Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Research findings on the effectiveness of framing marriage equality for lesbian and gay couples in terms of affect/love versus rights and equality. The study uses randomized field experiments to test the impact of these frames on support for marriage equality. The findings suggest that while shifting the framing from rights to love does not increase support, using an affective frame with a visual cue of a heterosexual couple reduces support. Videos emphasizing love may be more effective at generating support, but this could be due to other factors such as attractiveness or conformity.
Typology: Study notes
1 / 35
This page cannot be seen from the preview
Don't miss anything!
Everybody Wants Somebody to Love: Emotion, Rationality, and Framing LGBT rights
Brian F. Harrison Yale University^1 Melissa R. Michelson Menlo College
(^1) Prepared for delivery at the 2015 annual meeting of the Western Political Science Association, Las Vegas, NV. We would like to thank Ben Bishin, Andrew Flores, Tony Smith, and Ken Sherrill, for their helpful and insightfulcomments on earlier drafts. All errors, of course, remain our own.
Abstract Mass attitude change is a rare event, and numerous studies show that people tend to maintain the attitudes they currently have. Yet, recent years have a seen a dramatic and rapid shift in attitudes about same-sex marriage—quite possibly one of the largest shifts in public opinion ever recorded. Advocacy organizations trying to change hearts and minds on the issue of same-sexmarriage have used a variety of appeals in their campaign. Some campaigns have focused on rationality by highlighting the idea of equal rights for LGBT individuals; other campaigns have focused on emotion, showing loving lesbian and gay couples in committed long-term relationships. Are differences in the framing of advocacy campaign messages responsible for observed shifts in public opinion? To answer this question, we partnered with a nationally-recognized LGBT advocacy group to test the effect of exposure to a short video that described the desire of a same-sex couple to marry either to ensure their legal rights or because they love one another. We then further tested the affect and rationality frames with an online survey experiment. Overall, our findings suggest that framing marriage equality for lesbian and gay couples in terms of affect/love is no more effective at generating support than messages thatframe the issue as about rights and equality.
KEYWORDS: Love, family, framing, LGBT rights, same-sex marriage, experiment
researchers have increasingly attempted to understand the psychological processes underlying framing effects (Iyengar 1991; Zaller 1992; Nelson, Clawson, and Oxley 1997; Chong and Druckman 2007). Druckman (2001: 1042) notes that “a framing effect is said to occur when, in the course of describing an issue or event, a speaker’s emphasis on a subset of potentially relevant considerations causes individuals to focus on these considerations when constructing their opinions.” In other words, framing effects occur when differences in the presentation of an issue or event changes one’s opinion about that issue or object; it is a process by which people orient their thinking (i.e. conceptualize) towards an object. This conceptualization process can have a substantial impact of the expected effects of messaging. Chong and Druckman (2007) present frames in an expectancy model in which different considerations towards an object are weighted by different considerations about that object. This leads to the individuals’ frame in thought. Frames in thought are often influenced by frames in communication , which is information (generally but not exclusively in the mass media) that attempts to alter the emphasis on different considerations toward an object. Thus, a framing effect is produced when a frame in communication influences an individuals’ frame in thought and subsequently their evaluation toward the attitude object. There is evidence from a variety of experiments, surveys and case studies that framing influences attitudes and behaviors across a wide variety of political and social issues. The seminal political framing study by Nelson, Clawson, and Oxley (1997) shows that framing a KKK rally as a free speech event garnered significantly more support and tolerance for the rally compared to framing it as a public disruption. Similarly, Sniderman and Theriault (2004) show that 85% of respondents support allowing a hate-group rally when a free-speech frame is emphasized whereas support drops to 45% when the issue is framed in terms of the potential for
violence. Similarly, Rasinki (1989) shows that support for government expenditures on welfare drops markedly when framed as “assistance to the poor” versus “welfare.” Other work focuses on policy areas like affirmative action (Gamson and Modigiani 1987); war (Dimitrova, et al. 2005); and stem cell research (Nisbett et al. 2003). Framing can also be an intentional endeavor; different groups can employ frames for mobilization and persuasion purposes (e.g., Snow & Benford 1992, Polletta & Ho 2006). Chong & Druckman (2007) underscore the importance of the issue framing because “discussion and debate over the appropriate frames for conceptualizing an issue lead ultimately to common perceptions and judgments about the consequences of a policy” (120). In other words, groups attempting to persuade or to advocate for a particular policy position often choose their frames carefully to dictate the terms and boundaries for a particular debate, with varying level of success. For example, Johnson (2012) finds that framing same-sex marriage in terms of equality drives opposition down, as does the degree to which equality frames are used in comparison to morality frames in the mass media. There are also important moderators of framing effects; Brewer (2003) shows that values moderate framing effects, finding that prior attitudes about gay rights shape evaluations and susceptibility to new frames.
LGBT Advocacy Groups & Recent Framing In 1996, only 27% of Americans supported marriage equality for same-sex couples; ten years later, support increased to 35%. More recent shifts have been more rapid, reaching majority support in various polls in 2013. The latest poll, from March 2015, pegs national support for same-sex marriage at 59 percent.^2 Many survey respondents openly admit to having changed their minds about marriage equality. For example, in a March 2013 Pew Research
(^2) http://blogs.wsj.com/washwire/2015/03/09/support-for-gay-marriage-hits-all-time-high-wsjnbc-news-poll/
families, and children in contrast to previous efforts that emphasized rational, rights-based appeals. In sum, LGBT rights organizations fighting for support for marriage equality have recently added affective frames to their messaging just as support for marriage equality has increased dramatically. Are these concurrent shifts a coincidence, or are affective frames more effective than rational frames? Scholarship on the power of affect suggests that they are generally more effective; we turn now to a discussion of that scholarship.
Message Persuasiveness: Affect and Rationality The use of emotional appeals by political elites and theorists is nothing new. Aristotle argued thousands of years ago in The Rhetoric that persuasive speech includes ethos (the speaker’s character), pathos (stirring the audience’s emotions), and logos (the logic of the argument). Brader (2006) quotes William James (1902): Emotional occasions… are extremely potent in precipitating mental rearrangements. The sudden and explosive ways in which love, jealousy, guilt, fear, remorse, or anger can seize upon one are known to everybody…. Andemotions that come in this explosive way seldom leave things as they found them.
Emotional messages dominate presidential television advertisements (Kaid and Johnston 2001) and have long been known to feature in political campaigns more generally (Lazarsfeld, Berelson and Gaudet 1944). Westen (2008) argues that the key to political success is to tug at voters’ emotions, not to rely on rational arguments or issue positions. Ramage and Bean (1998: 82) note that emotional appeals fall into pathos, the second of the three means of persuasion identified by Aristotle: An appeal to pathos causes an audience not just to respond emotionally but toidentify with the writer’s point of view—to feel what the writer feels…. Pathos thus refers to both the emotional and the imaginative impact of the message on an
audience, the power with which the writer’s message moves the audience to decision or action. Triadafilopoulos (1999) examines the powerful speakers of the Rev. Martin Luther King, Jr., noting that King’s rhetoric illustrates “the advantages of Aristotelian persuasion over rational deliberative and agnostic forms of public speech” (1999: 471). King succeeded because he recognized that “audiences judge claims to justice not only by their rationality, but also by their ability to touch the listener’s soul” (1999: 753). Young (1996) argues that communication is necessarily linked to desire, that “persuasion is partly seduction” (1996: 130). Other scholars view emotional appeals as central to political communication (Kern 1989; Perloff and Kinsey 1992). Critics of emotional appeals also consider them effective, even as they denounce them as manipulative and counter to democratic decision-making, or as likely to subvert rational thinking (Kamber 1997; Arterton 1992; Damasio 1994; Kinder 1994; Marcus 2000). Several scholars have recently employed field experiments to test the effectiveness of emotional appeals. Arceneaux (2012) finds that political messages that evoke loss aversion and fear are more likely to persuade. Brader (2005) uses experiments that demonstrate that cueing enthusiasm increases participation and activates partisan loyalties while cueing fear stimulates vigilance, increases reliance on contemporary evaluations, and facilitates persuasion (388). He notes that while existing work finds that citizens behave differently in different emotional states, it is less established whether political elites and persuasion campaigns can manipulate these emotions and change political attitudes and behavior as a result. In sum, for thousands of years, scholars examining the art of persuasion have recognized the power of emotional appeals designed to evoke a sense of shared identity with a speaker or writer of a persuasive message. The use of emotional appeals by LGBT rights organizations, often designed specifically to tug at listener’s heartstrings, are consistent with a long tradition of
receiving the message. It may be that individuals on the cusp of support for marriage equality— perhaps those who already support civil unions or who are questioning their beliefs about the issue—can be pushed over the line by an emotional appeal. Other individuals—perhaps those whose emotional responses to LGBT individuals tend towards disgust or whose anti-marriage equality opinions are more deeply held—will not be influenced by any change in framing. An alternative possibility is that prompting heterosexual Americans to think about love between LGBT people may prompt them to think about gay sex, which can trigger negative responses (the “ick” factor). Those opposed to same-sex marriage tend to maintain their beliefs primarily because they have an underlying, subconscious feeling of disgust at the thought of the sexual behavior between people of the same sex. Some research even suggests there are physiological changes when individuals view images they find disgusting, including exposure to LGBT individuals and pro-LGBT messages (Oxley, et al. 2008). In essence, rather than thinking of lesbian and gay individuals as people who love, it can be too personal, undermining the purpose of the messages (Rofes 1998; Mucciaroni 2008). In addition, highly salient, “easy” social issues should be the most resistant to attitude change, because attitudes are more clearly formed, limiting framing effects (Druckman & Leeper 2012). For all of these reasons, more rigorous empirical data are necessary to understand the effects of these appeals. In sum, there are several prominent examples of the choice to frame gay and lesbians’ ability to marry in terms of love and emotion and as an abstract and rational attempt to gain rights; however, an empirical test of framing effects has not been done. Is the recent shift to framing marriage equality in terms of love and family really more effective? Individuals may be more supportive of marriage equality when the issue is presented as about love, commitment and
family because that frame evokes a positive emotional response^5 , compared to when the issue is presented as about the dispassionate values of rights and equality. On the other hand, the “ick” factor may mean that a focus on love and family may have a backlash effect, causing decreased support for same-sex marriage. Given the demonstrated power of issue framing and the prevalence of existing frames in the battle over same-sex marriage in the real world, additional empirical evidence is needed to unpack the impact of these frames. Specifically, what are the effects of rational frames compared to emotional frames? Are the presently popular affective frames (which focus on love) more persuasive than rationality frames (which focus on rights)? Here, we leverage the unique properties of randomized experiments to contribute to the framing effects of the messages commonly used by advocacy organizations, with a specific focus on frames focusing on affect compared to those geared toward rationality.^6 We focus on the effectiveness of those messages in order to evaluate the persuasiveness of these real-world, contemporary discussions occurring in American politics and issue campaigns.
Experimental Design and Results Over the last decade, randomized experiments have experienced a renaissance in the social sciences and are considered the gold standard for impact assessment and program evaluation (Rossi, Freeman, and Lipsey 2003; Gerber and Green 2012). Generally, experiments in political science randomly assign individuals to receive a treatment expected to generate a change in attitude or behavior while others are randomly assigned to receive a placebo treatment
(^5) Parallel work has found that evoking feelings of disgust reduces support for marriage equality (see Adams, Stewart, and Blanchar 2014). (^6) Note that we chose these messages based on what advocacy organizations are using in their campaigns, not because of any underlying assumption about the appropriateness of messaging in the debate over LGBT rights. Ourintent is to maximize external validity while maintaining a high level of internal validity.
experiment were recruited using Amazon.com’s Mechanical Turk.^8 Participants were compensated $.25 for their participation in the experiment, consistent with MTurk market rates. Participants in the experiment were assigned to watch one of five short videos, each about a minute and a half long. One video was about recycling, serving as the baseline for the four treatment videos. In other words, we assume that individuals who were randomly assigned to watch the video about recycling are not persuaded by that video to change their attitude about marriage equality. The rate of support for marriage equality among that group can thus be compared to rates of support among individuals randomly assigned to watch one of the Lambda Legal videos. Because viewers are assigned randomly to the videos, any observed differences in rates of support for marriage equality can then be attributed to the viewing of the Lambda Legal video. The initial survey question asked respondents to choose a response (of A through E) to be randomly assigned a public policy video to view. While no such randomization took place, the question was designed to lessen social desirability and issue priming so respondents did not assume it was a survey about same-sex marriage.^9 Respondents were randomized, however, to one of five videos: the control (a video about the importance of recycling) and four treatment videos depicting a same-sex couple speaking about the importance of marriage equality: (1) a lesbian couple without children; (2) a gay male couple including children; (3) a gay male couple without children; and (4) a lesbian couple including children.^10
collection. Final results were sent to the organization as a courtesy but our findings were not altered in any way. Thestudies were also approved by the Institutional Review Boards at the authors’ home institutions. (^8) Others have noted the limitations and challenges associated with using MTurk workers for social science research (see Berinsky, Huber, and Lenz 2012; Casler, Bickel, and Hackett 2013). (^9) Arceneaux and Johnson (2013) also utilize this type of design not to control for social desirability but to gain leverage on selection bias and effects. (^10) Recall that the videos were produced by Lambda Legal prior to the planning for this research and thus the real families depicted sometimes have children included and sometimes do not. This sets up two embeddedmanipulations in the videos aside from the focus on affect or rationality: the gender of the same-sex couple, and the
After watching their randomly assigned video, participants were asked about their attitudes about several public policy issues like smoking and energy conservation. Within this battery, we included two questions about their position on LGBT rights: their support for same- sex marriage and their likelihood of voting for a hypothetical ballot initiative allowing same-sex marriage. No manipulation check was included, although the videos played automatically and the survey could not be advanced until enough time elapsed for the video to play in its entirety; thus, responses are most accurately considered intent-to-treat effects rather than average treatment effects. The videos differed in terms of the genders of the featured couples and whether or not they had children, but the crucial difference for our purposes was the degree to which the appeals were emotional. Two of the videos, #3 (the gay male couple with no children) and #4 (the lesbian couple with children), are more dispassionate. In Video #3, Tim and Ken mention their love for each other, but the bulk of the video is about their love for their home and about the legal protections they’ve put in place to support their relationship. Tim notes, “It’s important that our relationship is given the same respect, given the same privilege, given the same honor.” There is little visible love in the video, aside from one man’s hand on his partner’s thigh, although they are clearly a committed couple. Video #4, depicting the lesbian couple with children, is similar in its approach. Karen complains that the intake form at her new dentist’s office, where it asks for marital status, doesn’t include a box for civil unions. Later in the video, her partner Marcye notes, “When you have a little kid and you’re trying to explain to them why you’re not allowed to get married, that conversation is about injustice.” Again, as in Video #3, the couple is clearly in a committed long-term relationship and they care for each other very
presence or absence of children. We test for effects of these manipulations in the analyses below but find that thepresence or absence of children does not affect support for marriage equality. Screenshots and transcripts of the videos are in Figure 3.
different ways. Not all of these differences (variables) are controlled for in this experiment and couldn’t possibly be; the videos accurately test the sorts of messages being used by real-world advocacy organizations and how they are using true-life stories to fight for marriage equality but do not allow for rigorous testing of the hypothesis that a love frame is more persuasive than a rights frame. A total of 770 completed surveys were collected over a period of 17 days, from May 21 to June 4, 2012, with 146-169 individuals viewing each of the five videos (Table 1). Participants in each of the five conditions were relatively similar but as with all Mechanical Turk samples participants were not a representative sample of the U.S. adult population, with oversamples of Democrats, youth, and white non-Hispanics (see Table 1A in the Appendix for a randomization check).^11 [Table 1 and Figure 1 about here] For both dependent variables, support for marriage equality and willingness to vote yes on a marriage equality ballot initiative, support is consistently stronger among those cued with a Lambda Legal video, compared to the recycling video, confirming that individuals watched the videos and that they were influenced by the content. The most compelling are Videos #1 and #2, the videos that focus on love and commitment. Less compelling are Videos #3 and #4, which focus more on rights and being treated equally. These results suggest that framing marriage equality as an issue of love and commitment, rather than in a legal, dispassionate way, is somewhat more compelling. [Table 2 and Figure 2 about here]
(^11) Of those 770 respondents, 75 identified as a member of the LGBT community. However, while our LGBT respondents were more supportive of LGBT rights than the larger pool, they were not universally supportive: 75percent said that they supported same-sex marriage and 80 percent said that they would vote in favor of a state ballot initiative allowing same-sex marriage. Thus, we retain them in the results presented here. Analysis excluding those75 individuals generates very similar results, albeit with smaller Ns for each condition.
We then collapsed observations for Videos #1 and #2 into one category, and Videos # and #4 into another (Table 2). Support for marriage equality is strongest among those viewing the videos cueing love, by a difference of more than 7 percentage points compared to the other treatment videos—a large and statistically significant difference. Smaller differences, but in the hypothesized direction, are generated for the dependent variable item asking about a hypothetical ballot measure. [Figures 1 and 2 about here] These differences persist in a multivariate analysis controlling for other characteristics of the MTurk workers, including age, gender, education, partisanship, voter registration (reported), and race. Respondents shown the videos focusing on affect/love were more likely to say that they supported marriage equality than respondents shown the videos focusing on rights. In addition, support for marriage equality was stronger among younger respondents, women, Democrats, and white respondents (Table 3). [Table 3 about here]
While the differences from the MTurk experiment are large and statistically significant, the many moving parts in these videos mean that we cannot be certain that the differences are driven by the affect vs. rationality frames. To complement these videos with more controlled and internally valid experimental designs, we conducted a second experiment to more robustly test our theory.
Experiment #2: Survey Experiment In order to better focus on the effect of shifting the frame from rights to love, in September 2014 we conducted a randomized survey experiment on the Google Consumer Survey platform. Participants were exposed to a sentence that either described marriage equality as about
Overall, of the 3,034 individuals in our survey, 62.5 percent (1,896/3,034) said that they support same-sex marriage, a higher percentage than that obtained in most public opinion surveys. However, support did vary widely. Respondents viewing a heterosexual couple were the least likely to voice support for marriage equality (60.18%); respondents viewing the same-sex couples were more supportive (63.18% for those shown a gay male couple, 64.14% for those shown a lesbian couple). Here, we examine changes in support between matched pairs of sets of photos—those with an affective (love) prime and those with a rationality (rights) prime.^13 [Table 4 about here] In these studies, we find little evidence that the frame matters. Comparing responses to identical photos of lesbian and gay male couples but varying only the wording accompanying those photos generates negligible differences in support for marriage equality, as shown in Table
(^13) Differences in responses to those shown photos with and without children were negligible; thus, we collapse those categories in our analysis.
Discussion and Conclusion Marketing experts have known for years that appeals to emotion are more effective at generating sales of a product than are appeals to logic, and emotional appeals dominate television advertisements used in political campaigns. In a similar vein, recent campaigns by advocacy organizations including Human Rights Campaign and Freedom to Marry have focused on love and commitment, shifting away from an earlier focus on the legal rights of lesbian and gay individuals, in hopes that such a shift would increase the persuasiveness of their messages. Our experimental data provide little evidence of the effectiveness of that framing shift. Videos that emphasize love may be more effective at generating support for marriage equality than those that emphasize rights and equality, but those videos differ in many ways, and the persuasiveness of the love frame videos in our national MTurk experiment may be due to some other aspect of those videos that also varies, such as the attractiveness of the individuals or their conformity to expectations about lesbian and gay couples. Our survey experiments that varied only the frame and kept constant the presentation of the lesbian or gay couple found no evidence that the love frame was more effective. In fact, there may be a downside to humanizing lesbian and gay relationships as about love. It is possible that frames about loving same-sex couples may generate backlash against rather than support for marriage equality. There’s a so-called “ick” factor. If emphasizing the love and commitment of same-sex couples reminds heterosexual couples of their aversion to same-sex displays of affection (if not sex), then affective appeals may in fact reduce support for marriage equality and instead reinforce a more traditional and heteronormative conception of relationships. Further research is needed to better understand when and for whom affective appeals are more effective, or if advocacy organizations would be better served by a return to
