Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Apr 19.
Published in final edited form as: J Health Commun. 2009 Apr–May;14(3):262–275. doi: 10.1080/10810730902805838

Media Influence on Alcohol Control Policy Support in the U.S. Adult Population: The Intervening Role of Issue Concern and Risk Judgments

Michael D Slater 1, Frank Lawrence 2, Maria Leonora G Comello 3
PMCID: PMC2856665  NIHMSID: NIHMS192176  PMID: 19440909

Abstract

Data from a national random-digit dial survey (N=1272) were analyzed to examine the influence of news media use on alcohol-control policy support, and to test whether risk judgments and concern about alcohol-related risks mediated effects of news media use variables on support for various types of alcohol-control public policies. In so doing, we test the proposition that perceptions influenced by routine coverage of events such as crime or accidents may in part explain news effects on public policy support in the domain of health policy. Analyses indicated that the (positive) influence of attention to news about crime and accidents on support for laws increasing server liability and limiting marketing of alcohol products was mediated by concern about risks of alcohol-related injuries and by perceptions of the alcohol-attributable fraction of homicides and unintended injury fatalities. Tests of model fit suggest that concern precedes the more cognitive risk judgment in the mediation model.

Influencing health-related policy often is an effective option for improving public health, notably in the case of alcohol-related injuries (e.g., Toomey & Wagenaar, 1999; Wolfson et al., 1996). In turn, influencing media coverage has been proposed as a means of achieving changes in such health policy (Holder & Treno, 1997; Montgomery, 1993; Wallack & Dorfman, 2001). Relatively little attention has been paid, however, to studying how news coverage will influence support for health policies in general and with respect to alcohol control policies in particular among members of the general public.

We examine the role of news media use in influencing alcohol control policy support in a probability sample of U.S. adults, and hypothesize that such effects are mediated by concern about and judgments of risks associated with alcohol use.

News Influence on Health Policy Support

Agenda-setting research (McCombs & Shaw, 1991) has emphasized the role of the media in maintaining salience of given public issues. Research on these agenda-setting processes typically focuses on thematic coverage of public policy issues. However, much health-related news is focused on human drama or medical developments, not on policy concerns and controversies. For example, the vast majority of coverage relevant to alcohol risks is embedded in mentions of alcohol’s role in motor vehicle and other types of accidents and in instances of violent crime, or episodic coverage as Iyengar (1991) calls it—not in thematic stories about alcohol policy issues (Slater, Long, & Ford, 2006).

Since coverage of alcohol-related risk is largely episodic coverage of incidents of accident and crime, it is likely that such coverage provides exemplars of such risks. Exemplars—individual examples portrayed in the media—are typically as or more influential on a variety of judgments than the kind of statistical information typically reported in thematic coverage (see Zillmann & Brosius, 2000, for a review). Therefore, episodic coverage about health risks may provide a path for news media influence on health policy support.

Consistent with this argument, time-series studies examining the number of stories about drunk driving on the AP newswire found indirect relationships between amount of coverage and reduction in drunk driving behavior via the impact of this coverage on policy (Yanovitzky & Bennett, 1999; Yanovitzky, 2002). However, there is relatively little research exploring such relationships at the individual level.

Media and concern or judgments regarding alcohol-related risks

The proposition that media content will influence public judgments of public health risks, and their concern about those risks, is not new (Hargreaves, 2002). Certainly, there is a rich literature linking media use to increases in perceptions of social risks such as crime victimization (Gerbner, Gross, Morgan, Signorielli & Shanahan, 2002). Several studies have found that public rankings of public health risks track the prominence of these risks in news coverage, rather than their actual relative risk of causing death (Combs & Slovic, 1979; Kone & Mullet, 1994; Lichtenstein, Slovic, Fischhoff, Layman & Combs, 1978). A recent study provides evidence that attention to news media is linked to knowledge about cancer-related risks (Viswanath et al., 2006). However, little prior research of which we are aware has examined the relationship between individual news media use and policy support relevant to health and safety issues—much less examining mechanisms for such effects.

An earlier analysis of the present survey found direct influence of news media use, notably attention to crime, accident, and injury news content, on concern and risk judgments as part of a study exploring media exposure and attention to relevant story types as intervening variables between demographics and belief outcomes (Slater & Rasinski, 2005). In the present study, we extend that effort, and extend earlier aggregate-level research linking news coverage to alcohol-control policy, by testing the effects of news media exposure and attention to relevant content on alcohol control policy support.

Given that there is a great deal of diverse news covered in newspapers and broadcasts, most of it quite irrelevant to alcohol-related risks, attention to relevant coverage is more likely to predict policy support than is mere overall news exposure alone. However, if these effects are particularly robust, or if some types of alcohol-related coverage are prominent enough to be processed even if people are not actively interested in attending to such coverage, it may be that there will be a predictive relationship with exposure variables as well.

Hyp. 1. Support for alcohol control policies will be predicted by a) exposure to television news and newspapers and b) attention to news coverage about accidents, injuries, and violent crime.

As noted above, prior analyses support the proposition that news media use is predictive of alcohol-related risk judgments and issue concern (Slater & Rasinski, 2005). It seems likely that being concerned about the issue of alcohol-related injuries is likely to lead to support for alcohol-control policies. If concern is influenced by news media use and in turn influences policy support, this suggests a mediating relationship:

Hyp. 2: Predictive relationships found between a) news media exposure or b) attention to relevant news story content and support for alcohol-related public policies will be mediated by issue concern.

In the present case, issue concern about alcohol-related risks is defined as a perception that alcohol-related risks are ranked relatively high among a person’s other concerns or worries. Such salience of a health or safety risk, in our view, is inherently affective in nature (see Dunwoody & Neuwirth, 1991). However, exemplars in the news might also operate more cognitively, by directly impacting risk judgments such as the extent to which accident and crime risks are regarded as due to alcohol (i.e., the perceived alcohol-attributable fraction) and the severity of the outcomes of such incidents.

People who attend to more news stories about car crashes, accidental injury, and assaults and other violent crime are more likely to be exposed to stories that identify alcohol use as playing a role in these human tragedies. It may be that the availability of such exemplars in memory (Zillmann & Brosius, 2000) would result in increased judgments of severity and alcohol attributable fractions as well as increasing salience or concern, and that such judgments might encourage support for public policies intended to reduce such risks:

Hyp. 3: Predictive relationships found between a) news media exposure or b) attention to relevant news story content and support for alcohol-related public policies will be mediated by judgments of severity of risk and judgments of the extent of alcohol as a cause of mortality due to crime and accidents.

We acknowledge that these hypotheses presume a simple mediating role for concern, risk severity, and the perceived alcohol-attributable fraction that may in fact be more causally complex. It is possible, for example, that the process is instead a primarily cognitive one in which increased exposure or attention to exemplars increases perceptions of harm severity and the alcohol-attributable fraction, which increases concern about the issue, leading to policy support. Or, the process may be more affective, in which increased exposure and attention to exemplars in stories leads to increased concern; increases in perceived risk and alcohol-attributable fraction would then result from such concern as cognitions align with a more visceral response:

RQ: Which of the proposed forms of the mediating relationships between media use, concern and risk judgments, and policy is best supported in this study?

Other predictors of alcohol control policy support

Use of appropriate control variables is necessary to interpret the effects of self-reported media exposure and attention to various forms of media content, as variables associated with outcome variables are also likely predictive of media use choices and differences in amount of attention to relevant media content. Latimer, Harwood, Newcomb, and Wagenaar (2001) found that gender, drinking frequency, and concern for youth were the strongest predictors of alcohol policy support, whereas age, parental status, job status, and ethnicity had more limited predictive ability. In other research, political ideology is associated with suchsupport (Wagenaar, Harwood, Toomey, Denk, & Zander, 2000). Given the predictive value of the factors described above, we include most of them as control variables in the present study (job status was excluded, as it is relatively awkward to measure, had a minor effect in prior research, and we had education and income available as proxies). Moreover, we include sensation seeking (the search for novel experiences and a willingness to take risks; Zuckerman, 1994) as a control variable, as sensation-seekers are less attentive to such news (Slater & Rasinski, 2005).

Methods

Participants and design

The study employed a nationally-representative, random-digit-dial telephone survey of 1,272 adult respondents age 18 and older. The survey was conducted by the National Opinion Research Corporation (see Slater and Rasinski [2005] for details on survey methodology, cognitive interviews and pretesting). Response rate 1 or the minimum response rate, which includes all refusals and unreachable numbers in the denominator, was 39% (AAPOR, 2007). The sample contained 523 males (41%), and 748 females (59%); 1027 (81%) were white, 116 (9%) were African American, with 10% other. Subjects ranged in age from 18 to over 64 years old with the average being 47.

Control variables

Demographic variables included age, income, gender, education, ethnicity, children living in the household, and political orientation. Sensation-seeking was measured using two items asking about liking to do exciting but risky things and willingness to take chances without worrying about the consequences, α = .70. A single item was used to assess alcohol consumption frequency, 1 = more than once a day, 8 = never, reverse-coded.

Media use variables

Newspaper exposure was assessed by averaging questions asking about the number of days the respondent read a newspaper in a) a typical week and b) in the past week (possible responses from 0 to 7, α = .96). Parallel items were used for viewing of national TV (α = .84) and local TV (α = .91). Means (SDs) for typical news exposure were 4.60 (2.05), 4.93 (1.87), and 3.49 (2.83) days per week for national TV, local TV, and newspapers, respectively.

We created a latent variable as a measure of attention to crime- and accident-related news. The latent variable was indicated by responses to three survey items asking about attention to news stories about motor vehicle accidents (mean = 2.86, SD = .87), crime in general (mean = 3.11, SD = .81), and violent crime (mean = 2.81, SD = .82) 1=none and 4=a lot, α = .83.

Dependent variables

Eleven items based in part on prior studies of alcohol control policy support (notably Wagenaar et al., 2000) were included. An exploratory factor analysis using a varimax rotation partitioned those eleven items into three non-overlapping factors, using an eigenvalue of 1 as a cut-off. These items were subsequently treated as indicators of latent variables in the context of a structural equation model. The three variables were labeled enforcement, limit, and liable. Enforcement refers to support for greater enforcement of existing laws (eigenvalue = 2.13, variance explained = 23.7%), limitations refers to support for limitations to alcohol advertising (eigenvalue = 2.47, variance explained = 27.5%), and liable refers to support for laws holding servers liable for damage done by intoxicated customers (eigenvalue = 2.61, variance explained = 29%).

We created an enforcement latent variable that was identified by three items: “forbid open alcohol container in the vehicle” (mean = 4.16, SD = 1.12), “forbid liquor stores, bars, or restaurants to sell alcohol to someone who is visibly intoxicated” (mean = 4.27, SD = 1.09, and “stricter enforcement of laws in your community that forbid liquor stores, bars, or restaurants to sell alcohol to youth under age 21” (mean = 4.27, SD = 1.15). Responses were on a five point Likert type scale anchored at “strongly oppose” and “strongly support”, α = .79.

An advertising limitations latent variable was indicated by “laws in your community that would ban alcohol-related billboards” (mean = 3.44, SD = 1.23), “federal laws banning TV ads for hard liquor” (mean = 3.46, SD = 1.23), and “federal laws banning TV ads for beer and wine” (mean = 3.24, SD = 1.23), α = .90.

Finally, a server liability latent variable was identified by “state laws making bars and restaurants liable for lawsuits if someone becomes intoxicated there and causes a motor vehicle crash after leaving” (mean = 3.15, SD = 1.31), “state laws making bars and restaurants liable if someone becomes intoxicated there and injures someone else in an attack after leaving” (mean = 3.10, SD = 1.30), and “state laws making bars and restaurants liable if someone becomes intoxicated there and causes a non-motor vehicle incident, such as a fire, a fall, a drowning or an electrocution after leaving” (mean = 3.00, SD = 1.28), α = .93.

Intervening variables

Intervening variables included concern, perceived severity, and perceived proportion of fatalities associated with alcohol use for crime, motor vehicle, and other accidents. To measure concern regarding alcohol-related assault, car accidents, and violent crime, we first asked: “On a scale of 1 to 10, where 1 means ‘just a minor concern’, and 10 means ‘this is one of my biggest concerns, how concerned are you about the problem of attacks [car crashes/accidents such as fall, drowning, and electrocution caused] by people who are under the influence of alcohol?” Means (SDs) were 5.8 (3.2), 8.1 (2.4), and 5.7 (3.1) for concern about assaults, motor vehicle crashes, and other injury incidents, respectively. We then asked: “Using the same scale, how concerned are you about such an attack [crash/accident] happening to you or someone you know?” Means (SDs) were 5.5 (3.3), 7.8 (2.7), and 5.5 (3.4) for concern about assaults, motor vehicle crashes, and other injury incidents, respectively.

These six items were subsequently used to identify the latent variable concern (α = .82), via use of a higher-order factor (a higher-order factor is appropriate when other latent variables are closely related and in fact may be considered indicators of a more fundamental latent variable). That is, we first created a two-item factor for concern regarding each type of injury and then extracted a higher-order factor for the latent overall concern construct. Perceptions regarding severe injury risks were likewise treated as a latent variable that was identified by one item for each of the three risks: “Now let’s change the scale. Here 1 will mean ‘hardly ever’ and 10 will mean ‘almost always.’ On a scale from 1 to 10, how often do [attacks/motor vehicle accidents/other accidents such as falls, fire, or drowning] in which alcohol is involved result in severe physical harm other than death?”, α = .84 for the scale. Means (SDs) were 6.2 (2.4), 7.1 (2.1), and 5.9 (2.4) for concern about assaults, motor vehicle crashes, and other injury incidents, respectively.

Whereas the focus of these items was on non-fatal injury, the items indicating the latent variable perceived alcohol-attributable fraction (AAF in alcohol epidemiology) focus on fatal injuries as a result of accidents and crime in which alcohol was a causative factor: “Think of all the cases of deadly [attacks/motor vehicle accidents/other accidents such as fires, falls, or drowning] each year in this country. What percentage of them do you think involved someone who was under the influence of alcohol?” (α = .68 for the three-item scale). Means (SDs) were 46.4 (24.7), 51.1 (21.9), and 36.8 (23.0) for AAF for assaults, motor vehicle crashes, and other injury incidents, respectively.

Data analysis

Post stratification weights were used to adjust sample estimates for unequal selection probabilities and for non-response, and were based on region of the country, gender, age, race, and education (Korn & Graubard, 1999). All control variables described above were evaluated for use in assessing the relationships of exposure and attention with the proposed mediators (concern, severity, and perceived alcohol-attributable fraction) and the three outcomes. All control variables except for sensation-seeking were single-item measures and were included as observed variables; sensation-seeking was indicated by two items and was included as a latent variable. Non-significant control variables were excluded from the models, as inclusion of such non-significant controls would artifactually reduce model fit. The statistically significant control variables were retained in the models; they are not illustrated in figures showing tests of mediation for clarity of presentation.

All parameter estimates were obtained using weighted least square parameter estimates. Those estimates were derived using a diagonal weight matrix with standard errors and mean- and variance adjusted χ2 test statistics that used a full weight matrix. The hypotheses were tested using freeware R (R Development Core Team, 2008). We assessed model fit adapting the guidelines of Hu and Bentler (1999). A model was deemed as having excellent fit if CFI was .95 or greater and lower bound of RMSEA was .05 or less, consistent with those guidelines. A model had acceptable fit if CFI was at or above .90 but below .95, and lower bound of RMSEA was at or below .09 but above .05.

To test Hypotheses 1a, we created three separate models predicting each of the three policy outcomes from the various exposure measures. We did the same for Hypotheses 1b using the latent variable for attention as the predictor. Hypotheses 2a, 2b, 3a, and 3b involved tests for mediation. Hypotheses were assessed only where significant predictive relationships were found between exposure or attention and policy outcomes, consistent with Baron and Kenny’s (1986) model for studying mediation. We evaluated the mediational effect using the method proposed by MacKinnon, Lockwood, and Hoffman (1998). The implementation of the test of our indirect path was based on the work of MacKinnon, Lockwood, Hoffman, West, and Sheets (2002). The method used is the product of the parameter estimates divided by their respective standard errors, and confidence intervals are determined using resampling methods (MacKinnon, Lockwood, & Williams, 2004; Preacher & Hayes, 2004).

Results

Hypothesis 1a and 1b Tests

Hypothesis 1a proposed that support for alcohol control policies would be predicted by news exposure. Models were created for each of the three policy outcomes (enforcement: CFI = .98, RMSEA = .06; liability: CFI = .99, RMSEA = .03; advertising limits: CFI = .99, RMSEA = .04). No significant effects were found; Hypothesis 1a was not supported.

Hypothesis 1b proposed that support for alcohol control policies would be predicted by attention to news coverage of crime and unintended injuries. The CFI for all models was larger than 0.95 and the lower bound of the confidence interval for the RMSEA estimates were below 0.05 (see Table 1). Attention to crime and unintended injury news predicted support for enforcement (β = .102, 95% CI = .015, .194) and for marketing limitations (β = .122, CI = .009, .210), consistent with Hypothesis 1b, but not for server liability (β = .202, CI = −.83, 1.34).

Table 1.

Model Fit Statistics for Hypothesis 1b: Crime/Accident News Attention Predicting Alcohol-control Policy Support

χ2 df p.value CFI RMSEA
Policy outcome:
Enforce laws 74.8 22 <.001 0.99 0.04
Server liability 115 24 <.001 0.98 0.05
Limit alc. ads 98 23 <.001 0.98 0.06
Open in a new tab

Hypothesis 2a and 2b Tests

Because our hypothesis about direct exposure effects on policy support were not supported, we did not further pursue Hypothesis 2a, regarding mediation of such effects.

Hypothesis 2b proposed that concern would mediate the effect of attention on policy support. Model fit statistics for the mediational models are displayed in Table 2. The indirect path estimates for enforcement (β = .043, 95%CI = .011, .072) and for limiting alcohol advertising (β = .078, 95%CI = .049, .110) were statistically significant, consistent with Hypothesis 2b. Figures 2a and 2b display models for enforcement and for advertising limits.

Table 2.

Model Fit Statistics for Hypothesis 2b Mediation Models: Concern Mediates Effects of Crime/Accident News Attention on Alcohol-Control Policy

Policy outcome χ2 df p.value CFI RMSEA
Enforce laws 607 83 <.001 0.95 0.07
Limit alc. ads 519 75 <.001 0.96 0.07
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Figure 2a and 2b. Hypothesis 3b models testing mediation of crime/accident attention by alcohol-attributable fraction for a) enforcement and b) alcohol advertising limitations. All paths shown are statistically significant, as is the indirect path (see text). Control variables are not illustrated but were incorporated in analyses, see text.

To further characterize the mediation effect, we note that the path from crime and accident news attention to support for enforcement of alcohol-control laws decreased from γ = 0.1024 (p = 0.006) to γ = 0.0879 (p = 0.063) when concern was in the model and the path from attention to support for alcohol marketing and advertising limits decreased from γ = 0.095 (p = 0.005) to γ = 0.036 (p = 0.384).

Hypothesis 3a and 3b Tests

Hypothesis 3a was also not pursued in the absence of a direct effect of exposure on policy support, as it posited mediation of such an effect.

Hypothesis 3b proposed that perceived severity and alcohol-attributable fraction would mediate effects of crime/accident news attention on alcohol-control policy support. We tested only those outcomes—enforcement of alcohol laws and limitations on alcohol marketing and advertising—for which a significant direct effect between attention and the outcome were obtained. Statistically significant mediation paths (p < 0.05) were found through perceived alcohol-attributable fraction to enforcement (β = .054, 95%CI = .030, .085) and limiting alcohol advertising (β = .051, 95% CI = .029, .077). However, the paths from perceived severity to the policy outcomes were not significant. Hypothesis 3b, then, was supported with respect to perceived alcohol-attributable fraction but not perceived severity. The models and the path coefficients are displayed in Figure 2a and 2b. Since the indirect paths that included perceived severity were not statistically significant, we fixed those paths to zero before evaluating the fit of our mediation model. Model fit statistics are displayed in Table 3.

Table 3.

Model Fit Statistics for Hypothesis 3b: Perceived Alcohol-attributable Fraction Mediates Effects of Crime/Accident News Attention on Policy Outcomes

Policy outcome χ2 df p.value CFI RMSEA
Enforce laws 263 49 <.001 0.96 0.07
Limit alc. ads 234 49 <.001 0.97 0.06
Open in a new tab

Note: the indirect paths including severity were removed from these models, as severity did not meet necessary criteria for serving as a mediator (i.e., did not predict the outcome variables).

To further characterize the mediation effect, we also examined the direct effects of attention in the presence of the mediator (perceived alcohol-attributable fraction). The path from attention to crime/accident news to support for alcohol-control enforcement decreased from γ = 0.102 (p = 0.006) to γ = 0.094 (p = 0.015); the path to support for advertising limitations decreased from γ = 0.122 (p = 0.005) to γ = 0.079 (p= 0.02).

RQ: Which causal ordering of the mediators is best supported in this study?

We considered only those outcomes for which mediation was possible (enforcement and advertising limitations) and only the mediation variables that were along statistically significant mediation paths (concern and perceived alcohol-attributable fraction). The first option we evaluated was perceived alcohol-attributable fraction and concern acting simultaneously on the mediation path. The second model placed the perceived alcohol-attributable fraction prior to concern, while the third reversed the ordering. Model fit statistics are displayed in Table 4.

Table 4.

Model Fit Statistics for RQ: What Causal Order of Mediators Bests Fits These Data in a Combined Model?

χ2 df p.value CFI RMSEA
Model 1 (simultaneous mediation) 993 126 <.001 0.92 0.074
Model 2 (perceived AAF prior) 1071 128 <.001 0.91 0.076
Model 3 (concern prior) 983 128 <.001 0.92 0.073
Open in a new tab

Note: As reported in text, there is a statistically significant superiority in fit for Model 3 over Model 1, the next best-fitting model.

The best fit was achieved by model 3 where concern preceded the perceived alcohol-attributable fraction on the mediation path. A chi-square difference test comparing the two best fitting models, model 1 and model 3, suggests that model 3 was a significant improvement (p = 0.007) in model fit to the data over model 1.

Discussion

Our results suggest that attention to crime and accident news coverage predict support for enforcement of alcohol-control laws and restrictions on alcohol advertising. Our mediation tests suggest that the mechanisms underlying these relationships are the impact of attention on concern about the issue as well as on the perceived contribution of alcohol to crime and unintended injury risks.

Our analyses supported a salience- or affect-driven model, in which impact on overall concern was causally prior to more cognitive assessments of risk. Apparently, the more attention is paid to stories about violent crime and accidents, the greater the concern about alcohol-related injuries due to assault and accidents. It appears that this increased concern leads to generating cognitively consistent perceptions about the percentage of such injuries that are due to alcohol use; these perceptions of alcohol’s contribution to these injuries in turn predict support for various alcohol control policies. The model with the poorest fit was the cognitively-driven model, in which attention to such stories leads to increased estimates of the extent to which alcohol is a causal factor in crime and accidental death, leading in turn to increased concern. However, the cross-sectional data and exploratory nature of our modeling must be acknowledged. Causal paths should be retested using longitudinal data sets to be asserted with greater confidence.

The evidence that attention to news stories about crime and accident influences alcohol-control policy support via concern (and less certainly, risk perceptions) is reasonably clear given the cross-sectional nature of this data. Unfortunately, most news stories about accident and crime do not address alcohol’s role (Slater et al., 2006). Enough stories, however, do address this role that a person who attends to such coverage is likely to increase their concern regarding the risks associated with alcohol use. Substantively, the limitation of impact to readers who are relatively attentive to such coverage should be noted. This may work against direct impact on behavior, as attention to such coverage is negatively predicted by alcohol use (Slater & Rasinski, 2005). However, politicians and policy-active members of the public do tend to be attentive consumers of the news media (Grunig, 1983), and are more prone to such influence, increasing the likelihood of news coverage impact on alcohol control policy.

At present, we make inferences about causal effects of news coverage based on associational results. We believe it is probable, in fact, that the relationship between concern or risk judgments and attention to crime and accident stories is probably mutually reinforcing, consistent with spiral models of media influence (Slater, 2007).

We have conducted a follow-up study to further assess our causal interpretation. Results indicate that, when participants (albeit students, not general population) were exposed to crime and accident stories reporting a causal role of alcohol, concern about alcohol-related risks is predicted by attentiveness to crime and accident news. This relation became non-significant when news stories did not report a role of alcohol (Slater, Goodall, & Hayes, in press). These experimental results are consistent with our interpretation of these cross-sectional data: that attentiveness to crime and accident news, by increasing the impact of alcohol-relevant risk exemplars, increases alcohol-related concern.

We therefore extend the position taken by researchers on media advocacy (Wallack & Dorfman, 2001) regarding the importance of increased thematic coverage of public health problems. We suggest that coverage of crime and accidents that more consistently informs the reader as to when alcohol is involved apparently increases public concern about the risks imposed by alcohol misuse, and is likely to strengthen support for alcohol control policies among those attentive to such news coverage. Efforts by media advocates might usefully extend beyond attempts to influence thematic coverage, to efforts to encourage reporting of breaking crime and injury news that more accurately reflects the possible causal role of alcohol in the incident.

Figure 1.

Figure 1

Open in a new tab

Figure 1a and 1b. Hypothesis 2b models testing mediation of crime/accident attention by concern for the a) enforcement and b) advertising limitation policy support outcomes. Paths to and from concern are statistically significant as are the indirect effects (see text); the direct effect from attention to enforcement or advertising limitation is not significant in the presence of the mediators. Control variables are not illustrated but were incorporated in analyses. Direct effects on policy support for liability laws absent mediation were not significant and therefore models for liability support are not illustrated here or in Figure 2.

Acknowledgments

This research was supported by grant AA10377 from the National Institute on Alcohol Abuse and Alcoholism to the first author.

Contributor Information

Michael D. Slater, The Ohio State University

Frank Lawrence, The Pennsylvania State University.

Maria Leonora G. Comello, The Ohio State University

References

  1. AAPOR. Standard definitions: Final dispositions of case codes and outcome rates for surveys. [accessed March 29, 2007];2007 http://www.aapor.org/pdfs/standarddefs_4.pdf.
  2. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology. 1986;51:1173–1182. doi: 10.1037//0022-3514.51.6.1173. [DOI] [PubMed] [Google Scholar]
  3. Combs B, Slovic P. Newspaper coverage of causes of death. Journalism Quarterly. 1979;56:837–843. [Google Scholar]
  4. Dunwoody S, Neuwirth K. Coming to terms with the impact of communication on scientific and technological risk judgments. In: Wilkins L, Patterson P, editors. Risky business: Communicating issues of science, risk and public policy. NY: Greenwood; 1991. pp. 11–30. [Google Scholar]
  5. Gerbner G, Gross L, Morgan M, Signorielli N, Shanahan J. Growing up with television: Cultivation processes. In: Bryant J, Zillmann D, editors. Media Effects: Advances in Theory and Research. Mahwah, NJ: Lawrence Erlbaum; 2002. pp. 43–68. [Google Scholar]
  6. Grunig J. Communication behaviors and the attitudes of environmental publics: Two studies. Journalism Monographs. 1983;81:1–47. [Google Scholar]
  7. Hargeaves I. The role of the media. In: Weale A, editor. Risk, democratic citizenship, and public policy. NY: Oxford University Press; 2002. pp. 49–57. [Google Scholar]
  8. Holbert RL, Stephenson MT. The importance of indirect effects in media effects research: Testing for mediation in structural equation modeling. Journal of Broadcasting & Electronic Media. 2003;47(4):556–572. [Google Scholar]
  9. Holder HD, Treno AJ. Media advocacy in community prevention: News as a means to advance policy change. Addiction. 1997 June;92(2):S189–S199. [PubMed] [Google Scholar]
  10. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling. 1999;6(1):1–55. [Google Scholar]
  11. Iyengar S. Is anyone responsible? How television frames political issues. Chicago, IL: University of Chicago Press; 1991. [Google Scholar]
  12. Kone D, Mullet E. Societal risk perception and media coverage. Risk Analysis. 1994;14:21–24. doi: 10.1111/j.1539-6924.1994.tb00024.x. [DOI] [PubMed] [Google Scholar]
  13. Korn EL, Graubard BI. Analysis of health surveys. New York: Wiley; 1999. [Google Scholar]
  14. Latimer WW, Harwood EM, Newcomb MD, Wagenaar AC. Sociodemographic and individual predictors of alcohol policy attitudes: Results from a US probability sample. Alcoholism-Clinical and Experimental Research. 2001;25(4):549–556. [PubMed] [Google Scholar]
  15. Lichtenstein S, Slovic P, Fischhoff B, Layman M, Combs B. Judged frequency of lethal events. Journal of Experimental Psychology: Human Learning and Memory. 1978;4:551–578. [PubMed] [Google Scholar]
  16. MacKinnon DP, Lockwood CM, Hoffman JM. A new method to test for mediation. Annual meeting at the Society for Prevention Research; Park City, UT. 1998. [Google Scholar]
  17. MacKinnon DP, Lockwood CM, Williams J. Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research. 2004;39:99–128. doi: 10.1207/s15327906mbr3901_4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. MacKinnon DP, Lockwood CM, Hoffman JM, West SG, Sheets V. A comparison of methods to test mediation and other intervening variable effects. Psychological Methods. 2002;7:83–104. doi: 10.1037/1082-989x.7.1.83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. McCombs M, Shaw D. The agenda setting function of mass media. In: Protess DL, McCombs M, editors. Agenda setting: Readings on media, public opinion, and policymaking. Hillsdale, New Jersey: Lawrence Erlbaum Associates; 1991. pp. 17–26. [Google Scholar]
  20. Montgomery KC. The Harvard Alcohol Project: Promoting the designated driver on television. In: Backer TE, Rogers EM, editors. Organizational aspects of health communication campaigns: What works. Newbury Park, CA: Sage; 1993. pp. 178–202. [Google Scholar]
  21. Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers. 2004;36(4):717–731. doi: 10.3758/bf03206553. [DOI] [PubMed] [Google Scholar]
  22. R Development Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2008. URL http://www.R-project.org. [Google Scholar]
  23. Slater MD, Goodall CE, Hayes AF. Self-reported news attention does assess differential processing of media content: An experiment on risk perceptions utilizing a random sample of U.S. local crime and accident news. Journal of Communication (In press) [Google Scholar]
  24. Slater MD. Reinforcing spirals: The mutual influence of media selectivity and media effects and their impact on individual behavior and social identity. Communication Theory. 2007;17:281–303. [Google Scholar]
  25. Slater MD, Long MA, Ford VL. Alcohol and illegal drugs, violent crime, and accidents in U.S. local and national news. Journal of Studies on Alcohol. 2006;67:910–914. doi: 10.15288/jsa.2006.67.904. [DOI] [PubMed] [Google Scholar]
  26. Slater MD, Rasinski K. Media exposure and attention as mediators of exogenous influences on social risk judgments. Journal of Communication. 2005;55:810–827. [Google Scholar]
  27. Toomey TL, Wagenaar AC. Policy options for prevention: The case of alcohol. Journal of Public Health Policy. 1999;20(2):192–213. [PubMed] [Google Scholar]
  28. Viswanath K, Breen N, Meissner H, Moser RP, Hesse B, Steele WR, Rakowski W. Cancer knowledge and disparities in the information age. Journal of Health Communication. 2006;11:1–17. doi: 10.1080/10810730600637426. [DOI] [PubMed] [Google Scholar]
  29. Wagenaar AC, Harwood EM, Toomey TL, Denk CE, Zander KM. Public opinion on alcohol policies in the United States: Results from a national survey. Journal of Public Health Policy. 2000;21(3):303–327. [PubMed] [Google Scholar]
  30. Wallack L, Dorfman L. Putting policy into health communication: The role of media advocacy. In: Rice RE, Atkin CK, editors. Public communication campaigns. Thousand Oaks, CA: Sage Publications; 2001. pp. 389–401. [Google Scholar]
  31. Wolfson M, Toomey TL, Forster JL, Wagenaar AC, McGovern PG, Perry CL. Characteristics, policies and practices of alcohol outlets and sales to underage persons. Journal of Studies on Alcohol. 1996;57(6):670–674. doi: 10.15288/jsa.1996.57.670. [DOI] [PubMed] [Google Scholar]
  32. Yanovitzky I. Effect of news coverage on the prevalence of drunk-driving behavior: Evidence from a longitudinal study. Journal of Studies on Alcohol. 2002;63:342–351. doi: 10.15288/jsa.2002.63.342. [DOI] [PubMed] [Google Scholar]
  33. Yanovitzky I, Bennett C. Media attention, institutional response, and health behavior change: The case of drunk driving, 1978–1996. Communication Research. 1999;26:429–453. [Google Scholar]
  34. Zillmann D, Brosius HB. Exemplification in mass communication: The influence of case reports on the perception of issues. Mahwah, NJ: Erlbaum; 2000. [Google Scholar]
  35. Zuckerman M. Behavioral expressions and biosocial bases of sensation seeking. New York: Cambridge University Press; 1994. [Google Scholar]

RESOURCES