Regulatory Fit and Persuasion, p. Regulatory Fit, Rhetorical Strategy, and Political Persuasion Christopher D. Johnston Department of Political Science Duke University Durham, NC 27708 christopher.johnston@duke.edu Howard Lavine Department of Political Science University of Minnesota Minneapolis, MN 55455 lavine@umn.edu Paper prepared for the 2012 annual meeting of the Midwest Political Science Association, April 12th-15th, Chicago, IL. 1 Regulatory Fit and Persuasion, p. 2 Abstract We argue that citizen dispositions influence the relative success of different rhetorical strategies for persuasion, and thus function to constrain the behavior of elite actors in their attempts to win support for favored outcomes. Drawing on the psychological concept of regulatory fit, we propose that recent work on psychological dispositions and political preferences can be understood in terms of functional matches between the regulatory focus of citizens (prevention or promotion) and the framing of policies in terms of losses and gains (e.g. Cesario et al. 2004; Higgins 1998). In two experiments that rely on several policy domains, we find that persuasion is facilitated when a policy is framed so that it matches the chronic regulatory focus of the individual. We also provide some support for the idea that regulatory fit orientations constrain the advertising strategies of political campaigns. We conclude by considering the implications of these results for elite campaign strategy, and for the potential to micro-target persuasive communications on the basis of personality. Regulatory Fit and Persuasion, p. 3 Fully understanding the political importance of framing requires considering the two-way street of mutual interaction and influence among elites and mass publics. The common understanding of framing as a top-down, elite-driven process should be replaced with a more complex view whereby the beliefs and experiences of citizens affect what will resonate with them. - Smith (2009) A spate of recent research has shown that citizens’ stable psychological dispositions (e.g. traits, goals, needs) exert a substantial influence on their political preferences, identifications, and ideologies (e.g. Amodio et al. 2007; Carney et al. 2008; Hetherington and Weiler 2009; Gerber et al. 2010; Jost et al. 2003; Mondak 2010; Settle et al. 2010). In a widely-read theoretical integration of this literature, Jost, Glaser, Kruglanski and Sulloway (2003) argue that people embrace political conservatism because it serves a variety of motivational needs, including the reduction of “fear, anxiety, and uncertainty; [the avoidance of] change, disruption, and ambiguity; and to explain, order, and justify inequality among groups and individuals” (p. 340). From this point of view, political conservatism is simply a special case of motivated social cognition. In this research, we consider the broader political implications of such influence. Specifically, we argue that individual psychological dispositions constrain elite behavior by moderating the relative success of different rhetorical strategies on mass attitude change. Drawing on the concept of regulatory fit (Higgins 2005), we hypothesize that associations between psychological dispositions and political preferences emerge as a function of “matches” between the chronic regulatory focus of the individual (prevention or promotion), and the framing of policies in terms of losses or gains (Cesario et al. 2004; Higgins 1998). When the message frame matches the individual’s regulatory focus, persuasion is enhanced. This suggests the potential for elites to construct issue frames to fit the dispositional characteristics of an intended audience, or, in other words, to engage in personality-based micro-targeting of political Regulatory Fit and Persuasion, p. 4 messages. Such dynamics also imply a more powerful role for citizens in politics by suggesting one way in which they may be more than vessels to be filled with elite opinion. The remainder of the paper proceeds as follows: In the next section, we briefly touch on recent studies linking psychological dispositions and political preferences, and suggest that this work can be conceptually understood within the auspices of Higgins’ (1998) regulatory focus framework. We then outline the concept of regulatory fit, and discuss its implications for political persuasion. We test this framework with two experiments and aggregate (state-level) data, and conclude with a discussion of the broader implications of the theory and empirical findings. Dispositional Influence on Political Preferences Research on the role of stable individual differences on political preference draws on several subfields in psychology, including personality (e.g. Carney et al. 2008; Gerber et al. 2010; Mondak 2010), social cognition (e.g. Hetherington and Weiler 2009; Jost et al. 2003), cognitive neuroscience (e.g. Amodio et al. 2007; Oxley et al. 2008), and behavioral genetics (e.g. Alford, Funk and Hibbing 2005; Settle et al. 2010). Within personality psychology, the strongest effects on ideology are consistently found for the trait of openness to experience (which refers to a preference for novelty and stimulation versus the routine and familiar). Within the social cognitive literature, the need for non-specific cognitive closure (Kruglanski and Webster 1996; an indicator of aversion to ambiguity and need for certainty), is a strong predictor of conservative leanings (e.g. Chirumbolo et al. 2004; Thorisdottir and Jost 2011; Van Hiel, Pandelaere and Duriez 2004). Neuroscientific studies have linked conservatism to both cognitive rigidity (i.e. deficits in conflict monitoring; Amodio et al. 2007) and sensitivity to threatening visual and Regulatory Fit and Persuasion, p. 5 auditory stimuli (Oxley et al. 2008; Smith et al. 2011). And within behavioral genetics, Settle et al. (2010) have identified a specific gene associated with novelty seeking, a trait linked to liberalism (Carney et al. 2008; Mondak 2010). We believe it may be heuristically useful to subsume these empirical findings within a single theoretical framework that preserves their commonalities, but offers more straightforward guidance on the relevance of psychological predispositions to elite strategic behavior. Specifically, we adopt Higgins’ (1998) regulatory focus theory (hereafter, RFT), which holds that individuals chronically differ in the strategies by which hedonic principles (seeking pleasure and avoiding pain) are employed in the self-regulation of affect, cognition, and behavior. According to RFT, people either prefer a positive strategy focused on achieving gains (i.e., a promotion focus) or a negative strategy focused on avoiding losses (i.e., a prevention focus). Promotion-focused individuals attempt to maximize “hits” and minimize errors of omission, whereas prevention-focused individuals attempt to maximize correct rejections and minimize errors of commission.1 In this sense, promotion-focused individuals are oriented toward the achievement of maximal goals such as hopes, wishes, and aspirations (i.e. ideals), while prevention-focused individuals are oriented toward the achievement of minimal goals such as the attainment of security and the fulfillment of duties and obligations (i.e. oughts). The commonalities between regulatory focus and the work reviewed above should be readily apparent. A promotion-focused strategy entails openness to error and failure, comfort with risk-reward tradeoffs, and a relatively low need for security. Conversely, a preventionfocused strategy entails sensitivity to negative outcomes, an aversion to risk and loss, and a 1 It is important to note that regulatory focus is a general psychological principle, and thus operates both as a stable individual difference and, potentially, as a contextually-induced state (Higgins 1999). We see this as an additional advantage of this theoretical language, in contrast to, say, the “Big Five” which focuses on constructs unique to personality psychology. Regulatory Fit and Persuasion, p. 6 relatively high need for security. The greater concern among prevention-focused (versus promotion-focused) individuals with oughts (duties, obligations) is also consistent with recent work indicating that conservatives ascribe greater importance than liberals to the moral principles of in-group loyalty, submission to legitimate authority, and the protection of culturally-evolved institutions (Haidt 2012; Haidt and Graham 2007).2 Regulatory Fit and Persuasion Within the persuasion context, regulatory focus at the individual level operates by moderating the appeal of messages framed in terms of either the gains to be achieved from a recommended action, or losses to be incurred from not taking the action (see also Kahneman and Tversky 1979). Promotion-focused individuals are characterized by eagerness for gains and advancement, while prevention-focused individuals are characterized by their vigilance against losses. As Higgins (1998: 31) explains, “individuals with a strong promotion focus should be more motivated by incentives that are relevant to goals of accomplishment, whereas those with a strong prevention focus should be more motivated by incentives that are relevant to goals of safety” (p. 31). Shah, Higgins and Friedman (1998) provided empirical support for RFT by examining success in solving anagram problems conditional on chronic regulatory focus (promotion or prevention) and the frame of a monetary incentive. In one condition, participants were informed that they could earn a dollar (gain frame) if they solved 90 percent of the puzzles, while in a 2 Research also suggests that regulatory focus is strongly associated with asymmetrical activation in the frontal cortex (FC). Specifically, prevention-focused individuals show chronically higher levels of activation in the right FC relative to the left, and vice versa (Amodio et al. 2004). Asymmetric activation of the right FC is associated with a preference for the familiar and secure, sensitivity to threatening imagery, behavioral inhibition, and avoidancerelated emotions. Asymmetric left FC activation, by contrast, is associated with lesser sensitivity to threatening stimuli, behavioral activation, and approach-related emotions (see Harmon-Jones 2007 for a review). Regulatory Fit and Persuasion, p. 7 second condition participants were told they could avoid losing a dollar (loss frame) by not missing more than 10 percent of the puzzles. The authors found that participants with a chronic prevention focus performed better when exposed to the loss-focused monetary frame, whereas those with a chronic promotion focus performed better when exposed to the gains-focused frame. Cesario et al. (2004) extended RFT to persuasion by framing a message about an afterschool program either in terms of achieving gains (e.g. advancing children’s education) or preventing losses (preventing more children from failing). They found that attitude change in the direction of the message was enhanced when the frame of the message matched participants’ chronic regulatory focus. Work on persuasion in the political context similarly supports this “functional matching” hypothesis (Lavine and Snyder 1996; Lavine et al. 1999). For example, Lavine et al. (1999) found that citizens who are highly sensitive to threat (i.e. high authoritarians) were more persuaded by a get out the vote message when it was framed in terms of potential losses from non-voting than when it was framed in terms of the gains to be achieved by voting. By contrast, citizens who are low in threat sensitivity (i.e., low authoritarians) responded more favorably to a GOTV message when it was framed in terms of gains (than losses). In related work, Kam and Simas (2010) found that dispositionally risk-averse individuals showed greater loss aversion in a variety of prospect theory games. All of these results are consistent with Jost, Federico and Napier’s (2009) hypothesis that linkages between stable psychological dispositions and political preferences occur through a functional match between the motives and goals of the former and the instrumental and symbolic content of the latter. Regulatory Fit and Persuasion, p. 8 The Present Studies In the present research, we examine whether the effectiveness of persuasive communications on policy proposals depends simultaneously on (1) whether the policy frame highlights gains to be achieved or losses to be avoided; and (2) the chronic regulatory focus of the individual with respect to promotion or prevention. In particular, we hypothesize that: Those with a chronic prevention focus will be more persuaded by political communications focusing on potential losses from failure to enact a policy than potential gains from enacting a policy (H1). Those with a chronic promotion focus will be more persuaded by political communications focusing on potential gains from enacting a policy than potential losses from failure to enact a policy (H2). If these hypotheses receive support, individual differences in chronic regulatory focus may act to constraint the capacity of elites to move public opinion in electoral contests. In particular, if such dispositions cluster in political space, then elite actors have incentives to target their political advertising strategies to the predominant psychological profile of a given media market. In this paper, we examine the hypothesis that state-level differences in authoritarianism (Hetherington and Weiler 2009; Lavine et al. 2002; Stenner 2005) act to create variation in the extent to which political campaigns utilize positive- vs. negatively-toned advertisements – i.e., those that highlight gains or losses – in House, Senate, Gubernatorial and Presidential elections. If, as research amply shows (e.g., Lavine 2002), high authoritarians are more sensitive to threat than their less authoritarian counterparts (and thus more likely to be chronically focused on losses than gains), political campaigns should run more negatively-toned (i.e., prevention-focused) advertising in high authoritarian states, and more positively-toned (i.e., promotion-focused) advertising in less authoritarian states (H3). Regulatory Fit and Persuasion, p. 9 Persuasion Studies Two experiments were conducted to test the hypothesis that the effectiveness of persuasion depends on the conjunction of policy framing and individual differences in regulatory focus. The main study consists of 907 respondents collected through Amazon.com’s Mechanical Turk interface (see Appendix A for a description of the experimental procedure and descriptive statistics). A small college-student replication was conducted with 89 undergraduates at Duke University. The two studies are identical in design; we thus present them together. In both studies, respondents were presented with policy statements on eight distinct issues. For each issue, they were randomly assigned to receive one of three versions of a policy frame: control (no frame), gains frame, or losses frame (each respondent thus participated in eight separate experimental manipulations, and could be assigned to any of the three conditions for any issue). The order of presentation of the issues was randomized across respondents. In a pre-treatment survey, respondents completed demographic items, those pertaining to general political knowledge, and a scale of six items measuring chronic regulatory focus. Policy Treatments and Dependent Variable Respondents were presented with eight policy statements representing a range of ideological positions: government aid to help homeowners refinance underwater mortgages, tax cuts for large businesses, increased spending on infrastructure, greater funding for charter schools, required electronic verification of the citizenship of potential workers, greater spending on public services, restrictions on foreign imports, and expiration of the Bush tax cuts for wealthy individuals. For each issue, respondents were asked the following: “To what degree do you support or oppose the following policy” (strongly, support somewhat, support a little, oppose Regulatory Fit and Persuasion, p. 10 a little, oppose somewhat, or oppose strongly)? As described above, respondents were randomly assigned to receive one of three versions for each issue: a gain frame, a loss frame, or no frame. The exact wording for all issues and conditions is shown in Table 1. As the table shows, the frame manipulations are quite subtle, amounting to a change in a few words for each issue. The dependent variable for all analyses reported below is policy support, which ranges from 1 to 6, with higher values indicating greater support for the proposed policy. Table 1. Policy Treatments Issue Control Refinance Federal aid to struggling homeowners to refinance their mortgages at the current, lower rates Tax Cuts Additional tax cuts to large businesses Infrastructure Increased investments in our country’s infrastructure (e.g. roads, bridges) Charters Increased funding for charter schools (privately-owned, but publically-funded high schools that would compete with current schools) E-Verify Requiring all businesses to electronically verify the legality of their workers’ citizenship status Services Imports Bush Cuts Increased investments in public services such as teachers, cops and firefighters Increased restrictions on foreign imports Allowing the Bush tax cuts to expire for citizens making more than $250,000 per year Promotion Prevention Federal aid to struggling homeowners to refinance their mortgages at the current, lower rates to help them achieve financial stability Additional tax cuts to large businesses to promote the hiring of additional workers Increased investments to improve the quality of our country’s infrastructure (e.g. roads, bridges) Increased funding for charter schools (privately-owned, but publically-funded high schools that would compete with current schools) to incentivize greater achievement in our school systems Requiring all businesses to electronically verify the legality of their workers’ citizenship status to promote the hiring of U.S. citizens Federal aid to struggling homeowners to refinance their mortgages at the current, lower rates to prevent them from losing their homes Additional tax cuts to large businesses to prevent firing of additional workers Increased investments to prevent the deterioration of our country’s infrastructure (e.g. roads, bridges) Increased funding for charter schools (privately-owned, but publically-funded high schools that would compete with current schools) to prevent further deterioration of our school systems Requiring all businesses to electronically verify the legality of their workers’ citizenship status to prevent the hiring of non-U.S. citizens Increased investments in public services to prevent layoffs of teachers, cops and firefighters Increased restrictions on foreign imports to protect manufacturing from global competition Allowing the Bush tax cuts to expire for citizens making more than $250,000 per year to prevent a larger budget deficit Increased investments in public services to promote hiring of teachers, cops and firefighters Increased restrictions on foreign imports to promote domestic manufacturing in global competition Allowing the Bush tax cuts to expire for citizens making more than $250,000 per year to achieve a smaller budget deficit Regulatory Fit and Persuasion, p. 11 Regulatory Focus Respondents’ chronic regulatory focus (RF) was assessed with six items adapted from Lockwood, Jordan and Kunda (2002). These items appear first in the survey; thus, a gap was maintained in both time and thought between responses to the RF questions and the presentation of the policy items. For each RF item, respondents were asked, “Now we will briefly describe some people. Please indicate for each description whether that person is very much like you, like you, somewhat like you, a little like you, not like you, or not at all like you.” The six items included: “In general, I am focused on preventing negative events in my life,” “I am anxious that I will fall short of my responsibilities and obligations,” “Overall, I am more oriented toward preventing losses than I am toward achieving gains,” “In general, I am focused on achieving positive outcomes in my life,” “Overall, I am more oriented toward achieving success than preventing failure,” and “I frequently imagine how I will achieve my hopes and aspirations.” Greater agreement with the first three items indicates a relative prevention focus, and greater agreement with the last three items indicates a relative promotion focus. To remove variance associated with idiosyncratic use of the common measurement instrument, we calculated respondent-specific means across the items, and subtracted these from each of the six responses. We used these mean-deviated responses to construct a composite RF scale (α = .65 for the MTurk sample and .62 for the student sample). The scale was recoded to range from zero to one prior to analysis with higher values indicating a greater prevention focus. Analysis and Results In each study, we regressed policy support on two dummy variables representing membership in the two treatment conditions (with the unframed control group as the excluded Regulatory Fit and Persuasion, p. 12 category), RF, and the interaction of each treatment dummy with RF, using the OLS estimator. To account for autocorrelation (inherent to within-subject designs), we clustered standard errors by respondent.3 We expected to observe a negatively-signed interaction between RF and the gains frame dummy (indicating a greater persuasive effect of gains frames for promotion- than prevention-focused respondents), and a positively-signed interaction between RF and the loss frame dummy (indicating a greater persuasive effect of loss frames for prevention- than promotion-focused respondents). The statistical results for both studies are shown in Appendix B. Graphical depictions are shown in Figure 2a (MTurk Sample) and Figure 2b (undergraduate sample). The bars in each figure represent the estimated level of policy support for each experimental condition, separately for promotion- and prevention-focused respondents. The spikes represent 95% confidence bounds on these estimates. Results of MTurk Study. As Figure 2a shows, our hypotheses receive strong support in the MTurk study. Among respondents with a promotion focus (i.e., those who scored at the 10th percentile of the RF distribution), averaged policy support is highest in the frame highlighting gains to be achieved from acting on the proposed policies (M=4.34, vs. 4.12 and 4.23 in the control and loss conditions, respectively). By contrast, among respondents with a prevention focus (i.e., those scoring at the 90th percentile of the RF distribution), averaged policy support is highest in the frame highlighting losses to be avoided from enacting the proposed policies (M=4.56, vs. 4.22 and 4.36 for the control and gain conditions, respectively). The pattern of mean difference tests corresponds with expectations. For promotionfocused respondents, there was a significant effect for the gains treatment (vs. control; D=.21, 3 We replicated this analysis via restricted maximum likelihood, modeling the intercept as a function of individuallevel characteristics and a respondent-specific, normally distributed disturbance term. The results were largely identical. The findings also remain identical with or without clustering the standard errors. Regulatory Fit and Persuasion, p. 13 SE=.10, p<.05), but not for the loss treatment (vs. control; D=.11, SE=.10, n.s.). Conversely, for prevention-focused respondents, there was a significant effect for the loss treatment (vs. control; D=.34, SE=.10, p<.05), but not the gain treatment (vs. control; D=.14, SE=.10, n.s.). In substantive terms, framing a policy statement in terms of potential gains significantly heightened persuasion (i.e., policy agreement) among promotion- but not prevention-focused respondents, while framing a policy in terms of preventing potential losses significantly heightened persuasion among prevention- but not promotion-focused respondents. This is exactly the pattern predicted by theory. Results of Replication Study. We turn now to the results from the undergraduate study, which are shown in Figure 2b. As the figure shows, these results strongly replicate the main study’s findings. Among promotion-focused respondents, policy support was highest in the gains frame (M=4.41, vs. 3.90 and 4.11 in the control and loss conditions, respectively). By contrast, among prevention-focused respondents, policy support was highest in the loss frame (M=4.39, vs. 3.90 and 4.06 in the control and gain conditions, respectively). Once again, the pattern of mean difference tests corresponds with expectations. For promotion-focused respondents, the mean difference in policy support between the gains frame and the control condition is statistically significant (D=.52, SE=.23, p<.05), while the difference between the loss frame and the control condition is not (D=.21, SE=.24, n.s.). For prevention-focused respondents, the mean difference in policy support between the loss frame and the control condition is statistically significant (D=.47, SE=.23, p<.05), while the difference between the gain frame and the control condition is not (D=.16, SE=.23, n.s.). Regulatory Fit and Persuasion, p. Figure 2A. Policy Support by Condition and Regulatory Focus Prevention Focus (90%) 3.5 4 Policy Support 4.5 5 Promotion Focus (10%) Promote Control Prevent Promote Control Prevent Notes: Data from undergrad sample. Bars are estimated levels of policy support. Spikes are 95% CIs. Figure 2B. Policy Support by Condition and Regulatory Focus Prevention Focus (90%) 3.5 4 Policy Support 4.5 5 Promotion Focus (10%) Promote Control Prevent Promote Notes: Data from MTurk sample. Bars represent estimated policy support. Spikes are 95% CIs. Control Prevent 14 Regulatory Fit and Persuasion, p. 15 Dispositional Clustering and Elite Behavior The foregoing studies indicate that the effectiveness of a persuasive appeal depends on the functional match between the chronic regulatory focus of the individual and the frame of a message in terms of losses or gains. In this last empirical section, we extend our analysis to examine whether these findings have any import for the political advertising strategies of elites in Congressional, Gubernatorial, and Presidential elections. Specifically, if traits (e.g., RF, authoritarianism) are both geographically clustered and perceived to moderate the effectiveness of different types of ads (positive vs. negative tone), campaigns may choose to run a greater proportion of negatively-toned ads in states where levels of a prevention (vs. a promotion) focus (consciously or unconsciously understood) are perceived to be relatively high. State-Level Authoritarianism To obtain measures of state-level regulatory focus, we rely on four items used to measure authoritarianism (Feldman and Stenner 1997; Hetherington and Weiler 2009; Stenner 2005). These items are included in the 2000, 2004, and 2008 ANES studies, giving us enough data to generate reasonably efficient estimates of the state-level parameters. Our method of estimation is multilevel-regression and post-stratification (MRP; Gelman and Hill 2007; Lax and Phillips 2009; Park, Gelman and Bafumi 2006). MRP proceeds in two basic stages. First, individual-level responses to the variable of interest (in this case, authoritarianism) are modeled as a function of both individual and state-level predictors. The estimates from this model are then used to generate predicted values of the dependent variable for every possible combination of predictors (e.g., a white female between the ages of 18 and 29 with a high school degree residing in Oklahoma). These predicted values can then be utilized to estimate state-level opinion by Regulatory Fit and Persuasion, p. 16 weighting the predicted value for each combination of predictors by the actual percentage of people residing in a state with those characteristics, and then summing across these weighted estimates (i.e. the post-stratification in MRP). This is the basic procedure we use to generate state-level authoritarianism estimates. First, we pooled the data from each of the ANES studies. Our individual-level predictors included age (recoded into 4 categories, 18-29, 30-44, 45-64, and 65-up), the intersection of race and gender (six categories corresponding to the intersection of male/female with black, Hispanic and other), and education (four categories: high school degree or less, some college, college degree, post-graduate degree). At the state level, we included the following predictors: per capita income, proportion of the state’s vote that went to George W. Bush in 2004, a scale measuring the multiculturalism of the state (XXXX), the % of the state’s population living in urban areas, the state’ region, and the % of the state’s population with a high school diploma. The overall model was estimated as follows (Y=Authoritarianism): π΄ππ π πππ,πΊπππππ πΈππ’π ππ‘ππ‘π ππππ π¦π = π½ + πΌπ[π] + πΌπ[π] + πΌπ[π] + πΌπ[π] + πΌπ[π] , π΄ππ 2 πΌπ[π] ~π(0, ππ΄ππ ), for j = 1 to 4, π πππ,πΊπππππ 2 πΌπ[π] ~π(0, ππ πππ,πΊπππππ ), for k = 1 to 6, πΈππ’π 2 ), for l = 1 to 4, πΌπ[π] ~π(0, ππΈππ’π ππππ 2 ), for j = 1 to 3, πΌπ[π] ~π(0, πππππ π πππππ ππ‘ππ‘π πΌπ[π] ~π (πΌπ[π] π πππππ 2 + ∑5π=1 πΎπ ∗ π₯ππ , πππ‘ππ‘π ), where xi = State-Level Predictor ‘i’ 2 πΌπ[π] ~π(0, ππ πππππ ), for p = 1 to 5 Regulatory Fit and Persuasion, p. 17 To post-stratify, we utilize the 2000 Census’ 1% Public Use Micro-Data Sample, which gives estimates of the proportion of each citizen “type” within each state. The final state-level estimates of authoritarianism are calculated as: 4704 (1) ∑4704 π=1 ππ ∗ ππ / ∑π=1 ππ , where Nc is the number of individuals in cell “c” within a given state, and θc is the authoritarianism estimate for cell “c.” State-level authoritarianism estimates are plotted in Figure X (the authoritarianism scale ranges from zero to one). .75 .8 Figure X. State-Level Authoritarianism Estimates MS LA AR .7 WV KY SC AL GA .65 TN .6 ND NM OH .55 CT MA CO DC NH VT HI WI ID NV AZ CA MN IA ME MD NE NJ KS SD OK PA MO RI IN FL TX NC NY IL DE VA MI WY MT .5 AK UT OR WA As a “sanity-check” on these estimates, we examined the association between state-level authoritarianism and a recent measure of state-level “racial animus” derived from analyses of Google searches for racially charged words during the 2008 Presidential Election (StephensDavidowitz 2011). As Figure X indicates, there is a strong correspondence between the two state-level variables, suggesting that our measure of aggregate authoritarianism is not without a modicum of construct validity. Regulatory Fit and Persuasion, p. 18 .75 .8 Figure X. Relationship between State-Level Authoritarianism and Prejudice MS R2 = .56 AR AL SC .7 .65 GA .6 NM SD .55 ID HI CO .5 UT MN MT WY DC DE IL VA KSMD NE AZ ME WI NV IA CA CT NH MA OR AK KY WV TN NCFL TX IN MORI OK ND LA NY NJ OH MI PA VT .45 WA 20 40 60 State-Level Prejudice 80 100 Negativity of Campaign Advertising across States To measure strategic elite behavior related to rhetorical strategy, we utilize data from the Wisconsin Advertising Project at the University of Wisconsin (http://wiscadproject.wisc.edu/). We were able to obtain data for 2000, 2002 and 2008.4 For each year, the project codes every ad for House, Senate, Gubernatorial, and Presidential campaign with respect to “tone” (i.e., each ad is coded as negative, positive, or “contrast”). To estimate the overall negativity of campaign ads in each state (in each year), we divided the number of negative ads across all campaigns by the total number of ads for that year. Our measure of state-level campaign negativity is thus the proportion of all ads in a given state (within a given year) that were negative.5 4 The Wisconsin Advertising Project sent us a data disk for 2004, but it was unreadable. We have not yet received a replacement. 5 Not all states have data for every year, and thus our analyses are restricted to those with available ads. Regulatory Fit and Persuasion, p. 19 Analysis To evaluate our hypothesis, we present scatter plots of advertising negativity against state-level authoritarianism for each year. We include a loess curve for each plot, which gives a locally-weighted picture of the relationship between the two variables. As the panels in Figure X indicate, the results are mixed. In 2000, there appears to be a meaningful relationship between state-level authoritarianism and the proportion of negative ads in each state. The loess line slopes upward from about 15% negative ads to about 40% negative ads across the range of authoritarianism. This visual association is reinforced by a significant and substantively strong bivariate correlation (r=.42, p<.01). In 2002, the results are less clear, although there appears to be some suggestion of a positive relationship (r=.23, p<.15). In 2008, however, we find no clear evidence of any relationship between these variables (r=-.10, p>.45). .4 .3 .2 .1 0 Proportion of Negative Ads .5 Figure XX. Authoritarianism and Ad Negativity, 2000 Election .5 .55 .6 .65 State-Level Authoritarianism .7 .75 Regulatory Fit and Persuasion, p. .4 .2 0 Proportion of Negative Ads .6 Figure XX. Authoritarianism and Ad Negativity, 2002 Election .5 .55 .6 .65 State-Level Authoritarianism .7 .75 .7 .75 .6 .4 .2 0 Proportion of Negative Ads .8 Figure XX. Authoritarianism and Ad Negativity, 2008 Election .5 .55 .6 .65 State-Level Authoritarianism Discussion and Conclusions As psychologists have long noted, rationally equivalent descriptions of reality can be framed in terms of either gains or losses (e.g., 10% unemployment vs. 90% employment; Quattrone and Tversky 1982). Extensive work on prospect theory has shown that people are generally more sensitive to losses than gains, and that social judgments are often responsive to such framing effects (see Kahneman0 and Tversky 1979). Regulatory focus theory (RFT; Higgins 1998, 1999) builds on prospect theory by proposing that stable individual differences exist in the extent to which a person’s affect, cognition and behavior is motivated more by 20 Regulatory Fit and Persuasion, p. 21 achieving gains (a promotion focus) or avoiding losses (a prevention focus). In the present research, we relied on RFT to examine two questions about mass and elite politics: (1) Are individual differences in regulatory focus systematically related to the effectiveness of different types of policy frames? (2) Does geographic clustering in regulatory focus orientation constrain elite behavior in terms of campaign advertisement strategy? In two experiments that relied on eight different issues, we found that functional matches between individual dispositions and the framing of policy statements produce more persuasion than functional mismatches. Specifically, promotion-focused individuals were persuaded by the potential gains to be achieved from changes in public policy, whereas prevention-focused individuals were not. Equally, prevention-focused individuals were persuaded by the potential losses to be avoided by enacting a proposal, whereas promotion-focused were indifferent to such communications. Moreover, we found some evidence that aggregated clusters of RF orientations at the state level are linked to the degree of negativity in political advertising. In two of three years of elections examined, “prevention-oriented” states received a greater proportion of negative ads than “promotion-oriented” states. We believe that regulatory focus can be understood as an umbrella concept in personality with the potential to subsume a variety of dispositional factors that have been linked to political orientation. As Jost et al. (2003) explain: The promotion goals of accomplishment and advancement should naturally introduce a preference for change over stability, insofar as advancement requires change. The prevention goals of safety and security, on the other hand, should favor stability over change, to the extent that stability entails predictability and hence psychological security and control…To the extent that political conservatism is motivated, at least in part, by the desire for security and stability and the avoidance of threat and change, situations inducing a prevention-oriented regulatory focus might also induce a conservative shift in the general population. What is appealing about regulatory focus is its emphasis on the abstract concepts of approach and avoidance and pleasure and pain. Few principles in psychology are more basic than these Regulatory Fit and Persuasion, p. 22 distinctions, and they are critical to understanding human behavior in a variety of domains, including (as we have shown here), political persuasion. The fundamental nature of the distinction between gain and loss has also pervaded other subfields of political science, including the emphasis in political economy on risk aversion and political preference (e.g. Iversen 2005; Moene and Wallerstein 2001). The RF framework thus may prove to provide an integration point for two seemingly disparate literatures: the biological/psychological and the economic. That RF orientations moderate the effectiveness of different types of persuasion strategies has interesting implications for the strategies pursued by elites in policy and electoral contests. If dispositions cluster in political space, whether concrete (e.g., geographical) or contrived (e.g. opt-in networks such as TV and blogs), the opportunity for the microtargeting of political messages exists, and the present results suggest advantages in aggregate persuasion to such strategies. Future work should explore in greater depth the potential for elite micro-targeting on the basis of personality. Finally, our results speak to the literature on framing effects in political science. As indicated by the epigraph to this paper, the framing literature has largely ignored individual-level heterogeneity in the effectiveness of different types of frames. However, as citizens differ with respect to the types of goals – promotion vs. prevention – that are both accessible and important to them, frames focusing on gains vs. losses should be likely to attract differential levels of cognitive scrutiny and positive evaluation (Avnet and Higgins 2006; Freitas and Higgins 2002; Higgins 2005). A deeper understanding of the conditionalities of framing effects is therefore critical, and while recent work has begun to address this question (Druckman 2001; 2004; Kam and Simas 2010), much remains to be done. More generally, we hope that research on the joint effects of individual differences in abstract goal priorities and message framing leads to a greater Regulatory Fit and Persuasion, p. understanding of the ways in which elites and citizens both constrain, and are constrained by, one another. 23 Regulatory Fit and Persuasion, p. 24 APPENDIX A. DETAILS FOR MECHANICAL TURK EXPERIMENT Mechanical Turk is a service that Amazon.com provides to those who are interested in the performance of a large number of small tasks that require human judgment. Individuals performing the tasks create a user profile on the site and are paid a fee for their services that varies depending on the time, effort, and expertise involved in the task. Researchers post a human intelligence task (HIT) to the Amazon site and offer payment rate per HIT. A link to our survey was posted on this site, and advertised under the keywords “politics,” “survey,” “public policy,” and “2012 election.” The description of the task read, “You will be asked to answer several questions about yourself and watch and respond to a campaign ad.” Valid completions of the survey were compensated $1.00. The effective hourly rate was estimated to be $7.26. Summary statistics for the respondents are shown below. Median Age Male Black White Mean Political Interest (1-5) Mean Subjective Political Knowledge (1-5) Mean Partisanship (1-7, higher=Rep) Mean Ideology (1-7, higher=Con) Mean Importance of Religion to Self (1-5) Median Income Mean Income Median Education Unemployed or Temporarily Laid Off Student Working Now Central Tendencies and %s 28 60% 4% 81% 3.11 3.08 3.31 3.27 2.11 20K-30K 30K-40K Some College 13% 24% 53% Regulatory Fit and Persuasion, p. APPENDIX B. OLS REGRESSION ESTIMATES (MTurk and Undergraduate Studies) Variable Promotion Condition Prevention Condition Promote X RF Prevent X RF Regulatory Focus Constant MTurk .58 (.31) .16 (.34) -.53 (.64) .41 (.64) .02 (.41) 3.86 (.30) Undergrad. .22 (.13) .07 (.14) -.11 (.28) .35 (.28) .14 (.21) 3.88 (.11) Subjects Observations R2 F(7,86) and F(7,686) 87 694 .01 2.48 907 7247 .01 8.03 25 Regulatory Fit and Persuasion, p. 26 References Alford, John R., Carolyn L. Funk, and John R. Hibbing. 2005. “Are Political Orientations Genetically Transmitted?” American Political Science Review 99 (2): 153-167. Amodio, David M., James Y. Shah, Jonathan Sigelman, Paige C. Brazy, and Eddie Harmon Jones. 2004. “Implicit Regulatory Focus Associated with Asymmetrical Frontal Cortical Activity.” Journal of Experimental Social Psychology 40 (2): 225-232. Amodio, David M., John T. Jost, Sarah L. Master, and Cindy M. Yee. 2007. “Neurocognitive Correlates of Liberalism and Conservatism.” Nature Neuroscience 10: 1246-1247. Avnet, Tamar, and E. Tory Higgins. 2006. “How Regulatory Fit Affects Value in Consumer Choices and Opinions.” Journal of Marketing Research 43 (1): 1-10. Carney, Dana R., John T. Jost, Samuel D. Gosling, and Jeff Potter. 2008. “The Secret Lives of Liberals and Conservatives: Personality Profiles, Interaction Styles, and the Things They Leave Behind.” Political Psychology 29 (6): 807-840. Cesario, Joseph, Heidi Grant, and E. Tory Higgins. 2004. “Regulatory Fit and Persuasion: Transfer from ‘Feeling Right’.” Journal of Personality and Social Psychology 86 (3): 388-404. Chirumbolo, Antonio, Alessandra Areni, and Gilda Sensales. “Need for Cognitive Closure and Politics: Voting, Political Attitudes and Attributional Style.” International Journal of Psychology 39 (4): 245-253. Druckman, James N. 2001. “On the Limits of Framing Effects: Who Can Frame?” Journal of Politics 63 (4): 1041-1066. Druckman, James N. 2004. “Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects.” American Political Science Review 98 (4): 671-686. Ehrlich, Sean, and Cherie Maestas. 2010. “Risk Orientation, Risk Exposure, and Policy Opinions: The Case of Free Trade.” Political Psychology 31 (5): 657-684. Feldman, Stanley, and Karen Stenner. 1997. “Perceived Threat and Authoritarianism.” Political Psychology 18 (4): 741-770. Freitas, Antonio L., and E. Tory Higgins. 2002. “Enjoying Goal-Directed Action: The Role of Regulatory Fit.” Psychological Science 13 (1): 1-6. Haidt, Jonathan. 2012. The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Pantheon. Regulatory Fit and Persuasion, p. 27 Haidt, Jonathan, and Jesse Graham. 2007. “When Morality Opposes Justice: Conservatives Have Moral Intuitions That Liberals May Not Recognize.” Social Justice Research 20 (1): 98116. Harmon-Jones, Eddie. 2007. “Asymmetrical Frontal Cortical Activity, Affective Valence, and Motivational Direction.” In Social Neuroscience: Integrating Biological and Psychological Explanations of Social Behavior, Eddie Harmon-Jones and Piotr Winkielman (Eds.). New York: The Guilford Press. Hetherington, Marc J., and Elizabeth Suhay. 2011. “Authoritarianism, Threat, and Americans’ Support for the War on Terror.” American Journal of Political Science 55 (3): 546-560. Hetherington, Marc J., and Jonathan Weiler. 2009. Authoritarianism and Polarization in American Politics. New York: Cambridge University Press. Higgins, E. Tory. 1998. “Promotion and Prevention: Regulatory Focus as a Motivational Principle.” In Advances in Experimental Social Psychology, vol. 30, ed. Mark P. Zanna. San Diego: Academic Press. Higgins, E. Tory. 1999. “Persons and Situations: Unique Explanatory Principles or Variability in General Principles?” In The Coherence of Personality: Social-Cognitive Bases of Consistency, Variability, and Organization, Daniel Cervone and Yuichi Shoda (Eds.). New York: The Guilford Press. Higgins, E. Tory. 2005. “Value From Regulatory Fit.” Current Directions in Psychological Science 14 (4): 209-213. Iversen, Torben. 2005. Capitalism, Democracy, and Welfare. New York: Cambridge University Press. Gerber, Alan S., Gregory A. Huber, David Doherty, Conor M. Dowling, and Shang E. Ha. 2010. “Personality and Political Attitudes: Relationships across Issue Domains and Political Contexts.” American Political Science Review 104 (1): 111-133. Jost, John T., Christopher M. Federico, and Jaime L. Napier. 2009. “Political Ideology: Its Structure, Functions, and Elective Affinities.” Annual Review of Psychology 60: 307-337. Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. Sulloway. 2003. “Political Conservatism as Motivated Social Cognition.” Psychological Bulletin 129 (3): 339-375. Kahneman, Daniel, and Amos Tversky. 1979. “Prospect Theory: An Analysis of Decision Under Risk.” Econometrica 47 (2): 263-292. Kam, Cindy D., and Elizabeth N. Simas. 2010. “Risk Orientations and Policy Frames.” Journal of Politics 72 (2): 381-396. Regulatory Fit and Persuasion, p. 28 Kruglanski, Arie W. 1989. Lay Epistemics and Human Knowledge: Cognitive and Motivational Bases. New York: Plenum Press. Kruglanski, Arie W., and Donna M. Webster. 1996. “Motivated Closing of the Mind: ‘Seizing’ and ‘Freezing’.” Psychological Review 103 (2): 263-283. Kruglanski, Arie W., Donna M. Webster, and Adena Klem. 1993. “Motivated Resistance and Openness to Persuasion in the Presence or Absence of Prior Information.” Journal of Personality and Social Psychology 65 (5): 861-876. Lau, Richard R., Lee Sigelman, Caroline Heldman, and Paul Babbitt. 1999. “The Effects of Negative Political Advertisements: A Meta-Analytic Assessment.” American Political Science Review 93 (4): 851-875. Lavine, Howard, Diana Burgess, Mark Snyder, John Transue, John L. Sullivan, Beth Haney, and Stephen H. Wagner. 1999. “Threat, Authoritarianism, and Voting: An Investigation of Personality and Persuasion.” Personality and Social Psychology Bulletin 25 (3): 337-347. Lavine, Howard, and Mark Snyder. 1996. “Cognitive Processing and the Functional Matching Effect in Persuasion: The Mediating Role of Subjective Perceptions of Message Quality.” Journal of Experimental Social Psychology 32: 580-604. Lockwood, Penelope, Christian H. Jordan, and Ziva Kunda. 2002. “Motivation by Positive or Negative Role Models: Regulatory Focus Determines Who Will Best Inspire Us.” Journal of Personality and Social Psychology 83 (4): 854-864. Moene, Karl Ove, and Michael Wallerstein. 2001. “Inequality, Social Insurance, and Redistribution.” American Political Science Review 95 (4): 859-874. Mondak, Jeffery J. 2010. Personality and the Foundations of Political Behavior. New York: Cambridge University Press. Oxley, Douglas R., Kevin B. Smith, John R. Alford, Matthew V. Hibbing, Jennifer L. Miller, Mario Scalora, Peter K. Hatemi, and John R. Hibbing. 2008. “Political Attitudes Vary with Physiological Traits.” Science 321: 1667-1670. Settle, Jaime E., Christopher T. Dawes, Nicholas A. Christakis, and James H. Fowler. 2010. “Friendships Moderate an Association between a Dopamine Gene Variant and Political Ideology.” Journal of Politics 72 (4): 1189-1198. Shah, James, E. Tory Higgins, and Ronald S. Friedman. 1998. “Performance Incentives and Means: How Regulatory Focus Influences Goal Attainment.” Journal of Personality and Social Psychology 74 (2): 285-293. Smith, Kevin B., Douglas Oxley, Matthew V. Hibbing, John R. Alford, and John R. Hibbing. Regulatory Fit and Persuasion, p. 29 2011. “Disgust Sensitivity and the Neurophysiology of Left-Right Political Orientations.” PLoS ONE 6(10): 1-9. Stenner, Karen. 2005. The Authoritarian Dynamic. New York: Cambridge University Press. Thorisdottir, Hulda, and John T. Jost. 2011. “Motivated Closed-Mindedness Mediates the Effect of Threat on Political Conservatism.” Political Psychology 32 (5): 785-811. Van Hiel, Alain, Mario Pandelaere, and Bart Duriez. 2004. “The Impact of Need for Closure on Conservative Beliefs and Racism: Differential Mediation by Authoritarian Submission and Authoritarian Dominance.” Personality and Social Psychology Bulletin 30: 824-837.