How Industries Use Direct-to-Public Persuasion in Policy Conflicts: Asymmetries in Public Voting Responses

Abstract

Industries use persuasion strategies to gain public support when challenged by activist groups on consumer-relevant issues. This marketing practice, termed “direct-to-public persuasion,” has received limited attention in the field, and thus we have little understanding of when such campaigns fail or succeed. This research introduces a theoretically derived and empirically supported framework that draws from multiple areas, including marketing persuasion, political campaign strategy, sociopolitical legitimacy, and perceptual fit, to identify important differences in the effectiveness of these persuasion strategies on attitudes and voting behavior. The multimethod approach includes a field study of ballot measure voting during a national U.S. election and three experimental studies. The findings contribute new knowledge of asymmetries in public response to industry and activist arguments. Stronger arguments from both sides lead to more favorable outcomes, but activist groups benefit most. Suspicion of activist arguments weakens the impact on attitudes and voting; industry argument suspicion has limited impact, though it does increase the likelihood of voter switching. A financial argumentation strategy works best for the industry side, while societal argumentation is more effective for the activist side. The insights offer guidance for industries and activist groups as argument strategy success is contingent on the side that uses it.

Keywords

direct-to-public persuasion persuasion political campaign strategy political marketing sociopolitical legitimacy voting behavior

Industries often try to persuade the public at large to support their positions on prominent issues that are consequential for consumers and industry members. We term this phenomenon “direct-to-public persuasion,” a marketing-driven political influence strategy used to convince the public that the industry is on the correct side of an issue. Industries “talk” to the public not only to sustain or gain advantageous conditions, such as preventing policy actions viewed as costly or restrictive, but also to align the public with practices that further industry objectives.

Industries use direct-to-public persuasion to influence public voting on ballot measures in state and local referendums, public opinion in advance of legislative policy maker decisions, and ongoing public attitudes toward the industry. We focus on conflict scenarios in which industries are challenged by activist groups—collectives that advocate for the public interest—in ballot measures (e.g., Waterhouse 2014). Such conflicts are commonly high-profile and a strategic priority for industries, which spend substantial funds to secure favorable outcomes and protect industry practices (Maisel and Berry 2010; Sheingate 2016). The outcomes often have meaningful consequences for the public, including consumers, but the question of what drives voters to support one side or the other remains unanswered.

Table 1 presents examples of how industries and their activist opponents use direct-to-public persuasion to influence the public and gain support. The examples represent diverse issues, including food labeling, recycling, pharmaceutical drug pricing, and tobacco taxes. In the scenarios we examine in this article, industries defend the policy status quo to resist changing disputed practices (e.g., price limits on pharmaceuticals). However, in other settings, industries fight the status quo to expand or develop new markets.¹

Table 1.

Examples of Direct-to-Public Persuasion in Issue Referendums.

Issue, Location, Year	Overview of Activist (Vote Yes) Side on Issue and Industry (Vote No) Side on Issue	Top Donors	Total Campaign Contributions	Example of Financial Argumentation Strategy	Example of Societal Argumentation Strategy	Result (% of Votes)
Renewable Energy Standards (Proposition 127) Arizona 2018	A “yes” vote supported requiring electric utilities in Arizona to acquire a certain percentage of electricity from renewable resources each year, with the percentage increasing annually from 12% in 2020 to 50% in 2030.	NextGen Climate Action, League of Conservation Voters	$24.1 million	The price of renewable energy continues to drop every day and is already competitive with coal and gas.	Cleaner air and water means healthier families—and less respiratory illnesses like asthma.	31.4%
Renewable Energy Standards (Proposition 127) Arizona 2018	A “no” vote opposed requiring electric utilities in Arizona to acquire a certain percentage of electricity from renewable resources, thereby leaving in place the state’s existing renewable energy requirements of 15% by 2025.	Pinnacle West Capital Corporation, Grand Canyon State Electric	$40.9 million	Corporate and industrial rates would rise over 100%, and residential ratepayers would see an average annual increase of $1,250.	The measure would dramatically harm Arizona’s competitiveness and put our utilities’ reliable delivery of power at risk.	68.6%
Drug Price Standards (Proposition 61) California 2016	A “yes” vote supported regulating drug prices by requiring state agencies to pay no more than the U.S. Department of Veterans Affairs (VA) pays for prescription drugs.	AIDS Healthcare Foundation, California Nurses Association	$19.1 million	The proposition would save taxpayers billions of dollars in health care costs.	The proposition would provide better access to life-saving drugs.	46.8%
Drug Price Standards (Proposition 61) California 2016	A “no” vote opposed this regulation to require state agencies to pay no more than the VA pays for prescription drugs.	Pharmaceutical Research and Manufacturers of America, Merck & Co. Inc., Pfizer Inc.	$109.1 million	The proposition would increase state prescription drug costs.	The proposition would reduce patient access to medicines.	53.2%
Tobacco Tax Increase (Proposition 56) California 2016	A “yes” vote favored increasing the cigarette tax by $2 per pack, with equivalent increases on other tobacco products and electronic cigarettes	Tom Steyer, California Hospitals Committee on Issues, Million Voter Project Action Fund	$35.5 million	The proposition would reduce tobacco-related health care costs and would help pay for those costs.	The proposition would prevent youth smoking and address tobacco marketing targeting a youth market.	64.4%
Tobacco Tax Increase (Proposition 56) California 2016	A “no” vote opposed increasing the cigarette tax by $2 per pack, with equivalent increases on other tobacco products and electronic cigarettes.	Philip Morris USA, R.J. Reynolds Tobacco Company, Altria Client Services LLC	$71.0 million	The proposition would fund insurance companies and special interests more than it would fund treatments for smoking-related illnesses.	The proposition will not solve real problems like solving the drought and fighting violent crime.	35.6%
Expansion of Bottle Deposits (Question 2) Massachusetts 2014	A “yes” vote would expand the state’s beverage container deposit law to require deposits on all nonalcoholic drink containers, except beverages derived from dairy, infant formula, or medications.	Massachusetts Sierra Club, Elm Action Fund, Massachusetts Public Interest Research Group	$1.6 million	A yes vote equals big savings for towns’ waste management costs.	A yes vote equals more recycling and less trash and litter.	26.6%
	A “no” vote would keep the state’s beverage container deposit law in its current form.	American Beverage Association, Nestle Waters NA, Stop & Shop Supermarket Company	$9.6 million	The measure would cost nearly $60 million a year, more than three times the price of curbside programs.	Massachusetts should be a recycling leader, but Question 2 will keep us in the past.	73.4%

Notes: Information on ballot measures retrieved from https://ballotpedia.org/Ballot_Measures_overview.

Industries have resource advantages relative to activist opponents that suggest they should win policy issue battles, and they often do (Schlozman, Verba, and Brady 2012; Walker 2014). Because industries are often in the advantageous role of defending the policy status quo, they can leverage the uncertainty of any policy change, whereas activist groups often have the more difficult task of convincing the public that the status quo is failing (Drutman 2015). Despite their advantages, industries do not always prevail, which may be due to an underlying predisposition of the public to be more skeptical of the industry and its motives (Berry and Wilcox 2018; McKinsey 2021; Trumbull 2012). The fact that industries do frequently win suggests that they succeed in overcoming the public’s skepticism with effective persuasion arguments (McFadden and Lusk 2013). Our research examines the dynamics of competing campaigns to better understand and explain voting outcomes.

The industry-level practice of direct-to-public persuasion has received limited attention in the marketing field, and we know little about persuasion involving political campaigns with competing positions on an issue. The marketing literature has examined consumer responses to persuasion attempts in a variety of important contexts, and advertising research has examined aspects of argument effectiveness (e.g., Meyers-Levy and Malaviya 1999; Wan et al. 2010), but political arguments are believed to differ from commercial arguments (Gordon et al. 2012). Political marketing research has largely focused on candidate elections rather than election voting on policy issues (see Jung and Crichter 2018; Klein and Ahluwalia 2005), and the lobbying literature addresses influencing policy makers rather than influencing the public (see Martin et al. 2018; Waterhouse 2014). Further, although studies show that voters often switch their support during campaigns, and even a nominal shift can change the outcome (McFadden and Lusk 2013; Phillips, Urbany, and Reynolds 2008), knowledge of what causes the switching is limited.

To expand knowledge of this understudied topic, we develop a framework that draws from multiple research areas, including marketing persuasion, political campaign strategy, sociopolitical legitimacy, and perceptual fit. We apply legitimacy theory to explain our predicted asymmetry of argument effectiveness across the competing sides and draw on perceptual fit theory to explain the efficacy of specific argumentation strategies on the public’s attitudes and voting behavior. We use a multimethod approach that includes a field study of ballot measure voting on two policy issues during a national U.S. election and three experimental studies involving additional issues to test our ideas. The policy issues included prescription drug pricing, tobacco product taxes, renewable energy standards, and container recycling laws.

Our findings offer insight into how the public’s attitudes and voting intentions change during the campaign when they are exposed to competing arguments from industry and activist groups. First, we uncover important asymmetries that should be considered in understanding and in using direct-to-public persuasion strategies. Specifically, stronger arguments from both industry and activist groups lead to more favorable attitudes and voting outcomes, but activist groups benefit more from stronger arguments. Greater argument suspicion weakens the impact on attitudes and voting, but primarily for activist groups. Therefore, overall, voters’ evaluations of activist-side arguments have a greater impact than industry-side arguments on outcomes. Second, an exploration of vote switching found that industry argument suspicion has limited impact, though it does increase the likelihood of voter switching, particularly for voters who switch their support to the activist side. Finally, using follow-up experiments, we show that a financially focused argumentation strategy works best for the industry side, whereas a societally focused strategy is more effective for the activist side.

We contribute to marketing research with new knowledge of a marketing practice that influences voters and critical political outcomes. We add to the scope of persuasion research with insight into how voters respond to industry versus activist political campaigns, finding novel evidence of an asymmetric public response, where voter judgments of competing arguments—and the degree to which those arguments are strong or suspicious—help predict the relative impact of each side’s campaign. We use sociopolitical legitimacy theory in a new way, proposing that an industry-activist legitimacy gap helps explain why industry argument strength has less impact, and apply persuasion theory to explain why voters give greater weight to suspicion of activist arguments. Our pre- and postelection recontact field study design captures individual voter switching, which is not commonly examined, and provides a rare indication that determinants differ for voter segments that flipped their allegiance from one side to the other. Finally, we offer the first known concrete guidance for the use of direct-to-public persuasion strategies with recommendations that differ for industries and activist groups, as we find that a given strategy’s success is contingent on the side that uses it.

Direct-to-Public Persuasion Model and Hypothesis Development

Figure 1 depicts the conceptual relationships we investigate. Our model predicts that the public’s evaluations of industry- and activist-side arguments, in terms of argument strength and argument suspicion, play a central role in predicting outcomes including attitudes toward the issue, voting on the issue, and vote switching. These evaluations depend on each side’s argumentation strategy. We begin by describing the two competing sides in the issue conflict scenarios we study and key distinctions between them, and then we develop the hypothesized asymmetric effects of argument strength and suspicion on the voting outcomes.

Figure 1.

Conceptual model.

Industry and Activist Conflicts over Issues

We define “competing sides on the issue” as the competitors in a policy issue conflict and focus on industry groups and activist groups, traditional policy adversaries (Berry and Wilcox 2018). While industries comprise firms that produce a certain class of goods or services, public interest activist groups are political constituents working in a range of policy areas who claim to represent the public or collective good (e.g., Scott 2018; Waterhouse 2014). Evidence suggests that the public responds differently to the industry and activist sides because of enduring skepticism about the undue influence of business interests (Gallup 2019; Pew Research Center 2019). While attitudes toward individual industries vary, public favorability ratings of business have trended downward for decades and eroded further since the 2008 financial crisis (ANES 2020; Werner 2015).

A legitimacy theory perspective indicates that response to direct-to-public persuasion from the industry and activist sides should be asymmetric due to a legitimacy gap caused by the belief that industry-serving interests drive industry-side motives whereas commitment to the public good drives activist-side motives (Schlozman, Verba, and Brady 2012). Normative and sociopolitical legitimacy underlie the public’s judgment of whether an industry practice is acceptable—meeting cultural and political norms—or should be sanctioned (e.g., Lamin and Zaheer 2012; Suddaby, Bitektine, and Haack 2017). Legitimacy judgments are likely to discourage many voters from initially supporting the industry side, but election polling and outcomes suggest that voters come to judge the industry more favorably during the course of a campaign. For example, in recent ballot measures that sought genetically modified organism labeling on food products (Oregon Measure 92 in 2014), regulations on dialysis clinic pricing (California Proposition 8 in 2018), and expansion of bottle deposit laws (Massachusetts Question 2 in 2014), the majority initially supported the activist side but switched to support the industry position, allowing the industry to win the vote (see https://ballotpedia.org/List_of_ballot_measures_by_year).

The Influence of Competing Persuasion Campaign Arguments

Argumentation is defined as a strategic approach used to justify a particular policy or political position and promote or challenge its implications (Baumgartner et al. 2009; Dunn 2012). Effective argumentation is critical for industries battling controversy and for activist groups that confront them, as both strive to legitimize their positions (Eesley, DeCelles, and Lenox 2016; Klein, Smith, and John 2004). We explore the effectiveness of each side’s argumentation on two key dimensions: argument strength and argument suspicion. Research on argument strength, including studies in competitive political contexts (e.g., Chong and Druckman 2010), has not concurrently examined the effects of argument suspicion.

We define “argument strength” as perceived persuasiveness determined by an individual’s cognitive response, either favorable or unfavorable, to an argument (Areni and Lutz 1988; Zhao et al. 2011). Empirical findings show that strong arguments consistently shift attitudes and beliefs but to varying degrees (e.g., Arceneaux 2012). Although argument strength has had limited study in real-world scenarios of competing arguments over time, some research reports that in the presence of a strong argument from one side, a weaker argument from the opposing side is viewed even more negatively and can actually backfire (Chong and Druckman 2010). As baseline predictions, we hypothesize:

H₁: Industry argument strength increases (a) favorable attitudes toward and (b) voting in favor of the industry side on the issue.

H₂: Activist argument strength increases (a) favorable attitudes toward and (b) voting in favor of the activist side on the issue.

We define “argument suspicion” as the perception that an argument is implausible or has a hidden intent (e.g., Campbell and Kirmani 2000; Fransen, Smit, and Verlegh 2015). Although there is far less theoretical and empirical knowledge on the influence of argument suspicion than argument strength, marketing research on persuasion knowledge and skepticism toward advertising offers rich evidence that suspicion-driven responses to persuasion can have significant negative effects on attitudes and behaviors.

Consumers draw on persuasion knowledge, beliefs about how marketers influence consumers, to determine whether something is a persuasion attempt (e.g., Campbell and Kirmani 2000). In our research scenario, arguments have clear intent to persuade to vote no or vote yes, but whether the arguments’ tactics are viewed as appropriate impacts their effectiveness (e.g., Isaac and Grayson 2017). In some cases, voters may infer finer-level motives in an argument’s tactics that trigger suspicion (e.g., Is the “vote no” side using arguments that demonize the “vote yes” side because it lacks more valid arguments?; Kirmani and Campbell 2009). Research on skepticism toward advertising messages has examined dispositional ad skepticism (e.g., Obermiller, Spangenberg, and MacLachlan 2005) as well as situational skepticism (e.g., Forehand and Grier 2003). Marketing studies suggest that when voters respond to an industry’s arguments with skepticism, they are likely to scrutinize its underlying intentions, suspicious that they are deceptive as well as self-serving (see, e.g., Forehand and Grier 2003).

While this suggests that the public is inclined to view industry arguments as suspicious, other research suggests that activist arguments can also face public doubt, such as when they appear to exaggerate the problems and risks related to contested issues (Baur and Palazzo 2011; Trumbull 2012). As such, we expect voter evaluations of argument suspicion to be relevant for both sides on the issue. We predict the following:

H₃: Industry argument suspicion decreases (a) favorable attitudes toward and (b) voting in favor of the industry side on the issue.

H₄: Activist argument suspicion decreases (a) favorable attitudes toward and (b) voting in favor of the activist side on the issue.

Asymmetric Response to Competing Industry and Activist Arguments

The impact of argument strength

While we predict that argument strength will have a favorable effect for both the industry and activist sides, there is some evidence suggesting that effects are not equally favorable. We suggest, based on the difference in sociopolitical legitimacy (Bitektine 2011; Suddaby, Gustavson, and Bitektine 2017), that strong industry arguments will have a weaker effect relative to strong activist arguments. Due to their negative views of industry practices, voters’ legitimacy judgments may trigger doubt of the underlying intent of industry arguments. As a result, voters may not be as influenced even by strong arguments to support the industry’s side as by strong arguments to support the activist side. For example, voters assessing a proposition to cap drug prices (e.g., Proposition 61 in Table 1) may mistrust the motives underlying pharmaceutical industry pricing practices, which could diminish the impact of a strong industry argument. In contrast, because voters are likely to have more positive legitimacy judgments of activist groups (e.g., the AIDS Healthcare Association, which works to make life-saving drugs more affordable), the strength of their arguments will not be diminished. This type of voter response is consistent with research showing that when a prior ad triggers doubts about that ad’s intent, the effects of argument strength in subsequent ads are diminished (Darke and Ritchie 2007). We hypothesize:

H₅: Industry argument strength has a weaker positive effect than activist argument strength on (a) attitudes toward the issue and (b) voting on the issue.

The impact of argument suspicion

Voters’ responses to influence attempts reflect their persuasion beliefs, and they are likely to be far more guarded against industry arguments than activist arguments (see Kirmani and Campbell 2009). This initial disadvantage for the industry side may ultimately offer it an opportunity to use campaign tactics that lead voters to put more weight on activist argument suspicion and/or less weight on industry argument suspicion.

A method used by industries to increase voter suspicion of activist opponents is to intensify the public’s doubt about activist-supported policy changes. Industries often argue that activist-side policy positions are impractical, ill-conceived, unrealistic, and costly (Baumgartner et al. 2009; Hiatt, Grandy, and Lee 2015). For instance, industry arguments might warn of hidden complexities or expenses, implying that activist arguments are not as genuine and aboveboard as they may seem (Drutman 2015). The tactic of using cogent, rational challenges to activist arguments is likely viewed as credible, as it seems cautious and shows due diligence. With this approach, the industry can heighten voters’ formerly low persuasion knowledge of the activist side, increasing the weight of activist argument suspicion in voting decisions. This aligns with findings that previously trusted persuasion claims are further scrutinized when they are revealed to have hidden intentions (e.g., Campbell and Kirmani 2000), particularly in competitive contexts such as ours (e.g., Chong and Druckman 2010).

McFadden and Lusk’s (2013) study of a 2012 California ballot measure (Proposition 37) to require labeling genetically engineered foods supports this view. They found that industry advertising helped flip voter support for the activist side (around 70% less than four weeks prior to the election) to majority support for the industry. Industry advertising focused on loopholes that would make implementation difficult and costly, while activist advertising (which lost traction over time) focused on the ruthlessness and deceit of the large industries involved. Although research in ballot measure, referendum, and other issue-focused election scenarios is limited, similar tactics are described in other sources (e.g., Hussain 2020; Richards, Allender and Di Fang 2013). Following this logic and evidence, we hypothesize:

H₆: Industry argument suspicion has a weaker negative effect relative to activist argument suspicion on (a) attitudes toward the issue and (b) voting on the issue.

Exploring Responses of Vote-Switching Segments

Direct-to-public persuasion is most effective when it actually switches voter support from one side to the other. While brand switching is well examined, less is known about vote switching, when persuasion campaigns are short in duration and have a common deadline (Shachar 2009). An analysis of vote switching is needed to show whether voting outcomes are due to each side’s arguments solidifying voters’ initial support or causing them to switch their support over the course of a campaign (Klein and Ahluwalia 2005).

There is limited guidance on whether the response asymmetry our model predicts would hold for voters who switched their support compared with those who did not or whether the drivers of switching could differ for those initially supporting the industry side compared with those initially supporting the activist side. Swings in public support are well documented by public opinion research, yet there is limited knowledge of how competing campaign arguments motivate voters to switch allegiance. Because research provides little insight on which to base predictions, we do not formally hypothesize effects on switching behavior but rather allow the empirical results to address the knowledge gaps.

Field Study of Ballot Measures

To test our hypotheses, we conducted a large field study examining attitudes toward and voting on two California ballot measures involving distinct issues from different industries: Proposition 56, which proposed a tobacco tax increase, and Proposition 61, which proposed price limits on prescription drugs purchased for state programs (for examples of campaign media, see Web Appendix W1). These propositions featured highly salient, consumer-relevant issues in which the industries invested significantly in direct-to-public persuasion campaigns to defeat the measures. By investigating different issues and industries, we reduce the potential that our analyses capture idiosyncratic factors particular to one issue or one industry.

Methodology

We examined the impact of industry and activist arguments on attitudes and voting using survey data we collected using a national consumer research panel of California voters in two waves: five weeks before (measuring respondent characteristics, preelection attitudes, and voting intentions) and one week after (measuring argument strength, argument suspicion, postelection attitudes, and self-reported voting) the November 2016 national election. We collected data for the same respondents in both waves to analyze the extent to which industry- and activist-side persuasion campaigns shifted individual voters’ attitudes and voting.

Sample

We recruited respondents through Qualtrics, an online market research panel aggregator whose quality certification includes checking each person’s IP address to exclude duplication and replacing those who finish the survey in less than one-third of the average completion time. Respondents were registered voters in California who indicated “I will vote” in the election in a screening question at the beginning of the preelection survey. After the election, we invited all 1,806 respondents of the preelection survey to answer the postelection survey. Of the 904 postelection survey respondents who voted in the election,² 58.4% were female, the mean age was 51 years old, 92.1% had some college education or higher, and 70.7% reported their socioeconomic status (SES) as middle to upper class.

Measures

In the preelection survey, we captured preelection attitudes toward the issue in the context of the drug price and tobacco tax propositions using a three-item (“wrong/right,” “harmful/beneficial,” and “unacceptable/acceptable”), seven-point, previously validated scale and voting intention on the issue for both propositions as a binary variable (0 = intend to vote “no,” in favor of the industry side; 1 = intend to vote “yes,” in favor of the activist side). In addition, we measured several individual-level characteristics, including political affiliation, need for orientation, and demographic characteristics (age, gender, education, and SES).

In the postelection survey, we measured our two dependent variables, attitudes toward the issue and voting on the issue, using the same multi-item scale used to capture preelection attitudes and a binary variable (0 = voted “no,” 1 = voted “yes”), respectively. Respondents were then presented with two principal arguments from both the industry and activist sides for both propositions. To emphasize external validity, we identified the principal arguments from the California Official Voter Information Guide published by the office of the California Secretary of State (2016), which includes a set of Official Arguments formally submitted by each side (see Web Appendix W2). Respondents were asked to indicate their perceptions of argument strength and argument suspicion for each pair of industry arguments and activist arguments. We measured the two variables with two-item, seven-point, previously validated scales measuring the extent to which each of the arguments was convincing and aroused suspicion.³ Finally, we measured familiarity with the arguments. For all multi-item scales, the mean of the items was used in the model estimation. Table 2 presents the descriptive statistics for all variables, and measurement details are provided in Web Appendix W2.

Table 2.

Descriptive Statistics for Field Study Variables.

Variables	Drug Price Proposition M (SD)	Tobacco Tax Proposition M (SD)	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Attitudes toward issue	4.25 (2.09)	5.10 (2.10)	1	−.408	.753	.006	−.561	.724	−.240	.162	−.224	.064	−.126	.100	.118	.010	−.865	.672
2. Industry argument strength	4.02 (1.78)	3.64 (1.80)	−.448	1	−.257	−.157	.406	−.347	.394	−.009	.099	.027	−.031	−.071	−.048	.027	.385	−.345
3. Activist argument strength	4.41 (1.85)	4.54 (1.93)	.739	−.332	1	.109	−.546	.628	−.147	.319	−.205	.161	−.150	.040	.117	−.011	−.707	.555
4. Industry argument suspicion	3.96 (1.73)	4.16 (1.79)	.220	−.255	.289	1	.045	.079	.062	.025	−.045	−.020	.008	.041	−.022	−.073	−.078	.072
5. Activist argument suspicion	3.84 (1.90)	3.51 (2.01)	−.535	.478	−.488	−.006	1	−.438	.228	−.068	.150	−.040	.001	−.023	−.085	.000	.524	−.419
6. Preelection attitudes toward issue	4.48 (1.86)	4.94 (2.11)	.448	−.272	.433	.184	−.273	1	−.215	.091	−.218	.088	−.120	.112	.134	.012	−.708	.832
7. Familiarity with industry arguments	4.88 (1.71)	4.73 (1.79)	−.178	.401	−.091	−.007	.237	−.112	1	.378	.046	.120	.138	−.004	.017	−.118	.268	−.261
8. Familiarity with activist arguments	5.24 (1.52)	5.46 (1.53)	.255	−.042	.400	.110	−.081	.087	.359	1	−.018	.208	.060	.002	.010	−.028	−.086	.040
9. Political affiliation	3.31 (2.15)	3.30 (2.14)	−.240	.140	−.273	−.139	.181	−.244	.037	−.115	1	−.097	.083	−.111	.011	−.102	.224	−.182
10. Need for orientation	5.32 (1.42)	4.41 (1.78)	.057	.025	.106	.059	−.001	.176	.128	.226	−.057	1	−.010	−.008	−.011	−.037	−.053	.030
11. Age	51.33 (15.60)	51.02 (15.70)	−.109	−.067	−.156	−.037	.020	−.134	.060	.055	.069	.115	1	.003	.066	−.054	.122	−.134
12. Education	3.23 (.58)	3.23 (.58)	.044	.025	−.032	.023	.028	.038	−.007	−.038	−.109	−.040	−.001	1	.321	−.078	−.087	.107
13. SES	2.83 (.88)	2.85 (.88)	.020	.043	−.028	−.008	.013	.005	.042	−.014	.026	−.064	.083	.305	1	−.083	−.145	.129
14. Gender	M = 41.6%, F = 58.4%	M = 41.6%, F = 58.4%	−.032	−.018	.023	−.028	−.064	−.035	−.093	−.014	−.088	.066	−.052	−.092	−.087	1	.026	.006
15. Voting on the issue	Y = 52.3% N = 47.7%	Y = 68.2% N = 31.8%	−.789	.437	−.660	−.191	.482	−.381	.196	−.207	.249	−.057	.124	−.070	.019	.026	1	−.704
16. Preelection voting intention on issue	Y = 62.0% N = 38.0%	Y = 67.2% N = 32.8%	.381	−.248	.360	.130	−.274	.722	−.099	.059	−.211	.135	−.188	−.027	−.036	−.021	−.394	1

Notes: Correlation coefficients below diagonal relate to drug price proposition (>|.067| are significant at p < .05; N = 851). Correlation coefficients above diagonal relate to tobacco tax proposition (>|.066| are significant at p < .05; N = 871).SES = socioeconomic status.

Model

We tested our hypotheses by estimating attitudes toward and voting on the issue as a function of argument strength and argument suspicion. For postelection attitudes toward the issue (Attitudes_t,i), we used an ordinary least squares (OLS) approach, specified as follows:

\begin{aligned} Attitude s_{t, i} & = β_{0} + β_{1} Industry_Argument_Strengt h_{t, i} \\ + β_{2} Activist_Argument_Strengt h_{t, i} \\ + β_{3} Industry_Argument_Suspicio n_{t, i} \\ + β_{4} Activist_Argument_Suspicio n_{t, i} \\ + β_{5} Attitude s_{t - 1, i} + β_{6} Familiar_Industr y_{t, i} \\ + β_{7} Familiar_Activis t_{t, i} + β_{8} Political_Affiliatio n_{i} \\ + β_{9} Orientatio n_{i} + β_{10} Ag e_{i} + β_{11} Gende r_{i} \\ + β_{12} Educatio n_{i} + β_{13} SE S_{i} + ε_{i}, \end{aligned}

where the key predictor variables are industry argument strength (Industry_Argument_Strength_t,i), activist argument strength (Activist_Argument_Strength_t,i), industry argument suspicion (Industry_Argument_Suspicion_t,i), and activist argument suspicion (Activist_Argument_Suspicion_t,i). Covariates include preelection attitudes toward the issue (Attitudes_{t − 1,i}), familiarity with industry and activist arguments (Familiar_Industry_t,i, Familiar_Activist_t,i), political affiliation (Political_Affiliation_i), need for orientation (Orientation_i), age (Age_i), gender (Gender_i), education (Education_i), and socioeconomic status (SES_i). For our voting on the issue outcome (Vote_t,i), we used a logit model with the same predictors and covariates as the attitudes toward the issue model, substituting preelection voting intentions (Vote_{t − 1,i}) in place of preelection attitudes.

Our analysis was subject to several forms of bias, which we wanted to resolve. First, because we measured our dependent variables and key predictor variables using the same instrument, we adopted Lindell and Whitney’s (2001) approach to test for common method variance using a method factor capturing need for orientation on an unrelated ballot measure included in the survey instrument. Correlation coefficients corrected for common method variance indicate that this does not significantly bias our data (see Web Appendix W3).

Second, we addressed endogeneity due to either reverse causality between argument strength and attitudes toward the issue or omitted variables using two-stage least squares estimation and comparing the estimated coefficients with the OLS estimated coefficients. We specified the first-stage equation to include a set of exogenous variables (the covariates from the Attitudes_t,i model) plus an instrumental variable, Reactance_i, capturing individual i’s resistance to a perceived threat to free choice or behavior. The instrument was relevant (i.e., correlated with argument strength) and met the exclusion criterion (i.e., uncorrelated with the error term of the Attitudes_t,i equation). In the second-stage equation, the estimated residuals from the first-stage regression were included in the Attitudes_t,i model to control for the endogenous regressors. Our test indicates that endogeneity does not bias the OLS coefficients (for details regarding the appropriateness of our instrument and model results, see Web Appendix W4).

Third, because a subset of respondents from the preelection survey opted to complete the postelection survey, we addressed selection bias using Heckman’s (1979) two-step estimation of the effect of argument strength on attitudes toward and voting on the issue. In the first-stage selection equation, Voted_t,i (binary variable indicating whether or not the respondent voted on the proposition in the election) is regressed on our covariates plus the conceptually unrelated variable Reactance_i. The second-stage outcome equations for attitudes and voting were estimated using the Attitudes_t,i and Vote_t,i model specifications, respectively, conditioned on whether the respondent voted on the proposition in the election. From the estimated coefficients for the second-stage model, including the inverse Mills ratio, we conclude that selection bias did not adversely affect model estimation (see Web Appendix W5).

Results

The Influence of Competing Persuasion Campaign Arguments

A model-free examination of the data offers preliminary evidence of our predicted relationships (see Web Appendix W6). Table 3 contains the results related to our hypothesized effects. Note that for our analysis, as industry is opposed to the proposition, negative coefficients indicate favorable attitudes toward and voting in favor of the industry side. We find consistent support for H₁ and H₂. Stronger industry arguments led to more favorable attitudes toward the industry side on both the drug price (β_{Industry_Argument_Strength} = −.178, p < .01) and tobacco tax (β_{Industry_Argument_Strength} = −.128, p < .01) propositions and increased the likelihood of voting in favor of the industry side on the drug price (β_{Industry_Argument_Strength} = −.623, p < .01; 65.1% likelihood of voting no⁴) and tobacco tax (β_{Industry_Argument_Strength} = −.476, p < .01; 61.7% likelihood of voting no) propositions. Similarly, stronger activist arguments led to more favorable attitudes toward this side for the drug price (β_{Activist_Argument_Strength} = .630, p < .01) and tobacco tax (β_{Activist_Argument_Strength} = .471, p < .01) propositions and increased the likelihood of voting in favor of this side for the drug price (β_{Activist_Argument_Strength} = 1.300, p < .01; 78.6% likelihood of voting yes) and tobacco tax (β_{Activist_Argument_Strength} = 1.206, p < .01; 77.0% likelihood of voting yes) propositions.

Table 3.

Impact of Industry and Activist Argument Strength and Suspicion on Attitudes Toward and Voting on the Issue.

Hyp.	Variables	Attitudes Toward the Issue				Voting on the Issue
		Drug Price Proposition		Tobacco Tax Proposition		Drug Price Proposition		Tobacco Tax Proposition
		Coefficient	Effect Size $η_{p}^{2}$	Coefficient	Effect Size $η_{p}^{2}$	Coefficient	Odds Ratio	Coefficient	Odds Ratio
	Constant	1.672***		2.413***		–3.126***		–1.672*
	Predictor Variables
H₁	Industry argument strength	−.178***	.035	−.128***	.026	−.623***	.536	−.476***	.621
H₂	Activist argument strength	.630***	.285	.471***	.202	1.300***	3.657	1.206***	3.339
H₃	Industry argument suspicion	−.008	.000	−.026	.002	.194**	1.214	.078	1.082
H₄	Activist argument suspicion	−.172***	.037	−.120***	.026	−.391***	.676	−.283***	.753
	Covariates
	Preelection attitudes toward the issue	.129***	.025	.349***	.185
	Preelection voting intention on the issue					1.233***	3.430	2.604***	13.520
	Familiarity: industry arguments	−.032	.001	−.027	.001	−.088	.916	−.092	.912
	Familiarity: activist arguments	.023	.001	.003	.000	−.177*	.838	−.221**	.801
	Political affiliation	−.007	.000	−.028*	.002	−.072*	.931	−.145**	.865
	Need for orientation	−.015	.000	−.047**	.005	−.069	.934	−.005	1.005
	Age	−.002	.001	−.002	.001	−.006	.994	.002	1.002
	Gender	.213**	.006	−.028	.000	.284	1.328	.424*	1.527
	Education	.192**	.007	.117*	.003	.692***	1.998	−.090	.914
	SES	.060	.002	−.009	.000	−.090	.914	.218	1.244
	N	851		871		851		871
	R²	.628		.703		.534		.574
	AIC					1179.760		1091.398
H₅	Test of equality of industry vs. activist argument strength coefficients (t-value)	−4.077***		−2.821***		−3.551***		−3.400***
H₆	Test of equality of industry vs. activist argument suspicion coefficients (t-value)	−2.618***		−1.969**		−2.249**		−2.192**

*p < .10.

**p < .05.

***p < .01.

Notes: One-tailed tests of significance. AIC = Akaike information criterion. SES = socioeconomic status. For attitudes toward the issue, negative coefficients indicate favorable attitudes toward the industry side on the issue (as industry opposes the proposed policy); positive coefficients indicate favorable attitudes toward the activist side on the issue (as activist side supports the proposed policy). For voting on the issue, negative coefficients indicate voting in favor of the industry side on the issue (as industry opposes the proposed policy); positive coefficients indicate voting in favor of the activist side on the issue (as the activist side supports the proposed policy).

Our results offer mixed support for H₃, as suspicion of industry arguments had little impact on voting outcomes. It had no effect on attitudes toward either proposition, contrary to H_3a. It significantly decreased the likelihood of voting in favor of the industry side for the drug price proposition (β_{Industry_Argument_Suspicion} = .194, p < .05; 45.2% likelihood of voting no), but not for the tobacco tax proposition, partially supporting H_3b. By contrast, suspicion of activist arguments played a consistent role, fully supporting H₄, resulting in less favorable attitudes toward that side for the drug price (β_{Activist_Argument_Suspicion} = −.172, p < .01) and tobacco tax (β_{Activist_Argument_Suspicion} = −.120, p < .01) propositions and decreased likelihood of voting in favor of that side for the drug price (β_{Activist_Argument_Suspicion} = −.391, p < .01; 40.3% likelihood of voting yes on the proposition) and tobacco tax (β_{Activist_Argument_Suspicion} = −.283, p < .01; 43.0% likelihood of voting yes) propositions.

Asymmetric Response to Competing Industry and Activist Arguments

The impact of argument strength

To compare the relative impact of industry and activist argument strength in predicting our key outcomes, we used the delta method to test the equality of the estimated coefficients for the two variables (Greene 2003). Specifically, we calculated t = (β_{Industry_Argument_Strength} − β_{Activist_Argument_Strength})/SE, where β_{Industry_Argument_Strength} and β_{Activist_Argument_Strength} are the absolute value of the two parameter estimates and SE is the standard error of their difference calculated as Sqrt[Var(β_{Industry_Argument_Strength}) + Var(β_{Activist_Argument _Strength}) − 2 × Cov(β_{Industry_Argument_Strength}, β_{Activist_Argument_Strength})]. The results provide full support for H₅, indicating that activist argument strength has a significantly greater impact than industry argument strength on attitudes toward the drug price (t = −4.077, p < .01) and tobacco tax (t = −2.821, p < .01) propositions and on voting for the drug price (t = −3.551, p < .01) and tobacco tax (t = −3.400, p < .01) propositions. This is consistent with a comparison of effect sizes, which indicates that activist argument strength has a strong effect on attitudes toward both the drug price ( $η_{p}^{2}$ = .285) and tobacco tax ( $η_{p}^{2}$ = .202) propositions, while industry argument strength has a moderate-small effect for the drug price ( $η_{p}^{2}$ = .035) and tobacco tax ( $η_{p}^{2}$ = .026) propositions.⁵

The impact of argument suspicion

Using the delta method to compare the relative impact of industry and activist argument suspicion in predicting attitudes and voting revealed that activist argument suspicion has a significantly greater impact than industry argument suspicion on attitudes toward the drug price (t = −2.618, p < .01) and tobacco tax (t = −1.969, p < .05) propositions and on voting for the drug price (t = −2.249, p < .05) and tobacco tax (t = −2.192, p < .05) propositions, in support of H₆. A comparison of effect sizes reflects this difference, such that activist argument suspicion has a moderate-small effect on attitudes toward the issue for both the drug price ( $η_{p}^{2}$ = .037) and tobacco tax ( $η_{p}^{2}$ = .026) propositions, but industry argument suspicion has a negligible effect for the drug price ( $η_{p}^{2}$ = .000) and tobacco tax ( $η_{p}^{2}$ = .002) propositions. Effect sizes for all parameters and overall model statistics are reported in Table 3.

Robustness Checks

Although our survey respondents’ characteristics generally align with the demographics of California voters in the 2016 election based on U.S. Census Bureau data in terms of gender, age, and income, our sample included higher proportions of individuals with higher education levels and whites/Caucasians and lower proportions of Hispanics (see Web Appendix W7). To help ensure that our results are representative of California voters, we performed a robustness check in which we weighted observations using an iterative proportional fitting procedure, which takes the marginal distributions of demographic variables from the California voter population and returns weights such that the marginal distributions of the demographic variables in the weighted survey data match those in the population (Wonnacott and Wonnacott 1990). Compared with results for the unweighted models, our analysis finds consistent effects for 15 of the 16 total hypothesized effects of the impact of argument strength and argument suspicion on attitudes toward and voting on the issue for the two propositions (see Web Appendix W8).

Exploration of Vote Switching Segments

Of the two ballot measures we study, the drug price proposition experienced significant swings in public support. Political polling in California leading up to the election indicated that 73% of registered voters supported the proposition in late July (Tulchin Research and Strategic Consulting 2016), but this decreased to 66% in mid-September (USC Dornslife/LA Times) and further to 51% in mid-October (Hoover Institute/YouGov). Following intensive pharmaceutical industry media spending in the six weeks before the election, the proposition was defeated in the November election with 48.6% of support (all polling data retrieved from https://ballotpedia.org/California_Proposition_61,_Drug_Price_Standards_(2016)).⁶ Such a swing in support did not occur for the tobacco tax proposition.

We considered two scenarios. First, we focused on voters who intended to vote in favor of the activist side in the preelection survey (i.e., would vote yes on the proposition) and measured our dependent variable by identifying those who actually voted in favor of that side (voted yes) versus those who switched and actually voted in favor of the industry side (voted no) as measured in the postelection survey (1 = switched vote, 0 = did not switch vote). Second, we focused on those who intended to vote in favor of the industry side and measured the dependent variable by identifying those who actually voted in favor of the industry side versus those who switched and voted in favor of the activist side.

As Table 4 shows, argument strength and argument suspicion play important roles in vote switching (all results apply to both the drug price and tobacco tax propositions). Strong industry (activist) arguments significantly increased (decreased) the likelihood of voters switching support from the activist side to the industry side and decreased (increased) the likelihood of switching support from the industry side to the activist side. Consistent with our previous analyses, activist argument strength had a greater impact than industry argument strength in both scenarios. The results related to argument suspicion reveal a notable effect that may have been masked in our previous analyses. Consistent with our previous findings, suspicion of activist arguments significantly increased (decreased) voter switching to the industry (activist) side. However, in contrast with our previous findings, suspicion of industry arguments significantly increased the likelihood that voters who initially supported the industry switched their support to the activist side. This distinctive finding is bolstered by a test of equality on the impact of argument suspicion for the opposing sides, which indicates that suspicion of activist arguments does not play a significantly greater role than suspicion of industry arguments in this scenario.

Table 4.

Impact of Industry and Activist Argument Strength and Suspicion on Vote Switching.

Variables	Switched from Intending to Vote in Favor of Activist Side to Voting in Favor of Industry Side				Switched from Intending to Vote in Favor of Industry Side to Voting in Favor of Activist Side
	Drug Price Proposition		Tobacco Tax Proposition		Drug Price Proposition		Tobacco Tax Proposition
	Coefficient	Odds Ratio	Coefficient	Odds Ratio	Coefficient	Odds Ratio	Coefficient	Odds Ratio
Constant	.633		.553		−5.874***		−.140
Predictor Variables
Industry argument strength	.635***	1.887	.394**	1.483	−.592***	.554	−.642***	.526
Activist argument strength	−1.275***	.280	−1.315***	.268	1.360***	3.895	1.306**	3.689
Industry argument suspicion	−.155	.856	.118	1.126	.294**	1.342	.400**	1.487
Activist argument suspicion	.463***	1.589	.287**	1.332	−.256**	.774	−.346**	.707
Covariates
Familiarity: industry arguments	.202*	1.224	.130	1.138	.056	1.057	−.044	.957
Familiarity: activist arguments	.144	1.155	.243	1.275	−.239	.787	−.203	.817
Political affiliation	.120**	1.127	.011	1.011	−.018	.982	.367***	.693
Need for orientation	.057	1.059	−.010	.990	−.054	.948	.077	1.080
Age	.007	1.007	−.005	.995	−.006	.994	−.002	.998
Gender	−.163	.850	−.726**	.484	.671**	1.956	.091	.913
Education	−.730***	.482	−.448	.639	.692**	1.998	−1.191**	.304
SES	.169	1.184	−.001	.999	.104	1.109	.773**	2.166
N	851		871		851		871
R²	.460		.237		.464		.432
AIC	665.533		352.953		380.347		298.456
Test of equality of industry vs. activist argument strength coefficients (t-value)	−2.525**		−3.140***		−2.551**		−1.971**
Test of equality of industry vs. activist argument suspicion coefficients (t-value)	−1.970**		1.691*		.151		.191

*p < .10.

**p < .05.

***p < .01.

Our field study findings affirm the relationships predicted in our framework and hypotheses. In the next section, we develop hypotheses predicting how two distinctive and prominent argumentation strategies, one focused on financial factors and one focused on societal factors, vary in effectiveness for the industry and activist sides. We then test our predictions using three experimental studies.

Comparing Argumentation Strategy Effectiveness for the Industry and Activist Sides

Prior research on perceptual fit suggests that argument effectiveness would vary depending on fit with the persuader (e.g., Simmons and Becker-Olsen 2006). The positive effects of high fit between a firm or a brand and entities with which it is associated, such as sponsorships and cause-related marketing, are well established (Robinson, Irmak, and Jayachandran 2012; Simmons and Becker-Olsen 2006). Despite this, Baumgartner et al. (2009) document that opposing sides, including industry and activist groups, tend to use the same few strategies in the same policy battle, indicating that the fit between the persuader and the argument appears to be often overlooked. This raises an important question: Would a more focused approach be more effective and, if so, how should it be chosen?

To address this question, we examine argumentation strategies consistent with the image of the two sides in the conflicts we study. Financial argumentation is based on factors related to imposing or reducing financial benefits or costs, whereas societal argumentation is based on (nonfinancial) factors related to inhibiting or promoting shared societal goals or values. With financial argumentation, an industry may claim that there will be onerous financial costs for the public if current policy changes, whereas an activist group may argue that current policy needlessly burdens taxpayers (Drutman 2015; Waterhouse 2014). Using societal argumentation, an industry may claim that a policy change would weaken consumer access to important products or services, whereas an activist group may argue that current policy serves only a privileged few (Scott 2018).

Fit between an organization and an argument is high when the two are perceived as congruent because they share similar intangible associations (Simmons and Becker-Olsen 2006). Argumentation based on financial factors is congruent with how industries are viewed, as an industry’s image is characterized by economic and operational aspects (Boltanski and Thévenot 2006). Analysts’ reports profile industries in terms of financial and market performance and industry actions directly affect the economic status of consumers (e.g., Amiram, Kalay, and Sadka 2017). Because financial argumentation is more congruent with public expectations of industry, this strategy is a better fit with the industry side and will lead to more effective persuasion (Robinson, Irmak, and Jayachandran 2012). As such, we predict that financial argumentation will be perceived as stronger and less suspicious when used by the industry side.

In contrast, argumentation based on societal factors is congruent with the image of activist groups, which generally are perceived as organizations that represent the collective good for citizens (Sheingate 2016). These organizations build consensus on higher-level principles and fundamental values such as social justice (Baur and Palazzo 2011; Schlozman, Verba, and Brady 2012). Because the activist side is focused on promoting shared societal goals or values, its image is more congruent with societal argumentation. As such, we predict that societal argumentation will be perceived as stronger and less suspicious when used by the activist side.

Although empirical evidence to inform our predictions is limited, research on organizational legitimacy provides some conceptual support by suggesting that industries battling to control policy most often assume a market and economic orientation, expressed by referencing price, competitiveness, and efficiency, whereas groups representing the public interest often assume a civic orientation, expressed by referencing collective interest, solidarity, and integrity (Boltanski and Thévenot 2006). In summary, we predict:

H₇: The effect of argumentation strategy on argument strength and argument suspicion is moderated by the competing side on the issue, such that

For the industry side, financial argumentation leads to higher argument strength and lower argument suspicion than societal argumentation.

For the activist side, societal argumentation leads to higher argument strength and lower argument suspicion than financial argumentation.

For financial argumentation, use by the industry side leads to higher argument strength and lower argument suspicion than use by the activist side.

For societal argumentation, use by the activist side leads to higher argument strength and lower argument suspicion than use by the industry side.

Building on our previous predictions that argument strength and argument suspicion impact attitudes toward and voting on the issue, we expect these evaluations to mediate the effect of an argumentation strategy on voting outcomes. Specifically:

H₈: The mediating role of argument strength and argument suspicion depends on the competing side on the issue, such that

For the industry side, financial (vs. societal) argumentation increases argument strength and decreases argument suspicion, which then increases favorable attitudes toward and voting in favor of the industry side on the issue.

For the activist side, societal (vs. financial) argumentation increases argument strength and decreases argument suspicion, which then increases favorable attitudes towards and voting in favor of the activist side on the issue.

Experimental Studies of Argumentation Strategy Effects

We test H₇ and H₈ in three experimental studies on different issues addressed in recent ballot measures (see Table 1): pharmaceutical drug price standards, which aligns with one of our field study contexts; recyclable bottle deposits; and renewable energy standards. In a controlled, randomized setting, we exposed participants to two opposing messages (one from each side on the issue) that used either financial or societal argumentation strategies, thus separating the effects of argumentation strategy from those of frequency of exposure (Chong and Druckman 2010).

Experimental Study 1

Methodology

We used a mixed experimental design with two manipulations administered at two points in time within the same survey. Participants were first introduced to a hypothetical scenario in which the residents of a U.S. state are scheduled to vote on a proposition to adopt statewide pharmaceutical drug price standards and were sequentially shown two messages related to the proposition. First, participants saw one of four messages in a 2 (argumentation strategy: financial vs. societal) × 2 (side on the issue: industry vs. activist) between-subjects experimental design (time 1). Next, in the same survey, we manipulated argumentation strategy within subject; depending on which argumentation strategy participants saw at time 1, they saw one of the two strategies (financial or societal) for the opposing side (time 2). In each condition, participants saw a different combination of arguments from each side (both financial; both societal; or one of each, in a different order), enabling us to examine perceptions of argument strength and suspicion of each strategy in the presence of opposing arguments.

In the introduction, participants were told that Proposition 25 asks voters to decide whether their state should adopt standards to restrict the amount that some state agencies and programs pay for selected prescription drugs (for stimuli, see Web Appendix W9). We then introduced each message one at a time, indicating that it is a message designed by supporters (opponents) of Proposition 25 to persuade voters to vote yes (no) and support (defeat) the proposition. We derived the arguments used in our manipulations from the official arguments in the California Official Voter Information Guide for Proposition 61 (Drug Price Standards, November 8, 2016).⁷

After participants viewed the two arguments, they indicated their attitudes toward the proposition and how likely they were to vote yes or no in the referendum (1 = “I will definitely vote NO and oppose this proposition,” and 7 = “I will definitely vote YES and support this proposition”). Next, they saw the message viewed at time 1 and evaluated its argument strength and suspicion. After this, they undertook the same evaluations for the message viewed at time 2. We then assessed political affiliation, need for orientation, age, gender, education, and SES. Detailed descriptions of each of the measures appear in Web Appendix W10.

To enhance external validity, we recruited a nationally representative sample of registered voters. We partnered with Qualtrics to recruit 659 registered voters (49.68% female; M_age = 52.05 years; age range: 18–91 years). Qualtrics enforced quotas so that the sample was proportional to the U.S. population in terms of region, gender, age, household income, education, and ethnicity based on census data. Because the issue was voted on in California in 2016 and Ohio in 2017, we dropped 33 participants who indicated residency in these states, which left a sample of 626 participants for analyses. We conducted a pretest to confirm that our argumentation strategy manipulation was successful (for details, see Web Appendix W11).

Results

Given the within-subject nature of our design, in which we measure strength and suspicion of the second argument after participants evaluate the first argument, carryover and anchoring effects inherent to the design limit the interpretation of the second message evaluation (Podsakoff et al. 2003). Thus, to test the hypothesized effects in the presence of arguments from both sides, while ensuring that the argument strength and suspicion measures are not subject to bias, we use participants’ evaluations of the first argument and control for the argument evaluated second.

We used PROCESS Model 58 (Hayes 2018) to test our full conceptual model (see Table 5). An OLS regression on argument strength, with argumentation strategy (1 = financial, 0 = societal), side on the issue (1 = industry, 0 = activist), and their interaction as predictors. Covariates were individual-level characteristics and the argumentation strategy participants saw at time 2. The analysis revealed significant negative main effects of argumentation strategy (b_financial = −.529, p < .01) and side on the issue (b_industry = −.607, p < .01) and the predicted interaction effect (b_{financial × industry} = .841, p < .01). A similar regression on argument suspicion revealed significant positive main effects of argumentation strategy (b_financial = .384, p <= .05) and side on the issue (b_industry = .627, p < .01) and the predicted interaction effect (b_{financial × industry} = −.744, p < .01).

Table 5.

Response to Argumentation Strategies from Competing Sides: Drug Price Standards Proposition.

		Dependent Variables
Hyp.	Predictor Variables	Coefficient (Effect Size, $η_{p}^{2}$ )
	Moderating Effect of Issue Side on Argumentation Strategy	Argument Strength
	Financial argumentation strategy (time 1, 0 −1)	−.529*** (.001)
	Industry side (time 1, 0−1)	−.607*** (.003)
	Financial argumentation strategy × Industry side	.841*** (.026)
	Intercept	4.104***
	R²	.123
	Model significance	F(10, 611) = 8.581***
	Moderating Effect of Issue Side on Argumentation Strategy	Argument Suspicion
	Financial argumentation strategy (time 1, 0−1)	.384** (.000)
	Industry side (time 1, 0−1)	.627*** (.007)
	Financial argumentation strategy × Industry side	−.744*** (.017)
	Intercept	3.747***
	R²	.044
	Model significance	F(10, 611) = 2.851***
	Effects of Argument Strength and Argument Suspicion	Attitudes Toward the Issue	Voting on the Issue
	Financial argumentation strategy (time 1, 0−1)	.014 (.000)	.019 (.000)
	Argument strength	.589*** (.032)	.631*** (.045)
	Argument suspicion	−.266*** (.051)	−.211*** (.050)
	Industry side (time 1, 0 −1)	.766 (.000)	.621 (.004)
	Argument strength × Industry side	−.683*** (.073)	−.681*** (.065)
	Argument suspicion × Industry side	.533*** (.210)	.509*** (.186)
	Intercept	2.134***	2.016***
	R²	.330	.299
	Model significance	F(13, 608) = 23.063***	F(13, 608) = 19.977***
H_8a	Conditional indirect effect through argument strength, industry	−.029 [−.120, .032]	−.015 [−.100, .053]
H_8a	Conditional indirect effect through argument suspicion, industry	−.096 [−.230, −.006]	−.107 [−.261, −.007]
H_8b	Conditional indirect effect through argument strength, activist	−.312 [−.564, −.101]	−.334 [−.578, −.110]
H_8b	Conditional indirect effect through argument suspicion, activist	−.102 [−.245, −.005]	−.081 [−.207, −.001]
	Index of moderated mediation (argument strength)	.282 [.059, .541]	.319 [.086, .576]
	Index of moderated mediation (argument suspicion)	.006 [−.158, .174]	−.026 [−.203, .130]
	Simple Effects of Issue Side on Argument Strength and Suspicion	Argument Strength	Argument Suspicion
H_7a	Simple effects of financial (vs. societal) argumentation for industry side	.312*	−.360**
H_7b	Simple effects of financial (vs. societal) argumentation for activist side	−.529***	.384**
H_7c	Simple effects of industry (vs. activist) side for financial argumentation	.234	−.117
H_7d	Simple effects of industry (vs. activist) side for societal argumentation	−.607***	.627***

*p < .10.

**p < .05.

***p < .01.

Notes: Results from PROCESS Model 58, which includes need for orientation, political affiliation, age, gender (male), education, SES, and argumentation strategy at time 2 (0−1) as covariates; 95% confidence intervals reported for conditional indirect effects; n = 622 for these analyses due to missing values on age and SES covariates.

Tests of the simple effects of argumentation strategy for each side on the issue showed that for the industry side, financial argumentation led to marginally significantly higher argument strength (b_financial = .312, p < .10) and significantly lower argument suspicion (b_financial = −.360, p < .05), as compared with societal argumentation, in support of H_7a. For the activist side, societal argumentation led to significantly higher argument strength (b_financial = −.529, p < .01) and lower argument suspicion (b_financial = .384, p < .05), as compared with financial argumentation, supporting H_7b (negative coefficients indicate enhanced effects and positive coefficients indicate diminished effects because side on the issue is coded as industry = 1 and activist = 0).

Tests of the simple effects of side on the issue for each argumentation strategy revealed that, for financial argumentation, use by the industry side was not significantly different than use by the activist side for both argument strength (b_industry = .234, p > .10) and argument suspicion (b_industry = −.117, p > .10), contrary to H_7c. However, in support of H_7d, for societal argumentation, use by the activist side led to significantly higher argument strength (b_industry = −.607, p < .01) and significantly lower argument suspicion (b_industry = .627, p < .01) than use by the industry side.

For the full moderated mediation model, for the industry side, argument suspicion mediated the effects of argumentation strategy on attitudes and voting (conditional indirect effect on attitude: b = −.096, 95% confidence interval [CI]: [−.230, −.006]; conditional indirect effect on voting: b = −.107, 95% CI: [−.261, −.007]), but argument strength did not (conditional indirect effect on attitude: b = −.029, 95% CI: [−.120, .032]; conditional indirect effect on voting: b = −.015, 95% CI: [−.100, .053]), partially supporting H_8a. For the activist side, both argument strength (conditional indirect effect on attitude: b = −.312, 95% CI: [−.564, −.101]; conditional indirect effect on voting: b = −.334, 95% CI: [−.578, −.110]) and argument suspicion (conditional indirect effect on attitude: b = −.102, 95% CI: [−.245, −.005]; conditional indirect effect on voting: b = −.081, 95% CI: [−.207, −.001]) mediated the effects of argumentation strategy on attitudes and voting, fully supporting H_8b.⁸

Experimental Studies 2 and 3

To assess the generalizability of our findings across different issues, we conducted two additional experimental studies. Both studies used the same design as Experimental Study 1. Experimental Study 2 (331 U.S.-based participants from the Prolific Academic panel; 44.11% female; M_age = 33.19 years, age range: 18–73 years) focused on expanding a state’s bottle deposit law to require deposits for all nonalcoholic, noncarbonated drinks (Expansion of Bottle Deposits Initiative, Massachusetts, November 4, 2014). Experimental Study 3 (340 U.S.-based participants from Prolific Academic; 50.88% female; M_age = 34.70 years, age range: 18–73 years) focused on increasing a state’s renewable energy standards, requiring electricity providers to obtain at least 50% of their electricity from renewable sources by 2030 (Renewable Energy Standards Initiative, Arizona, November 6, 2018). The stimuli are presented in Web Appendix W9 and pretests to confirm our argumentation strategy manipulations are included in Web Appendix W11.

Results from the two studies were largely identical to those of Experimental Study 1 (see Table 6). For both propositions, as expected and in line with Experimental Study 1, for the industry side, financial argumentation led to higher argument strength and lower argument suspicion than societal argumentation, while for the activist side, societal argumentation led to higher argument strength and lower argument suspicion than financial argumentation. Comparing side on the issue effects for each argumentation strategy, for societal argumentation, use by the activist side led to higher argument strength and lower suspicion than use by the industry side but, as in Experimental Study 1 and contrary to predictions, participants did not perceive financial argumentation as significantly stronger or less suspicious when used by the industry versus the activist side. Across both propositions, argument strength mediated the effect of argumentation strategy on attitudes toward and voting on the issue for both the industry and activist sides. Argument suspicion also had significant mediating effects for the industry and activist sides across both propositions with two exceptions. For the renewable energy standards proposition, argument suspicion mediated the effect of argumentation strategy on voting but not on attitudes toward the industry side, and argument suspicion had a significant mediating effect on attitudes but not on voting for the activist side.⁹

Table 6.

Response to Argumentation Strategies from Competing Sides: Bottle Deposits and Energy Standards Propositions.

		Bottle Deposits Proposition		Energy Standards Proposition
Hyp.	Predictor Variables	DV: Coefficient (Effect Size, $η_{p}^{2}$ )		DV: Coefficient (Effect Size, $η_{p}^{2}$ )
	Moderating Effect of Issue Side on Argumentation Strategy	Argument Strength		Argument Strength
	Financial argumentation strategy (time 1, 0 −1)	−.911*** (.001)		−.747*** (.040)
	Industry side (time 1, 0−1)	–1.959*** (.138)		–2.566*** (.168)
	Financial argumentation strategy × Industry side	1.575*** (.075)		2.573*** (.165)
	Intercept	3.842***		5.186***
	R²	.232		.322
	Model significance	F(10, 310) = 9.386***		F(10, 322) = 15.355***
	Moderating Effect of Issue Side on Argumentation Strategy	Argument Suspicion		Argument Suspicion
	Financial argumentation strategy (time 1, 0 −1)	.587*** (.001)		.538** (.004)
	Industry side (time 1, 0−1)	1.658*** (.169)		2.054*** (.195)
	Financial argumentation strategy × Industry side	−.982*** (.032)		–1.334*** (.054)
	Intercept	2.837***		2.287***
	R²	.201		.242
	Model significance	F(10, 310) = 7.819***		F(10, 322) = 10.300***
	Effects of Argument Strength and Argument Suspicion	Attitudes	Voting	Attitudes	Voting
	Financial argumentation strategy (time 1, 0−1)	.451*** (.041)	.503*** (.029)	.205 (.024)	.024 (.007)
	Argument strength	.428*** (.008)	.602*** (.008)	.208** (.024)	.430*** (.017)
	Argument suspicion	−.273*** (.010)	−.234** (.002)	−.271*** (.005)	−.068 (.001)
	Industry side (time 1, 0 −1)	.546 (.015)	1.180 (.003)	.675 (.008)	2.215** (.012)
	Argument strength × Industry side	−.817*** (.145)	–1.068*** (.149)	−.522***(.126)	−.833*** (.192)
	Argument suspicion × Industry side	.648*** (.376)	.756*** (.358)	.383*** (.320)	.255* (.311)
	Intercept	3.579***	2.172**	5.173***	3.057**
	R²	.477	.460	.480	.497
	Model significance	F(13, 307) = 21.558***	F(13, 307) = 20.119***	F(13, 319) = 22.654***	F(13, 319) = 24.249***
H_8a	Conditional indirect effect through argument strength, industry	−.258 [−.492, −.081]	−.309 [−.623, −.075]	−.573 [−.900, −.270]	−.735 [–1.144, −.365]
H_8a	Conditional indirect effect through argument suspicion, industry	−.148 [−.326, −0010]	−.205 [−.461, −.010]	−.088 [−.238, .035]	−.148 [−.353, −.006]
H_8b	Conditional indirect effect through argument strength, activist	−.390 [−.672, −.158]	−.549 [−.895, −.222]	−.155 [−.335, −.024]	−.321 [−.607, .097]
H_8b	Conditional indirect effect through argument suspicion, activist	−.161 [−.344, −.021]	−.137 [−.350, −.004]	−.146 [−.319, −.017]	−.037 [−.177, .108]
	Index of moderated mediation (argument strength)	.132 [−197, .473]	.239 [−.209, .669]	−.417 [−.775, −.086]	−.413 [−.891, 026]
	Index of moderated mediation (argument suspicion)	.013 [−.213, .245]	−.068 [−.365, .226]	.057 [−.140, .263]	−.111 [−.365, .088]
	Simple Effects of Issue Side on Argument Strength and Suspicion	Argument Strength	Argument Suspicion	Argument Strength	Argument Suspicion
H_7a	Simple effects of financial argumentation for industry side	.664***	−.394*	1.826***	−.796***
H_7b	Simple effects of financial argumentation for activist side	−.911***	.587***	−.747***	.538**
H_7c	Simple effects of industry side for financial argumentation	−.383	.676 ***	.006	.719***
H_7d	Simple effects of industry side for societal argumentation	–1.959 ***	1.658 ***	–2.566***	2.054 ***

Notes: Results from PROCESS Model 58, which includes need for orientation, political affiliation, age, gender (male), education, SES, and argumentation strategy at time 2 (0 −1) as covariates; for bottle deposits study, 10 participants from MA, where proposition was voted on, were dropped from analyses; for the energy standards study, 7 participants from AZ were dropped from analyses; 95% confidence intervals reported for conditional indirect effects; * p < .10, ** p < .05, *** p < .01.

In summary, across the three propositions, we obtained consistent support for the mediating role of argument strength and argument suspicion (H_8aand H_8b). Our results consistently show that societal argumentation is more effective for the activist side: it is more effective than financial argumentation for the activist side (H_7b), and is perceived as stronger and less suspicious when used by the activist side than the industry side (H_7d). For financial argumentation, results were somewhat mixed: as expected, financial argumentation was more effective than societal argumentation for the industry side (H_7a). However, it was not perceived as significantly stronger and less suspicious when used by the industry versus the activist side (H_7c). We summarize the results across all experimental studies in Web Appendix W12.

Exploration of the Role of Political Affiliation

We contend that battles between industry and activist groups over consumer-related issues tend to be less partisan than other political scenarios such as candidate elections. For example, major early polls reported 77% of Democrats and 70% of Republicans supported Proposition 61, the drug price standards proposition examined in our field study and Experimental Study 1 (Public Policy Institute of California 2016; Tulchin Research and Strategic Consulting 2016).

To gain insight into the impact of partisanship in our context, we performed additional regression analyses to examine the extent to which political affiliation moderates the effects of side on the issue and argumentation strategy on argument strength and argument suspicion across the three issue contexts. Overall, our results indicate that only one of the six interaction effects between political affiliation and argumentation strategy was significant, but all six interaction effects of political affiliation and side on the issue were significant (see Web Appendix W13). The strongest Democrats perceived activist argumentation more favorably than industry argumentation, but there were no significant differences in how Republicans perceived the two sides’ argumentation. These results suggest that political affiliation may play a role in how voters view the industry and activist sides. This is consistent with recent polls indicating that the majority of conservatives (liberals) have a positive (negative) view of big business (Saad 2019) and is in line with current polarization along party lines in many domains in the United States (Berry and Wilcox 2018; Scott 2018).

Discussion

We find support for our direct-to-public persuasion model across a field study and three experimental studies examining four policy conflict scenarios recently voted on in U.S. state ballot measures. Our findings show that industry and activist arguments play a key role in voting decisions, but industry arguments have less impact than activist arguments. Stronger arguments from both sides lead to more favorable outcomes, but activist groups benefit most. Similarly, industry argument suspicion has limited influence, except for voters who switch their support to the activist side. While societal argumentation is the preferred strategy for the activist side and financial argumentation is preferred for the industry side, the industry’s competitive advantage is far less pronounced than expected.

Theoretical Implications

Our exploration of direct-to-public persuasion increases the breadth of persuasion theory within the marketing domain. Persuasion campaigns in political settings—in which opinions are often strongly held, outcomes are win or lose, and consequences are societal rather than tied to individual consumers—are more complex in their effects than persuasion in commercial settings (Gordon et al. 2012; Shachar 2009). As there is little theoretical grounding to build on, our research offers new ways of understanding this unique form of persuasion and adds to the scope of sociopolitical legitimacy theory, research on voting behavior, and perceptual fit theory.

Our findings related to an asymmetric public response in industry versus activist conflicts caution against assumptions that competing campaigns have an equal opportunity to persuade the public. By addressing unexplored differences in public response to competing campaigns, including legitimacy differences, our research updates persuasion theory by examining how and why arguments from competing sides affect voting behavior. Our results show that competition in politics is uneven, encompassing not just resource imbalances that may advantage one side, but also asymmetries across sides in the extent to which persuasion knowledge plays a role in voters’ response (see Druckman, Fein, and Leeper 2012; Kirmani and Campbell 2009). Because the effects of dueling campaign arguments over time are not well known, yet are central to political outcomes, our findings highlight an important direction for future persuasion research.

Although theoretical guidance on the impact of argumentation on vote switching behavior is scarce, our results reveal an unexpected role of industry argument suspicion in that, although it has limited effect when voters are considered in aggregate, it is a key driver for the segment of voters who switch their support to the activist side. This finding is novel for persuasion research, as it indicates that the industry argument suspicion effects that drive a change in voter support over the course of a campaign differ somewhat from those driving overall voter support. This implies that response to the competing sides is not only asymmetric but nuanced and that research needs to consider various outcome measures to capture underlying factors.

Our finding that the industry side is most effective with a financial argumentation strategy while the activist side is most effective with societal argumentation offers new theoretical support regarding the need for fit between the persuader’s identity and arguments. This complements the perceptual fit literature that has traditionally addressed the fit between a company’s actions and claims (Wagner, Lutz, and Weitz 2009) or a company’s identity and cause-related marketing choices (Robinson, Irmak, and Jayachandran 2012). We contribute to perceptual fit theory by finding that fit does not have equal importance for both sides, as an image-congruent argumentation strategy is vital for the industry, but voters may be more tolerant of a broader strategy for the activist side.

Implications for Practice

Our results offer practical guidance that differs for the industry versus activist sides related to the substance of their arguments and their goals of acquiring or retaining supporters (see Table 7). The findings are relevant for marketers and practitioners who design and implement direct-to-public campaigns with industry associations, corporations, public interest advocacy and activist groups, consultancies, and other policy-focused coalitions (Sheingate 2016; Walker 2014).

Table 7.

Key Findings and Guidance for Industries and Activist Groups.

Key Findings for Winning Public Support	Guidance for Industry Practice	Guidance for Activist Group Practice
Substance of Principle Arguments
Strong arguments benefit both sides, but activist argument strength has a greater impact than industry argument strength	Develop campaign arguments focused on diminishing the strength of activist arguments	Develop campaign arguments with ironclad strength and the potential to preempt industry arguments
Suspicion of activist arguments negatively affects outcomes, but industry argument suspicion has limited impact	Take an aggressive approach, raising skepticism of activist tactics, but avoid triggering a backlash	Take a defensive approach, preserving the legitimacy of activist group practices and arguments
Industry arguments with a financial focus are perceived as stronger and less suspicious; activist group arguments with a societal focus are stronger and less suspicious	Focus on a financial argumentation strategy based on factors related to imposing or reducing financial benefits or costs	Focus on a societal argumentation strategy based on (nonfinancial) factors related to inhibiting or promoting shared societal goals or values
Acquiring and Retaining Voter Support
Industry argument suspicion is only a key factor for voters who switch their support to the activist side	Focus on acquiring supporters by targeting the segment of voters whose initial support for the activist side is tentative	Emphasize retention of early supporters by reenforcing public perceptions that activist group legitimacy is higher
When both sides use financial argumentation, the industry side’s competitive advantage over the activist side is limited	Build voter support with a narrow argumentation strategy focused on specific financial implications to voters	Retain voter support with a broader argumentation strategy focused on societal factors and counter industry claims with financial arguments

While intuitive that both sides benefit from strong arguments, it is most critical for the activist side to be certain that voters will view its arguments as strong, as this is the effect with the highest impact on voting outcomes. This implies that it should be more effective for the industry to focus on diminishing the strength and increasing suspicion of activist arguments. As suspicion of industry arguments has limited impact, the industry has some license to encourage skepticism of activist tactics. This approach may require finesse, however, so as not to trigger a backlash if the issue is salient and the industry is controversial (Scott 2018).

Preelection polls often indicate majority support for the activist side, suggesting that the public is initially drawn to a public interest perspective. When the activist side has this early advantage, it needs to retain initial supporters by ensuring that its arguments are strong and by utilizing counterarguments to protect its legitimacy. In contrast, the industry side must focus on acquiring voters and build support with arguments that highlight the hidden complexities and costs of the proposed policy change. As industries often succeed in acquiring supporters during a campaign, as voting nears, the activist side may need to shift its focus in order to cultivate a switching segment.

Our results show that the fit between an argumentation strategy and the identity of the persuader is key, especially for the industry side, which is safer emphasizing financial arguments closely aligned to its business sector profile. The activist side has some degrees of freedom as it is viewed positively by the public with either a financial or societal argumentation strategy. While societal argumentation is likely to have the most favorable impact for the activist side, financial argumentation may be necessary to challenge overstated cost analyses used by industry.

In summary, our results suggest the industry side does best to follow an aggressive approach of attacking activist-side tactics while using a narrow argumentation strategy. Conversely, the activist side does best by vigilantly preserving its legitimacy in reaction to any industry attacks, while using a broader argumentation strategy.

Implications for Public Policy

Given the consequential outcomes of the ballot measure venue we study and continued increases in funding ($1.24 billion in 2020; https://ballotpedia.org/2020_ballot_measures), it is surprising that there is limited emphasis on the transparency of industry and organizational support, such as comprehensive information about donors and the extent of their financial support. Our findings indicate that response to campaign arguments may be driven by legitimacy perceptions of the opposing sides, suggesting that policy makers should facilitate increased information to voters. As the public is primarily informed and persuaded by advertising during political campaigns, including the ballot initiative process, advertising law changes might be the most effective approach to reform (Richards, Allender and Di Fang 2013; Xu et al. 2020).

The Federal Election Commission regulates campaign finance law and political advertising, but its authority does not extend to ballot measure advertising, which is largely a state issue (Glazer and Haggin 2019). California, under the jurisdiction of the California Fair Political Practices Commission (www.fppc.ca.gov), has among the strictest regulation for ballot measure advertising, following the logic that it is less clear who is responsible for these ads compared with candidate election ads. Because California already requires that top donors be explicitly listed in television, electronic media, and print advertisement as well as in mass mailings and robocalls, it can serve as a model for other states’ ballot measure commissions to develop more comprehensive disclosure and increased transparency requirements. Alternatively, the Federal Election Commission could take steps to facilitate more unified federal standards for ballot measure advertisements. Without federal guidelines, consumers will continue to experience a patchwork approach to information that is unequal across states, making it difficult to assess the impact of direct-to-public persuasion on voters’ decisions and policy outcomes.

Limitations and Directions for Future Research

Our investigation addresses one type of direct-to-public industry persuasion but not others, such as the more indirect route of shaping legislator decisions by influencing public opinion (for an examination of this issue, see Burstein [2003]) or expensive direct-to-public industry image campaigns. In our field study, both ballot measures took place in one U.S. state and, thus, lack geographic diversity to mitigate potential unobserved influences. While our experimental studies address well-established argumentation strategies, they represent a partial picture of argumentation used to justify (or sanction) industry practice to the public.

There is an important need for more field studies, natural experiments, and other real-world persuasion research initiatives, as well as more research explaining asymmetric public response (Druckman, Fein, and Leeper 2012). We conducted a posttest to investigate our position that the public perceives that activist groups are more motivated than the industry side to serve the public interest (see Web Appendix W14). While our findings confirm this assertion, future experimental research should examine the extent to which this perception is a driver of voting outcomes. Future studies should explore whether our results indicating that societal argumentation strategies create substantially higher suspicion when used by the industry side could be attenuated for industries with a high proportion of companies with strong corporate social responsibility reputations. Although our findings across four consumer-relevant issues suggest that our hypotheses are likely to generalize across other high-salience issues, further research should investigate whether asymmetry may be muted in contests over lower salience issues.

We believe our findings will largely generalize to other contested issue scenarios, including situations where (1) the industry is challenging rather than defending the status quo and (2) industries battle other industries. When an industry tries to establish or expand a market for growth, it may challenge the policy status quo. A recent example includes the petroleum and gas industry using political clout to expand fracking onto previously protected Native American land in New Mexico (Nelson 2020). In these cases, the asymmetric impact of activist arguments having greater impact than industry arguments may be even more pronounced because the industry’s self-interested motivation for market expansion may be more obvious to the public. Further, financial argumentation may actually backfire by magnifying the industry-serving objectives unless public attitudes have shifted over time to align with industry initiatives—for example, to promote casino development and marijuana legalization across the United States (Humphreys 2010; Kees et al. 2020). In situations where industries battle other industries, when there is a legitimacy gap and one industry’s public image is more favorable than the other’s (e.g., renewable energy industries fighting fossil fuels industries over energy subsidies and standards; see Gallup 2019), arguments by the higher legitimacy industry could be more influential in driving outcomes. However, if there is little or no legitimacy gap (e.g., two agricultural industries battling over source of origin labeling laws), asymmetrical effects would not manifest.

Despite the fact that direct-to-public persuasion occurs across a variety of industries and scores of issues and has been prevalent for decades, further research is needed because the many dimensions that characterize this form of marketing may represent significant boundary conditions on the impact of argumentation on voting outcomes. Our research offers an important step forward in advancing knowledge of this underexamined area of marketing.

Footnotes

Acknowledgments

The authors thank the JM review team for their insight and guidance throughout the review process. They also thank Leonard Berry, Jeffrey Berry, and Steven Lacey for their input.

Associate Editor

Jan-Benedict E.M. Steenkamp

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Online supplement:

Notes

References

American National Election Studies, University of Michigan, and Stanford University (2020), “ANES 2020 Time Series Study,” Inter-University Consortium for Political and Social Research (July 19, 2021), https://electionstudies.org/data-center/2020-time-series-study/.

Amiram

Dan

Kalay

Alon

Sadka

Gil

(2017), “Industry Characteristics, Risk Premiums, and Debt Pricing,” Accounting Review, 92 (1), 1–27.

Arceneaux

Kevin

(2012), “Cognitive Biases and the Strength of Political Arguments,” American Journal of Political Science, 56 (2), 271–85.

Areni

Charles S.

Lutz

Richard J.

(1988), “The Role of Argument Quality in the Elaboration Likelihood Model,” in Advances in Consumer Research, Vol. 15, Houston

Michael J.

, ed. Provo, UT: Association for Consumer Research, 197–202.

Bagozzi

Richard P.

Youjae

(1988), “On the Evaluation of Structural Equation Models,” Journal of the Academy of Marketing Science, 16 (1), 74–94.

Baumgartner

Frank R.

Berry

Jeffrey M.

Hojnacki

Marie

Leech

Beth L.

Kimball

David C.

(2009), Lobbying and Policy Change: Who Wins, Who Loses, and Why. Chicago: University of Chicago Press.

Baur

Dorothea

Palazzo

Guido

(2011), “The Moral Legitimacy of NGOs as Partners of Corporations,” Business Ethics Quarterly, 21 (4), 579–604.

Berry

Jeffrey M.

Wilcox

Clyde

(2018), The Interest Group Society, 6th ed. New York: Routledge.

Bitektine

Alex

(2011), “Toward a Theory of Social Judgments of Organizations: The Case of Legitimacy, Reputation, and Status,” Academy of Management Review, 36 (1), 151–79.

10.

Boltanski

Luc

Thévenot

Laurent

(2006), On Justification: Economies of Worth. Princeton, NJ: Princeton University Press.

11.

Burstein

Paul

(2003), “The Impact of Public Opinion on Public Policy: A Review and an Agenda,” Political Research Quarterly, 56 (1), 29–40.

12.

California Secretary of State (2016), Official Voter Information Guide. Sacramento: California Secretary of State Elections Division.

13.

Campbell

Margaret C.

Kirmani

Amna

(2000), “Consumers’ Use of Persuasion Knowledge: The Effects of Accessibility and Cognitive Capacity on Perceptions of an Influence Agent,” Journal of Consumer Research, 27 (1), 69–83.

14.

Chong

Dennis

Druckman

James N.

(2010), “Dynamic Public Opinion: Communication Effects over Time,” American Political Science Review, 104 (4), 663–80.

15.

Cohen

Joshua

(1988), Statistical Power Analysis for the Behavioral Sciences. New York: Routledge Academic.

16.

Darke

Peter R.

Ritchie

Robin B.

(2007), “The Defensive Consumer: Advertising Deception, Defensive Processing, and Distrust,” Journal of Marketing Research, 44 (1), 114–27.

17.

Druckman

James N.

Fein

Jordan

Leeper

Thomas J.

(2012), “A Source of Bias in Public Opinion Stability,” American Political Science Review, 106 (2), 430–54.

18.

Druckman

James N.

Peterson

Erik

Slothuus

Rune

(2013), “How Elite Partisan Polarization Affects Public Opinion Formation,” American Political Science Review, 107 (1), 57–79.

19.

Drutman

Lee

(2015), The Business of America Is Lobbying: How Corporations Became Politicized and Politics Became More Corporate. Oxford, UK: Oxford University Press.

20.

Dunn

William N

. (2012), Public Policy Analysis, 5th ed. London: Routledge.

21.

Eesley

Charles

DeCelles

Katherine A.

Lenox

Michael

(2016), “Through the Mud or in the Boardroom: Examining Activist Types and Their Strategies in Targeting Firms for Social Change,” Strategic Management Journal, 37 (12), 2425–440.

22.

Forehand

Mark R.

Grier

Sonya

(2003), “When Is Honesty the Best Policy? The Effect of Stated Company Intent on Consumer Skepticism,” Journal of Consumer Psychology, 13 (3), 349–56.

23.

Fornell

Claes

Larcker

David F.

(1981), “Evaluating Structural Equation Models with Unobservable Variables and Measurement Error,” Journal of Marketing Research, 18 (1), 39–50.

24.

Fransen

Marieke L.

Smit

Edith G.

Verlegh

Peeter W.J.

(2015), “Strategies and Motives for Resistance to Persuasion: An Integrative Framework,” Frontiers in Psychology, 6, 1201.

25.

Gallup (2019), “Confidence in Institutions” (accessed February 15, 2021), https://news.gallup.com/poll/1597/confidence-institutions.aspx.

26.

Glazer

Emily

Haggin

Patience

(2019), “Political Ads Are Flourishing Online. Few Agree How to Regulate Them,” The Wall Street Journal (November 15), https://www.wsj.com/articles/as-political-ad-spending-balloons-online-consensus-on-regulation-is-elusive-11573813803.

27.

Gordon

Brett R.

Lovett

Mitchell J.

Shachar

Ron

Arceneaux

Kevin

Moorthy

Sridhar

Peress

Michael

, et al. (2012), “Marketing and Politics: Models, Behavior, and Policy Implications,” Marketing Letters, 23 (2), 391–403.

28.

Greene

William H

. (2003), Econometric Analysis, 5th ed. Upper Saddle River, NJ: Prentice Hall.

29.

Hayes

Andrew F

. (2018), Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression Based Approach, 2nd ed. New York: The Guilford Press.

30.

Heckman

James J

. (1979), “Sample Selection Bias as a Specification Error,” Econometrica, 47 (1), 153–61.

31.

Hiatt

Shon R.

Grandy

Jake B.

Lee

Brandon H.

(2015), “Organizational Responses to Public and Private Politics: An Analysis of Climate Change Activists and U.S. Oil and Gas Firms,” Organization Science, 26 (6), 1769–86.

32.

Humphreys

Ashlee

(2010), “Megamarketing: The Creation of Markets as a Social Process,” Journal of Marketing, 74 (2), 1–19.

33.

Hussain

Suhauna

(2020), “Uber, Lyft Push Prop. 22 Message Where You Can’t Escape It: Your Phone,” Los Angeles Times (October 8), https://www.aol.com/news/uber-lyft-push-prop-22-120027093.html.

34.

Isaac

Mathew S.

Grayson

Kent

(2017), “Beyond Skepticism: Can Accessing Persuasion Knowledge Bolster Credibility?” Journal of Consumer Research, 43 (6), 895–912.

35.

Jung

Minah H.

Critcher

Clayton R.

(2018), “How Encouraging Niceness Can Incentivize Nastiness: An Unintended Consequence of Advertising Reform,” Journal of Marketing Research, 55 (1), 147–61.

36.

Kees

Jeremy

Fitzgerald

Paula

Dorsey

Joshua D.

Hill

Ronald Paul

(2020), “Evidence-Based Cannabis Policy: A Framework to Guide Marketing and Public Policy Research,” Journal of Public Policy & Marketing, 39 (1), 76–92.

37.

Kirmani

Amna

Campbell

Margaret C.

(2009), “Taking the Target’s Perspective: The Persuasion Knowledge Model,” Social Psychology of Consumer Behavior, 297–316.

38.

Klein

Jill G.

Ahluwalia

Rohini

(2005), “Negativity in the Evaluation of Political Candidates,” Journal of Marketing, 69 (1), 131–42.

39.

Klein

Jill Gabrielle

Craig Smith

John

Andrew

(2004), “Why We Boycott: Consumer Motivations for Boycott Participation,” Journal of Marketing, 68 (3), 92–109.

40.

Lamin

Anna

Zaheer

Srilata

(2012), “Wall Street vs. Main Street: Firm Strategies for Defending Legitimacy and Their Impact on Different Stakeholders,” Organization Science, 23 (1), 47–66.

41.

Lindell

Michael K.

Whitney

David J.

(2001), “Accounting for Common Method Variance in Cross-Sectional Research Designs,” Journal of Applied Psychology, 86 (1), 114–21.

42.

Maisel

L. Sandy

Berry

Jeffrey M.

(2010), “The State of Research on Political Parties and Interest Groups,” in The Oxford Handbook of American Political Parties and Interest Groups, Sandy Maisel

Berry

Jeffrey M.

, eds. Oxford, UK: Oxford University Press, 3–18.

43.

Martin

Kelly D.

Josephson

Brett W.

Vadakkepatt

Gautham G.

Johnson

Jean L.

(2018), “Political Management, Research and Development, and Advertising Capital in the Pharmaceutical Industry: A Good Prognosis?” Journal of Marketing, 82 (3), 87–107.

44.

McFadden

Brandon R.

Lusk

Jayson L.

(2013), “Effects of Cost and Campaign Advertising on Support for California’s Proposition 37,” Journal of Agricultural and Resource Economics, 38 (2), 174–86.

45.

McKinsey and Company (2021), “The Next Normal Arrives: Trends that will Define 2021 – and Beyond,” in McKinsey Global Publishing. New York: McKinsey & Co.

46.

Meyers-Levy

Joan

Malaviya

Prashant

(1999), “Consumers’ Processing of Persuasive Advertisements: An Integrative Framework of Persuasion Theories,” Journal of Marketing, 63 (4_suppl1), 45–60.

47.

Nelson

Cody

(2020), “‘Their Greed Is Gonna Kill Us’: Indian Country Fights Against More Fracking,” The Guardian (June 10), https://www.theguardian.com/us-news/2020/jun/10/new-mexico-fracking-navajo-indian-country.

48.

Obermiller

Carl

Spangenberg

Eric

MacLachlan

Douglas L.

(2005), “Ad Skepticism: The Consequences of Disbelief,” Journal of Advertising, 34 (3), 7–17.

49.

Pew Research Center (2019), “Views of the Economic System and Social Safety Net,” in American Trends Panel Wave 53. Washington DC: Pew Research Center.

50.

Phillips

Joan M.

Urbany

Joel E.

Reynolds

Thomas J.

(2008), “Confirmation and the Effects of Valenced Political Advertising: A Field Experiment,” Journal of Consumer Research, 34 (6), 794–806.

51.

Podsakoff

Philip M.

MacKenzie

Scott B.

Lee

Jeong-Yeon

Podsakoff

Nathan P.

(2003), “Common Method Biases in Behavioral Research: A Critical Review of the Literature and Recommended Remedies,” Journal of Applied Psychology, 88 (5), 879–903.

52.

Public Policy Institute of California (2016), PPIC Statewide Survey: Californians and Their Government. San Francisco: Public Policy Institute of California.

53.

Richards

Timothy

Allender

William

Fang

(2013), “Media Advertising and Ballot Initiatives: The Case of Animal Welfare Regulation,” Contemporary Economic Policy, 31 (1), 145–62.

54.

Robinson

Stefanie Rosen

Irmak

Caglar

Jayachandran

Satish

(2012), “Choice of Cause in Cause-Related Marketing,” Journal of Marketing, 76 (4), 126–39.

55.

Saad

Lydia

(2019), “Do Americans Like or Dislike ‘Big Business’?” Gallup (accessed December 26, 2019), https://news.gallup.com/poll/270296/americans-dislike-big-business.aspx?.

56.

Schlozman

Kay L.

Verba

Sidney

Brady

Henry E.

(2012), The Unheavenly Chorus: Unequal Political Voice Political Voice and the Broken Promise of American Democracy. Princeton, NJ: Princeton University Press.

57.

Scott

John C

. (2018), Lobbying and Society: A Political Sociology of Interest Groups. Cambridge, UK: Polity Press.

58.

Shachar

Ron

(2009), “The Political Participation Puzzle and Marketing,” Journal of Marketing Research, 46 (6), 798–815.

59.

Sheingate

Adam

(2016), Building a Business of Politics: The Rise of Political Consulting. Oxford, UK: Oxford University Press.

60.

Simmons

Carolyn J.

Becker-Olsen

Karen L.

(2006), “Achieving Marketing Objectives Through Social Sponsorships,” Journal of Marketing, 70 (4), 154–69.

61.

Suddaby

Roy

Bitektine

Alex

Haack

Patrick

(2017), “Legitimacy,” Academy of Management Annals, 11 (1), 451–78.

62.

Trumbull

Gunnar

(2012), Strength in Numbers: The Political Power of Weak Interests. Cambridge, MA: Harvard University Press.

63.

Tulchin Research and Strategic Consulting (2016), “California Statewide Survey Shows Strong Support for Proposition 61,” Tulchin Research Poll Results (July 28), https://tulchinresearch.com/wp-content/uploads/2016/08/California-RX-Drug-Pricing-362-A-PUBLIC-MEMO-7-16-Final.pdf.

64.

Wagner

Tillmann

Lutz

Richard J.

Weitz

Barton A.

(2009), “Corporate Hypocrisy: Overcoming the Threat of Inconsistent Corporate Social Responsibility Perceptions,” Journal of Marketing, 73 (6), 77–91.

65.

Walker

Edward T

. (2014), Grassroots for Hire: Public Affairs Consultants in American Democracy. Cambridge, UK: Cambridge University Press.

66.

Wan

Echo Wen

Rucker

Derek D.

Tormala

Zakary L.

Clarkson

Joshua J.

(2010), “The Effect of Regulatory Depletion on Attitude Certainty,” Journal of Marketing Research, 47 (3), 531–41.

67.

Waterhouse

Benjamin C

. (2014), Lobbying America. Princeton NJ: Princeton University Press.

68.

Werner

Timothy

(2015), “Gaining Access by Doing Good: The Effect of Sociopolitical Reputation on Firm Participation in Policy Making,” Management Science, 61 (8), 1989–2011.

69.

Wonnacott

Thomas H.

Wonnacott

Ronald J.

(1990), Introductory Statistics, Vol. 5. New York: John Wiley & Sons.

70.

Alison Jing

Moorman

Christine

Qin

Vivian Yue

Rao

Akshay

(2020), “Four More Years: Presidential Elections, Comparative Mind-Set, and Managerial Decisions,” Academy of Management Journal, 63 (5), 1370–94.

71.

Zhao

Xiaoquan

Strasser

Andrew

Cappella

Joseph N.

Lerman

Caryn

Fishbein

Martin

(2011), “A Measure of Perceived Argument Strength: Reliability and Validity,” Communication Methods and Measures, 5 (1), 48–75.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.22 MB