Pathways of Global Cultural Diffusion: Mass Media and Stated Attitudes about Men Violence against Women

By: Jeffrey Swindle, Population Research Center, University of Texas at Austin, [email protected]

Revised and Resubmitted to the American Journal of Sociology

Appendices | Data & Code

Version: July 1, 2021

Abstract: Current theories of global cultural diffusion outline abstract mechanisms. I identify specific pathways through which public cultural scripts can flow to individual people, focusing he ead f ci ab e abe ad e hgh a edia i Maai. I begin with a mass media content analysis, relying on multiple sources including a new dataset of newspaper articles. I show that transnational organizations worked with Malawian journalists to dce ce cdeig e iece agai e, heea eeaie edia companies disseminated content portraying gender stereotypes. Linking this information to survey data, I find that exposure to these distinct types of content were divergently associated ih ee aed ejeci f e iece agai e. The e ie a multifaceted portrait of global cultural diffusion with diverse messaging streams simultaneously reaching individual people.

Keywords: gbaiai, ce, diffi, a edia, e iece agai e, Maai

Funding: I received research support from NIA training grant T32AG000221 and NICHD training grant T32HD007339 awarded to the Population Studies Center at the University of Michigan, as well as a NICHD training grant (T32HD007081) awarded to the Population Research Center at the University of Texas at Austin. I also gratefully acknowledge use of the services and facilities of the Population Studies Center at the University of Michigan funded by NICHD center grant P2CHD041028 and the Population Research Center at the University of Texas at Austin funded by NICHD center grant P2CHD041028. I further received research grants from the Department of Afroamerican and African Studies, Department of Sociology, African Studies Center, International Institute, Population Studies Center, and School of Natural Resources and Environment at the University of Michigan. The research content is solely my responsibility and does not necessarily represent the official views of any funders.

Acknowledgements: I thank Deirdre Bloome, Margaret Frye, Arland Thornton, and Kiyoteru Tsutsui for their detailed comments on earlier versions of this paper. I am grateful for advice I received from members of the Inequality, Demography, and Family working group at the University of Michigan, as well as Adam Ashforth, Cheyney Dobson, Emma Kaliya, Japhet Mchakulu, Limbani Moya, Henry Mphande, Florence Nkhoma, Limbani Phiri, Louisa Roberts, Sarah Seelye, Ann Swidler, Susan Watkins, Brady West, and many others. Introduction Much of the sociological literature regarding global cultural diffusion emphasizes the worldwide spread of public cultural scripts about societal development and human rights. Scholars identify numerous scripts promoted under the umbrella of “development,” such as the importance of economic growth for advancing social tolerance, the benefits of small families, and the notion of inalienable human rights (Ferguson 1994; Joas 2013; Merry 2006; Meyer et al. 1997; Thornton 2005). Many leaders at powerful transnational organizations seek to universally promote these scripts, including intergovernmental organizations like the United Nations and World Bank, international nongovernmental organizations (NGOs) such as CARE and Save the Children, transnational social movement organizations like Amnesty International, and foreign aid agencies like the United States Agency for International Development. They circulate these scripts across national governments worldwide, shaping their policies and, in some cases, their practices (Boli and Thomas 1999; Cole and Ramirez 2013; Hafner-Burton and Tsutsui 2005). Evidence suggests developmental scripts may also reach individual people, as substantial majorities of citizens in a diverse range of countries declare attitudes that match many of these scripts (Thornton et al. 2015). Moreover, scholars argue that people with exposure to key institutions that transnational organizations often influence, like education and mass media, are more likely to express such attitudes (Hadler 2017). In this literature, what people say their attitudes are is useful not because it necessarily indicates how people will act, but rather because it reveals global cultural diffusion at the individual level (Mohr et al. 2020:21-22, 34-43). The theorization and measurement of individuals’ exposure to public cultural scripts about development and human rights, however, often remains indirect. Several studies document that education institutions circulate developmental scripts in their curricula (e.g., Kamens et al. 1996; Russell et al. 2018), but mass media often is assumed to contain similar messaging without formal assessment of its content, a problem several scholars recognize (Boyle et al. 2002:26; Hadler 2017:178; Hendi 2017:235; Pierotti 2013:261; Thornton et al. 2015:290-292). Recent work highlights the problem with assuming long-theorized institutional sources of global cultural diffusion serve exclusively as conduits for human rights messaging, showing that illiberal and other actors can use similar strategies to transmit alternative scripts (Boyle et al. 2015; 2017a; Frank 1997; Schofer et al. 2019; Velasco 2020b). Accordingly, I examine the specific pathways through which transnational organizations and other powerful actors circulate public cultural scripts about gender relations through mass media. Scholarship consistently depicts mass media as one of the main theorized sources of global cultural diffusion (Ayoub and Garretson 2019; Boyle and Hoeschen 2001; Clark 2012; Essary 2007; Inkeles and Smith 1974), and transnational organizations use mass media campaigns to widely declare that men’s violent action toward women is a violation of human rights (Htun and Weldon 2012; Keck and Sikkink 1998, chapter 5; Merry 2006; Montoya 2013). Mass media in these studies means one-way communication streams produced for mass audiences rather than social media, which are interactive and likely feature a messier array of public cultural material. Cross-national survey analyses show that personal use of mass media, broadly speaking, is often positively associated with women’s stated attitudinal rejection of men’s violence against women (Pierotti 2013). At the same time, entertainment media companies create and distribute a flood of content that perpetuates gender stereotypes, and other studies find that exposure to such content is associated with expressed justification of men’s abuse of women (Bleakley et al. 2012; Flynn et al. 2016; Ward 2016). This raises two related issues. First, what public cultural scripts about gender relations and men’s abuse toward women in particular flow

2 through mass media? Second, how do these scripts reach individuals and inform their stated views respecting the morality of men’s violence against women? In addressing these questions, I draw upon literature showing how targeted media programs may shape people’s attitudinal expressions (e.g., Singhal et al. 2004) and theories about how broad media coverage of a social issue can percolate to individuals through their conversation networks (e.g., Katz and Lazarsfeld 1955). Moreover, mass media’s potential reach may fall along gendered lines given variation in women and men’s content consumption and social position. I address these issues in the setting of Malawi, where a relatively high percentage of people express objection to men’s violence against women, specifically 83.7 percent of women and 87.3 percent of men in 2015-16 (ICF 2020). Malawi is of distinct interest because both transnational organizations and foreign entertainment media companies heavily influence the available mass media programming, through which they may spread unique scripts about gender relations (Englund 2011; Gray 2011). I draw upon a variety of data sources dating from 2000 to 2016, including reports from the national government and other organizations, key informant interviews, four cross-sectional national surveys, and a new dataset of newspaper articles that discussed male violence against women in the two most common daily newspapers in Malawi, which a team of research assistants and I constructed. Beginning with content analysis, I show that transnational organizations facilitated the production of mass media disseminating public cultural scripts critical of men’s abusive treatment of women. They established personal relationships with journalists and directly provided them with potential stories, leading to numerous newspaper articles featuring specific cases of men abusing women and covering transnational organizations’ activities to combat such violence. Conversely, foreign entertainment media companies and some domestic journalists spread programming across Malawi containing gender stereotypes that portrayed men as aggressive and women as submissive. This content was tailored for heterosexual men. Turning next to logistic regression, I uncover results that suggest people’s stated rejection of men’s violence against women is generally positively associated with the recent publication of—as well as personal exposure to—mass media content with scripts denouncing such violence. At the same time, I show that men’s unique personal exposure to mass media comparatively laden with gender stereotypes is negatively associated with their declared rejection of men’s violence against women. Additional analyses and robustness checks confirm the opposing associations between the type of mass media content and individuals’ stated attitudes. Broadly, these results collectively indicate that public cultural scripts flow from multiple transnational sources to individual people. The worldwide dissemination of developmental and human rights scripts is an important component to global cultural diffusion and their reach extends beyond national governments to ordinary citizens. Moreover, developmental scripts circulated via mass media can spill across the population beyond mass media users. At the same time, other important actors, such as but not limited to entertainment media companies, concurrently promote alternative scripts that also reach individual people. Individuals’ social positions, such as their gender identity, introduce variation in the reach of certain public cultural scripts. Tracing the pathways of global cultural diffusion shows the multifaceted nature of such processes and the immense distance public cultural scripts can travel.

Theorizing Global Cultural Diffusion The Global Spread of Public Cultural Scripts to Individuals

3 Sociological research on global cultural diffusion generally has analyzed national level trends and associations, especially world society theory (Drori and Krücken 2009). At the same time, it is recognized that the same institutional forces that encourage convergence across national governments likely extend their reach to individual people (Frank and Meyer 2002; Lerch et al. 2017; Meyer et al. 1975:228). Scholars observe that large majorities of citizens across diverse societies generally express many attitudes, beliefs, and values aligned with a range of development and human rights topics, a manifestation of the global spread of “developmental idealism” (Thornton et al. 2012a; see also Dorius and Swindle 2019). Nonetheless, religious and geographic ideological clustering, especially remains when it comes to issues like homosexuality (Roberts 2019) and less so for comparatively older issues of concern like women’s economic, educational, and political rights (Pandian 2018). Scholars approximate people’s exposure to institutions through which transnational organizations and many other actors promote developmental scripts. In practice, this research relies on empirical measures like a person’s level of education, mass media use, urban living experience, Christian religious membership, and work outside the home, as well as the number of international NGOs present in their country, each of which are shown to be positively related with several attitudinal declarations consistent with developmental and human rights messaging (for a review, see Hadler 2017). The public cultural content that these institutions circulate is often not directly measured, however. As scholars have indicated, this necessitates a more precise understanding of global cultural diffusion by explicitly measuring the flow of scripts about a social issue across public cultural content (Charles 2020:112; Givens and Jorgenson 2013:421; Hadler 2016:343-349; Hadler 2017:37-44; Pierotti 2013:241-242). With this degree of specificity, there is also a need to scrutinize people’s coinciding exposure to alternative scripts promoted by other globalizing organizations. For example, illiberal social movements and many national governments now present a formidable global network of antagonism toward democracy, higher education, and liberalism in general (Carothers and O’Donohue 2019; Schofer et al. 2019). Many religious organizations also lead international campaigns protesting the LGBT rights and access to abortion (Boyle et al. 2015; Ferguson 2019; Velasco 2020b). Besides these direct challenges to liberal scripts about development and human rights, there are numerous other transnational flows of cultural messaging (Bail et al. 2019). In analyses of global cultural diffusion that do include the individual level in their empirical analyses, scholars often rely on survey measures of a person’s stated “attitude” as their outcome of interest. Attitudes have enjoyed prominence in social science research for decades as rough measures of personal evaluations, since the attitudes people express during fixed-response survey questions are often positively associated with their related behaviors despite many instances of inconsistencies (Glasman and Albarracín 2006; Vaisey 2009; see also Johnson- Hanks et al. 2011:69-73, 141-145; Yeatman et al. 2020). Yet, a person’s attitudinal expressions often are mixed with statements of personal identity (Frye 2012; Joas 2013:85-86). Some survey respondents say they espouse attitudes that align with developmental scripts because they think the interviewer wants to hear such answers and may provide subsequent opportunities for them and their communities (Angotti and Kaler 2013). However, when a person expresses an attitudinal response that matches developmental scripts for reasons of identity or appearance, it demonstrates their knowledge of such scripts and perception that they are important (Frye 2017:949-951; see also Lizardo 2017:103-104). People’s stated attitudes, therefore, are elements of “declarative personal culture” (Lizardo et al. 2016), and as such they are useful signals of the individuals’ awareness of public cultural scripts about development and human rights (Behrman

4 and Frye 2021:23-24; Meyer 1986a:216; Pierotti 2013:261-262; Thornton et al. 2012b:337). Extending this logic, group level patterns across people’s stated attitudes can be indicative of public-to-personal cultural diffusion without assuming that people’s stated attitudes directly reflect their subsequent behavior. Men’s Violence against Women Of the many social issues for which transnational organizations and other powerful organizations circulate public cultural scripts, I focus on men’s violence toward women. For parsimony, I refer to scripts that denounce men’s violence against women (MVAW) as anti- MVAW scripts. The United Nations’ 1993 Declaration on the Elimination of Violence against Women codified anti-MVAW scripts as part of development and human rights, stating, “violence against women is an obstacle to the achievement of equality, development and peace” (UN 1993:1). This sentiment was further expanded in the 1995 Fourth World Conference on Women in Beijing, which declared, “violence against women both violates and impairs or nullifies the enjoyment by women of human rights and fundamental freedoms” (UN 1995:92). Many transnational organizations, national governments, academic and policy experts, and political activists champion anti-MVAW scripts, and, as of 2020, public databases indicate that 155 countries had national legislation specifically outlawing domestic violence (World Bank 2021) and 191 had ratified the Convention on the Elimination of All Forms of Discrimination against Women (UN 2021). Several theoretical frameworks outline possible predictors of individuals’ stated rejection of MVAW (for a review, see Ranganathan et al. 2021). Household resource theories emphasize the psychological toll of limited assets (e.g., Goode 1971), while household bargaining theories focus on partner’s relative status and power within the household (e.g., Lundberg and Pollak 1996). These economic approaches conceptualize household wealth, women’s income and education (absolute and relative to their male partner), household decision-making, and marital practices as key factors that shape intimate partner violence. Related “backlash” theories contend that when women outpace their male partners or are viewed as infiltrating a male-dominated social sphere like the workplace, they face retribution (Svec and Andic 2018; Weitzman 2014). Socioecological theories of intimate partner violence, which are especially common among public health scholars, consider factors at the individual, household, community, and societal levels (Heise 1998; Uthman et al. 2010; c.f., Fulu and Miedema 2015). They differ from household dynamics theories in that they emphasize social factors like the number of women in leadership positions and locally-situated norms supporting gender inequality. Global cultural diffusion theories extend this discussion by noting that discourse about development and human rights, including the condemnation of MVAW, is also a form of public culture, though not tied to a single country or society. It follows that the diffusion of anti-MVAW scripts into specific social environments may reach people there and inform their stated attitudes. Building on this latter approach, Pierotti’s (2013) cross-national analysis of the Demographic and Health Surveys (DHS) demonstrated that women’s declarative attitudes shifted toward greater rejection of MVAW during the first decade of the twenty-first century across 23 of the 26 nations countries sampled across Africa, Asia, Latin America, and the Middle East (see also Cools and Kotsadam 2017; Kurzman et al. 2019). What made her analysis exceptionally persuasive evidence for the global cultural diffusion theory, rather, was that the temporal shift toward increasing expressions of rejecting MVAW that she observed was not due to cohort effects or socio-demographic changes, and at the same time there was steady growth in the number of international conferences addressing gender inequality worldwide.

5 It has been several years since Pierotti’s study, so in Figure 1 I update her temporal analysis of stated attitudes toward MVAW to include more recent surveys and to examine men as well as women (ICF 2020).1 The global trend toward increasing declarative rejection of MVAW has continued for women and also is found among men: since 2000, women’s stated rejection of MVAW has increased across 41 of 45 countries and men’s stated rejection has increased across 28 of 33 countries (Appendix A). Consistent with other studies (Sardinha and Catalán 2018; Tran et al. 2016; Uthman et al. 2009; see also Charles 2019), men are more likely than women to express rejection of MVAW in most countries. Though men’s and women’s differing levels of exposure to anti-MVAW scripts likely play a role in this gender gap, social performance and identity likely play an especially salient factor in men’s statements. Aligned with this view, Lawson et al. (2021) show—in a sample of 317 heterosexual couples from Mwanza, Tanzania—that women are much more likely to say their male partner accepts MVAW than the men are to say this themselves. [Figure 1] Mass Media and Global Cultural Diffusion Pierotti (2013) posited that transnational organizations circulate anti-MVAW scripts in the mass media through the policy debates they inspire in addition to their targeted mass media programs. This theorizing is consistent with several experimental studies that observe positive treatments effects of anti-MVAW media programs on recipients’ stated rejection of MVAW (Chang et al. 2020:44-45), including evidence from specific areas of Mexico (Arias 2019), Nigeria (Banerjee et al. 2019), and Uganda (Abramsky et al. 2014). The health communication literature further outlines the effects of various campaign strategies, such as viewers’ ability to identify with drama characters (Polletta et al. 2013; Smith et al. 2007), the materiality and symbolism of campaign objects (McDonnell 2010; 2016), and the interpersonal communication about a social issue that a campaign generates (Jeong and Bae 2018). In her cross-national analysis, Pierotti found a significant positive association between women’s at least weekly mass media use (coded as either newspaper, radio, or television use) and their stated rejection of MVAW in just over half of the 26 countries and a significant negative association only in India. These cross-sectional results provided fair but not outright support for the argument that anti-MVAW scripts likely flow through mass media. Pierotti (p. 261) lamented that her exclusive reliance on the DHS meant not having direct measures of the presence of anti-MVAW scripts in mass media nor of specific media campaigns and their presentation styles. A series of more recent cross-sectional studies that distinguished between people’s types of mass media use—newspaper, radio, and television—interestingly found a variety of positive, negative, and null relationships with people’s expressed objection of MVAW (Forsyth and Ward 2021; Jesmin and Amin 2017; Oyediran and Isiugo-Abanihe 2005; Uthman et al. 2009; Wang 2018). These findings, paired with Pierotti’s somewhat non-absolute conclusions, underscore that there could be different public cultural scripts presented across unique forms of mass media. To that point, powerful entertainment media companies based in the United States, various European nations, China, India, and Nigeria all produce content filled with consistent gender stereotypes about masculinity and femininity, as demonstrated by various large-scale content analyses of movies, television shows, and popular music (Flynn et al. 2016; Ghaznavi et al. 2017; Liebler et al. 2015). These large-scale studies find that male characters are especially common in entertainment-focused media and they generally are portrayed as aggressive and

1 Figures rely on coding packages from Bischof (2017) and Jann (2014).

6 violent, while women are more likely to be depicted as submissive, though the tide may be changing in very recent years. Importantly, several meta-analyses have established that people’s exposure to music videos, movies, and television shows featuring such stereotypical depictions of women and men was associated with expressing support for MVAW (Malamuth and Check 1981; Rhodes et al. 2018; Ward 2016). Content analyses, therefore, are vital for understanding how transnational organizations and entertainment media companies can circulate scripts to individuals through mass media. I propose that by first examining the prevalence of anti-MVAW scripts and gender stereotypes across different types of mass media, one can more accurately interpret the associations between various types of personal mass media use and stated attitudes about MVAW. In addition, one can consider the implications of media modality. Text-based media provides relatively straightforward messaging, whereas visuals elicit greater immediate emotional responses but are more open in interpretation (Powell et al. 2015; Valkenburg et al. 2016:322). Moreover, data on mass media content can be temporally linked to survey data, thereby creating variables that measure the frequency at which particular scripts appeared in mass media in the weeks or months prior to when survey respondents were personally interviewed (de Vreese et al. 2017). One important theoretical benefit to directly linking content and survey data is that much of the empirical research on global cultural diffusion and intimate partner violence has overlooked that salient issues in the media can orient what people discuss in public conversation, which in turn can indirectly influence individuals (Katz and Lazarsfeld 1955; Solovei and van den Putte 2020). People’s friends and family may bring up an issue they heard about on the radio, work colleagues might discuss a newspaper article they read over lunch, or a religious leader might orient their weekly discourse in response to recent media coverage of a social issue, among many possibilities (Thornton 2005:227). Restricting mass media influence to personal use ignores social networks as well as mass media’s role as an issue-framing institution (McCombs and Shaw 1972; Schildkraut and Elsass 2017). Public cultural scripts present in mass media content, then, can reach people through their personal consumption or second-hand through the social environment (Carbone and Mijs 2021; Jeong and Bae 2018; Zaller 1992). All of these reasons point to the benefits of tracing mass media content flows in a specific society and then linking that information to surveys.

The Empirical Case of Malawi I examine the pathways of global cultural diffusion across mass media and their relations with people’s stated attitudes about MVAW in the context of contemporary Malawi, focusing on the timeline of 2000-2016. Malawi is of particular interest for analyzing diffusion pathways because it has one of the highest rates of stated rejection of MVAW across African countries, which implies that anti-MVAW scripts are likely in circulation and, relatively speaking, well- received (ICF 2020).2 Returning to Figure 1, Malawians’ rates of stated rejection of MVAW increased over time, rising from 64.7 percent for women and 74.8 percent for men in 2000 to 87.4 percent for both in 2010. The flattening rate of stated rejection of MVAW in 2015-16 is consistent with other nations with high rates of expressed rejection. The Moral Background Malawi features a historically “layered” (Mahoney and Thelen 2009; Wimmer 2021) public cultural ecology upon which a person might draw upon when declaring their attitudes

2 Only two recent DHS, Mozambique 2015 and South Africa 2016 find higher rates of stated attitudinal rejection of MVAW than Malawi 2010 or Malawi 2015-16 (ICF 2020).

7 about MVAW. Men historically have exerted dominance in politics, religion, and many areas of social life in Malawi; land, however, is generally passed down through the mother’s line and men generally move into the women’s family compound upon intimate partnership, except in the northern region (Phiri 1983). During colonization, women’s relative power and autonomy in Malawi diminished due to European colonizers’ and Christian missionaries’ emphasis on patriarchal familial structures (Ibik 1970; Kudo 2017; Peters 1997). Male authority continued to expand under the leadership of Malawi’s first President, Hastings Kamuzu Banda, who led the country from its political independence in 1963-65 until 1994. Banda imbued what he described as “Pan-Africanist” ideals with the practices of his Chewa people (the largest ethnic group in Malawi) to foster a national identity (Kambalame et al. 2008). This was reflected in language: Banda always spoke in English personally, yet he mandated English and Chichewa as national languages. Like many male authoritarians, Banda envisioned women as the embodiment of a static Malawian culture. He mandated that women perform dances before his speeches, decreed that they must wear long chitenge skirts to their ankles, and shunned foreign intervention promoting gender equality (Gilman 2011; Thornton et al. 2014). In this way, both the colonial and post-Independence periods strengthened narratives of male authority, female subordination, and the “danger” of female autonomy (Frye and Gheihman 2018). Toward the end of Banda’s administration, however, the national government began to sign off on international treaties related to women’s rights (Semu 2002). The government ratified the Convention on Discrimination against Women in 1987 (UN 1991) and adopted their first legislation to legally protect women from discrimination and men’s violence in 1997 (Barkvoll 2009).3 This paved the way for the more expansive Protection against Domestic Violence Act in 2006, which provided substantive legal remedies and social services to victims (Kanyongolo and White 2017). Today, the national government presents itself as firmly committed to international principles of gender equality and challenging what it describes as, “strong attitudes about women being subservient to men” that are “institutionalized” (Malawi Ministry of Gender, Children, Disability and Social Welfare 2014:36). In other words, new public cultural scripts denouncing MVAW are being promoted in contemporary Malawi amidst narratives of patriarchal authority inherited from the colonial and post-Independence periods, all of which is overlaid upon enduring matrilineal practices in much of the country. This layered “moral background” (Abend 2014) provides people with various ideals when considering gender relations. Malawian social scientists and foreign ethnographers note two keywords that citizens regularly invoke in discussions about disagreements and violent action. The first is nkhanza, which means unjustified abuse or cruelty (Saur et al. 2005; see also Johnson 2018:18-19). People use this word to describe the actions of a physically violent husband but also more broadly, such as a parent that neglects to pay their children’s school fees (“‘Abuse’ of School Children?” The Nation, December 10, 2012). When transnational organizations began highlighting MVAW as an important social issue in Malawi during the late 1990s and 2000s, Malawian policymakers and citizens translated anti-MVAW scripts using this existing concept of nkhanza. The second principle many citizens turn to is kulangiza, which means necessary advising, counseling, and disciplining. This is usually considered to be verbal in nature, but physical violence is sometimes classified as kulangiza too (Chepuka et al. 2014; Nthala 2013:65). Widely understood notions of unjustified abuse (nkhanza) and requisite discipline (kulangiza) offer important normative

3 The government’s reluctance was manifest in their first report to the Convention on Discrimination against Women, in which they said Malawi did not “consider itself bound by such provisions … [which] require immediate eradication of such traditional customs and practices” (UN 1991, see note 40).

8 foundations for Malawians that—alongside the historical layers of matrilineal lineage systems, European-introduced patriarchal authority and gender stereotypes—likely inform people’s stated views about MVAW. Mass Media Forms and Usage Mass media is a primary medium for circulating public cultural scripts in Malawi, and it includes newspapers, radio, television, and movies. Internet use is on the rise, but in 2015-16 less than five percent of women and about fourteen percent of men in Malawi said they used the internet at least weekly (NSO and ICF 2017:41-42), and such use was predominantly on mobile phones with more limited capacities than computers (NSO 2015:27-28). During Banda’s rule, mass media production in Malawi was mostly limited to the government-owned Malawi Broadcasting Corporation radio station, which largely aired only informational content and government propaganda as well as music and sports (Harris 2017:28- 31; Lwanda and Kanjo 2013:32; for exceptions, see: Kamwendo 2008:276; Mitchell 2002:6). After Banda, the country transitioned to multiparty democracy and switched the Malawi Broadcasting Corporation from an official branch of the federal government to a formally independent, but tax-funded, public service media outlet (Harris 2017). They further permitted media freedom, which led to the creation of additional media outlets. By 2014, there were two major daily newspapers—The Daily Times and The Nation, one weekly tabloid—The Weekend Times, 34 regional or national radio stations, ten Malawian television channels, and three pan- African satellite television subscription services (NSO 2015:56-60, 70-73). Numerous makeshift video parlors in outdoor markets also appeared nationwide, which used digital video disc players to show copies of movies on television sets to audiences up to a few dozen people for a marginal fee (Geldof 2011; Gray 2011). The DHS provide statistics on individuals’ personal mass media use in Malawi over time, as visualized in Figure 2. At least weekly newspaper use has remained moderately low in Malawi during the twenty-first century compared to other countries, lingering between only 8.3-12.9 percent for women and 15-26.5 percent for men. One major limiting factor is English literacy because newspapers are in English: in 2008, 35 percent of men and 26 percent of women between ages 15-49 stated that they could read and write in English (MPC 2019).4 [Figure 2] Radio is the most used type of mass media in Malawi. It is affordable, consistent with oral history practices, and generally broadcast in the widely-spoken Chichewa dialect and sometimes in other common regional languages (Kamwendo 2008). Weekly radio use was highest in 2004 at 66.5 percent of women and 85.3 percent of men. Like in many other countries, radio use in Malawi has declined in recent years, falling in 2015-16 to 30 percent for women and 49.3 percent for men. This decline is likely due in part to the increasing frequency of electrical blackouts: surveyed businesses reported 0.8 blackout incidents per month in 2009 yet the rate rose to 6.9 in 2014 (World Bank 2019; see also NSO 2014; Taulo et al. 2015). The other factor in radio’s decline is the rise of “television,” which most people in Malawi define as any visual content displayed on a television device, including movies. Television use jumped dramatically between 2000 and 2010 and peaked at 15.9 and 34.3 percent among women and men, respectively, before decreasing somewhat in 2015-16 when electrical blackouts increased. The high cost of a television subscription and limited electricity access across Malawi’s vastly rural population make television use lower than in many other countries.

4 The Nation does publish short supplements in Chichewa and Chitumbuka biweekly, but circulation is limited and the content is very similar (Angotti et al. 2014).

9 The few domestic television stations feature content in Chichewa and English, as well as some other native languages, while satellite stations come almost entirely from the South African company MultiChoice and are in English (NSO 2015:72).5 Video parlors only show foreign movies, which are in English as well as Mandarin, Hindi, and Yoruba, among other languages; there is not a domestic movie industry nor any operating movie theaters in Malawi (Chimbuto 2016:80-81; Magalasi 2015; see also The Nation, November 21, 2014). Many people watch shows and movies without knowing the language in use. Notably, the movies shown at video parlors are unlicensed copies and many people historically have accessed satellite television using free-to-air decoders from China (e.g., Tech Dot Africa, July 2, 2010). In other words, people are accessing content produced by foreign entertainment media companies without those companies formally selling them that content. Beyond their personal mass media consumption, people often learn of ideas and news presented in mass media through casual discussion with others or in public settings. Prior ethnographic research on Malawians’ discussions about AIDS demonstrates the importance of local conversation networks and overhearing impromptu talk at the market, bar, water hole, bus station, or simply in passing (Kaler et al. 2015; Swidler and Watkins 2009; 2015). One locally embedded Malawian ethnographer recorded: “we then chatted a bit over this issue, [she] saying Balaka is also the highest district affected by the AIDS as she heard from the radio one day” (Watkins and Ashforth 2015).6 Another observed a community leader state, “the message of HIV and AIDS is spread through many ways, like through the radio, in the prayer houses, at the market with friends, in the news papers, at schools and many more places that a day cannot pass without hearing the message of HIV and AIDS” (Watkins and Ashforth 2015).7 These reports are consistent with NGO reports from Malawi in which the majority of people reported on surveys that they had heard about AIDS directly from mass media sources as well as through friends, at community meetings, and in other public settings (MHRRC 2013:41-44). Conversation networks and the social environment likely distill mass media messages about MVAW just like they do about AIDS. In other words, scripts about MVAW and gender relations present in the media may inform the topics of informal conversations and the discourse of community and national authorities in Malawi. Additionally, there are accessibility barriers to personal mass media consumption like English fluency, literacy, cost, and geography. Mass Media Content Styles Mass media’s potential influence on people in Malawi hinges on its content. Since mass media options in Malawi are relatively limited compared to countries with more substantial domestic media production, transnational organizations and foreign entertainment media companies can more easily shape the content available. As outlined by Englund (2011:25-31, 40), the basic theme in Malawian mass media over the years has been “developmentalism,” which is the coverage of a range of topics posited to contribute to societal development, such as education, human rights, HIV/AIDS prevention, and gender equality (see also Angotti et al. 2014; Harris 2017:133-137). A recent cross-national study found that journalists in Malawi saw “supporting national development” as one of their primary goals, whereas journalists from nineteen Western European and North American countries saw their job as keeping the government in check (Kalyango et al. 2017). Malawian

5 The first Malawian show on MultiChoice satellite television began in 2017 (Nyasa Times, July 28, 2017). 6 Malawi Journals Project, April 18, 2010, Simon_100404. The Malawi Journals Project data are publicly available at: https://deepblue.lib.umich.edu/handle/2027.42/113269. 7 Malawi Journals Project, December 17, 2011, Patuma_111202.

10 journalists’ views on this point are consistent with their education and training, as foreign aid initiatives that sought to instill this developmentalist logic in domestic journalists have long played a key role in directing local news companies in Malawi and many other aid-receiving nations (Harris 2017; see also Lugo-Ocando 2020:119-136; Shoemaker and Reese 2014). The mass media focus on developmentalism in Malawi stems in large measure from transnational organizations, who—together with the domestic NGOs they fund—provide domestic journalists with regular opportunities to report on one of their projects in exchange for “brown bag” payments of $10-15 USD (African Media Barometer 2012:58; Nyasa Times, December 27, 2019; see also Lodamo and Skjerdal 2009). Since journalists are paid very little, developing relationships with transnational organizations and domestic NGOs can be vital (Manda and Kufaine 2013). In the words of one journalist interviewed by Harris (2017:136): “The NGO gives money so the reporter will write. So they indirectly co-determine what is published and what is aired.” Following this logic, gender equality and anti-MVAW scripts may be among the developmental and human rights messages integrated into Malawian mass media content, especially across newspaper articles given newspaper journalists’ connections to transnational organizations. Targeted radio and television programs that foreign donors fund are additionally likely to contain such messaging. Some Malawian journalists find creative ways to present competing or multivalent messaging in mass media. Englund (2011) describes how in one of the most popular radio programs—Nkhani Zam’maboma (News from the Districts)—listeners called into the show and discussed human rights issues. This format allowed the show to give voice to callers that challenged the foreign imposition of human rights without requiring the show administrators and hosts to explicitly challenge developmentalist scripts themselves. When intended messaging becomes altered like this, it can have unexpected downstream impacts, akin to McDonnell’s (2010; 2016) observations on AIDS campaigns in Ghana. Other mass media content in Malawi, such as the sports pages of the major newspapers and popular songs on the radio, does not revolve around human rights and instead focuses on entertainment. Music, sports, celebrity gossip, and other entertainment-focused media appears across nearly all media outlets, but is especially prominent in satellite television and the movies available at video parlors, which solely provide foreign-produced entertainment media (Geston 2006:41-47, 76-83; Gray 2014:987). The Weekend Times tabloid, popularly known as the “weekly scandal sheet,” is also entertainment-focused and attention-grabbing; it covers stories and rumors of government corruption, cheating among celebrities, sexual abuse, witchcraft, and gruesome violence (Mchakulu 2018). Thus, developmental messages like anti-MVAW scripts may be less likely to appear in entertainment media, and the portrayal of gender stereotypes may be more common. Differences between the two dominant styles of available mass media content in Malawi—developmentalist and entertainment-based—imply that they might share anti-MVAW scripts and gender stereotypes, respectively. This would unsettle the assumption that mass media exposure is positively associated with stated rejection of MVAW. Rather, I expect mass media exposure to anti-MVAW scripts to be positively associated with people’s expressed rejection of MVAW, while mass media exposure to public cultural scripts that bolster gender stereotypes to be negatively associated with people’s stated rejection of MVAW. Women’s and men’s unique social positions also likely play into their stated attitudes, as well as their ability to consume different types of mass media. To disentangle these forces, I turn to a mixed methods approach.

11 Data and Methods I conduct a sequential mixed methods analysis (Small 2011:67-68). For me, the association between mass media and stated attitudes about MVAW hinges on the content of different forms of media, thus necessitating a two-step design for linked research questions. I begin with a content analysis to assess how transnational organizations and entertainment media companies influence the production and availability of mass media related to gender and MVAW in Malawi. In doing so, I trace the pathways through which different styles of content are circulated. This lays the groundwork for analyzing whether distinct mass media content and therefore particular public cultural scripts likely reach ordinary citizens. I employ regression analyses examining the relationship between different forms of mass media exposure and stated rejection of MVAW among female and male respondents. I include a replication package that contains copies of the official reports and newspaper articles cited in the content analysis as well as all necessary information and syntax to replicate the quantitative portion of the content analysis and the full statistical analysis.8 Content Analysis Official Reports, Newsletters, and Interviews. I use several sources of qualitative data to examine the flows of public cultural scripts about MVAW and gender across different forms of mass media in Malawi. The first sources of data for my content analysis are official reports and newsletters about media content and use published by transnational organizations, domestic NGOs, and several branches of the Malawi government. I also searched across domestic and regional newspapers for articles about the media industry in Malawi and people’s access to different types of media content. Besides outlining the general sources of available mass media, many of these reports, newsletters, and articles discussed the portrayal of women in media and coverage of MVAW. I located many of these documents online while others I collected when visiting the offices of transnational organizations, domestic NGOs, and government branches in Malawi. To further understand the pathways of cultural diffusion in mass media, I interviewed sixteen key informants in government, social movement, and transnational organization leaders that were personally involved in national programs to combat MVAW. I identified these programs from a novel national database of foreign aid-funded projects, the Malawi Aid Management Platform (Malawi Ministry of Finance, Economic Planning and Development 2017; see also Peratsakis et al. 2012). I also interviewed Limbani Moya, the original editor of The Weekend Times. Appendix B contains a list of the specific official reports and newspaper articles that I cite in the findings presented (though not all those I consulted) in addition to a list of key informants I interviewed. Newspaper Articles. With the help of Malawian research assistants Florence Nkhoma and Henry Mphande, I created a new dataset of all newspaper articles that discussed interpersonal violence that were published in the Monday through Friday editions of Malawi daily newspapers The Daily Times and The Nation between January 1, 2000 and February 14, 2016 (the interview date of the final respondent in the last survey I use). Despite being the most long-standing and popular daily newspapers in Malawi, digital formats of their historical archives were limited, so we located hard copies in Malawi at several locations across the country.9 To identify relevant

8 Replication package: https://www.dropbox.com/s/t08c87a08aockf0/Pathways.zip?dl=0 9 We visited the newspapers’ official archives in Blantyre, two libraries in Lilongwe located at the Malawi Government’s Human Rights Commission and the Malawi Human Rights Resource Centre, and one library in Zomba at the Center for Social Research at the University of Malawi. This was necessary to ensure full coverage of the two newspapers, as each archive or library was missing records due to water damage or staff transition.

12 articles, we carefully looked through all pages of these newspapers during the time period and read the article titles on each page. We then read the full text of all articles that had titles related to violence, threats, gender relations, or sex. Through our reading we identified 2,760 newspaper articles that documented or abstractly discussed interpersonal violence, which we defined as a person forcefully undertaking any harmful or undesired action against another person. We took pictures of these articles and organized them for coding. I then trained a team of trained undergraduate research assistants—Emma Bruder, Madeleine Conrad, Madeleine Danes, Samantha Galbavy, Yi Gao, Stephanie Green, Kaila Graham, Zhuang Han, Rebecca Jacobs, Trevor Orginski, Emily Schwartz, Molly Sproul, and Leah Weinstein—to work with me to re-read all identified articles and code them for a variety of characteristics. In addition to basic descriptors (article date, title, author, etcetera), we coded each event of interpersonal violence listed, including the demographic characteristics of the victims and perpetrators and the nature of the offences described. Our coding included placing all articles into one of four categories: MVAW campaigns, other campaigns, MVAW cases, other cases. Articles that explicitly shared anti-MVAW scripts were coded as articles about MVAW campaigns. They generally either described specific initiatives to combat MVAW by the national government, foreign aid agencies, NGOs, or activists, or they were opinion articles condemning MVAW. Articles coded as other campaigns, conversely, either fully or partially justified MVAW or they reported on campaigns about other types of violence (e.g., the killing of people with albinism). Articles about MVAW cases documented at least one specific incident of MVAW whereas articles about other cases described events in which someone abused a man or when there was violence between women. Seventeen percent of the 2,760 articles we identified described a campaign against MVAW and a specific case of MVAW. We coded these articles in the MVAW campaigns category because they explicitly shared anti-MVAW scripts. Articles that only described MVAW cases, on the other hand, implied abuse was not “normal” and they often noted the perpetrator’s sentence, but this presentation format only implicitly criticized MVAW. Analytical Strategy. Combining these multiple sources of data, I conduct a content analysis to outline the specific pathways within mass media through which anti-MVAW scripts and alternative scripts about powerful aggressive men and submissive sexualized women are disseminated. I emphasize the role of transnational organizations and entertainment media companies in circulating such content. I also report descriptive statistics for the frequency of the four categories of newspaper articles about interpersonal violence, and I examine the temporal frequency of newspaper articles that document cases of MVAW or campaigns about MVAW. The findings from the content analysis provide indications of what forms of mass media content contain anti-MVAW scripts or gender stereotypes, which informs my quantitative analysis. Regression Analysis Survey Data. To test the relationships between different types of mass media exposure and people’s stated attitudes about MVAW, I use four cross-sectional surveys conducted in Malawi: the 2000, 2004, 2010, and 2015-16 Demographic and Health Surveys (DHS) (Boyle et al. 2017b). The four surveys each featured a two-stage cluster design in which households were selected from enumeration areas that were selected from districts.10 All women from selected

10 There are 28 districts in Malawi, however, the district of Likoma Island—population of 10,714 in the 2008 census—was only included in the 2015-16 DHS and the district of Neno—population of 52,751 in 2008—was created in 2003. Additionally, the 2000 and 2004 Malawi DHS were representative across Malawi’s three regions rather than across districts (though the eleven most populous districts were oversampled so that representative

13 households between the ages of 15-49 were invited to participate. In one-third of the selected households (one-fourth in the case of the 2000 DHS), all men ages 15-54 were asked to be interviewed. Pooled together, the four surveys sampled 72,500 women and 21,006 men. Aligned with DHS recommendations, I keep the women’s and men’s samples separate. In the analysis this helps ascertain whether anti-MVAW scripts and alternative scripts supportive of gender stereotypes appear to have reached women and men differently via mass media exposure. Dependent Variable. My dependent variable, stated rejection of MVAW, measures individuals’ stated rejection of the notion that a husband is justified in beating his wife. Here I draw from the following survey question: “Sometimes a husband is annoyed or angered by things that his wife does. In your opinion, is a husband justified in hitting or beating his wife in the following situations: (1) If she goes out without telling him? (2) If she neglects the children? (3) If she argues with him? (4) If she refuses to have sex with him? (5) If the food is not properly cooked?”11 Eighty-four percent of men and 78 percent of women say that a husband is not justified in beating his wife in any of the five proposed situations. The coefficient alpha from these five questions is 0.85 for women and 0.78 for men. Loadings from a single factor principal- component analysis range between 0.76-0.82 for women and 0.71-0.75 for men, indicating that the five questions tap into a single latent construct. Following others (e.g., Pierotti 2013), I use a binary variable identifying respondents that reject all five scenarios. Independent Variables. I exploit the temporal variation across the publication dates of newspaper articles and respondents’ survey interview dates, as each survey was conducted over a period of five to six months. I construct an individual level measure of the number of MVAW newspaper articles published one month (30 days) prior to a respondent’s personal interview. For example, if a person were interviewed on April 20, 2014, then their assigned value would be the total number of MVAW articles published between March 20 and April 19, 2014. Respondents interviewed on the same day share the same values, but those interviewed on the next day could differ. I include articles about MVAW campaigns or specific cases of MVAW as “MVAW newspaper articles,” though in additional analyses I use two measures for articles detailing MVAW campaigns or MVAW cases. To assess respondents’ radio exposure to anti-MVAW scripts, I leverage survey questions asked in the 2000, 2004, and 2010 surveys stating, “In the last few months, have you listened to any of the following program series about family planning or health on the radio? Uchembere Wabwino (Safe Motherhood), Phukusi Lamoyo (Bag of Life), Umoyo M’Malawi (Health in Malawi), Dokotala Wapawailesi (Radio Doctor), or Chitukuko M’Malawi (Development in Malawi)? Respondents were then asked to report each of these five radio programs they listened to. Though the survey question refers to these radio programs as being about family planning and health, all of the programs regularly discuss MVAW and share anti-MVAW scripts except the last one, which focuses on economic matters. I create a count measure of the number of the first four of these radio programs that respondents said they heard in the last few months. Given that the survey questions I use to create this count variable are not included in the 2015-16 survey, analyses using this variable are limited to the 2000-2010 time period.

descriptive statistics for them could be reported using the survey weights) (NSO and ORC Macro 2001; 2005; 2011; NSO and ICF 2017; see also Grant 2015:footnote 2). 11 Respondents are not asked whether their husband was justified in beating them or whether they were justified in beating their wife. The questions instead refer to a husband and a wife.

14 Additionally, I use three binary variables capturing respondents’ at least weekly newspaper, radio, and television use, respectively. The content analysis guides my interpretation of these variables as mediums for exposure to anti-MVAW scripts or gender stereotypes. Control Variables. I include additional covariates that often are used in the literature as predictors of stated attitudes toward MVAW and that likely also predict individuals’ mass media exposure. By conditioning on these confounding variables, I avoid potential colliders influenced by both expressed rejection of MVAW and mass media exposure (Cunningham 2021; Elwert and Winship 2014). This allows me to evaluate the associations between declarations of rejecting MVAW and forms of mass media exposure without backdoor pathways influencing their relationship. In Appendix D, I examine more conservative modeling criteria. To begin, I employ four measures of individuals’ exposure to scripts about development and human rights employed in the literature: a continuous measure of individuals’ years of schooling completed and binary measures for urban living, Christian religious identity, and working outside the home in a non-agricultural job (Boyle et al. 2002; Charles 2020; Pierotti 2013). Each of these factors potentially may also increase people’s mass media exposure as they move people outside the home, increase their opportunity for mass media consumption, and inflate their social networks. Recent empirical work on the latter of these four variables, employment, finds that it has a negative association with declarative attitudes about MVAW despite theories that employment may expose people to development scripts (Cools and Kotsadam 2017; Behrman and Frye 2021). Education, for example, is especially important to control given that it is generally positively associated with both stated MVAW rejection and personal mass media use. I next draw inspiration from household resource and bargaining theories of gender relations (Goode 1971; Lundberg and Pollack 1996). I rely on the household wealth index factor score measures provided by the DHS because household wealth facilitates one’s ability to consume media and it may decrease family members’ emotional uncertainty and distress about finances, arguably reducing support for norms of domestic conflict and force. I additionally use a categorical measure of heterosexual partnership history. Malawi legally prohibits homosexuality and such partnerships are not collected in the DHS. The heterosexual partnership categories are never partnered, formerly but not currently partnered, currently in a monogamous partnership, and currently in a polygamous partnership. Partnership in this case refers to cohabiting or married intimate relationships. This is helpful given prior work showing that marriage has a strong, positive association with expressed rejection of MVAW in most African countries (Hindin 2014), but in some societies polygamy is associated with lower odds of stated rejection than monogamy (Rani et al. 2004). Motivated by socioecological approaches to community norms about gender relations (Heise 1998), I construct a binary variable for matrilineal lineage by collapsing Malawians’ ethnicities into patrilineal or matrilineal categories based on Ibik’s (1970) ethnographic research. I also use a continuous variable for age because people’s connection to youth norms promoting physical violence as a means of conflict resolution likely decline as they grow older, whereas people’s lifetime exposure to developmental scripts expands, and prior research shows that age has a reliably positive association with expressions of rejecting MVAW (e.g., Kurzman et al. 2019). Modeling Strategy. I estimate several logistic regression models measuring the associations between mass media exposure to both anti-MVAW scripts and gender stereotypes and Malawians’ expressed rejection of MVAW. My strategy of linking media content data to individual level surveys is similar to a few recent studies on media persuasion (de Vreese et al.

15 2017; Huang et al. 2021). In particular, Brosius et al. (2019) match individuals’ interview dates with the publication dates of newspaper articles. Researchers have not applied this approach to the DHS nor to stated attitudes about gender relations or MVAW. Consistent with the literature (Cools and Kotsadam 2017; Kurzman et al. 2019; see also Behrman and Frye 2021), I use fixed effects (indicator variables) for surveys and Malawi’s geographic districts. These factors may shape both individuals’ mass media exposure and stated rejection of MVAW yet they are not dependent on either. Survey fixed effects are important because of legal reforms that could have influenced people’s stated attitudes about MVAW over time (see Figure 1; Pierotti 2013), as well as temporal changes in mass media accessibility and people’s use of different types of mass media (see Figure 2). District fixed effects are important given unobserved differences across districts—such as electricity blackouts or the number of informal video parlors (African Media Barometer 2012:27, 45; see also Grimm et al. 2015:1783)—that could bias the estimates for mass media variables in my models. Another reason to use district fixed effects are to account for other time-invariant or historical factors that could influence stated attitudes about MVAW in Malawi, namely the original settlement areas of the many societies present prior to European colonization as well as geographic variation in longstanding foreign influences from Christian missionaries, Islamic trade networks, and labor migration (Kudo 2017; Pike 1965; Sicard 2000; Van Kol 2008).12 To account for the complex survey design of the DHS, I calculate robust standard errors by clustering at the primary sampling unit and I assign individual probability weights provided with each of the women’s and men’s surveys. The models include the full set of control variables and I mean-center the continuous controls—education, wealth, and age—within each survey. Since comparing logistic regression coefficients within and across models is a well- established problem (Allison 1999; Ai and Norton 2003; Mood 2010), I convert all coefficient estimates to average marginal effects (AMEs), holding all other independent and control variables at their actual value for each observation (Breen et al. 2018; Long and Mustillo 2018). I interpret all AMEs as associations, and I consider the AMEs for variables respecting mass media as indications of the reach of public cultural scripts to individual people rather than causal effects of mass media exposure on stated attitudes. Some mass media users likely seek out content consistent with their pre-treatment views, so AMEs for personal newspaper, radio, and television use likely reflect a mixture of mass media effects, selection processes, and reinforcement effects (Bartels 1993; Zaller 1996). I note that selection processes do not cloud the association between a respondent’s expressed rejection of MVAW and the number of MVAW newspaper articles published one month prior to each respondent’s interview date.

Content Analysis of Mass Media Flows Mass Media Critical of Men’s Violence against Women Figure 3 summarizes the major mass media diffusion pathways of anti-MVAW scripts in Malawi. Transnational organizations promoted anti-MVAW scripts to Malawian media companies and journalists, who then produced specific newspaper articles, radio programs, and

12 I also replicate my findings using region fixed effects (southern, central, and northern regions of Malawi) rather than district fixed effects and find no substantive differences. In my case, district fixed effects are preferable to random effects since I do not have explicit interest in quantifying the unexplained variance among districts. There are also some methodological limitations to using random effects with DHS data given its complex survey design and lack of district level probability weights, which I outline in Appendix D (Bollen et al. 2016; Elkasabi et al. 2020; West et al. 2018).

16 television shows for public consumption. Malawian journalists generally dispersed anti-MVAW scripts in ways that were consistent with transnational organizations’ expectations, but at times they exerted independence in modifying transnational organizations’ messages, so the lines in Figure 3 do not imply claims of non-adjusted message diffusion or that all anti-MVAW scripts stemmed from transnational organizations. [Figure 3] Representatives from Malawi’s offices of United Nations Women, United Nations Population Fund, and several foreign aid agencies, in partnership with some pan-African and Malawian NGOs, led numerous “training” meetings for journalists, editors, and producers.13 They coached Malawian media personnel about the importance of covering MVAW, avoiding victim-blaming, and strategically denouncing MVAW in their journalism (The Daily Times, November 29, 2004; The Nation, November 27, 2016; see also Gender Links and Malawi Institute of Journalism 2001; Gender Links 2011; Morna 2010; Public Media Alliance 2012). These trainings led newspaper journalists, radio producers and television writers to create content covering MVAW and explicitly denouncing it as a human rights violation. The predominance of the topic of MVAW in Malawi newspapers is demonstrated in Figure 4, which gives counts of newspaper articles about different types of interpersonal violence that two research assistants and I identified in The Daily Times and The Nation daily newspapers between January 1, 2000 and February 14, 2016. The first two bars show that there were 1,241 articles about cases of MVAW compared to 580 articles documenting cases of violence against men or committed by women.14 The next two bars indicate that among articles that covered campaigns about interpersonal violence or discussed it in an abstract way, 890 explicitly condemned MVAW while only 117 did not. In many of the 890 articles about MVAW campaigns, journalists glowingly described interventions to combat MVAW that transnational organizations funded (e.g., The Daily Times, March 27, 2008). Other MVAW campaign articles summarized public speeches denouncing MVAW made by Malawi’s President or another government or transnational leader (e.g., The Daily Times, November 18, 2014; The Daily Times, December 11, 2014) or featured opinion editorials critical of MVAW (e.g., “Malawians Should Always Claim Their Human Rights,” The Nation, December 12, 2012). [Figure 4] Overall, there were 1,979 newspaper articles that covered cases of MVAW or described campaigns against MVAW, which is nearly half the total number of weekdays (4,206) during the time period examined. The regularity of MVAW articles, and the comparative scarcity of other types of articles about interpersonal violence, speaks to the power of transnational organizations in shaping newspaper content. Notably, there was no reactive increase in content challenging human rights messages, as Koo and Choi (2019) observe occurred in South Korea. This may have been due to Malawian journalists’ reliance on financial support from transnational organizations and journalists’ common—though not universal—stated view that their key objective was to raise awareness of human rights (Kalyango et al. 2017). The influence of transnational organizations over the content in the daily newspapers is further apparent in the monthly frequency at which MVAW articles appeared in The Daily Times and The Nation, as depicted in Figure 5. The enactment of Malawi’s first National Gender Policy

13 Interviews with Emma Kaliya and Limbani Phiri. 14 The fact that stories about MVAW are more common than violence committed against men is significant because men disproportionately dominate journalism jobs in Malawi (The Nation, December 3, 2009; African Media Barometer 2012:55; Gender Links 2003; 2010).

17 in March 2000 was likely one primary reason for the substantial increase in the number of articles during 2000 (Semu 2002:89). The rapid increase during late 2005 and through the first several months of 2006 was due primarily to domestic activists’ and transnational organizations’ combined efforts to provide journalists with material for articles about MVAW, which they hoped would sway public opinion and parliamentary approval in anticipation of the upcoming vote on the Protection against Domestic Violence Act.15 The effort was successful and the law was enacted on April 26, 2006. The smaller number of articles in 2004, 2007, and 2014 likely resulted in part from temporary lapses in foreign aid funding for domestic NGOs’ efforts to advocate against MVAW, which included providing trainings for journalists (The Daily Times, December 8, 2004).16 [Figure 5] Further evidence of transnational organizations’ influence on newspaper content comes from the publication of newspaper articles about MVAW campaigns surrounding the annual 16 Days of Activism against Gender Violence campaign, which occurred each year between November 25 and December 10. The earliest documentation of the campaign in Malawi is from 2000, though it increased in size and prominence after the Protection against Domestic Violence Act in 2006 and when Norway began funding the campaign in 2008 (Arise!: A Newsletter of the Network against Gender Violence, November-December 2000; MHRRC 2007; 2013). Various international donors subsequently provided aid for efforts by the Malawi government and domestic NGOs to spread awareness of MVAW during this annual campaign through protest marches, official gatherings, and public presentations in primarily rural areas (FEMNET 2003; 2009; Malawi Ministry of Gender, Children, Disability and Social Welfare 2014; MEGEN 2011; 2012; 2013; UN Women 2015). During the campaigns, government leaders, Malawian activists and NGO leaders, and members of the police rented three large touring buses and traveled around the country giving presentations about MVAW (The Daily Times, December 6, 2000; The Nation, “Stocktaking 16 Days of Activism,” December 10, 2012). Campaign organizers made sure that several radio, television, and print journalists from the Malawi Broadcasting Corporation, Zodiak, and other media companies accompanied their caravan and they provided journalists with their own large van.17 In most years, journalists covered campaign activities in depth. Figure 6 is one such article (The Nation, December 1, 2010). [Figure 6] Figure 7 features sixteen graphs documenting the weekly number of newspaper articles about MVAW campaigns between October and January annually between 2000 and 2015-16. The vertical grey section in each graph represents the 16 Days campaign period from November 25 through December 10. The graphs show that in the early years of the campaign the number of MVAW campaign articles during the 16 Days period generally was not higher than the weeks preceding or following it. This changed in 2006—the year the Protection against Domestic Violence Act passed—and continued annually thereafter, with articles about MVAW campaigns appearing more during the 16 Days period (or the following week after journalists returned home to write) than the preceding or following weeks, with the exception of 2011. [Figure 7] Transnational organizations similarly extended their efforts to shape media content to radio and television outlets, though the lack of publicly available historical recordings or

15 Interviews with Lugede Chiphwafu Chiumya and Seodi White. See also Kanyongolo and White (2017:193-194). 16 Interview with Emma Kaliya. 17 Interviews with Emma Kaliya and Limbani Phiri.

18 transcripts prevents the quantification of such efforts.18 Radio and television journalists similarly attended the media “training” workshops, and they also developed working relationships with transnational organizations to cover their programs in return for small payments. Moreover, transnational organizations had longstanding efforts to provide funding to develop specific radio and television programs about family planning, contraception, and HIV/AIDS (Lwanda 2010:393-395, 402-403; Manyozo 2008:34; Mhagama 2015). As transnational organizations extended their vision to the social problem of MVAW, they ensured that these same programs covered MVAW as well (e.g., The Nation, June 4, 2012; National AIDS Commission 2010:66). Examples of such programs included the popular informational radio program Uchembere Wabwino (Safe Motherhood), the radio drama Tichitenji (What Do We Do), and the television drama Tikuferanji (Why Are We Dying). The national government, with financial support from transnational organizations, published regular monitoring reports in which they listened to the targeted radio programs and watched the targeted television programs to ensure the content did not deviate from the proscribed developmental scripts (e.g., National AIDS Commission 2010). Gender Stereotypes from Entertainment Media Companies Public cultural scripts that portrayed women as passive objects and men as powerful aggressors also were present in Malawi mass media, sometimes imported directly from foreign entertainment media companies. Figure 8 summarizes several major pathways through which scripts encouraging gender stereotypes were circulated to individual people. The connecting lines in Figure 8 show the general pathways of diffusion, but foreign entertainment media companies inspired yet did not determine the content that domestic media companies produced. [Figure 8] One important source of gender stereotypical scripts was The Weekend Times tabloid newspaper. Limbani Moya, the tabloid’s founding editor, explained that he was inspired to create The Weekend Times after touring the offices of foreign tabloids in the United States while participating in a visiting program for African journalists.19 In this way, the tabloid exemplified hybridity: a domestic publication that directed its content toward Malawian citizens yet was similar to foreign entertainment media in presentation style and depiction of gender stereotypes. After its founding in 2009, The Weekend Times was an immediate success and rapidly blossomed to a circulation on par with its parent newspaper, The Daily Times (Englund 2011:31; Mchakulu 2018). The tabloid’s portrayal of gender stereotypes was apparent in their headlines and photographs alone, such as two examples of moderate cover pages shown in Figure 9. Moya expressed great pride in tabloid journalists’ painstaking efforts to uncover the truth about what he saw as unreported social issues, like cases of grisly domestic abuse and adultery among political leaders. He also acknowledged the tabloid’s reliance on sexualized photographs of women’s bodies, including the popular weekly full page photograph of a new “Action Girl” model.20 This was largely unprecedented for media content in Malawi given stringent censorship laws and strict conservative clothing norms placed on women (The Nation, January 20, 2012; The Nation, December 18, 2013). The tabloid thus gained major public attention among elite policymakers,

18 Interviews with Panji Harawa and Jean Mwandira. 19 Interview with Limbani Moya. He also spent time in England and Germany for similar programs, and he made a point to personally study the national tabloids there. 20 While readers could interpret the “Action Girl” photographs in multiple ways (e.g., empowering versus objectifying), their consistency in lighting, pose, and clothing challenged their ambiguous potential. Johannson’s (2007) analysis of similar portraits in British tabloids applies well: “the female sexuality presented is clean, controlled and non-threatening. With its daily repetitions of this one ideal, the feature fetishes this particular kind of female sexuality and femininity, which…would signal the ultimate desire of men to control women” (p. 103-104).

19 who attempted to ban the paper, as well as lay citizens that competed for a copy of the tabloid each Friday afternoon, much to Moya’s satisfaction (Nyasa Times, February 24, 2012; Panapress, November 4, 2010; see also Nyasa Times, February 21, 2012).21 [Figure 9] The visual components of the tabloid likely helped it reach a broader audience, especially among heterosexual men, and made it more memorable (Mchakulu 2018:12; see also Powell et al. 2019). Many women, on the other hand, likely avoided the tabloid because they opposed its gender stereotypes or feared they would be labeled a “dangerous” woman or prostitute for expressing interest in its sexualized content. Several researchers have documented the danger for women in Malawi of expressing interest in condoms for this latter reason (e.g., Chimbiri 2007; Tavory and Swidler 2009). Being labeled in this way could lead to stigma, divorce, abuse, and losing the money they receive from their partner to take care of themselves and their children (Frye and Gheihman 2018; Swidler and Watkins 2007). Besides the tabloid, gender stereotypes were further spread through satellite television. This programming featured reality television shows, including Big Brother Africa, which so captivated the attention of the public when it started in 2003 that some Malawian parliamentarians tried to ban the show, and temporarily did so, because it included nudity and sex scenes involving the show participant from Malawi (BBC News, August 15, 2003). The popular actionX satellite television channel featured graphic violence alongside very gendered content, as indicated by its advertisement tagline: “High impact action movies and series. Explosions, bombs, and bullets, flying fists and feet, dangerous men and far more dangerous women!” (Geston 2006:77). Other television stations primarily showed popular music videos by American, Malawian, and other African artists, many of which featured gender stereotypes as well as physical fighting between couples (Gray 2014). The most prevalent types of foreign movies available in video parlors were violent Chinese martial arts and American action movies, such as the James Bond and Die Hard series. In interviews with Malawian viewers at the video parlors, Gray (2014:990) found that many concluded that America and China must be exceptionally violent and dangerous societies (see also Magalasi 2015). Citizens and government elites alike worried that male movie-watchers would become vicious toward women (The Nation, August 19, 2015). Additional movies that were available for viewing or purchase, especially in the later years of my analysis, were dramas from India and Nigeria. Despite being less violent, the movies still depicted women as submissive to men (Chimbuto 2016:92-98; see also Gray 2014:989). There were also many reports of video parlors regularly showing foreign pornographic films, usually in secret late at night due to strict censorship laws but other times more openly (The Chief Censorship Officer, Malawi Government Censorship Board 2000; The Nation, August 30, 2013). As with the tabloid, many people expressed concerns about this content, and some women may have avoided video parlors because being seen there could lead to rumors and suspicion (Mpondaminga 2001). Some popular songs by Malawian artists and domestic commercials on the radio were not modeled after foreign entertainment media yet nonetheless justified MVAW (Gender Links 2003; see also Englund 2011:117-118; Lwanda 2010:385-386). For example, the song Choncho Ndi Amunanga (All the Same, He is My Husband) drew from the principle of kulangiza and stated bluntly: “even though he beats me, I don’t mind,” and “beating is medicine for marriage” (Nthala 2013:187; see also Mlenga 2011). Gender stereotypes were also present on the margins

21 After the government and several citizens brought lawsuits against The Weekend Times, it succumbed to pressure and closed on January 31, 2014 (Nyasa Times, January 31, 2014).

20 in other types of entertainment-based media, such as commentary by sports announcers, speeches by religious icons, and advertisements. However, a fair amount of the available entertainment- based mass media content in Malawi that featured gender stereotypes either came directly from foreign entertainment media companies or was partly based on foreign models.

The Relationship between Mass Media Exposure and Stated Attitudes about Men’s Violence against Women Having established the potential pathways via mass media for the diffusion of different public cultural scripts about MVAW, I next examine the reach of these scripts at the individual level. I use logistic regression to examine the relationship between people’s mass media exposure and stated rejection of MVAW. Table 1 lists the descriptive statistics for all variables included in my models.22 78 percent of women and 84 percent of men stated that they rejected MVAW. There is considerable variation in the number of MVAW newspaper articles published one month (30 days) prior to respondent’s personal interview dates, ranging from one to 29 with a mean of 13.42 (SD=6.54) among women and 13.83 (SD=6.50) among men. Just over 21 percent of men reported at least weekly newspaper use whereas only 10 percent of women did. Percentages were nearly equivalent for television use. Radio listening was higher: on average women reported having heard just over two MVAW radio programs in the past few months, and 49 percent of women said they listened to the radio at least weekly. Men reported hearing just under three such programs on average, and 67 percent were weekly radio listeners. [Table 1] I conduct four logistic regression models predicting Malawians’ expressed rejection of MVAW, each replicated separately for women and men. In what follows, I focus my discussion on the magnitude and direction of the observed average marginal effects of independent variables capturing mass media exposure. Thereafter, I review the results for known predictors stemming from theories of global cultural diffusion, household resource and bargaining, and socioecological approaches. Newspapers Results from the four models are provided in Table 2. Starting with Model 1, I observe average marginal effects (AMEs) of 0.005 (CI=0.004, 0.007) for women and 0.003 (CI=0.002, 0.005) for men for each additional MVAW newspaper article published one month prior to when a respondent was interviewed.23 In other words, each MVAW article is associated with a 0.5 percentage point increase in the probability of expressing rejection of MVAW for women and a 0.3 percentage point increase for men. This relationship is better appreciated in Figure 10, which displays the predicted probabilities of stated rejection by the number of MVAW articles. Women’s probability of declaring they reject MVAW if only one such article had been published is 71 percent. It gradually increases with each additional article, reaching 79.4 percent when 15 MVAW articles were published and 86 percent when 29 MVAW articles appeared (the highest value observed). Men’s probability of stating they reject MVAW progressively rises from 79.5 percent when one MVAW article had been published to 88.8 percent when 29 such articles were.

22 Correlation matrices are in Appendix C. 23 Bivariate logistic regression models testing the association between MVAW newspaper articles and respondents’ stated rejection of MVAW report AMEs of 0.011 (CI=0.010, 0.012) for women and 0.006 (CI=0.004, 0.007) for men. When I include district and survey fixed effects, I find AMEs of 0.006 (CI=0.005, 0.007) for women and 0.004 (CI=0.002, 0.006) for men.

21 These increases are substantial and larger than, for example, those associated with the full range of educational achievement, a commonly used indicator for exposure to developmental scripts.24 [Table 2] [Figure 10] In Model 1, I also find a small yet positive AME of 0.015 (CI=0.002, 0.028) for weekly newspaper use on women’s declarative rejection of MVAW. The direction of this association is consistent with the interpretation that women who read the newspaper weekly would be especially likely to read articles about MVAW that in turn inform their attitudinal expression. However, there is no clear association between these variables for men. One explanation for this difference is that men may be comparatively more likely to read The Weekend Times tabloid (Mchakulu 2018), thus conflating the types of scripts about MVAW that men in particular would be exposed to when reading newspapers. I further replicate Model 1 and add an interaction effect between MVAW newspaper articles and weekly newspaper use (Appendix D). This tests whether the association between expressed rejection and MVAW articles is heightened for regular newspaper readers, who in theory may be more directly exposed to such articles. I find no evidence of an interaction effect. Personal newspaper use may capture people’s consumption of the daily newspapers and The Weekend Times tabloid, especially for men, so it may not serve as a clear indicator of either exposure to anti-MVAW scripts or gender stereotypes. Moreover, the influence of MVAW articles may extend beyond people who potentially could have read these articles themselves— weekly newspaper users—to those that do not read the newspaper weekly. In this interpretation, newspaper consumers, who would have to be literate in English, communicate the information they read to others and it then spreads across the social environment. This aligns with qualitative research showing that many people in Malawi are very proactive in seeking out, sharing, and discussing novel information, especially when it related to developmental scripts (Morfit 2011; Swidler and Watkins 2009). In theory, temporal fluctuations in the number of actual cases of MVAW could underlie the variation in MVAW articles, but it is important to recall how such abuse often goes untold. In additional analyses reported in Appendix D, I explore only examining articles about either MVAW campaigns or MVAW cases. Campaign articles explicitly state anti-MVAW scripts but generally treat MVAW abstractly, whereas case articles personalize MVAW but only implicitly share anti-MVAW scripts. My analyses show that, for women, both types of articles are positively associated with declared rejection of MVAW. For men only MVAW cases are positively associated with stated rejection when both types of articles are included in the model. This suggests men in Malawi may gravitate to stories that blame specific male perpetrators for particular violent events rather than generalized calls for gender equality or broad criticisms of men’s abuse (see Riley and Dodson 2016). Radio and Television I next turn to the cross-sectional associations between people’s radio and television use and their expressions of rejecting MVAW. Model 2 reports the association between an individual hearing MVAW radio programs and their stated rejection. I rely only on the 2000, 2004, and 2010 surveys for this model, as the relevant questions were not asked in the 2015-16 survey. For

24 Model 1 indicates that each additional year of schooling is associated with strong and robust AMEs on stated rejection of MVAW of 0.009 (CI=0.007, 0.010) for women and 0.007 (CI=0.005, 0.009) for men. This translates to 9.2 and 6.9 percentage point increases in stated rejection for women and men, respectively, across the full educational trajectory measured (no education to some tertiary or more).

22 women, there is no noticeable association. There is for men, for whom listening to a MVAW radio program has an AME of 0.009 (CI=0.003, 0.015), which translates to a 3.6 percentage point increase in declaring one’s rejection of MVAW for those that listen to all four programs. Interestingly, generic weekly radio use is associated with a 2.1 percentage point increase in stated rejection for women, and a lower and noisy 0.8 percentage point increase for men (with a negative lower bound in the estimate’s confidence interval). It is important to recall the multivalent nature of radio content, especially popular songs that directly justify MVAW (Englund 2011; Nthala 2013). I further consider whether hearing MVAW radio programs is associated with a higher increase in the probability of expressing rejection of MVAW among those who listen to the radio regularly than those who do not. I replicate Model 2 and include an interaction effect between MVAW radio programs heard and weekly radio use. To provide an accurate interpretation, I present results from interactions visually and in terms of predicted probabilities in Figure 11 (Long and Mustillo 2018; Mize 2019). For women and men that do not use the radio weekly, there is little difference in the probability of stated rejection based on the number of MVAW radio programs they listen to. Among weekly radio users, though, those that listen to MVAW radio programs have higher probabilities of stated rejection than those who do not. For women, however, the difference is subtle (77.1 versus 75.4). For men, this difference equates to 5.4 percentage points (84.5 versus 79.1). This result resembles Smith et al.’s (2007) study of HIV transmission behaviors and radio programming, in which they found that consistently listening to HIV radio programs led people to identify with the drama characters. Similarly, weekly radio users that also have heard the MVAW radio programs in Malawi would be most likely to have consistent exposure to the characters, which could facilitate their ability to relate to them. [Figure 11] I now turn to the relationship between individuals’ expressed rejection of MVAW and their “television” use, which refers to content aired on domestic and satellite television stations as well as movies. Before presenting these results, it is important to recall gender differences in accessible television programming from the content analysis. Men’s television consumption disproportionately favors content sustaining gender stereotypes—such as available movies— while women’s television consumption likely exposes them to comparatively mixed content including MVAW television programs, satellite channels, and—to a lesser extent—available movies. As shown in Model 3, there is a negligible association between television use and stated rejection for women, perhaps reflective of the array of anti-MVAW scripts and gender stereotypes that they may be exposed to given the varied forms of television they may watch. For men, conversely, there is a robust 2.5 percentage point decrease in stated rejection for men. Importantly, the cross-sectional nature of the data means selection processes likely explain at least part of the observed association. While the control variables partially help mitigate selection effects, men that are more accepting of MVAW might seek out opportunities to view television knowing that much of the content features gender stereotypes consistent with their views. In such cases, men likely have their stated attitudes strengthened and reinforced (Zaller 1996). Model 4 builds on the prior models and includes data from all four surveys. Each of the independent variables related to mass media exposure is included except the number of MVAW radio programs heard because these questions were not asked in the 2015-16 Malawi DHS. For women and men, the AMEs for MVAW newspaper articles and television use in Model 4 are nearly identical to Models 1 and 3, respectively. The associations for radio use change somewhat between Models 2 and 4, with the percentage point increase in expressed rejection of MVAW

23 among women that use the radio weekly halves, while this association for men greatly increases in the absence of the more specific measure of having heard MVAW radio programs. In terms of comparative magnitude, the most important results from Models 1-4 are the substantial increase in the probability of stated rejection across the range of MVAW articles shown in Figure 10, and negative association between stated rejection and television use for men.25 Controls and Robustness Checks The results for control variables are quite consistent across Models 1-4. Education and urban living are associated with substantial increases in the probability of declaring that one rejects MVAW. Christian identification is positively associated with expressed rejection, but the relationship is not very robust. I further find that working in a non-agricultural job is negatively associated with stated rejection among Malawian women, similar to Cools and Kotsadam’s (2017) observation in their cross-national analysis of African countries (see also Charles 2020; Rani and Bonu 2009). Reports from Malawian NGOs and the daily newspapers document pervasive gender discrimination at work, including male bosses demanding sex from female employees (e.g., The Nation, May 1, 2009; Chirwa 1999). Household wealth is positively associated with declarations of rejection, consistent with the literature (Cools and Kotsadam 2017; Kurzman et al. 2019). So is matrilineal lineage, but the association is much weaker perhaps due to changing patterns of land ownership rights or greater variation in lineage practices than can be captured with a single dichotomous variable (see Chiweza 2005; Kaarhus 2010). Alternatively, individuals’ matrilineal lineage might have less purchase on their stated attitudes than the dominant lineage practices at the community or district level. The relationship between marital status and stated rejection of MVAW is positive and especially strong. For women, having been or currently being married—whether a monogamous or polygamous relationship—is consistently associated with a 4.4-7.0 percentage point increase in expressed rejection compared to never being married. For men, being currently in a monogamous marriage, as opposed to never being married, is associated with a 5.7-5.8 percentage point increase in stated rejection, while being formerly married or currently in a polygamous relationship is positively associated with stated rejection, but the AMEs are comparatively small and less robust. The AMEs for the final control variable of age confirm prior literature that older age is strongly associated with greater expression of MVAW rejection in Malawi. Further research is necessary to examine whether this is due to gradual lifetime exposure to anti-MVAW scripts or other life course factors like temporal distance from youth fighting norms or experience being a parent, especially of older daughters. In robustness checks reported in Appendix D, I alter how I measure the mass media variables, including examining multiple ways of capturing the recent publication of MVAW newspaper articles, using distinct measures for each MVAW radio program, and employing categorical variables for personal newspaper, radio, and television use. I also adjust my modeling approach by leveraging a more limited set of controls to avoid potential collider biases and also using random effects for districts and MVAW newspaper articles. Though informative, the results from these robustness checks do not alter my substantive conclusions.

Discussion and Conclusion

25 I also examine whether the AMEs associated with exposure to anti-MVAW scripts might be mitigated by exposure to gender stereotypes. To do so, I replicate Model 4 and add an interaction effect between MVAW newspaper articles and weekly television use. There is no evidence of mitigation.

24 A long line of sociological research has established the importance of global cultural diffusion (Boyle et al. 2002; Drori and Krücken 2009; Thornton et al. 2015). Transnational organizations such as the United Nations play a central role in the circulation of public cultural scripts about development and human rights, which have contributed—alongside many other factors—to many countries’ adoption of similar laws and policies across numerous social domains (Boli and Thomas 1999; Merry 2006). Though decoupling is rampant, many nations’ practices have also gradually shifted in response to their rhetorical commitments (Cole and Ramirez 2013). The most recent work in this area emphasizes the emergence of illiberal institutions and their role in the spread of alternative scripts competing with or circulating alongside that of transnational organizations (Boyle et al. 2015; 2017a; Ferguson 2019; Schofer et al. 2019; Velasco 2020b). Lingering in the background is whether and how such scripts reach ordinary citizens (Pope and Meyer 2016:296-298). Drawing upon work on developmental idealism (Thornton 2005), I seek to identify and describe the pathways of global cultural diffusion that flow at the individual level. Diffusion Pathways in Malawi I focused on the case of public cultural scripts about men’s violence against women (MVAW) as circulated via mass media in Malawi between 2000 and 2016. Combining content analysis and logistic regression, I traced some of the mass media pathways that bring scripts condemning such violence and supporting gender stereotypes to people in Malawi. I found that transnational organizations partnered with domestic NGOs to host training meetings for journalists in which they promoted anti-MVAW scripts and paid journalists for producing such content. As a result, coverage of MVAW was high in the daily newspapers throughout the time period and often rose when transnational organizations and Malawian activists were especially engaged in initiatives to publicly denounce MVAW, such as the annual 16 Days of Activism against Gender Violence campaign. At the same time, entertainment media companies—which were often foreign—distributed alternative scripts that portrayed gender stereotypes across several mass media sources in Malawi, namely: satellite television, movies, and The Weekend Times tabloid. These forms of entertainment media were very popular among men, whereas women likely avoided such content given its messaging and the negative implications that being seen as one who desires this content could have for them. In the statistical analysis, people’s exposures to mass media content favoring anti- MVAW scripts or gender stereotypes were divergently associated with their expressed rejection of MVAW on national surveys. When more MVAW newspaper articles were published, people were much more likely to say they rejected MVAW. In general, radio use and having heard MVAW radio programs was associated with greater probabilities of stated rejection. In contrast, I observed a negative relationship between men’s personal “television” use and their expressed rejection of MVAW, which is consistent with the finding from the content analysis that of all forms of available mass media, gender stereotypes were most circulated in the television programming men consumed. The divergent associations are not causal effects on people’s expressed attitudes, but they nevertheless suggest that exposure to anti-MVAW scripts encourages survey respondents to express rejection of MVAW whereas exposure to gender stereotypes may reinforce the justification of MVAW. Given the general zealousness with which many people in Malawi seek to learn the language of public cultural scripts about development and human rights, it is an especially useful case for examining the diffusion pathways of such scripts. In addition, the layered moral background in Malawi is amenable to the seemingly contradictory circulation of both anti-

25 MVAW scripts and gender stereotypes. Other contexts with more opposition to foreign models or greater stated acceptance of MVAW could help identify the necessary institutional factors for diffusion to occur. In addition, other contextual factors related to gender, social class, ethnicity, and socioeconomic status that differ from Malawi could substantially alter the patterns I observe. For example, diffusion pathways might be somewhat constrained across castes in India or linguistic groups in Indonesia. Situational dynamics are always relevant, and there are undoubtedly many stories of global cultural diffusion, not one (Watkins and Hodgson 2019). The case of Malawi is especially useful for theorizing how global cultural diffusion may occur at the individual level and the multifaceted landscape of scripts that are circulated worldwide. Extending Global Cultural Diffusion to the Micro Level Theoretical claims about macro-micro links are present in world society theory as well as developmental idealism accounts of global cultural diffusion (e.g., Meyer 1986a; 1986b; Thornton 2005), but I connect these literatures in showing how transnational organizations’ efforts to globally circulate public cultural scripts about development and human rights via mass media can reach beyond national governments and down to the level of individual people. However, all forms of mass media exposure are not associated with declarative attitudes aligned with developmental scripts. Powerful entertainment media companies are globally dispersing their own set of scripts that also reach individuals. Moreover, people’s social identities, such as their gender, shape their possible exposure. Similar elaborations about macro-micro diffusion pathways are equally necessary for other theorized sources of global cultural diffusion like education, urban environments, NGO networks, and foreign aid projects. Internet content with its individually-tailored, multi-modal content, and unique interactive features (Cacciatore et al. 2016) is also exceptionally relevant. Internet access and social media use are still low in Malawi and somewhat limited in several other places, but globally they are increasing exponentially. Though such pathways reach ordinary citizens, this does not mean that their influence hinges on individuals’ having direct exposure. In fact, the largest AMEs on stated rejection of MVAW that I observed were for the number of MVAW newspaper articles published one month prior to respondent’s personal survey interview date, regardless of whether they personally read the newspaper. This resembles Durkheimian arguments by Jepperson and Meyer (2011) that public cultural material can orient the social environment in which individuals are embedded (see also Katz and Lazarfeld 1955). It points to a need for further inquiry into the role of group dynamics in spreading imported scripts (Prentice and Paluck 2020; Vaisey and Lizardo 2010) and how people’s personal life experiences, current circumstances, and relationships inform their understanding of different scripts (McDonnell et al. 2017). When foreign public cultural scripts are imported to a society, they may not reach people equally, as my analysis demonstrated with respect to gender identity. Since women in Malawi would risk being labeled as promiscuous for consuming some entertainment media like The Weekend Times or being seen hanging out regularly at the video parlor, they likely are less exposed than men to the cultural content circulated via these forms of mass media. This played out in the regression results, as women’s weekly newspaper use—likely mostly consisting of the daily newspapers that are littered with anti-MVAW scripts—was positively associated with expressing rejection of MVAW, while this was not the case for men’s weekly newspaper use, which was more likely a combination of the daily newspapers and the tabloid with its explicit portrayal of gender stereotypes. Correspondingly, television use was negatively associated with stated rejection only for men, who would have been more likely than women to attend the video parlors that showed foreign movies with strong gender stereotypes.

26 Scripts’ potential reach across individuals also can vary based on how they are presented. I found that women’s declarative rejection of MVAW was associated with MVAW newspaper articles regardless of whether those articles explicitly supported anti-MVAW scripts or implicitly did so by discussing particular instances of abuse. Among men, only articles about MVAW cases were positively associated with their expressions of rejecting MVAW. It could be that some men felt resentful of discussions questioning men’s patriarchal authority (Riley and Dodson 2016). Conversely, some men might not have felt comfortable bringing up such normative arguments, so articles about MVAW campaigns would not have stimulated supportive discussion within men’s conversation networks. Regardless of the mechanism, the result highlights the importance of message presentation for different audiences: narratives versus structural arguments might fall differently on the minds of the privileged and discriminated (see also Davis and Robinson 1991; de Graaf et al. 2012; Robinson and Bell 1978). This insight could inform current research streams on the comparative efficiency of narratives versus facts in political persuasion (e.g., Hagmann et al. 2021; Kubin et al. 2021; Murrar and Brauer 2019). My analysis also adds to the literature documenting how intermediary brokers often act as local translators of imported scripts for individuals in a receiving population (Jijon 2019; Peters 2020; Swidler and Watkins 2017; Ward 2020). In Malawi, journalists’ relationships with transnational organizations and the domestic NGOs they contract were very important for the incorporation of developmental scripts in domestic mass media. The Malawian intermediaries involved in the diffusion of foreign entertainment media somewhat curated available content, such as the video parlor owners that made decisions about which movies to show publicly. Research on how intermediary brokers share, block, or modify foreign scripts is paramount, as well as the conditions under which brokers do and do not proliferate (Footitt et al. 2020). A Multifaceted View of Global Cultural Diffusion Specifying diffusion pathways broadens theoretical considerations of what types of public cultural scripts are transmitted on a global scale. Within sociological research on global cultural diffusion, theoretical interest primarily has been focused on the dissemination of scripts about development and human rights. This is apparent in the long-standing use of the terms “global cultural scripts” and “world cultural models” in the literature to refer to scripts about development and human rights. As is being increasingly recognized given the recent expansion of illiberal politics worldwide (e.g., Schofer et al. 2019), transnational organizations like the United Nations and the World Bank are not the only types of players that circulating scripts on a global scale, and institutional explanations of global cultural diffusion must be broadened to examine multiple types of cultural content (Almeida and Chase-Dunn 2018; Lounsbury and Wang 2020; Wimmer 2021:1419-1420). This includes the creation of empirical measures of the spread of other types of cross-national organizations besides the commonly used proxy indicators of the number of international NGO memberships and the KOF globalization index (Dreher 2006; see Velasco 2020b). For mass media, this could be content-based measures like the number of recent MVAW newspaper articles I constructed, or higher level variables capturing the circulation of content produced by large entertainment media companies like MultiChoice. Besides recognizing that a variety of organizations and actors can circulate public ideas cross-nationally, it is equally important to stress differences in societies’ abilities to export public ideas and shape cultural dynamics in other societies. Malawi is not in this position and thus is particularly subject to foreign influence. As a contrast, consider that, as of 2017, four of the five largest global media companies were still based in the United States (Thussu 2018:53). Of course, there are some signs of change. Bollywood in India, Nollywood in Nigeria, and other

27 major industry sources of entertainment media from China, South Korea, and Brazil are growing rapidly, but they are still small in comparison (Boyd-Barrett and Mirrlees 2019). Outside of mass media, a similar power dynamic is apparent. For example, when it comes to deciding the International Monetary Fund’s programs and policies, the United States wields extraordinary influence even though there are 190 member countries (Kentikelenis and Babb 2019; Kentikelenis and Seabrooke 2017). This familiar structure points to a conception of cultural imperialism in which powerful societies may transnationally circulate the same culture wars present in their own societies including the dominant public cultural frames (e.g., liberal versus conservative in the United States) as well as which social issues should be a matter of public concern in the first place. For example, Velasco (2020b) shows how public debates about LGBT rights in Europe and the United States are reproduced across societies worldwide as organizations evangelize their views abroad. Likewise, civil conflicts over globalist versus nationalist politics in Brazil, England, India, and the United States are now informing similar debates in other locales (Carothers and O’Donohue 2019). Moving Beyond Declarative Attitudes Besides capturing the reach of transnationally circulated scripts regarding what people say, an especially pressing avenue for future research on global cultural diffusion at the individual level is how exposure to such scripts shapes how people act, especially with respect to gender relations (Allendorf and Thornton 2015; Cloward 2016). How does exposure to anti- MVAW scripts, gender stereotypes, and other forms of cultural messaging affect whether men abuse women, the age at which people first become partnered, if people use contraceptives, and how labor is distributed across household members? Moreover, what roles do declarative and nondeclarative personal culture play in such behavior? Regarding MVAW specifically, distinct temporal trends present a puzzling global paradox: in nearly all countries with DHS and other comparable survey data, people’s probability of saying they reject MVAW increased during the twenty-first century, but women’s self-reports of being abused by men remained steady and in Malawi they actually increased from 21 to 26 percent between 2004 and 2015-16 (ICF 2020; WHO 2021). Research on the national level conditions of “decoupling” between human rights commitments and practices offers a useful lens for thinking through this paradox. Changes in rhetorical commitments at first outpace changes in practices (Hafner-Burton and Tsutsui 2005). The commitment-practice gap initially appears to widen, as awareness of practices that are not aligned with these commitments become more monitored (Fariss 2014). Over time, that gap can narrow among democratic-leaning governments when they feel domestic and international pressure to comply with their human rights commitments and they receive aid that helps them to do so (Cole and Ramirez 2013; Wei and Swiss 2020). Coercive compliance on matters of gender and sexuality, though, sometimes can backfire (Cole 2013; Velasco 2020a). The paradoxical trends in stated attitudes toward MVAW and women’s self-reports of experiencing MVAW might be following a parallel process, and currently many societies might be undergoing a moment of growing consciousness leading to an increase in women’s ability to say they have been abused, such as activists in Malawi suggest is happening there (The Daily Times, June 6, 2012). At the same time, actual abuse could be increasing, or at least steady, if efforts to diffuse anti-MVAW scripts are causing a retaliatory backlash effect among men. These possibilities are difficult to evaluate because self-reported experience of violence can often vary from reality for a variety of reasons (Acosta 2021). Women can face retribution for acknowledging the abuse they experience or they may want to avoid the label of “victim” (Boyle

28 and Rogers 2020; Hansen et al. 2021; Khan et al. 2018). Conversely, after exposure to anti- MVAW scripts women might feel newly emboldened to be honest about their partner’s abuse or they might recall their interactions and relationship in a different light. Future research must carefully analyze the connections between people’s exposure to varying scripts about MVAW and gender relations, expressed attitudes, self-reported life experiences, and behaviors over time. Triangulating survey-based measures with data from administrative records, court case databases, and online search data and other computational sources, as well as the use of multiple methods like list experiments, in-depth interviews, and ethnographic observations where possible, will be crucial for deciphering between changes in reporting about violence and actual violence (e.g., Cullen 2020; Gibson et al. 2020; McDougall et al. 2018; Medie 2020; Schuler et al. 2011; c.f., Agüero and Frisancho 2021; Agüero et al. 2020). Overall, researchers of MVAW have made massive strides in measurement, but improvement is imperative (WHO 2021:38-40). I uncover in this study how multiple public cultural scripts about MVAW are disseminated through mass media to people in Malawi. Transnational organizations like the United Nations as well as entertainment media companies—sometimes working through Malawian intermediaries—spread unique scripts through media content to individual people. These concurrent sources of global cultural diffusion constitute a dynamic social environment in which multiple scripts about a single issue are promoted on a global scale and reach individual people.

29 References

Abend, Gabriel. 2014. The Moral Background: An Inquiry into the History of Business Ethics. Princeton and Oxford: Princeton University Press. Abramsky, Tanya, Karen Devries, Ligia Kiss, Janet Nakuti, Nambusi Kyegombe, Elizabeth Starmann, Bonnie Cundill, Leilani Francisco, Dan Kaye, Tina Musuya, Lori Michau, and Charlotte Watts. 2014. “Findings from the SASA! Study: A Cluster Randomized Controlled Trial to Assess the Impact of a Community Mobilization Intervention to Prevent Violence against Women and Reduce HIV Risk in Kampala, Uganda.” BMC Medicine 12(122):1–17. Acosta, Laura. 2021. “Victimhood Dissociation and Conflict Resolution: Evidence from the Colombian Peace Plebiscite.” Theory and Society. Agüero, Jorge M., Úrsula Aldana, Erica Field, Veronica Frisancho, and Javier Romero. 2020. “Is Community-Based Targeting Effective in Identifying Intimate Partner Violence?” AEA Papers and Proceedings 110:605–9. Agüero, Jorge M., and Veronica Frisancho. 2021. “Measuring Violence against Women with Experimental Methods.” Economic Development and Cultural Change. Ai, Chunrong, and Edward C. Norton. 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters 80(1):123–29. Allendorf, Keera, and Arland Thornton. 2015. “Caste and Choice: The Influence of Developmental Idealism on Marriage Behavior.” American Journal of Sociology 121(1):243– 87. Allison, Paul D. 1999. “Comparing Logit and Probit Coefficients across Groups.” Sociological Methods & Research 28(2):186–208. Almeida, Paul, and Chris Chase-Dunn. 2018. “Globalization and Social Movements.” Annual Review of Sociology 44(1):189–211. Angotti, Nicole, Margaret Frye, Amy Kaler, Michelle Poulin, Susan Cotts Watkins, and Sara Yeatman. 2014. “Popular Moralities and Institutional Rationalities in Malawi’s Struggle against AIDS.” Population and Development Review 40(3):447–473. Angotti, Nicole, and Amy Kaler. 2013. “The More You Learn the Less You Know? Interpretive Ambiguity across Three Modes of Qualitative Data.” Demographic Research 28(33):951–80. Arias, Eric. 2019. “How Does Media Influence Social Norms? Experimental Evidence on the Role of Common Knowledge.” Political Science Research and Methods 7(3):561–78. Ayoub, Phillip M., and Jeremiah Garretson. 2017. “Getting the Message Out: Media Context and Global Changes in Attitudes toward Homosexuality.” Comparative Political Studies 50(8):1055–85. Bail, Christopher A., Taylor W. Brown, and Andreas Wimmer. 2019. “Prestige, Proximity, and Prejudice: How Google Search Terms Diffuse across the World.” American Journal of Sociology 124(5):1496–1548. Banerjee, Abhijit, Eliana La Ferrara, and Victor Orozco. 2019. “Entertainment, Education, and Attitudes toward Domestic Violence.” AEA Papers and Proceedings 109:133–37. Barkvoll, Benedicte. 2009. “Intimate Partner Violence in Malawi and the Gender Order: A Quantitative Study Examining the Association between IPV in Malawi and the Gender Order.” University of Oslo. Bartels, Larry M. 1993. “Messages Received: The Political Impact of Media Exposure.” American Political Science Review 87(2):267–85.

30 Behrman, Julia, and Margaret Frye. 2021. “Attitudes toward Intimate Partner Violence in Dyadic Perspective: Evidence from Sub-Saharan Africa.” Demography 9115955. Bischof, Daniel. 2017. “New Graphic Schemes for Stata: Plotplain and Plottig.” The Stata Journal 17(3):748–59. Bleakley, Amy, Patrick E. Jamieson, and Daniel Romer. 2012. “Trends of Sexual and Violent Content by Gender in Top-Grossing U.S. Films, 1950–2006.” Journal of Adolescent Health 51(1):73–79. Boli, John, and George M. Thomas, eds. 1999. Constructing World Culture: International Non- Governmental Organizations since 1875. Palo Alto, CA: Stanford University Press. Bollen, Kenneth A., Paul P. Biemer, Alan F. Karr, Stephen Tueller, and Marcus E. Berzofsky. 2016. “Are Survey Weights Needed? A Review of Diagnostic Tests in Regression Analysis.” Annual Review of Statistics and Its Application 3(1):375–92. Boyd-Barrett, Oliver, and Tanner Mirrlees. 2019. Media Imperialism: Continuity and Change. Rowman & Littlefield. Boyle, Elizabeth H., Minzee Kim, and Wesley Longhofer. 2015. “Abortion Liberalization in World Society, 1960–2009.” American Journal of Sociology 121(3):882–913. Boyle, Elizabeth Heger, Shannon Golden, and Wenjie Liao. 2017. “The Catholic Church and International Law.” Annual Review of Law and Social Science 13(1):395–411. Boyle, Elizabeth Heger, and Andrea Hoeschen. 2001. “Theorizing the Form of Media Coverage Over Time.” Sociological Quarterly 42(4):511–27. Boyle, Elizabeth Heger, Miriam King, and Matthew Sobek. 2017. IPUMS-Demographic and Health Surveys: Version 4.0 [Dataset]. Minnesota Population Center and ICF International: NICHD Grant Number R01HD069471. Boyle, Elizabeth Heger, Barbara J. McMorris, and Mayra Gomez. 2002. “Local Conformity to International Norms: The Case of Female Genital Cutting.” International Sociology 17(1):5– 33. Boyle, Kaitlin M., and Kimberly B. Rogers. 2020. “Beyond the Rape ‘Victim’–‘Survivor’ Binary: How Race, Gender, and Identity Processes Interact to Shape Distress.” Sociological Forum 35(2):323–45. Breen, Richard, Kristian Bernt Karlson, and Anders Holm. 2018. “Interpreting and Understanding Logits, Probits, and Other Nonlinear Probability Models.” Annual Review of Sociology 44(1):39–54. Brosius, Anna, Erika J. van Elsas, and Claes H. de Vreese. 2019. “Trust in the European Union: Effects of the Information Environment.” European Journal of Communication 34(1):57–73. Bursztyn, Leonardo, Alessandra L. González, and David Yanagizawa-Drott. 2020. “Misperceived Social Norms: Women Working Outside the Home in Saudi Arabia.” American Economic Review 110(10):2997–3029. Cacciatore, Michael A., Dietram A. Scheufele, and Shanto Iyengar. 2016. “The End of Framing as We Know It … and the Future of Media Effects.” Mass Communication and Society 19(1):7–23. Carbone, Luca, and Jonathan Mijs. 2021. “Sounds like Meritocracy to My Ears: Exploring the Link between Inequality in Popular Music and Private Culture.” SocArXiv. Carle, Adam C. 2009. “Fitting Multilevel Models in Complex Survey Data with Design Weights: Recommendations.” BMC Medical Research Methodology 9:49. Carothers, Thomas, and Andrew O’Donohue, eds. 2019. Democracies Divided: The Global Challenge of Political Polarization. Washington, D.C: Brookings Institution Press.

31 Chang, Wei, Lucía Díaz-Martin, Akshara Gopalan, Eleonora Guarnieri, Seema Jayachandran, and Claire Walsh. 2020. “What Works to Enhance Women’s Agency: Cross-Cutting Lessons from Experimental and Quasi-Experimental Studies.” Abdul Latif Jameel Poverty Action Lab (J-PAL) Working Paper. Charles, Maria. 2020. “Gender Attitudes in Africa: Liberal Egalitarianism across 34 Countries.” Social Forces 99(1):86–125. Chepuka, Lignet, Miriam Taegtmeyer, Genesis Chorwe-Sungani, Janet Mambulasa, Ellen Chirwa, and Rachel Tolhurst. 2014. “Perceptions of the Mental Health Impact of Intimate Partner Violence and Health Service Responses in Malawi.” Global Health Action 7(1):24816. Chimbiri, Agnes M. 2007. “The Condom Is an ‘Intruder’ in Marriage: Evidence from Rural Malawi.” Social Science & Medicine 64(5):1102–15. Chimbuto, Joseph George Salijeni. 2016. “The Representation of Women in Nollywood Films: An Investigation of Its Impact on Audiences in Malawi.” University of Roehampton. Chiweza, Asiyati Lorraine. 2005. “Women’s Inheritance Rights in Malawi: The Role of District Assemblies.” Development in Practice 15(1):83–89. Clark, Rob. 2012. “Bringing the Media in: Newspaper Readership and Human Rights.” Sociological Inquiry 82(4):532–56. Cloward, Karisa. 2016. When Norms Collide: Local Responses to Activism against Female Genital Mutilation and Early Marriage. New York: Oxford University Press. Cole, Wade M. 2013. “Government Respect for Gendered Rights: The Effect of the Convention on the Elimination of Discrimination against Women on Women’s Rights Outcomes, 1981– 2004.” International Studies Quarterly 57(2):233–249. Cole, Wade M., and Francisco O. Ramirez. 2013. “Conditional Decoupling: Assessing the Impact of National Human Rights Institutions, 1981–2004.” American Sociological Review 78(4):702–25. Cools, Sara, and Andreas Kotsadam. 2017. “Resources and Intimate Partner Violence in Sub- Saharan Africa.” World Development 95:211–30. Correll, Shelley J., Cecilia L. Ridgeway, Ezra W. Zuckerman, Sharon Jank, Sara Jordan-Bloch, and Sandra Nakagawa. 2017. “It’s the Conventional Thought That Counts: How Third-Order Inference Produces Status Advantage.” American Sociological Review 82(2):297–327. Cullen, Claire. 2020. “Method Matters: Underreporting of Intimate Partner Violence in Nigeria and Rwanda.” World Bank Policy Research Working Paper 9274. Cunningham, Scott. 2021. Causal Inference: The Mixtape. New Haven: Yale University Press. Davis, Nancy J., and Robert V. Robinson. 1991. “Men’s and Women’s Consciousness of Gender Inequality: Austria, West Germany, Great Britain, and the United States.” American Sociological Review 56(1):72–84. de Graaf, Anneke, Hans Hoeken, José Sanders, and Johannes W. J. Beentjes. 2012. “Identification as a Mechanism of Narrative Persuasion.” Communication Research 39(6):802– 23. de Vreese, Claes H., Mark Boukes, Andreas Schuck, Rens Vliegenthart, Linda Bos, and Yph Lelkes. 2017. “Linking Survey and Media Content Data: Opportunities, Considerations, and Pitfalls.” Communication Methods and Measures 11(4):221–44. Dreher, Axel. 2006. “Does Globalization Affect Growth? Evidence from a New Index of Globalization.” Applied Economics 38(10):1091–1110. Dorius, Shawn F., and Jeffrey Swindle. 2019. “Developmental Idealism in Internet Search Data.” Sociology of Development 5(3):286–313.

32 Drori, Gili S., and Georg Krücken. 2009. “World Society: A Theory and a Research Program in Context.” Pp. 3–35 in World Society: The Writings of John W. Meyer, edited by G. Krücken and G.S. Drori. Oxford University Press. Elkasabi, Mahmoud, Ruilin Ren, and Thomas W. Pullum. 2020. Multilevel Modeling Using DHS Surveys: A Framework to Approximate Level-Weights. DHS Methodological Reports 27. United States Agency for International Development. Elwert, Felix, and Christopher Winship. 2014. “Endogenous Selection Bias: The Problem of Conditioning on a Collider Variable.” Annual Review of Sociology 40(1):31–53. Englund, Harri. 2011. Human Rights and African Airwaves: Mediating Equality on the Chichewa Radio. Indianapolis, IN: Indiana University Press. Essary, Elizabeth Helen. 2007. “Speaking of Globalization...: Frame Analysis and the World Society.” International Journal of Comparative Sociology 48(6):509–26. Fariss, Christopher J. 2014. “Respect for Human Rights Has Improved Over Time: Modeling the Changing Standard of Accountability.” American Political Science Review 108(2):297–318. Ferguson, James. 1994. The Anti-Politics Machine: “Development,” Depoliticization, and Bureaucratic Power in Lesotho. Minneapolis, MN: University of Minnesota Press. Ferguson, Jason. 2019. “The Great Refusal: The West, the Rest, and the Geopolitics of Homosexuality.” Paper Presented at the Annual Meetings of the American Sociological Association, New York. Flynn, Mark A., Clay M. Craig, Christina N. Anderson, and Kyle J. Holody. 2016. “Objectification in Popular Music Lyrics: An Examination of Gender and Genre Differences.” Sex Roles 75(3):164–76. Footitt, Hilary, Angela Crack, and Wine Tesseur. 2020. Development NGOs and Languages: Listening, Power and Inclusion. Palgrave Macmillan. Forsyth, Selina, and Kaitlin P. Ward. 2021. “Media Use and Men’s Approval of Intimate Partner Violence in Honduras.” Journal of Interpersonal Violence 0886260521993926. Frank, David J. 1997. “Science, Nature, and the Globalization of the Environment: 1870-1990.” Social Forces 76(2):409–37. Frank, David John, and John W. Meyer. 2002. “The Profusion of Individual Roles and Identities in the Postwar Period.” Sociological Theory 20(1):86–105. Frye, Margaret. 2012. “Bright Futures in Malawi’s New Dawn: Educational Aspirations as Assertions of Identity.” American Journal of Sociology 117(6):1565–1624. Frye, Margaret. 2017. “Cultural Meanings and the Aggregation of Actions: The Case of Sex and Schooling in Malawi.” American Sociological Review 82(5):945–76. Frye, Margaret, and Nina Gheihman. 2018. “Like Bees to a Flower: Attractiveness, Risk, and Collective Sexual Life in an AIDS Epidemic.” Sociological Science 5:596–627. Fulu, Emma, and Stephanie Miedema. 2015. “Violence against Women: Globalizing the Integrated Ecological Model.” Violence against Women 21(12):1431–55. Geldof, Marije. 2011. “Earphones Are Not for Women: Gendered ICT Use among Youths in Ethiopia and Malawi.” Information Technologies & International Development 7(4):69–80. Geston, Jesse. 2006. “How Has MultiChoice Africa Affected the Way People View Television in African Countries?” Master’s Thesis, Grand Valley State University. Ghaznavi, Jannath, Katherine L. Grasso, and Laramie D. Taylor. 2017. “Increasingly Violent but Still Sexy: A Decade of Central Female Characters in Top-Grossing Hollywood and Bollywood Film Promotional Material.” International Journal of Communication 11:23–47.

33 Gibson, Mhairi A., Eshetu Gurmu, Beatriz Cobo, María M. Rueda, and Isabel M. Scott. 2020. “Measuring Hidden Support for Physical Intimate Partner Violence: A List Randomization Experiment in South-Central Ethiopia.” Journal of Interpersonal Violence 0886260520914546. Gilman, Lisa. 2011. The Dance of Politics: Gender, Performance, and Democratization in Malawi. Philadelphia, PA: Temple University Press. Givens, Jennifer E., and Andrew K. Jorgenson. 2013. “Individual Environmental Concern in the World Polity: A Multilevel Analysis.” Social Science Research 42(2):418–31. Glasman, Laura R., and Dolores Albarracín. 2006. “Forming Attitudes That Predict Future Behavior: A Meta-Analysis of the Attitude–Behavior Relation.” Psychological Bulletin 132(5):778–822. Goode, William J. 1971. “Force and Violence in the Family.” Journal of Marriage and Family 33(4):624–36. Grant, Monica J. 2015. “The Demographic Promise of Expanded Female Education: Trends in the Age at First Birth in Malawi.” Population and Development Review 41(3):409–38. Gray, Jonathan. 2011. “Mobility through Piracy, or How Steven Seagal Got to Malawi.” Popular Communication 9(2):99–113. Gray, Jonathan. 2014. “Scales of Cultural Influence: Malawian Consumption of Foreign Media.” Media, Culture & Society 36(7):982–97. Grimm, Michael, Robert Sparrow, and Luca Tasciotti. 2015. “Does Electrification Spur the Fertility Transition? Evidence from Indonesia.” Demography 52(5):1773–96. Hadler, Markus. 2016. “Individual Action, World Society, and Environmental Change: 1993– 2010.” European Journal of Cultural and Political Sociology 3(2–3):341–74. Hadler, Markus. 2017. The Influence of Global Ideas on Environmentalism and Human Rights: World Society and the Individual. London: Palgrave Macmillan. Hafner-Burton, Emilie M., and Kiyoteru Tsutsui. 2005. “Human Rights in a Globalizing World: The Paradox of Empty Promises.” American Journal of Sociology 110(5):1373–1411. Hagmann, David, Julia A. Minson, and Catherine H. Tinsley. 2021. “Personal Narratives Build Trust Across Ideological Divides.” OSF Preprints. Hansen, Maria, Kari Stefansen, and May-Len Skilbrei. 2020. “Non-Reporting of Sexual Violence as Action: Acts, Selves, Futures in the Making.” Nordic Journal of Criminology. Harris, Suzanne Temwa. 2017. “Synthesising Media, Politics and Foreign Intervention: An Examination into Malawi’s Media System Transformation.” Doctor of Philosophy in Communications, Hong Kong Baptist University. Hayford, Sarah R., Sarah Garver, Abdramane B. Soura, Yuk Fai Cheong, Rose Grace Grose, and Kathryn M. Yount. 2020. “Community Influences on Female Genital Mutilation/Cutting: A Comparison of Four Francophone West African Countries.” Studies in Family Planning 51(1):3–32. Hendi, Arun S. 2017. “Globalization and Contemporary Fertility Convergence.” Social Forces 96(1):215–38. Heise, Lori L. 1998. “Violence against Women: An Integrated, Ecological Framework.” Violence against Women 4(3):262–90. Hill, Seth J., James Lo, Lynn Vavreck, and John Zaller. 2013. “How Quickly We Forget: The Duration of Persuasion Effects from Mass Communication.” Political Communication 30(4):521–47.

34 Hindin, Michelle J. 2014. “Adolescent Childbearing and Women’s Attitudes towards Wife Beating in 25 Sub-Saharan African Countries.” Maternal and Child Health Journal 18(6):1488–95. Htun, Mala, and S. Laurel Weldon. 2012. “The Civic Origins of Progressive Policy Change: Combating Violence against Women in Global Perspective, 1975–2005.” American Political Science Review 106(03):548–569. Huang, Junming, Gavin Cook, and Yu Xie. 2021. “Large-Scale Quantitative Evidence of Media Impact on Public Opinion toward China.” ArXiv:2012.07575 [Physics]. Ibik, J. O. 1970. “Law of Marriage and Divorce: Malawi.” in Restatement of African Law, edited by A. N. Allott. London: Sweet & Maxwell. ICF. 2020. “The DHS Program STATcompiler.” Retrieved January 16, 2020 (http://www.statcompiler.com). Inkeles, Alex, and David H. Smith. 1974. Becoming Modern: Individual Change in Six Developing Countries. Cambridge, MA: Harvard University Press. Jann, Ben. 2014. “Plotting Regression Coefficients and Other Estimates.” The Stata Journal 14(4):708–37. Jeong, Michelle, and Rosie Eungyuhl Bae. 2018. “The Effect of Campaign-Generated Interpersonal Communication on Campaign-Targeted Health Outcomes: A Meta-Analysis.” Health Communication 33(8):988–1003. Jepperson, Ronald, and John W. Meyer. 2011. “Multiple Levels of Analysis and the Limitations of Methodological Individualisms.” Sociological Theory 29(1):54–73. Jesmin, Syeda S., and Iftekhar Amin. 2017. “Impact of the Mass Media in Changing Attitudes towards Violence against Women in Bangladesh: Findings from a National Survey.” Journal of Family Violence 32(5):525–34. Jijon, Isabel. 2019. “Toward a Hermeneutic Model of Cultural Globalization: Four Lessons from Translation Studies.” Sociological Theory 37(2):142–61. Joas, Hans. 2013. The Sacredness of the Person: A New Genealogy of Human Rights. Washington, D.C: Georgetown University Press. Johansson, Sofia. 2007. Reading Tabloids: Tabloid Newspapers and their Readers. Huddinge: Södertörn Högskola. Johnson, Jessica. 2018. In Search of Gender Justice: Rights and Relationships in Matrilineal Malawi. Cambridge University Press. Johnson-Hanks, Jennifer A., Christine A. Bachrach, S. Phillip Morgan, Hans-Peter Kohler, and Lynette Hoelter. 2011. Understanding Family Change and Variation: Toward a Theory of Conjunctural Action. London: Springer Verlag. Kaarhus, Randi. 2010. “Women’s Land Rights and Land Tenure Reforms in Malawi: What Difference Does Matriliny Make?” Forum for Development Studies 37(2):171–92. Kaler, Amy, Susan Cotts Watkins, and Nicole Angotti. 2015. “Making Meaning in the Time of AIDS: Longitudinal Narratives from the Malawi Journals Project.” African Journal of AIDS Research 14(4):303–14. Kalyango, Yusuf, Folker Hanusch, Jyotika Ramaprasad, Terje Skjerdal, Mohd Safar Hasim, Nurhaya Muchtar, Mohammad Sahid Ullah, Levi Zeleza Manda, and Sarah Bomkapre Kamara. 2017. “Journalists’ Development Journalism Role Perceptions.” Journalism Studies 18(5):576– 94.

35 Kambalame, John, E. P. Chidzalo, and J. W. M. Chadangalara. 2008. Our African Way of Life. Reissued with illustrations and a new foreword by Dr. Desmond D. Phiri, edited by C. Young and H. K. Banda. Blantyre, Malawi: Central Africana. Kamens, David H., John W. Meyer, and Aaron Benavot. 1996. “Worldwide Patterns in Academic Secondary Education Curricula.” Comparative Education Review 40(2):116–138. Kamwendo, Gregory. 2008. “Globalization, Linguistic Diversity, and Information Dissemination in Malawi.” Perspectives on Global Development and Technology 7(3–4):271–80. Kanyongolo, Ngeyi Ruth, and Seodi White. 2017. “Legislating against the Odds: Lessons Learned from Efforts to Legislate against Marital Rape in Malawi.” in The Right to Say No: Marital Rape and Law Reform in Canada, Ghana, Kenya and Malawi, edited by M. Randall, J. Koshan, and P. Nyaundi. Oxford and Portland, OR: Hart Publishing. Katz, Elihu, and Paul F. Lazarsfeld. 1955. Personal Influence: The Part Played by People in the Flow of Mass Communications. New York, NY, US: Free Press. Keck, Margaret E., and Kathryn Sikkink. 1998. Activists Beyond Borders: Advocacy Networks in International Politics. Ithaca, NY: Cornell University Press. Kentikelenis, Alexander E., and Sarah Babb. 2019. “The Making of Neoliberal Globalization: Norm Substitution and the Politics of Clandestine Institutional Change.” American Journal of Sociology 124(6):1720–62. Kentikelenis, Alexander E., and Leonard Seabrooke. 2017. “The Politics of World Polity: Script- Writing in International Organizations.” American Sociological Review 82(5):1065–92. Khan, Shamus R., Jennifer S. Hirsch, Alexander Wambold, and Claude A. Mellins. 2018. “‘I Didn’t Want to Be “That Girl”’: The Social Risks of Labeling, Telling, and Reporting Sexual Assault.” Sociological Science 5:432–60. Koo, Jeong-Woo, and Jaesung Choi. 2019. “Polarized Embrace: South Korean Media Coverage of Human Rights, 1990–2016.” Journal of Human Rights 18(4):455–73. Koziol, Natalie A., James A. Bovaird, and Sonia Suarez. 2017. “A Comparison of Population- Averaged and Cluster-Specific Approaches in the Context of Unequal Probabilities of Selection.” Multivariate Behavioral Research 52(3):325–49. Kubin, Emily, Curtis Puryear, Chelsea Schein, and Kurt Gray. 2021. “Personal Experiences Bridge Moral and Political Divides Better Than Facts.” Proceedings of the National Academy of Sciences 118(6). Kudo, Yuya. 2017. “Missionary Influence on Marriage Practices: Evidence from the Livingstonia Mission in Malawi.” Journal of African Economies 26(3):372–431. Kurzman, Charles, Willa Dong, Brandon Gorman, Karam Hwang, Renee Ryberg, and Batool Zaidi. 2019. “Women’s Assessments of Gender Equality.” Socius 5:2378023119872387. Lawson, David W., Susan B. Schaffnit, Joseph A. Kilgallen, Yusufu Kumogola, Anthony Galura, and Mark Urassa. 2021. “He for She? Variation and Exaggeration in Men’s Support for Women’s Empowerment in Northern Tanzania.” Evolutionary Human Sciences. Lerch, Julia, Patricia Bromley, Francisco O. Ramirez, and John W. Meyer. 2017. “The Rise of Individual Agency in Conceptions of Society: Textbooks Worldwide, 1950–2011.” International Sociology 32(1):38–60. Liebler, Carol M., Wei Jiang, and Li Chen. 2015. “Beauty, Binaries, and the Big Screen in China: Character Gender in Feature Films.” Asian Journal of Communication 25(6):584–99. Lizardo, Omar. 2017. “Improving Cultural Analysis: Considering Personal Culture in Its Declarative and Nondeclarative Modes.” American Sociological Review 82(1):88–115.

36 Lizardo, Omar, Robert Mowry, Brandon Sepulvado, Dustin S. Stoltz, Marshall A. Taylor, Justin Van Ness, and Michael Wood. 2016. “What Are Dual Process Models? Implications for Cultural Analysis in Sociology.” Sociological Theory 34(4):287–310. Lodamo, Berhanu, and Terje S. Skjerdal. 2009. “Freebies and Brown Envelopes in Ethiopian Journalism.” Ecquid Novi: African Journalism Studies 30(2):134–54. Long, J. Scott, and Sarah A. Mustillo. 2018. “Using Predictions and Marginal Effects to Compare Groups in Regression Models for Binary Outcomes.” Sociological Methods & Research 0049124118799374. Lounsbury, Michael, and Milo Shaoqing Wang. 2020. “Into the Clearing: Back to the Future of Constitutive Institutional Analysis.” Organization Theory 1:1–27. Lugo-Ocando, Jairo. 2020. Foreign Aid and Journalism in the Global South: A Mouthpiece for Truth. Rowman & Littlefield. Lundberg, Shelly, and Robert A. Pollak. 1996. “Bargaining and Distribution in Marriage.” Journal of Economic Perspectives 10(4):139–58. Lwanda, John Chipembere. 2010. “‘EDZI Ndi Dolo’ ('AIDS Is Mighty’): Singing HIV/AIDS in Malawi, 1980–2008.” Pp. 384–403 in The Culture of AIDS in Africa: Hope and Healing through Music and the Arts, edited by G. Barz and J. M. Cohen. New York: Oxford University Press. Lwanda, John, and Chipo Kanjo. 2013. “Computers, Culture and Music: The History of the Recording Industry in Malawi.” The Society of Malawi Journal 66(1):23–42. Magalasi, Mufunanji. 2015. “Factors Affecting the Growth of the Malawian Film Industry.” Journal of Development and Communication Studies 4(1):18–32. Mahoney, James, and Kathleen Thelen. 2009. “A Theory of Gradual Institutional Change.” Pp. 1–37 in Explaining Institutional Change: Ambiguity, Agency, and Power, edited by J. Mahoney and K. Thelen. Cambridge: Cambridge University Press. Malamuth, Neil M., and James V. P. Check. 1981. “The Effects of Mass Media Exposure on Acceptance of Violence against Women: A Field Experiment.” Journal of Research in Personality 15(4):436–46. Manda, Levi Zeleza, and Noel Drake Kufaine. 2013. “Starving the Messenger: A Study of Journalists’ Conditions of Service in Malawi.” Journal of Development and Communication Studies 2(2–3):301-311–311. McCombs, Maxwell E., and Donald L. Shaw. 1972. “The Agenda-Setting Function of Mass Media.” The Public Opinion Quarterly 36(2):176–87. McDonnell, Terence E. 2010. “Cultural Objects as Objects: Materiality, Urban Space, and the Interpretation of AIDS Campaigns in Accra, Ghana.” American Journal of Sociology 115(6):1800–1852. McDonnell, Terence E. 2016. Best Laid Plans: Cultural Entropy and the Unraveling of AIDS Media Campaigns. University of Chicago Press. McDonnell, Terence E., Christopher A. Bail, and Iddo Tavory. 2017. “A Theory of Resonance.” Sociological Theory 35(1):1–14. McDougal, Lotus, Samuel Krumholz, Nandita Bhan, Prashant Bharadwaj, and Anita Raj. 2018. “Releasing the Tide: How Has a Shock to the Acceptability of Gender-Based Sexual Violence Affected Rape Reporting to Police in India?” Journal of Interpersonal Violence 0886260518811421. Mchakulu, Japhet Ezra July. 2018. “Mediating an Alternative Public Sphere: Malawian Readers Attitudes and Perceptions towards a Tabloid.” Cogent Social Sciences 4(1):1452841.

37 Medie, Peace A. 2020. Global Norms and Local Action: The Campaigns to End Violence against Women in Africa. Oxford, New York: Oxford University Press. Merry, Sally Engle. 2006. Human Rights and Gender Violence: Translating International Law into Local Justice. Chicago: University of Chicago Press. Meyer, John W. 1977. “The Effects of Education as an Institution.” American Journal of Sociology 83(1):55–77. Meyer, John W. 1986a. “Myths of Socialization and Personality.” Pp. 212–25 in Reconstructing Individualism: Autonomy, Individuality, and the Self in Western Thought, edited by T. C. Heller, D. E. Wellbery, and M. Sosna. Stanford, CA: Stanford University Press. Meyer, John W. 1986b. “The Self and the Life Course: Institutionalization and Its Effects.” Pp. 199–216 in Human development and the life course: multidisciplinary perspectives, edited by A. B. Sørensen, F. E. Weinert, and L. R. Sherrod. Hillsdale, NJ: L. Erlbaum Associates. Meyer, John W., John Boli, George M. Thomas, and Francisco O. Ramirez. 1997. “World Society and the Nation‐State.” American Journal of Sociology 103(1):144–81. Meyer, John W., John Boli-Bennett, and Christopher Chase-Dunn. 1975. “Convergence and Divergence in Development.” Annual Review of Sociology 1(1):223–46. Mhagama, Peter. 2015. “Donor Funding to Community Radio Stations in Malawi and Its Impact on Their Performance.” Journal of Southern African Studies 41(6):1301–14. Minnesota Population Center (MPC). 2019. Integrated Public Use Microdata Series, International: Version 7.2 [Machine-Readable Database]. Minneapolis, MN: University of Minnesota. Mitchell, Maura. 2002. “‘Living Our Faith’: The Lenten Pastoral Letter of the Bishops of Malawi and the Shift to Multiparty Democracy, 1992–1993.” Journal for the Scientific Study of Religion 41(1):5–18. Mize, Trenton D. 2019. “Best Practices for Estimating, Interpreting, and Presenting Nonlinear Interaction Effects.” Sociological Science 6:81–117. Mohr, John, Christopher A. Bail, Margaret Frye, Jennifer C. Lena, Omar Lizardo, Terence E. McDonnell, Ann Mische, Iddo Tavory, and Frederick F. Wherry. 2020. Measuring Culture. New York, NY: Columbia University Press. Montoya, Celeste. 2013. From Global to Grassroots: The European Union, Transnational Advocacy, and Combating Violence against Women. Oxford: Oxford University Press. Mood, Carina. 2010. “Logistic Regression: Why We Cannot Do What We Think We Can Do, and What We Can Do about It.” European Sociological Review 26(1):67–82. Morfit, N. Simon. 2011. “‘AIDS Is Money’: How Donor Preferences Reconfigure Local Realities.” World Development 39(1):64–76. Murrar, Sohad, and Markus Brauer. 2019. “Overcoming Resistance to Change: Using Narratives to Create More Positive Intergroup Attitudes.” Current Directions in Psychological Science 28(2):164–69. Nthala, Grant Macloly Moloko. 2013. “The Role of the Musical Arts in HIV/AIDS Intervention in Malawi.” Thesis, University of the Free State. Oyediran, Kolawole Azeez, and Uche C. Isiugo-Abanihe. 2005. “Perceptions of Nigerian Women on Domestic Violence: Evidence from 2003 Nigeria Demographic and Health Survey.” African Journal of Reproductive Health 9(2):38–53. Pandian, Roshan K. 2018. “World Society Integration and Gender Attitudes in Cross-National Context.” Social Forces 97(3):1095–1126.

38 Peters, Pauline E. 1997. “Against the Odds: Matriliny, Land and Gender in the Shire Highlands of Malawi.” Critique of Anthropology 17(2):189–210. Peters, Rebecca Warne. 2020. Implementing Inequality: The Invisible Labor of International Development. New Brunswick: Rutgers University Press. Phiri, Kings M. 1983. “Some Changes in the Matrilineal Family System among the Chewa of Malawi since the Nineteenth Century.” The Journal of African History 24(2):257–74. Pierotti, Rachael S. 2013. “Increasing Rejection of Intimate Partner Violence: Evidence of Global Cultural Diffusion.” American Sociological Review 78(2):240–65. Pike, John G. 1965. “A Pre-Colonial History of Malawi.” The Nyasaland Journal 18(1):22–54. Polletta, Francesca, Monica Trigoso, Britni Adams, and Amanda Ebner. 2013. “The Limits of Plot: Accounting for How Women Interpret Stories of Sexual Assault.” American Journal of Cultural Sociology 1(3):289–320. Pope, Shawn, and John W. Meyer. 2016. “Local Variation in World Society: Six Characteristics of Global Diffusion.” European Journal of Cultural and Political Sociology 3(2–3):280–305. Powell, Thomas E., Hajo G. Boomgaarden, Knut De Swert, and Claes H. de Vreese. 2015. “A Clearer Picture: The Contribution of Visuals and Text to Framing Effects.” Journal of Communication 65(6):997–1017. Powell, Thomas E., Hajo G. Boomgaarden, Knut De Swert, and Claes H. de Vreese. 2019. “Framing Fast and Slow: A Dual Processing Account of Multimodal Framing Effects.” Media Psychology 22(4):572–600. Prentice, Deborah, and Elizabeth Levy Paluck. 2020. “Engineering Social Change Using Social Norms: Lessons from the Study of Collective Action.” Current Opinion in Psychology 35:138– 42. Ranganathan, Meghna, Lori Heise, Amber Peterman, Shalini Roy, and Melissa Hidrobo. 2021. “Cross-Disciplinary Intersections between Public Health and Economics in Intimate Partner Violence Research.” SSM - Population Health 14:100822. Rani, Manju, and Sekhar Bonu. 2009. “Attitudes toward Wife Beating: A Cross-Country Study in Asia.” Journal of Interpersonal Violence 24(8):1371–97. Rani, Manju, Sekhar Bonu, and Nafissatou Diop-Sidibe. 2004. “An Empirical Investigation of Attitudes towards Wife-Beating among Men and Women in Seven Sub-Saharan African Countries.” African Journal of Reproductive Health 8(3):116–36. Rhodes, Nancy, Bridget Potocki, and Desirae S. Masterson. 2018. “Portrayals of Intimate Partner Violence in Music Videos: Effects on Perceptions of IPV Warning Signs.” Media Psychology 21(1):137–56. Riley, Liam, and Belinda Dodson. 2016. “‘Gender Hates Men’: Untangling Gender and Development Discourses in Food Security Fieldwork in Urban Malawi.” Gender, Place & Culture 23(7):1047–60. Roberts, Louisa L. 2019. “Changing Worldwide Attitudes toward Homosexuality: The Influence of Global and Region-Specific Cultures, 1981–2012.” Social Science Research 80:114–31. Robinson, Robert V., and Wendell Bell. 1978. “Equality, Success, and Social Justice in England and the United States.” American Sociological Review 43(2):125–43. Russell, S. Garnett, Julia C. Lerch, and Christine Min Wotipka. 2018. “The Making of a Human Rights Issue: A Cross-National Analysis of Gender-Based Violence in Textbooks, 1950-2011.” Gender & Society 32(5):713–38.

39 Sardinha, LynnMarie, and Héctor E. Nájera Catalán. 2018. “Attitudes Towards Domestic Violence in 49 Low- and Middle-Income Countries: A Gendered Analysis of Prevalence and Country-Level Correlates.” PLOS ONE 13(10):e0206101. Saur, Maria, Linda Semu, and Stella Hauya Ndau. 2005. Nkhanza: Listening to People’s Voices. Zomba, Malawi: Kachere Series. Schildkraut, Jaclyn, and H. Jaymi Elsass. 2017. “The Influence of Media on Public Attitudes.” Pp. 115–35 in The Wiley Handbook of the Psychology of Mass Shootings. John Wiley & Sons, Ltd. Schofer, Evan, John W. Meyer, and Julia Lerch. 2019. “Illiberal Attacks on the University: Political Repression of Higher Education 1975-2018.” Paper Presented at the Annual Meetings of the American Sociological Association, New York. Schuler, Sidney Ruth, Rachel Lenzi, and Kathryn M. Yount. 2011. “Justification of Intimate Partner Violence in Rural Bangladesh: What Survey Questions Fail to Capture.” Studies in Family Planning 42(1):21–28. Semu, Linda. 2002. “Kamuzu’s Mbumba: Malawi Women’s Embeddedness to Culture in the Face of International Political Pressure and Internal Legal Change.” Africa Today 49(2):77–99. Shoemaker, Pamela J., and Stephen D. Reese. 2014. Mediating the Message in the 21st Century: A Media Sociology Perspective. Third Edition. New York: Routledge/Taylor & Francis Group. Sicard, S. V. 2000. “The Arrival of Islam in Malawi and the Muslim Contribution to Development.” Journal of Muslim Minority Affairs 20(2):291–311. Singhal, Arvind, Michael J. Cody, Everett M. Rogers, and Miguel Sabido, eds. 2004. Entertainment-Education and Social Change: History, Research, and Practice. Mahwah, New Jersey: Lawrence Erlbaum. Small, Mario Luis. 2011. “How to Conduct a Mixed Methods Study: Recent Trends in a Rapidly Growing Literature.” Annual Review of Sociology 37(1):57–86. Smith, Rachel A., Edward Downs, and Kim Witte. 2007. “Drama Theory and Entertainment Education: Exploring the Effects of a Radio Drama on Behavioral Intentions to Limit HIV Transmission in Ethiopia.” Communication Monographs 74(2):133–53. Smith, Rachel A., Youllee Kim, Xun Zhu, Dimi Théodore Doudou, Eleanore D. Sternberg, and Matthew B. Thomas. 2018. “Integrating Models of Diffusion and Behavior to Predict Innovation Adoption, Maintenance, and Social Diffusion.” Journal of Health Communication 23(3):264–71. Solovei, Adriana, and Bas van den Putte. 2020. “The Effects of Five Public Information Campaigns: The Role of Interpersonal Communication.” Communications 45(s1):586–602. Svec, Joseph, and Tanja Andic. 2018. “Cooperative Decision-Making and Intimate Partner Violence in Peru.” Population and Development Review 44(1):63–85. Swidler, Ann, and Susan Cotts Watkins. 2007. “Ties of Dependence: AIDS and Transactional Sex in Rural Malawi.” Studies in Family Planning 38(3):147–62. Swidler, Ann, and Susan Cotts Watkins. 2009. “‘Teach a Man to Fish’: The Sustainability Doctrine and Its Social Consequences.” World Development 37(7):1182–96. Swidler, Ann, and Susan Watkins. 2015. “Practices of Deliberation in Rural Malawi.” Pp. 133– 66 in Deliberation and Development: Rethinking the Role of Voice and Collective Action in Unequal Societies, Equity and Development. The World Bank. Swidler, Ann, and Susan Cotts Watkins. 2017. A Fraught Embrace: The Romance and Reality of AIDS Altruism in Africa. Princeton, NJ: Princeton University Press.

40 Taulo, John L., Kenneth Joseph Gondwe, and Adoniya Ben Sebitosi. 2015. “Energy Supply in Malawi: Options and Issues.” Journal of Energy in Southern Africa 26(2):19–32. Tavory, Iddo, and Ann Swidler. 2009. “Condom Semiotics: Meaning and Condom Use in Rural Malawi.” American Sociological Review 74(2):171–89. Thornton, Arland. 2005. Reading History Sideways: The Fallacy and Enduring Impact of the Developmental Paradigm on Family Life. Chicago, IL: University of Chicago Press. Thornton, Arland, Georgina Binstock, Mohammad Jalal Abbasi-Shavazi, Dirgha Ghimire, Arjan Gjonca, Attila Melegh, Colter Mitchell, Mansoor Moaddel, Yu Xie, Li-shou Yang, Linda Young-DeMarco, and Kathryn M. Yount. 2012a. “Knowledge and Beliefs about National Development and Developmental Hierarchies: The Viewpoints of Ordinary People in Thirteen Countries.” Social Science Research 41(5):1053–68. Thornton, Arland, Dirgha J. Ghimire, and Colter Mitchell. 2012b. “The Measurement and Prevalence of an Ideational Model of Family and Economic Development in Nepal.” Population Studies 66(3):329–45. Thornton, Arland, Rachael S. Pierotti, Linda Young-DeMarco, and Susan Watkins. 2014. “Developmental Idealism and Cultural Models of the Family in Malawi.” Population Research and Policy Review 33(5):693–716. Thornton, Arland, Shawn F. Dorius, and Jeffrey Swindle. 2015. “Developmental Idealism: The Cultural Foundations of World Development Programs.” Sociology of Development 1(2):277– 320. Thussu, Daya. 2018. “A New Global Communication Order for a Multipolar World.” Communication Research and Practice 4(1):52–66. Tran, Thach Duc, Hau Nguyen, and Jane Fisher. 2016. “Attitudes towards Intimate Partner Violence against Women among Women and Men in 39 Low- and Middle-Income Countries.” PLOS ONE 11(11):e0167438. United Nations (UN). 1991. Convention on the Elimination of All Forms of Discrimination against Women (CEDAW): Declarations, Reservations and Objections to CEDAW. United Nations (UN). 1993. A/RES/48/104. Declaration on the Elimination of Violence against Women. United Nations (UN). 1995. Report of the Fourth World Conference on Women. Beijing. United Nations (UN). 2021. “United Nations Treaty Collection Status of Treaties: Convention on the Elimination of All Forms of Discrimination against Women New York, 18 December 1979.” Retrieved January 20, 2021 (https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=IV- 8&chapter=4&lang=en). Uthman, Olalekan A., Stephen Lawoko, and Tahereh Moradi. 2009. “Factors Associated with Attitudes towards Intimate Partner Violence against Women: A Comparative Analysis of 17 Sub-Saharan Countries.” BMC International Health and Human Rights 9:14. Uthman, Olalekan A., Stephen Lawoko, and Tahereh Moradi. 2010. “Sex Disparities in Attitudes Towards Intimate Partner Violence against Women in Sub-Saharan Africa: A Socio-Ecological Analysis.” BMC Public Health 10(1):223. Vaisey, Stephen. 2009. “Motivation and Justification: A Dual-Process Model of Culture in Action.” American Journal of Sociology 114(6):1675–1715. Vaisey, Stephen, and Omar Lizardo. 2010. “Can Cultural Worldviews Influence Network Composition?” Social Forces 88(4):1595–1618.

41 Van Kol, Willemijn. 2008. “Ummah in Zomba: Transnational Influences on Reformist Muslims in Malawi.” Journal of Muslim Minority Affairs 28(3):435–52. Velasco, Kristopher. 2020a. “A Growing Queer Divide: The Divergence between Transnational Advocacy Networks and Foreign Aid in Diffusing LGBT Policies.” International Studies Quarterly 64(1):120–32. Velasco, Kristopher. 2020b. “Queering the World Society: Global Norms, Rival Transnational Networks, and the Contested Case of LGBT Rights.” SocArXiv. Wang, Lili. 2018. “Influences of Media Exposure on Chinese University Students’ Attitudes and Perceptions of Intimate Partner Violence.” Deviant Behavior 39(9):1202–16. Ward, L. Monique. 2016. “Media and Sexualization: State of Empirical Research, 1995–2015.” The Journal of Sex Research 53(4–5):560–77. Ward, Patricia. 2020. “‘What Do They Want from Us?’ How Locals Work to Meet Their Employers’ Expectations in Jordan’s Aid Sector.” Sociology of Development 6(3):318–37. Watkins, Susan Cotts, and Adam Ashforth. 2015. “Malawi Journals Project.” Deepblue. Retrieved January 10, 2016 (https://deepblue.lib.umich.edu/handle/2027.42/113269). Watkins, Susan Cotts, and Dennis Hodgson. 2019. “Developmental Idealism, the International Population Movement, and the Transformation of Population Ideology in Kenya.” Sociology of Development 5(3):229–47. Wei, Qian, and Liam Swiss. 2020. “Filling Empty Promises? Foreign Aid and Human Rights Decoupling, 1981-2011.” The Sociological Quarterly. Weitzman, Abigail. 2014. “Women’s and Men’s Relative Status and Intimate Partner Violence in India.” Population and Development Review 40(1):55–75. West, Brady T. 2020. “Survey Weights” edited by P. Atkinson, S. Delamont, A. Cernat, J. W. Sakshaug, and R. A. Williams. SAGE Research Methods Foundations. West, Brady T., Linda Beer, Garrett W. Gremel, John Weiser, Christopher H. Johnson, Shikha Garg, and Jacek Skarbinski. 2015. “Weighted Multilevel Models: A Case Study.” American Journal of Public Health 105(11):2214–15. West, Brady T., Joseph W. Sakshaug, and Guy Alain S. Aurelien. 2018. “Accounting for Complex Sampling in Survey Estimation: A Review of Current Software Tools.” Journal of Official Statistics 34(3):721–52. Wimmer, Andreas. 2021. “Domains of Diffusion: How Culture and Institutions Travel around the World and with What Consequences.” American Journal of Sociology 126(6):1389–1438. World Bank. 2019. “Power Outages in Firms in a Typical Month (Number) - Malawi.” Databank. Retrieved February 11, 2019 (https://data.worldbank.org/indicator/IC.ELC.OUTG?locations=MW). World Bank. 2021. “Women, Business and the Law Database.” Retrieved January 20, 2021 (https://wbl.worldbank.org/en/wbl-data). World Health Organization. 2021. Violence against Women Prevalence Estimates, 2018: Global, Regional and National Prevalence Estimates for Intimate Partner Violence against Women and Global and Regional Prevalence Estimates for Non-Partner Sexual Violence against Women. Geneva: World Health Organization, on behalf of the United Nations Inter-Agency Working Group on Violence against Women Estimation and Data (UNICEF, UNFPA, UNODC, UNSD, UN Women). Yeatman, Sara, Jenny Trinitapoli, and Sarah Garver. 2020. “The Enduring Case for Fertility Desires.” Demography 57(6):2047–56.

42 Zaller, John. 1992. The Nature and Origins of Mass Opinion. Cambridge, England: Cambridge University Press. Zaller, John. 1996. “The Myth of Massive Media Impact Revived: New Support for a Discredited Idea.” Pp. 17–78 in Political Persuasion and Attitude Change, edited by D. C. Mutz, P. M. Sniderman, and R. A. Brody. Ann Arbor, MI: University of Michigan Press.

43 Figures and Tables

Figure 1. Cross-national trends in stated rejection of MVAW. Wme Me 100 100

M M W

A 75 75 V M

50 50 R

S

:

25 25 P

0 0 2000 2009 2018 2000 2009 2018

Note: Data are from Statcompiler, which aggregates estimates from the cross-national Demographic and Health Surveys (ICF 2020). Individual lines represent countries, including Malawi and 44 other countries across Africa, Asia, Central and South America, and Eastern Europe (see Appendix A for further details).

44 3/6/2021 ReecMedia_gbaedia_f_20210305_Fig.SVG Figure 2. Cross-national trends in at least weekly newspaper, radio, and television use.

Wme Me 100 100 y l

k 75 75 e e W

r e p a p s 50 50 w e N

: t n e c r

e 25 25 P Malawi Malawi 0 0 2000 2009 2018 2000 2009 2018 Year Year

Wme Me 100 100

75 75 y l k e e W

o i d

a 50 50 Malawi R

: t n e c r

e Malawi P 25 25

0 0 2000 2009 2018 2000 2009 2018 Year Year

Wme Me 100 100 y l

k 75 75 e e W

n o i s i v 50 50 e l e T

: t n e c r

e 25 25 P Malawi Malawi

0 0 2000 2009 2018 2000 2009 2018 Year Year

Note: Data are from Statcompiler, which aggregates estimates from the cross-national Demographic and Health Surveys (ICF 2020). Individual lines represent countries, including Malawi and 44 othere:///Ue/ide/Liba/Mbie countries across Dce/caeCdDc/Sch/Reeach/Maai/W Africa, Asia, Central and South America,iig/Paha an fd Gba Eastern Ca Diffi/Fige/Media_20 1/1 Europe (see Appendix A for further details).

45 Figure 3. Mass media diffusion pathways of anti-MVAW scripts in Malawi.

Transnational Organizations

Malawian Media Companies

Newspaper Articles Radio Programs TV Programs e.g., Tearing down e.g., What Do We Do? e.g., Why are We Dying? Sexual Violence

Individual People

46 Figure 4. Total number of newspaper articles published in The Nation and The Daily Times between January 1, 2000 and February 14, 2016, categorized by the types of interpersonal violence cases and campaigns discussed.

MVAW Ca

O Ca

MVAW Caa

O Caa

0 350 700 1,050 1,400 Na A

47 Figure 5. Monthly number of MVAW newspaper articles published in The Nation and The Daily Times.

45 A

30 a N

W A

V 15 M

M 0

2000 2005 2010 2015 Da

Ra S

Note: Smoothed totals are averaged across a period of 13 months: 6 months before and after a given month.

48 Figure 6. Newspaper article from December 1, 2010 reporting on activities during the 16 Days of Activism against Gender Violence campaign.

49 Figure 7. Weekly number of newspaper articles about MVAW campaigns that were published in The Nation and The Daily Times during October through January of each year. : e

l 11 11 11 11 g c i i a l A

k m e a e 5.5 5.5 5.5 5.5 e C

a W W A V

e 0 0 0 0 M N Oc 1-7, Ja 21-27, Oc 7-13, Ja 20-26, Oc 6-12, Ja 19-25, Oc 5-11, Ja 25-31, 2000 2001 2001 2002 2002 2003 2003 2004 : e

l 11 11 11 11 g c i i a l A

k m e a e 5.5 5.5 5.5 5.5 e C

a W W A V

e 0 0 0 0 M N Oc 3-9, Ja 23-29, Oc 2-8, Ja 22-28, Oc 1-7, Ja 21-27, Oc 7-13, Ja 20-26, 2004 2005 2005 2006 2006 2007 2007 2008 : e

l 11 11 11 11 g c i i a l A

k m e a e 5.5 5.5 5.5 5.5 e C

a W W A V

e 0 0 0 0 M N Oc 5-11, Ja 25-31, Oc 4-10, Ja 24-30, Oc 3-9, Ja 23-29, Oc 2-8, Ja 22-28, 2008 2009 2009 2010 2010 2011 2011 2012 : e

l 11 11 11 11 g c i i a l A

k m e a e 5.5 5.5 5.5 5.5 e C

a W W A V

e 0 0 0 0 M N Oc 7-13, Ja 20-26, Oc 6-12, Ja 19-25, Oc 5-11, Ja 25-31, Oc 4-10, Ja 24-30, 2012 2013 2013 2014 2014 2015 2015 2016 Dae Dae Dae Dae

Note: The varying start and end dates for each graph reflect the first full calendar week in October and the last full calendar week in January. The grey area in each sub-figure marks November 25 through December 10, the annual period in which the 16 Days of Activism against Gender Violence campaign occurred.

50 Figure 8. Mass media diffusion pathways of cultural scripts that portray gender stereotypes in Malawi.

Foreign Media Entertainment Organizations

Malawian Media South African Satellite Pirated DVDs and Companies TV Public Video Parlors

Music Videos and The Weekend Times Radio Songs Foreign Movies e.g., All the Same, Reality TV Tabloid e.g., James Bond He is My Husband e.g., Big Brother

Individual People

51 Figure 9. Two exemplary cover pages of The Weekend Times tabloid from April 19-21, 2013 and December 21-23, 2012.

52 Figure 10. Probability of stated rejection of MVAW by the number of MVAW newspaper articles published one month prior to a respondent’s survey interview date. Wme Me 1 1 W A

V .9 .9 M

f o

n o i c e j e R

.8 .8 d e Sa

: i l i b a

b .7 .7 o Pr

.6 .6 0 15 30 0 15 30 MVAW Nepaper Aricle MVAW Nepaper Aricle

Note: Results obtained from Table 2, Model 1. Shading indicates 95 percent confidence intervals.

53 Figure 11. Probability of stated rejection of MVAW for weekly radio users and others by the number of MVAW radio programs they have heard. Wme Me 1 1 W A V

M .9 .9

f o

n o i c e j e R

.8 .8 d e a S

: i l i b

a .7 .7 b o P

.6 .6 < Radio Weekl Radio Weekl < Radio Weekl Radio Weekl

MVAW Radio Pogam: None All 4

Note: Results obtained from a replication of Table 2, Model 2 with an additional interaction between the number of MVAW radio programs heard and at least weekly radio use. Error bars indicate 95 percent confidence intervals.

54 Table 1. Descriptive statistics from women’s and men’s samples for all variables.

Women Men % Mean S.D. Min Max % Mean S.D. Min Max Dependent Variable Stated Rejection of MVAW 78.00 0 1 84.01 0 1

Independent Variables MVAW Newspaper Articles 13.42 6.54 1 29 13.83 6.50 1 29 Newspaper Weekly 10.42 0 1 21.11 0 1 MVAW Radio Programsa 2.07 1.72 0 4 2.89 1.43 0 4 Radio Weekly 48.66 0 1 67.17 0 1 Television Weekly 11.05 0 1 21.98 0 1

Control Variables Education 5.21 3.74 0 13 6.52 3.69 0 13 Urban 17.90 0 1 19.58 0 1 Christian 86.25 0 1 84.63 0 1 Working 25.51 0 1 42.79 0 1 Household Wealth 2.08 1.44 0 5 2.19 1.41 0 5 Matrilineal 77.20 0 1 77.86 0 1 Marital History Never 19.20 0 1 36.73 0 1 Formerly Married 12.63 0 1 3.53 0 1 Currently Married: Monogamy 57.89 0 1 54.63 0 1 Currently Married: Polygamy 9.93 0 1 5.07 0 1 Age 27.96 9.26 15 49 29.01 10.58 15 54 Note: Statistics include data from the 2000, 2004, 2010, and 2015-16 Malawi Demographic and Health Surveys except for MVAW newspaper articles, which I derive by linking my database of newspaper articles with survey respondents’ interview dates. a = Includes data from the 2000, 2004, and 2010 Demographic and Health Surveys.

55

Table 2. Average marginal effects with 95 percent confidence intervals from logistic regression models predicting stated rejection of MVAW.

Model 1 Model 2 Model 3 Model 4 2000-2016 2000-2010 2000-2016 2000-2016 Women Men Women Men Women Men Women Men MVAW Newspaper 0.005*** 0.003*** 0.005*** 0.003*** Articles [0.004, 0.007] [0.002, 0.005] [0.004, 0.007] [0.002, 0.005] Newspaper Weekly 0.015* -0.005 0.013 -0.003 [0.002, 0.028] [-0.022, 0.012] [-0.000, 0.027] [-0.020, 0.014] MVAW Radio 0.000 0.009** Programs [-0.003, 0.004] [0.003, 0.015] Radio Weekly 0.021*** 0.008 0.009* 0.015* [0.010, 0.031] [-0.012, 0.028] [0.001, 0.018] [0.001, 0.030] Television Weekly 0.003 -0.025* 0.000 -0.025* [-0.012, 0.017] [-0.045, -0.006] [-0.015, 0.015] [-0.044, -0.005] Education 0.009*** 0.007*** 0.008*** 0.005*** 0.009*** 0.007*** 0.008*** 0.007*** [0.007, 0.010] [0.005, 0.009] [0.006, 0.010] [0.003, 0.008] [0.008, 0.010] [0.005, 0.009] [0.007, 0.010] [0.005, 0.009] Urban 0.041*** 0.032** 0.042*** 0.033* 0.050*** 0.040*** 0.042*** 0.036** [0.025, 0.058] [0.008, 0.056] [0.022, 0.062] [0.001, 0.065] [0.033, 0.067] [0.017, 0.064] [0.025, 0.058] [0.012, 0.059] Christian 0.009 0.017 0.002 0.017 0.008 0.017 0.008 0.016 [-0.005, 0.023] [-0.004, 0.038] [-0.015, 0.019] [-0.008, 0.042] [-0.006, 0.022] [-0.004, 0.038] [-0.006, 0.023] [-0.005, 0.037] Working -0.011* -0.002 -0.018** -0.001 -0.012* -0.004 -0.012* -0.002 [-0.022, -0.001] [-0.015, 0.012] [-0.031, -0.005] [-0.019, 0.016] [-0.022, -0.002] [-0.017, 0.010] [-0.022, -0.001] [-0.016, 0.011] Wealth 0.007*** 0.006* 0.004 0.004 0.007*** 0.007* 0.006*** 0.006* [0.004, 0.011] [0.001, 0.011] [-0.000, 0.008] [-0.002, 0.011] [0.004, 0.011] [0.002, 0.012] [0.003, 0.010] [0.001, 0.012] Matrilineal 0.009 0.012 0.001 0.016 0.010 0.012 0.009 0.011 [-0.004, 0.022] [-0.009, 0.033] [-0.014, 0.017] [-0.011, 0.043] [-0.003, 0.023] [-0.009, 0.033] [-0.004, 0.022] [-0.010, 0.032] Marital History Never (Ref.) Formerly Married 0.059*** 0.015 0.070*** 0.018 0.057*** 0.015 0.060*** 0.016 [0.044, 0.074] [-0.022, 0.052] [0.051, 0.089] [-0.031, 0.067] [0.042, 0.072] [-0.022, 0.053] [0.045, 0.075] [-0.022, 0.054] Currently Married: 0.057*** 0.058*** 0.060*** 0.057*** 0.055*** 0.058*** 0.056*** 0.057*** Monogamy [0.045, 0.068] [0.039, 0.077] [0.046, 0.075] [0.033, 0.081] [0.043, 0.066] [0.039, 0.077] [0.044, 0.068] [0.038, 0.077] Currently Married: 0.046*** 0.006 0.055*** -0.003 0.044*** 0.005 0.045*** 0.004

56

Polygamy [0.030, 0.062] [-0.032, 0.043] [0.036, 0.074] [-0.049, 0.043] [0.028, 0.060] [-0.032, 0.043] [0.030, 0.061] [-0.033, 0.042] Age 0.003*** 0.004*** 0.003*** 0.004*** 0.003*** 0.004*** 0.003*** 0.004*** [0.003, 0.004] [0.003, 0.005] [0.003, 0.004] [0.002, 0.005] [0.003, 0.004] [0.003, 0.005] [0.003, 0.004] [0.003, 0.005] Survey Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes District Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Respondents (N) 72324 20870 47702 13368 72367 20876 72221 20832 Note: Significance tests (p-value): * .05, ** .01, *** .001.

Appendices | Data & Code

57