comparable yet fit this structure. Likewise, we chose to edit for highly comparable messages in terms of tone, production, and content to assess impact on our measures of empathy and affect, rather than comparing ads with different persuasive strategies, production, actors, and so forth. We believe that such stimuli are more comparable and provide a more rigorous test of hypotheses, but the modest difference in length remains a limitation of this study. Given this limitation, more confident inference can be made from our analyses of mediation than from analyses of direct effects of the ads on outcomes.

In sum, then, our findings suggest that mediated messages concerning alcohol-related risks can increase responses consistent with an empathetic self-concept with concomitant increases in support for potentially beneficial health policy.

Declaration of interest

The authors declare that they have no conflict of interest. The authors alone are responsible for the content and writing of the article.