Investigating the relationship between social media consumption and fear of crime: A partial analysis of mostly young adults

Computers in Human Behavior 77 (2017) 158e168

Contents lists available at ScienceDirect

Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh

Full length article

Investigating the relationship between social media consumption and fear of crime: A partial analysis of mostly young adults Jonathan Intravia a, *, Kevin T. Wolff b, Rocio Paez a, Benjamin R. Gibbs a a b

Ball State University, Department of Criminal Justice and Criminology, North Quad, 282, Muncie, IN 47306, USA John Jay College of Criminal Justice, City University of New York, 524W. 59th Street, Room 63104T, NY 10019, New York, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 June 2017 Received in revised form 23 August 2017 Accepted 31 August 2017 Available online 1 September 2017

Theories of media effects have long established a link between media consumption and fear of crime. To date, prior investigations have almost exclusively focused on traditional types of media content (e.g., television news) or entertainment media (e.g., crime-related shows). However, less is known how social media consumption may influence individuals' levels of fear. Using data collected from a multisite sample of mostly young adults, the present study assesses: (1) the relationship between various types of social media consumption (overall, general news, and crime-related content) and fear of crime and (2) whether these relationships differ based on key audience characteristics. Findings reveal that overall social media consumption is significantly related to fear of crime and this relationship varies by perceptions of safety. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Fear of crime Cultivation theory Audience effects Social media

1. Introduction Individuals' fear of crime and violence is well established in the social science literature. Decades of prior research has linked numerous individual- and neighborhood-level factors to fear of crime, which include, but not limited to, women, non-whites, older individuals, education level, victimization experiences, perceived neighborhood conditions, and community crime rates (BruntonSmith & Sturgis, 2011; Gainey, Alper, & Chappell, 2011; Hindelang, Gottfredson, & Garofalo, 1978; Liska & Warner, 1991; Ross, 1993; Rountree & Land, 1996; Taylor & Hale, 1986; Wyant, 2008). In addition, the consequences of worrying about criminal victimization are well documented. For example, previous investigations have connected fear of crime to reduced neighborhood cohesion, altered habits such as staying indoors more frequently and withdrawing from social activities, increased target hardening, and impaired physical and psychological health (Britto, Van Slyke, & Francis, 2011; Cobbina, Miller, & Brunson, 2008; Conklin, 1975; Garofalo, 1981; Perkins & Taylor, 1996; Skogan & Maxfield, 1981; Stafford, Chandola, & Marmot, 2007; Warr, 2000). The fascination with crime and violence among the general public is widespread. Perhaps no bigger explanation for such

* Corresponding author. E-mail address: [email protected] (J. Intravia). http://dx.doi.org/10.1016/j.chb.2017.08.047 0747-5632/© 2017 Elsevier Ltd. All rights reserved.

interest in crime can be traced to increasing media, entertainment, and news coverage on topics associated with criminal and violent acts (McCall, 2007; Surette, 2007). This is evident today with high profiled events involving individuals such as Philando Castile, Michael Brown, and others, as well as increased news coverage surrounding violence found in communities such as Chicago, IL, and Orlando, FL. Extant research has shown that not only do individuals receive most of their information about crime from media content, but their attitudes and perceptions toward crime are also shaped by what they consume from media (Gross & Aday, 2003; Roberts & Stalans, 1997; Surette, 2007). Early attention on the effects of media consumption and fear of crime can be traced to Gerbner and colleagues' cultivation research on how widespread television broadcasting influenced consumers' attitudes and perceptions about social reality (Gerbner & Gross, 1976; Gerbner, Gross, Jackson-Beeck, Jeffries-Fox, & Signorielli, 1978; Gerbner, Gross, Morgan, & Signorielli, 1980). Specifically, the authors were interested in understanding how prolonged exposure to violent and aggressive content on television would cultivate fear and mistrust in audiencesda pattern coined as “Mean World Syndrome.” Presently, much of the available evidence suggests that various types of media and news consumption, such as the frequency of watching television news, listening to the radio, and viewing crime-related entertainment programming, significantly increases citizens' fear of crime (Chiricos, Eschholz, & Gertz, 1997; Dowler, 2003; Weitzer & Kubrin, 2004). In addition, previous

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

studies suggest that the media and fear relationship differs based on audience characteristics such as age, race/ethnicity, sex, and prior victimization to name a few (Chiricos, Padgett, & Gertz, 2000; Chiricos et al., 1997; Kort-Butler & Hartshorn, 2011).1 While important advancements have been made between the media consumption and fear relationship, notably absent from this area of research is the impact of social media consumption on individuals' fear of crime. This is an oversight in the extant literature given that scholars recently argued the cultivation framework “can be applied to any dominant medium” such as social networking sites (Morgan, Shanahan, & Signorielli, 2014, p. 481). Understanding this gap in the literature is important for several reasons. First, social media platforms have shown a tremendous growth in the number of users since the early 2000s (Perrin, 2015). Although this growth has been observed across all ages, according to recent statistics, younger (ages 18e29) and middle-aged individuals (ages 30e49) are more likely to use social media compared to older adults (ages 50e64 and 65 and older) (Greenwood, Perrin, & Duggan, 2016). Correspondingly, citizens are frequently turning to social media networks to obtain news and other information (Gottfried & Shearer, 2016). In addition, users of social media can disseminate news and information to others, which provides a unique method for how content is shared, consumed, and interpreted. Second, advocates and critics of cultivation argue that research needs to adapt the theory to understand new and important independent variables, or “technologies of message production” (Morgan & Shananhan, 2010, pp. 350e51). Therefore, identifying whether a major form of communication, such as social media, not only has important implications for understanding how the future of cultivation may adjust to contemporary modes of media, but also remains an unidentified source in the construction of social issues such as fear (Roche et al., 2016, p. 233). Third, regardless of official recorded crime levels, individuals who fear criminal victimization surpasses the actual number of crime victims on a yearly basis (Hale, 1996). In fact, public apprehension about crime remains high today. Despite crime statistics, such as the National Crime Victimization Survey (NCVS), illustrating that only a small segment of the population have personal experiences with crime or violence,2 the 2016 Gallup Poll illustrated that Americans' level of concern about crime is the highest in the past 15 years with approximately 79 percent of adults worrying either “a great deal” or “a fair amount” about crime and violence. Thus, given the unique characteristics of social networking platforms, individuals who are interested in learning more about crime and violence events may turn todor engage indconversations and posts on social media more frequently (Weber, 2014). As a result, social media consumption and engagement may also cultivate, or increase, fear among individuals. In the present study, our aim is to contribute to and expand the existing body of work that focuses on media consumption and fear of crime in two important ways. First, we examine whether social

1 A host of research on cultivation theory and audience effects (reception research) illustrates that media effects are applicable to various crime- and justicerelated outcomes beyond fear of crime. For example, prior efforts have examined the effect of various types media consumption on attitudes and perceptions directed toward punitiveness, criminal justice policy, policing, juvenile crime, drug issues, victim characteristics, school shootings, terrorism, and white-collar crime (see Callanan & Rosenberger 2011; Donley & Gualtieri 2017; Gilliam & Iyengar, 2000; Goidel, Freeman, & Procopio 2006; Menifield, Rose, Homa, & Cunningham 2001; Nellis & Savage, 2012; Nielsen & Bonn 2008; Roche, Pickett, & Gertz, 2016; Slingerland, Copes & Sloan , 2006; Thompson, 2010). 2 For example, according to the most recent published criminal victimization statistics from 2015, less than one percent (0.98) of all persons age 12 or older experienced a violent victimization, whereas approximately eight percent (7.60) of all households experienced property victimization (Truman & Morgan, 2016).

159

media consumption is related to individuals' level of fear, controlling for other major forms of media usage (e.g., television news, crime programming, and the Internet), demographics, and correlates of fear. Specifically, we assess the effects of three types of social media exposure on fear: overall consumption, general news consumption, and consumption of news/stories involving crime and violence. Second, we examine whether the relationship between social media exposure and fear of crime is conditioned by a number of well-known audience characteristicsdsex, race, residential area, prior victimization, feelings of safety, and perceptions of neighborhood problemsdto understand whether the impact of social media consumption on fear of crime is more (or less) prominent among specific subgroups of social media users. Before presenting the results of the current study, we first present an overview of cultivation theory and the key perspectives in reception research. Next, we discuss the empirical research related to media consumption and fear of crime. From there, we provide information on the unique characteristics of social media. 2. Theoretical background Gerbner and colleagues' cultivation theory investigated how the increasing growth of television viewership affected, or shaped, viewers' conceptions of social reality (Gerbner & Gross, 1976; Gerbner, 1969; Gerbner et al., 1978). According to their research, the more time spent consuming media (e.g., heavy exposure to television), the greater likelihood users' perceptions of the real world will align with what is depicted in the media (Gerbner et al., 1980; Weitzer & Kubrin, 2004). Despite modest support illustrating that heavy television consumption increases fear of victimization, critics of the theory argued the cultivation framework is limited in scope because it failed to account for the variation in viewers' demographic characteristics and social backgrounds (see Eschholz, 1997; Hirsch, 1980; Hughes, 1980). Because of these early limitations, scholars begun to examine how characteristics, social contexts, and past experiences of media audiences may impact their attitudes and perceptions. Specified broadly as “reception research” or “audience reception theory,” there are six key perspectives that are salient in the cultivation process (e.g., media consumption and fear of crime). First, the mainstreaming hypothesis illustrates that despite differences in consumers' attitudes and views (e.g., social and/or political), heavy media consumption homogenizes individuals to share similar perspectives (Gerbner et al., 1980; Morgan et al., 2014). Second, the affinity hypothesis suggests media effects will be stronger for consumers who share similar characteristics to the victims portrayed in the media (Hirsch, 1980). Third, the vulnerability hypothesis argues that media effects will be more responsive among individuals who feel defenseless to criminal victimization (e.g., women and the elderly) (Skogan & Maxfield, 1981). Fourth, the resonance hypothesis contends that the combination of media messages and personal (or perceived) experiences may provide a “double dose” to some users and amplify the effect of cultivation (Doob & Macdonald, 1979; Gerbner et al., 1980). Fifth, the substitution hypothesis argues that cultivation may be more pronounced among consumers without personal experiences to the messages shown in the media (e.g., non-victims) (Gerbner et al., 1980; Gunter, 1987; Liska & Baccaglini, 1990). Lastly, a “ceiling effect” may occur for certain subgroups such as women and minorities (Chiricos et al., 1997). For example, Heath and Petraitis (1987) found that media exposure affected fear of crime among males but not females. To explain this divergent finding, the authors suggest that women's fear levels may already be at “maximum level” and additional influences, such as media messages, does not affect this group any further (pp. 120e21).

160

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

Overall, theoretical accounts of media effects suggest that consumption of media may impact individuals' fear of crime. Moreover, media effects may differ based on audience demographics, personal experiences, and social backgrounds. Surprisingly, less known is how social media consumption may affect individuals' fear levels. We next provide an overview of the previous research on media consumption and fear of crime. Prior efforts in this domain provide promising evidence to the audience characteristics that are important in cultivating consumers of media. 3. Previous research on media consumption and fear of crime For the past 50 years, fear of crime has been recognized as a major issue facing society. Traditionally, prior research has found that one's environment or perceived surroundings may signal cues of danger and increase fear (Ferraro, 1995; Intravia, Stewart, Warren, & Wolff, 2016; Warr, 1990, 2000). A different approach to understanding how fear is cultivated among individuals is analyzing media content. Recognizing that mass media are a primary source of crime information (Graber, 1980; Skogan & Maxfield, 1981), scholarship has identified patterns associated with crime-related content covered in media outlets. Collectively, prior efforts illustrate that regardless of the media source (e.g., television, newspaper), news coverage of crime presents a distorted image by disproportionately publicizing violent (or serious) events over non-violent (or non-serious) reports (Lipschultz & Hilt, 2014; Marsh, 1991). This is not surprising given that news value theory contends that the production of news (e.g., probability of an event being reported) and consumers' engagement (e.g., commenting on stories) is based on a series of news factors such as proximity, continuity, influence, personalization, reach, unexpectedness, damage/negative consequences, and controversy (Eilders, 2006; Ziegele, Breiner, & Quiring, 2014). It is well documented that media consumption increases levels of fear among citizens. To date, prior efforts illustrate that various types of news consumption such as television news, newspapers/ magazines, and the Internet are related to fear (Chiricos et al., 2000, 1997; Kohm, Waid-Lindberg, Weinrath, Shelley, & Dobbs, 2012; Roche et al., 2016; Romer, Jamieson, & Aday, 2003). However, less is certain for which news sources are the most salient in cultivating fear due to the many different ways researchers have operationalized (e.g., frequency such as hours/days consumed, salience/credibility of news source/stories) and/or examined news content (e.g., local vs. television news) (see Chiricos et al., 2000; Gross & Aday, 2003; Kort-Butler & Hartshorn, 2011; Weitzer & Kurbin, 2004). Further, some studies show that the specific types of genre/content consumed (as opposed to overall television consumption) is important for understanding cultivation effects. For example, prior research has found that television programming involving politics, nonfiction crime shows, tabloid news, and reality policing shows significantly increases fear of crime/victimization (Cohen & Weimann, 2000; Dowler, 2003; Eschholz, Chiricos, & Gertz, 2003; Jamieson & Romer, 2014; Kort-Butler & Hartshorn, 2011; for lack of support see; Roche et al., 2016). Despite the many ways by which media consumption has been examined with fear of crime, scholarship has acknowledged that the messages depicted from media usage may not be uniform; rather, cultivation may affect individuals differently based on their backgrounds and characteristics. Collectively, previous investigations have found the relationship between media consumption and fear of crime differs by race/ethnicity as well as sex (Callanan, 2012; Chiricos et al., 2000, 1997; Weitzer & Kubrin, 2004). Others have suggested that media consumption on fear of crime varies by characteristics such as victimization and perceived neighborhood conditions. For instance, a number of studies

illustrate that media consumption was significantly related to elevated levels of fear among individuals who perceive risk/ victimization and/or were victimized in the past (Chiricos et al., 1997; Weitzer & Kubrin, 2004); whereas others found the media/ fear relationship is more pronounced among non-victims (Eschholz et al., 2003). Finally, research also illustrates that the media consumption and fear of crime relationship may differ by neighborhood context such as those residing in areas where the perceived African American population is high or where criminal activity is elevated (Chiricos et al., 2000, 1997; Eschholz et al., 2003; Weitzer & Kubrin, 2004). Overall, the above research illustrates that media consumption impacts citizens' fear of crime. Further, previous studies show that audience characteristics and backgrounds are important to consider when examining the media consumption and fear relationship. Yet, the results are mixed with respect to which consumers' characteristics and backgrounds are most influential in the media consumption and fear relationship. In addition, previous efforts have almost exclusively examined traditional media content (e.g., television news, television programming shows, and newspapers) on fear of crime, overlooking the potential impact of social media consumption. In the next section, we illustrate the unique characteristics associated with social networking platforms. From there, we discuss the how social media consumption may increase fear of crime. 4. The unique characteristics of social media for accessing, engaging, and shaping news and information Due to technology becoming more mobile in the past few decades (e.g., cell/smart phones, tablets, laptop computers), accessing news and information on the Internet and social media has influenced how consumers obtain, absorb, and transmit content. For example, in 2016, 62% of U.S. adults received news from social networking sites (Gottfried & Shearer, 2016). With respect to social media engagement, a recent survey in 2016 found that 80% of social media news consumers clicked on links to news stories, 58% “liked” news stories, 49% shared or reposted news events, 37% commented on news stories, 36% posted news stories themselves, 31% discussed issues in the news, and 19% posted photos of videos of news events (Mitchell, Gottfried, Barthel, & Shearer, 2016). Regarding crimerelated content specifically, another survey found that 10% of social media users reported posting news stories involving crime on social networking sites for others to read and/or watch (Gottfried & Shearer, 2016). Social networking sites also allow users to experience unique characteristics over traditional forms of media (e.g., television news broadcasts and newspapers) such as accessing news shared or recommended by peers/family members as well as providing discretion to the types of news one may read, filter, or show engagement (e.g., post, share, “like,” or comment). For example, research on public opinion illustrates that the Internet and social media (as opposed to traditional media formats) allow users to participate in discussion and deliberation on topics/issues via comment sections (Anderson, Brossard, Scheufele, Xenos, & Ladwig, 2014). These discussions, in turn, may lead to rational or irrational discourse and ultimately influence or alter one's own perception or opinion about the information on the issue being discussed (Anderson et al., 2014; Lee & Jang, 2010). As a result of the unique characteristics found in social networking platforms, individuals utilizing social media have their own “personal network” for news, information, and informal discussion, which contributes to the uniqueness of shaping opinions, beliefs, and attitudes as opposed to solely relying on professional or major news networks and sources (Hermida, Fletcher, Korell, & Logan, 2012, pp. 815e16).

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

4.1. How social media characteristics and usage may be related to fear Because social media has many unique characteristics over traditional-based media sources, there are a few plausible ways that social media consumption may be related to fear. Theoretically, if individuals are interested in learning about crime and violence occurring in society (either at a local or national level), they may choosedor be motivateddto consume more content on social media as well as engage in information (e.g., commenting, sharing) pertaining to crime-related issues. Engagement via crime-related content, in turn, may foster or change users' beliefs and opinions about the information presented. Further, based on the heterogeneity of motivations and patterns that can be expected by social media users and social networking platforms (Hermida et al., 2012), individuals are not likely to receive identical content. As a result, one's frequency (or lack of) in using social media and engaging in crime and violence-related content may contribute to different levels of fear among users. The negative consequences of frequent social media consumption may also be related to increased levels of fear. For instance, numerous studies illustrate that social media usage is negatively related to how people feel, their level of satisfaction, their psychological health as well as increased levels of social withdrawal, depression, and anxiety (Kross, Verduyn, Demiralp, Park, Lee, Lin, Shablack, Jonides, & Ybarra, 2013; Martin, Bailey, Cicero, & Kerns, 2012; O'Keeffe & Clarke-Pearson, 2011). Although not a direct comparison with fear of crime, negative consequences of frequent social media consumption, such as anxiety, social withdrawal, poor psychological well-being, and low life satisfaction, are similar to individuals who exhibit higher levels of fear (Gabriel & Greve, 2003; Hale, 1996). Thus, it is plausible that individuals who engage in greater social media consumption may also exacerbate psychological and social anxieties that are associated with fear of crime and victimization. 5. Current study The purpose of this study is to build upon the contributions of prior research by examining whetherdand howdsocial media consumption is related to individuals' level of fear. First, we hypothesize that social media consumption will be positively associated with fear of crime. Second, similar to previous investigations, we hypothesize that the effects of social media consumption on fear of crime will differ by key audience and background characteristics. Due to inconsistent patterns found in previous studies testing audience effects on the media and fear relationship, we do not have clear expectations regarding the specific audience and background characteristics that may be most prominent. However, we do disaggregate various audience traits (e.g., demographics and correlates of fear) that are shown to influence an individual's level of fear. 6. Data and methods 6.1. Sample The data for this research was collected through a survey administered to adults across three college campuses during the Fall 2016 and Spring 2017 semesters. The instrument was disseminated to students in accordance with each university's Institutional Review Board (IRB). Researchers from the three universities attended various class sections, addressing students with an oral description and instructions for the survey. Students were informed of the voluntary nature of their participation and

161

anonymity of their responses. Researchers were prudent to explain there were no penalties for choosing not to participate and neither extra-credit nor incentives were given to those who filled out the survey. Identical information was provided in writing along with contact information for the lead researcher and the corresponding IRB if they had any questions or concerns regarding the survey or study. Overall, 34 classes were surveyed among the three institutions. We sought to increase the generalizability of our study by using a multisite sample, surveying adults from not only different regions of the country, but also diverse geographical landscapes (e.g., urban, suburban). The majority (51.6%) of our observations were respondents from a large Midwestern university located in a more traditional, semi-rural “college town.” At this site, we sampled 474 students (96.5% response rate) from 12 criminal justice classes. Our second site, representing the Northeast region (26.8% of the sample) of the United States is a large urban institution located in a densely populated metropolitan area. Here, we sampled 246 individuals (95.5% response rate) from a total of 9 classes e6 criminal justice, 2 sociology, and 1 anthropology. Lastly, we surveyed 198 adults (94.4% response rate) from 13 criminal justice classes at a medium-sized Southern university (21.6%) located in a mid-sized city. This university has a large “non-traditional” student population (e.g., older adults). The survey collection across the three sites yielded 918 adults with a total participation rate of 96.1% (918 surveys completed out of 955 collected). Just under half of the respondents (47.2%) reported being criminal justice and criminology majors.3 The final sample demographics consisted of 43.1% male, 54.4% white (18.8% African American, 16.1% Hispanic, 8.7% other race/ethnicity), and a mean age of 21.1 years old. 6.2. Measures 6.2.1. Dependent variable Fear of crime has been measured in many ways. Some scholars suggest that the construct should gauge feelings directly related to crime (Hale, 1996; Warr, 1990); whereas others believe fear should be measured by the behavioral actions of individuals who perceive a dangerous situation (e.g., “avoiding areas” or “walking alone at night”) (Gabriel & Greve, 2003). Consistent to previous media and fear studies (Chiricos et al., 2000; Eschholz et al., 2003), fear of crime was measured by asking respondents to indicate their level of fear (from 0 ¼ not fearful at all to 10 ¼ very fearful) for the following six crime-related events: (1) someone breaking into your home, (2) being robbed or mugged on the street, (3) being sexually assaulted or raped, (4) having your car or bicycle stolen, (5) being beaten up or assaulted by strangers, and (6) being murdered. A confirmatory factor analysis (CFA) illustrated that all six items loaded onto a single factor and fit the data adequately (CFI ¼ 0.996, TLI ¼ 0.979, RMSEA ¼ 0.057).4 The six items were summed to obtain a total score. Higher scores equate to greater levels of fear (a ¼ 0.91). Due to the cross-sectional nature of our data, it is important to note that there is no temporal ordering in our key measures. The relationship between media consumption and fear of crime may be reciprocal in nature. That is, individuals who are more fearful of crime and violence may be more motivated to turn to media content in order to learn, process, and understand crime-related issues (Eschholz et al., 2003; Gunter, 1987). Although not a direct

3 The Midwest sample ¼ 42.2% majors (57.8% non-majors), Northeast sample ¼ 45.1% majors (54.9% non-majors), and the South sample ¼ 61.7% majors (38.3% non-majors). 4 The average inter-item correlation was 0.644.

162

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

comparison between social media and fear, O'Keefe & Reid-Nash (1987) utilized two waves of panel data and found that television news consumption had a stronger effect on fear of crime than the reverse. 6.2.2. Independent variables Our primary independent variables consist of four social media measures that gauge social media consumption in three ways: overall consumption, general news consumption, and consumption with crime and violence news/stories on two major social media platformsdFacebook and Twitter. Consistent with prior research on media consumption (Chiricos et al., 1997; Donovan & Klahm, 2015), we measured overall social media consumption and general news consumption on social media by asking respondents “In a typical week, how much time do you spend …” (1) on social media (such as Facebook, Instagram, Twitter, or Reddit) and (2) reading or watching news stories on social media (such as Facebook, Instagram, Twitter, or Reddit). Response categories included “none”, “60 min or less”, “61 to 120 min”, “121 to 180 min”, “181 to 240 min”, and “241 min or more” and were coded so that higher scores indicated more consumption of the reported media variable. In addition, due to scholars stressing the importance of examining specific media content related to crime and justice as opposed to broad measures (e.g., overall media consumption and/or non-specific news consumption) (Roche et al., 2016), we created two measures that gauged the frequency of consuming crime and violencerelated information on social media. Specifically, we measured crime/violence consumption on Facebook and Twitter by asking respondents the following two questions: (1) On Facebook, how often do you read, watch, post, or interact with (such as share, like, or comment) stories or news involving crime or violence occurring in society? (2) On Twitter, how often do you read, watch, tweet, or interact with (such as retweet, like, or reply) stories or news involving crime or violence occurring in society? Response categories ranged from 1 ¼ never to 5 ¼ very often. 6.2.3. Control variables We also controlled for several additional media consumption measures, demographics, and correlates of fear. First, we controlled for five media-related measures to gauge overall television consumption, overall Internet consumption, television news consumption (local and national), and consumption of television crime shows. Similar to the media variables mentioned above, we asked respondents “In a typical week, how much time do you spend” engaging in the following types of media: (1) watching television (overall consumption), (2) using the Internet, (3) watching a local television news broadcast (such as the 5 P. M or 10 P. M news), (4) watching a national television news broadcast (such as CNN or Fox News), and (5) watching television crime shows (such as Criminal Minds, CSI, NCIS, or Law & Order Special Victims Unit). Response categories included “none”, “60 min or less”, “61 to 120 min”, “121 to 180 min”, “181 to 240 min”, and “241 min or more” and were coded so that higher scores indicated greater levels of consumption of the recorded media measure. Consistent with previous media consumption and fear of crime studies, we also controlled for several potential variables related to both audience effects and fear. Demographics included race (1 ¼ black), sex (1 ¼ female), age (measure continuously), political ideology (1 ¼ very liberal to 5 ¼ very conservative), and residential area (1 ¼ urban). We also controlled for victimization, perceptions of neighborhood problems, and perceived safety. Victimization was a dichotomous measure that asked respondents if they have been a victim of a crime in the past year (1 ¼ yes). Perceptions of neighborhood problems consisting of six items that asked respondents to indicate “how much of a problem” were the following conditions in

their “neighborhood”: (1) vandalism, (2) drunks and drug users, (3) abandoned buildings, (4) burglaries and thefts, (5) run down and poorly kept buildings, and (6) assaults and muggings. Response categories ranged from 1 ¼ a big problem to 3 ¼ not a problem and items were reversed coded and summed to illustrate that higher scores equate to more serious neighborhood problems (a ¼ 0.85). Lastly, we included a measure of perceived safety. Several lines of research indicate that feelings related to fear may be influenced by cognitive perceptions of a situation being threatening or dangerous (Gabriel & Greve, 2003; Rountree & Land, 1996). As a result, we asked respondents “How safe do you feel or would you feel being alone in your neighborhood at night?” Responses ranged from 1 ¼ very safe to 4 ¼ very unsafe. Descriptive statistics for all variables included in the current study are displayed in Table 1. 6.3. Analytic strategy In order to examine whether social media consumption is related to fear of crime, we used ordinary least squares regression (OLS) models in STATA (version 14). Prior to the multivariate analysis, we assessed the normality of our dependent variable. A preliminary examination of the dependent variable revealed that its distribution approached normality as the skew and kurtosis were within normal range (skew < 3.0; kurtosis <10.0). In addition, the normality of the residuals confirms that OLS is an appropriate analytic strategy and the results are likely to be unbiased and reliable. Specifically, we assessed the possibility that the error terms were heteroskedastic using White's (1980) test statistic to evaluate the null hypothesis that the residuals obtained are homoscedastic. Results of these diagnostic tests indicate that the disturbance terms produced in each model are homoscedastic. Further, variance inflation factors (VIF) were examined to assure that collinearity was not an issue in the current analysis with the highest VIF observed was 1.72. For the current analysis, it is advantageous to treat the key independent variables as continuous because it limits the space required within the tables as well as provides estimates that are easy to interpret. Prior to the estimation of the full models, we used a series of likelihood-ratio tests to assess whether our predictors should be treated as ordinal rather than continuous measures. The likelihood-ratio test compares the fit of the model with the simple continuous model to that of the more complex model (treating the variable as ordinal). If the ordinal results mimic that of a linear trend, the test statistic will indicate that the simpler model is acceptable. The relationship between each focal predictor variable and the dependent variable was assessed in a model that also included the control variables in order to assess which fit the data better. The likelihood-ratio tests require fitting both models and subtracting the log-likelihoods of the later model from the first to obtain the chi-square value and corresponding p-value. In all cases, the results of these diagnostic tests suggested including the measures as continuous, rather than ordinal, fit the data adequately, and did not result in a loss of information or poorer model fit. These tests, shown for our focal independent variables, are presented in Table 2. Table 2 includes two regression models for each of our key independent variables. The first model for each variable (Models 1, 3, 5, and 7) includes each of our control variables as well a measure of social media, measured continuously. The second model (Models 2, 4, 6, and 8) treats our measures of social media as if they were ordinal, using the modal category as the base. The estimates shown in Table 2 provide evidence that including each of our focal variables in the model as continuous measures did not significantly reduce model fit or take away from our understanding of the

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

163

Table 1 Descriptive statistics for key study variables. Variables

Mean (SD)

Range

Correlation with Fear

Fear of Crime Age Race (1 ¼ Black) Sex (1 ¼ Female) Political Status Residential Area (1 ¼ Urban) Victim (1 ¼ Yes) Safety Neighborhood Problems Media Measures Television (Overall) Internet (Overall) Local TV News National TV News Crime TV Shows Social Media (Overall) Social Media News (General) Facebook Crime (Specific) Twitter Crime (Specific)

32.218 (17.576) 21.103 (4.837) 0.188 (0.391) 0.568 (0.495) 2.915 (0.908) 0.267 (0.443) 0.181 (0.385) 1.989 (0.771) 8.444 (2.736)

0e60 18e68 0e1 0e1 1e5 0e1 0e1 1e4 6e18

e 0.105*** 0.127*** 0.387*** 0.118*** 0.152*** 0.066** 0.264*** 0.139***

2.581 4.196 0.835 0.854 1.531 3.275 1.935 2.734 2.075

0e5 0e5 0e5 0e5 0e5 0e5 0e5 1e5 1e5

0.052 0.038 0.027 0.047 0.116*** 0.196*** 0.111*** 0.134*** 0.058*

Note: *p < 0.10,

**

p < 0.05,

***

(1.543) (1.059) (1.031) (1.016) (1.537) (1.545) (1.431) (1.239) (1.240)

p < 0.01.

relationship between social media and fear of crime. First, in each of the secondary models, there is a fairly clear linear pattern to the results which corresponds to the direction of the linear estimate. Secondly, the Akaike information criterion (AIC) and Bayesian information criterion (BIC) scores suggest that the more complex models do not fit the data as well as the model containing only one continuous predictor. Stated differently, the amount of additional variance explained by the more complex model is not justified given the covariate space utilized by the ordinal predictors. Finally, the insignificant likelihood-ratio tests confirm that the simplified model fits the data adequately, and including our focal variables as categorical measures does not significantly improve model fit. Although not shown in tabular form, this process was repeated with each of our measures of traditional media consumption to be certain our models adequately fit the data as well. After this preliminary step, our analysis proceeded in several key stages. First, we examined whether social media consumption is related to fear of crime. We carried out the analysis in a stepwise fashion by first identifying the relationship between demographics and fear, followed by inclusion of media control measures, and finally a full model with the four measures of social media included. Next, we examined whether the relationship between our four measures of social media consumption (overall, general news, crime stories/news on Facebook and Twitter) and fear of crime differs by key demographics and individual characteristics such as race, sex, area of residence, victim, feelings of safety, and perceptions of neighborhood problems. We assessed this question in two specific ways. First, we estimated a single model that included an interaction term between our four social media variables and each of the key audience characteristics. In this model, a significant product term suggests the relationship between social media and fear is moderated by a given individual characteristic (i.e. race, sex, perceptions of safety). Second, we probed the conditional relationship by using a spotlight analysis to estimate the conditional effects at various values of the moderators shown to be significant (see Spiller, Fitzsimons, Lynch, & McClelland, 2013). The details of this analysis are provided below.

7. Results Table 3 presents the results of the multivariate analyses. We started by examining our key demographics and controls with our

outcome. Model 1 of Table 3 illustrates that individuals who are younger, Black, female, reside in an urban area, and feel unsafe were significantly more likely to report higher levels of fear. In contrast, political status and perceptions of neighborhood problems were not significantly related to fear and individuals who were victimized in the past were less fearful of crime; however, this later relationship was modest at p < 0.10. In Model 2 (Table 3), we included our key media control measures. None of the key media controls were significantly related to fear. Similar to Model 1, the effects age, race, sex, area of residence, and feelings safety remained significant; however, the magnitudes show a slight reduction. In Model 3 (Table 3), we introduced our four social media measures (overall social media consumption, general social media news consumption, consumption with crime stories/news on both Facebook and Twitter). As shown in Model 3, only overall social media consumption is significantly related to fear of crime. Once again, age, sex, race, residential area, and feelings of safety remained significantly related to fear of crime. In sum, and providing support for cultivation research more generally, overall social media consumption is significantly related to individuals' fear of crime. Specifically among the full sample of respondents, overall social media consumption appears to be related to fear of crime above and beyond the other media-related measures. Although general social media news consumption and consumption of crime/news stories via Facebook and Twitter were not significant in the full model, as stated above, reception research contends that the impact of media consumption differs based on users' characteristics and backgrounds. Thus, consumption of these social media-related measures may have a more pronounced relationship with fear among individuals with different characteristics. We turn to these results next.

7.1. Demographic subgroup analyses To assess whether the relationship between our key social media measures and fear of crime differs by race, sex, area of residence, victimization, feelings of safety, and perceived neighborhood problems, we first estimated a single regression model that included an interaction term between each of the social media variables and the individual characteristics noted above. Although the observed main effect of three of the four social media was insignificant, we include interaction terms for each of these

164

Table 2 Assessing model fit for social media predictors. Variables

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

b (SE)

b (SE)

b (SE)

b (SE)

b (SE)

b (SE)

b (SE)

b (SE)

0.418** 5.883** 12.781** 0.100 6.130** 1.633 1.002** e e e e e e e e e e e e e e e e e e e e e e e e e 27.672**

AIC/BIC Log-Likelihood Likelihood-Ratio Test

7601.51 7635.20 7603.77 3793.754 3790.885 X2 ¼ 5.74; p ¼ 0.220

Notes: N ¼ 908. *p < 0.10;

**

p < 0.05;

***

(0.112) (1.386) (1.091) (0.601) (1.198) (1.358) (0.354) e e e e e e e e e e e e e e e e e e e e e e e e e (3.433)

0.422** 5.674** 12.851** 0.086 5.774** 1.662 e 0.622 4.820** 4.995** 3.323* 1.610 Base Group e e e e e e e e e e e e e e e e e e e 33.486**

(0.112) (1.389) (1.094) (0.601) (1.210) (1.360) e (2.731) (1.812) (1.647) (1.552) (1.523) e e e e e e e e e e e e e e e e e e e e (3.103)

0.461** 5.985** 13.159** 0.017 6.078** 1.893

7656.72

7599.10 7632.78 7604.12 3792.549 3791.062 X2 ¼ 2.97; p ¼ 0.562

e e e e e e 0.553 e e e e e e e e e e e e e e e e e e 30.943**

p < 0.01. Unstandardized coefficients presented.

(0.110) (1.390) (1.081) (0.603) (1.204) (1.365) e e e e e e e (0.371) e e e e e e e e e e e e e e e e e e (3.187)

0.461** 5.841** 13.182** 0.038 6.079** 1.843 e e e e e e e e 0.647 Base Group 0.265 1.029 0.894 3.728 e e e e e e e e e e e e 31.948**

(0.111) (1.395) (1.083) (0.604) (1.209) (1.369) e e e e e e e e (1.761) e (1.410) (1.765) (1.971) (2.022) e e e e e e e e e e e e (3.156)

0.466** 5.877** 12.982** 0.006 6.282** 1.739 e e e e e e e e e e e e e e 1.044* e e e e e e e e e e e 29.279**

7657.06

7595.04 7628.73 7599.90 3790.522 3789.950 X2 ¼ 1.14; p ¼ 0.766

(0.110) (1.389) (1.082) (0.601) (1.200) (1.359) e e e e e e e e e e e e e e (0.427) e e e e e e e e e e e (3.289)

0.471** 5.801** 13.012** 0.012 6.244** 1.713 e e e e e e e e e e e e e e e 0.905 0.897 Base Group 1.783 3.676 e e e e e e 31.959**

(0.110) (1.395) (1.084) (0.602) (1.203) (1.362) e e e e e e e e e e e e e e e (1.480) (1.545) e (1.474) (2.155) e e e e e e (3.130)

0.429** 5.897** 13.365** 0.042 6.605** 1.792 e e e e e e e e e e e e e e e e e e e e 0.904* e e e e e 29.289**

0.411** 5.986** 13.325** 0.005 6.799** 1.798 e e e e e e e e e e e e e e e e e e e e e Base Group 3.217* 3.183* 3.555* 4.553* 29.098**

(0.113) (1.394) (1.073) (0.602) (1.221) (1.361) e e e e e e

7648.02

7596.80 7630.48 7598.58 3791.399 3789.290 X2 ¼ 4.22; p ¼ 0.239

7646.0

(0.112) (1.390) (1.073) (0.601) (1.216) (1.360) e e e e e e e e e e e e e e e e e e e e (0.435) e e e e e (3.378)

e e e e e e e e e e e e (1.515) (1.465) (1.692) (2.190) (3.283)

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

Age Race (1 ¼ Black) Sex (1 ¼ Female) Political Status Residential Area (1 ¼ Urban) Victim (1 ¼ Yes) Social Media (Overall) None 60 min or less 61e120 min 121e181 min 181e240 min 241 min or more Social Media News (General) None 60 min or less 61e120 min 121e181 min 181e240 min 241 min or more Facebook Crime (Specific) Never Rarely Sometimes Often Very Often Twitter Crime (Specific) Never Rarely Sometimes Often Very Often Constant

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

165

Table 3 OLS regression predicting fear of crime. Variables

Age Race (1 ¼ Black) Sex (1 ¼ Female) Political Status Residential Area (1 ¼ Urban) Victim (1 ¼ Yes) Safety Neighborhood Problems Television (Overall) Internet (Overall) Local TV News National TV News Crime TV Shows Social Media (Overall) X Age X Race X Sex X Political Status X Residential Area X Victim X Safety X Neighborhood Problems Social Media News (General) X Age X Race X Sex X Political Status X Residential Area X Victim X Safety X Neighborhood Problems Facebook Crime/Violence X Age X Race X Sex X Political Status X Residential Area X Victim X Safety X Neighborhood Problems Twitter Crime/Violence X Age X Race X Sex X Political Status X Residential Area X Victim X Safety X Neighborhood Problems Constant R-Squared Notes: N ¼ 908. *p < 0.10;

**

p < 0.05;

(1)

(2)

(3)

b (SE)

b (SE)

b (SE)

**

0.421 5.227** 12.065** 0.067 5.683** 2.551* 3.732** 0.042 e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e 24.582** 0.274 ***

(0.109) (1.384) (1.081) (0.591) (1.191) (1.350) (0.764) (0.211) e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e (3.470)

**

0.394 4.913** 11.867** 0.036 5.320** 2.442* 3.705** 0.052 0.388 0.291 0.844 1.059* 0.218 e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e 23.743** 0.223

(0.112) (1.399) (1.137) (0.596) (1.221) (1.358) (0.770) (0.213) (0.364) (0.496) (0.658) (0.650) (0.376) e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e (4.220)

(4) b (SE) **

0.302 4.501** 11.279** 0.120 5.727** 2.376* 3.581** 0.082 0.488 0.256 0.753 1.207* 0.077 0.957* e e e e e e e e 0.044 e e e e e e e e 0.584 e e e e e e e e 0.601 e e e e e e e e 18.714** 0.225

(0.116) (1.398) (1.157) (0.595) (1.235) (1.357) (0.769) (0.213) (0.367) (0.540) (0.659) (0.651) (0.381) (0.432) e e e e e e e e (0.447) e e e e e e e e (0.463) e e e e e e e e (0.461) e e e e e e e e (4.490)

0.060 8.938* 17.715** 2.840 6.954 2.620* 7.292** 0.396 0.709 0.137 0.636 1.132* 0.317 5.722* 0.046 0.861 0.166 0.857 0.092 0.346 1.858** 0.321 2.370 0.117 0.909 0.748 0.279 0.794 0.540 0.732 0.291 4.302 0.014 0.420 0.670 0.193 0.035 0.937 0.269 0.254 2.528 0.080 0.740 1.591 0.099 1.313 1.660 0.822 0.059 11.809 0.237

(0.300) (4.290) (3.377) (1.836) (3.756) (4.177) (2.288) (0.605) (0.377) (0.547) (0.675) (0.664) (0.389) (2.731) (0.087) (1.022) (0.830) (0.466) (0.908) (1.008) (0.590) (0.164) (2.855) (0.087) (1.113) (0.953) (0.477) (0.992) (1.159) (0.668) (0.180) (2.954) (0.085) (1.244) (1.013) (0.526) (1.063) (1.292) (0.673) (0.179) (3.159) (0.109) (1.134) (0.989) (0.501) (1.130) (1.198) (0.661) (0.182) (9.696)

p < 0.01. Unstandardized coefficients presented.

variables as well, to be sure there are no cross-over interactions present (a negative effect for one group and a positive effect for another, resulting in a null average effect). This approach resulted in a total of 32 interactions being entered into the regression model, shown in Model 4 of Table 3. Results of this analysis suggest that the relationship between overall social media consumption and fear of crime is conditioned by perceptions of safety. In order to probe this interaction further, we use a spotlight analysis to determine the simple effect of social media consumption at each possible value of perceived safety. This technique involves coding the moderator variable such that 0 represents the category of interest (i.e. those who feel very safe in their neighborhood, or those who feel very unsafe in their neighborhood) and estimating the simple effect of social media on fear of crime for each group (Spiller et al., 2013). More specifically, this is

done by recalculating the values of perceived safety, subtracting the focal value (1e4) from each observation and including the recoded measure (along its interaction with social media consumption) in the regression model. We use a spotlight to estimate the relationship between social media consumption and fear of crime for respondents who feel (1) “very safe,” (2) “safe,” (3) “unsafe,” and (4) “very unsafe” being alone in their neighborhood at night. The simple effects observed across the four models estimated suggest that as an individual feels less safe, the positive relationship between overall social media consumption and fear of crime decreases to the point of being nonsignificant. Specifically, for individuals reporting that they feel “very safe” being alone in their neighborhood at night there was a positive and significant relationship between media consumption and their fear of crime (b ¼ 2.31, SE ¼ 0.603, p < 0.05). The size of

166

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

the coefficient was reduced, but remained marginally significant and positive, for respondents who reported being “safe” (b ¼ 0.987, SE ¼ 0.430, p < 0.10). The magnitude of the relationship was reduced to the point of being negative and nonsignificant for those who report being “unsafe” (b ¼ 0.337, SE ¼ 0.590, p > 0.10), and was reduced further, yet was still nonsignificant, for those who feel very unsafe (b ¼ 1.662, SE ¼ 0.924, p > 0.10). Fig. 1 displays the results of these analyses graphically, holding all other variables at their means.

8. Discussion At the outset of our study, we were interested in expanding and contributing to the cultivation research on fear of crime. Specifically, we focused on social media as a platform given its unique characteristics and due to the proliferation in the number of users today (Greenwood et al., 2016; Perrin, 2015). To our knowledge, no prior investigations have considered social media consumption on attitudes related to fear of crime or victimization. From the four social media consumption measures considered (overall, general news, and crime/violence on Facebook and Twitter), our results suggest that overall social media consumption plays an important role in increasing fear among young adults. Although this finding is consistent to the arguments embedded in cultivation research (e.g., more frequent media consumption is associated with fear of crime), none of the distinct measures of social media consumption (e.g., general news consumption and consumption of crime/violence stories) considered in the current study were found to significantly related to fear of crime. From a theoretical standpoint, we find this result quite puzzling. Yet, there are a few possible explanations for our results. First, similar to the research on frequent social media consumption and negative consequences (e.g., anxiety, social withdrawal) (Kross et al., 2013; Martin et al., 2012; O'Keeffe & ClarkePearson, 2011), our findings allude to the possibility that greater time spent using social media intensifies social and psychological factors that generates fear. Second, it is possible that our measures of crime/violence consumption are not tapping into specific domains that may be related to fear among social media consumers. For example, there may be explicit types of crime/violence (e.g., murder, terrorism) or specific types of stories/information (e.g., breaking news) that may be related to fear among individuals who consume social media. Third, our measures do not specifically examine the unique characteristics, such as engaging in stories, found in social media platforms. Perhaps stories that involve crime/ violence that receive more attention (e.g., likes or shares) and/or

Fig. 1. The Relationship between Overall Social Media Consumption and Fear of Crime across Perceptions of Safety in Neighborhood.

discourse (comments/informal discussion) may affect how information is processed (e.g., leads to greater fear). More nuanced measures of social media engagement are needed to address why and how this is occurring. For instance, crime content that has more worried individuals commenting may affect how a story is perceived and generate fear among those who are engaged in the information. Further, with the exception of the modest relationship found with national television news consumption, none of the traditional or entertainment media variables were related to fear of crime. Although this may not be consistent to previous efforts (Callanan, 2012; Chiricos et al., 2000, 1997; Dowler, 2003; Eschholz et al., 2003; Jamieson & Romer, 2014; Kohm et al., 2012), our results are not surprising given that social media consumption is more prevalent among young adults and ultimately a more important and influential media source when compared to traditional or entertainment-related media content. This is evident in our sample of college students that reported using social media, on average, more than any traditional media format surveyed. When turning to the disaggregated results in our investigation, only one difference was found between our measures of social media consumption and fear of crime based on respondents' demographics and backgrounds. The results illustrated that overall social media consumption on fear of crime varies among individuals' perception of safety. Specifically, overall social media consumption was significantly related to fear among individuals who feel safe (as opposed to those who feel unsafe). Because of this pattern, this result may best be understood by the substitution perspective, which contends that media consumption may have stronger effects for individuals without personal experiences with crime and violence (Gerbner et al., 1980). Stated differently, individuals who feel safe may be more susceptible to the messages depicted in social media consumption, which is ultimately associated with their levels of fear. In sum, owing to the limited significant relationships observed in the current study, our results only find partial support for cultivation theory and reception research with respect to the social media consumption and fear of crime relationship. Our study is not without limitations. First, due to our sampling procedure, we were unable to collect information on the actual crime rate. Previous media-related studies show that city-level crime rates influence fear (Chiricos et al., 2000; Eschholz et al., 2003). Although not an exact proxy to actual crime levels, similar to Chiricos et al. (1997) and Dowler (2003), we did collect information pertaining to perceived neighborhood conditions. We recommend that future studies examine how the effect of social media consumption on fear of crime differs by city- and neighborhood-level measures (e.g., crime, disadvantage, and proportion of racial/ethnic minorities). Second, our sample was limited to mostly young adults attending universities in three distinct (but much different) settings. Although we believe a multisite sample is a strength when using university-based adults, survey data illustrates that individuals who utilize social networking sites for news are more likely to be younger and middle-aged (e.g., 18e29 and 30e49 years old) as well as have higher educational backgrounds (e.g., some college or college degree) (Gottfried & Shearer, 2016). Thus, we cannot be certain that our results would be replicable in other sampling frames (e.g., non-student and/or older populations) and we advocate for future studies to replicate our research with a representative sample. Third, because approximately 90% of the sample was under the age of 25, we were unable to disaggregate the results by younger and older respondents. Although age limitations are relatively common in media studies utilizing universitybased samples (see Britto & Noga-Styron, 2014; Kohm et al., 2012; Waid-Lindberg, Dobbs, & Mname, 2011), we encourage future

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

research to utilize a wider age distribution in order to determine whether the social media and fear relationship differs by age. Fourth, we did not collect information on personality characteristics. Prior studies illustrate that personality and emotional traits, such as neuroticism, low conscientiousness, and/or low extraversion, are related to both fear of crime and social media usage (Correa, Hinsley, & De Zuniga, 2010; Klama & Egan, 2011). We encourage future research to explore the nature and impact of personality traits in the social media and fear relationship. Fifth, while we are confident in the empirical direction of the relationships examined, our data are cross-sectional and we recommend that future assessments carry out such analyses in a longitudinal manner to address the issue of causality. For instance, it would be interesting to utilize longitudinal data in order to assess whether fear from high profiled events (e.g., police misconduct, school shootings, and terrorism) affect or alter individuals' habits to gather more information from media sources. Sixth, although overall social media consumption was found to be associated with fear of crime, consumption of crime/violence stories on Facebook and Twitter were not related to fear of crime. This, in part, could be due to the way these measures were operationalized (see methods section). As a result, additional predictors, or variations of the measures used, are needed to better understand the relationship between the unique characteristics of social media usage and fear of crime. For example, it would be beneficial to determine whether some forms of social media engagement (like, share, comment) are more salient than other types of engagement. Lastly, we did not measure the unique patterns of consumers' social media habits (both social networking platforms and individually selected content). It is important for future research to understand how users' differing motivations and patterns between social networking sites may affect fear of crime. For example, how do different social media platforms provide or allow users to access information related to crime and violence? In addition, do individuals' patterns in consuming and engaging in crime-related content differ across social media platforms? And if so, do users prefer or trust a specific social media site over another? In short, owing to data restrictions and limitations noted above, our study should be regarded as a preliminary investigation into the relationship between social media and crime and justice-related outcomes. Despite our limitations, the current study has provided a promising foundation for future cultivation research on social media consumption. Given the substantial increase in the number of individuals consuming “contemporary” technology/mass media such as the Internet and social media, it is important for researchers to continue examining how these growing media platforms influence individuals' attitudes, belief, and perceptions. Specifically, social media consumption may also influence attitudes directed toward criminal justice policies, the criminal justice system (e.g., police), and punitiveness. We encourage scholars to explore these additional avenues of research. References Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A., & Ladwig, P. (2014). The “nasty effect”: Online incivility and risk perceptions of emerging technologies. Journal of Computer-mediated Communication, 19(3), 373e387. Britto, S., & Noga-Styron, K. E. (2014). Media consumption and support for capital punishment. Criminal Justice Review, 39(1), 81e100. Britto, S., Van Slyke, D. M., & Francis, T. I. (2011). The role of fear of crime in donating and volunteering: A gendered analysis. Criminal Justice Review, 36(4), 414e434. Brunton-Smith, I., & Sturgis, P. (2011). Do neighborhoods generate fear of crime? An empirical test using the British Crime Survey. Criminology, 49(2), 331e369. Callanan, V. J. (2012). Media consumption, perceptions of crime risk and fear of crime: Examining race/ethnic differences. Sociological Perspectives, 55(1), 93e115. Callanan, V. J., & Rosenberger, J. S. (2011). Media and public perceptions of the police: Examining the impact of race and personal experience. Policing &

167

Society, 21(2), 167e189. Chiricos, T., Eschholz, S., & Gertz, M. (1997). Crime, news and fear of crime: Toward an identification of audience effects. Social Problems, 44(3), 342e357. Chiricos, T., Padgett, K., & Gertz, M. (2000). Fear, TV news, and the reality of crime. Criminology, 38(3), 755e786. Cobbina, J. E., Miller, J., & Brunson, R. K. (2008). Gender, neighborhood danger, and risk-avoidance strategies among urban African-American youths. Criminology, 46(3), 673e709. Cohen, J., & Weimann, G. (2000). Cultivation revisited: Some genres have some effects on some viewers. Communication Reports, 13(2), 99e114. Conklin, J. E. (1975). The impact of crime. New York: Macmillan. Correa, T., Hinsley, A. W., & De Zuniga, H. G. (2010). Who interacts on the Web?: the intersection of users' personality and social media use. Computers in Human Behavior, 26(2), 247e253. Donley, A. M., & Gualtieri, M. C. (2017). “Homeless killer”: An analysis of the media's portrayal of the victims of a serial killer. Deviant Behavior, 38(2), 226e239. Donovan, K. M., & Klahm, C. F. (2015). The role of entertainment media in perceptions of police use of force. Criminal Justice and Behavior, 42(12), 1261e1281. Doob, A. N., & Macdonald, G. E. (1979). Television viewing and fear of victimization: Is the relationship causal? Journal of Personality and Social Psychology, 37(2), 170. Dowler, K. (2003). Media consumption and public attitudes toward crime and justice: The relationship between fear of crime, punitive attitudes, and perceived police effectiveness. Journal of Criminal Justice and Popular Culture, 10(2), 109e126. Eilders, C. (2006). News factors and news decisions. Theoretical and methodological advances in Germany. Communications, 31(1), 5e24. Eschholz, S. (1997). Media and fear of crime: A survey of the research. The University of Florida Journal of Law & Public Policy, 9, 37e60. Eschholz, S., Chiricos, T., & Gertz, M. (2003). Television and fear of crime: Program types, audience traits, and the mediating effect of perceived neighborhood racial composition. Social Problems, 50(3), 395e415. Ferraro, K. F. (1995). Fear of crime: Interpreting victimization risk. SUNY Press. Gabriel, U., & Greve, W. (2003). The psychology of fear of crime. Conceptual and methodological perspectives. British Journal of Criminology, 43, 600e614. Gainey, R., Alper, M., & Chappell, A. T. (2011). Fear of crime revisited: Examining the direct and indirect effects of disorder, risk perception, and social capital. American Journal of Criminal Justice, 36(2), 120e137. Garofalo, J. (1981). The fear of crime: Causes and consequences. Journal of Criminal Law and Criminology, 72(2), 839e857. Gerbner, G. (1969). Toward “cultural indicators”: The analysis of mass mediated public message systems. Educational Technology Research and Development, 17(2), 137e148. Gerbner, G., & Gross, L. (1976). Living with television: The violence profile. Journal of Communication, 26(2), 172e194. Gerbner, G., Gross, L., Jackson-Beeck, M., Jeffries-Fox, S., & Signorielli, N. (1978). Cultural indicators: Violence profile no. 9. Journal of Communication, 28(3), 176e207. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1980). The “mainstreaming” of America: Violence profile no. 11. Journal of Communication, 30(3), 10e29. Gilliam, F. D., Jr., & Iyengar, S. (2000). Prime suspects: The influence of local television news on the viewing public. American Journal of Political Science, 560e573. Goidel, R. K., Freeman, C. M., & Procopio, S. T. (2006). The impact of television viewing on perceptions of juvenile crime. Journal of Broadcasting & Electronic Media, 50(1), 119e139. Gottfried, J., & Shearer, E. (2016). News use across social media platforms 2016. Washington: Pew Research Center. Graber, D. A. (1980). Crime news and the public. New York: Praeger. Greenwood, S., Perrin, A., & Duggan, M. (2016). Social media update 2016: Facebook usage and engagement is on the rise, while adoption of other platforms holds steady. Pew Research Center. Gross, K., & Aday, S. (2003). The scary world in your living room and neighborhood: Using local broadcast news, neighborhood crime rates, and personal experience to test agenda setting and cultivation. Journal of Communication, 53(3), 411e426. Gunter, B. (1987). Poor reception: Misunderstanding and forgetting broadcast news. Routledge. Hale, C. (1996). Fear of crime: A review of the literature 1. International Review of Victimology, 4(2), 79e150. Heath, L., & Petraitis, J. (1987). Television viewing and fear of crime: Where is the mean world? Basic and Applied Social Psychology, 8(1e2), 97e123. Hermida, A., Fletcher, F., Korell, D., & Logan, D. (2012). Share, like, recommend: Decoding the social media news consumer. Journalism Studies, 13(5e6), 815e824. Hindelang, M. J., Gottfredson, M. R., & Garofalo, J. (1978). Victims of personal crime: An empirical foundation for a theory of personal victimization. Cambridge, MA: Ballinger. Hirsch, P. M. (1980). The “scary world” of the nonviewer and other anomalies: A reanalysis of Gerbner et al.'s findings on cultivation analysis part I. Communication Research, 7(4), 403e456. Hughes, M. (1980). The fruits of cultivation analysis: A reexamination of some effects of television watching. Public Opinion Quarterly, 44(3), 287e302. Intravia, J., Stewart, E. A., Warren, P. Y., & Wolff, K. T. (2016). Neighborhood disorder and generalized trust: A multilevel mediation examination of social mechanisms. Journal of Criminal Justice, 46, 148e158.

168

J. Intravia et al. / Computers in Human Behavior 77 (2017) 158e168

Jamieson, P. E., & Romer, D. (2014). Violence in popular US prime time TV dramas and the cultivation of fear: A time series analysis. Media and Communication, 2(2), 31e41. Klama, E. K., & Egan, V. (2011). The Big-Five, sense of control, mental health and fear of crime as contributory factors to attitudes towards punishment. Personality and Individual Differences, 51(5), 613e617. Kohm, S. A., Waid-Lindberg, C. A., Weinrath, M., Shelley, T. O. C., & Dobbs, R. R. (2012). The impact of media on fear of crime among university students: A cross-national comparison 1. Canadian Journal of Criminology and Criminal Justice, 54(1), 67e100. Kort-Butler, L. A., & Hartshorn, K. J. S. (2011). Watching the detectives: Crime programming, fear of crime, and attitudes about the criminal justice system. The Sociological Quarterly, 52(1), 36e55. Kross, E., Verduyn, P., Demiralp, E., Park, J., Lee, D. S., Lin, N., et al. (2013). Facebook use predicts declines in subjective well-being in young adults. PloS One, 8(8), e69841. Lee, E. J., & Jang, Y. J. (2010). What do others' reactions to news on Internet portal sites tell us? Effects of presentation format and readers' need for cognition on reality perception. Communication Research, 37(6), 825e846. Lipschultz, J. H., & Hilt, M. L. (2014). Crime and local television news: Dramatic, breaking, and live from the scene. Routledge. Liska, A. E., & Baccaglini, W. (1990). Feeling safe by comparison: Crime in the newspapers. Social Problems, 37(3), 360e374. Liska, A. E., & Warner, B. D. (1991). Functions of crime: A paradoxical process. American Journal of Sociology, 96(6), 1441e1463. Marsh, H. L. (1991). A comparative analysis of crime coverage in newspapers in the United States and other countries from 1960e1989: A review of the literature. Journal of Criminal Justice, 19(1), 67e79. Martin, E. A., Bailey, D. H., Cicero, D. C., & Kerns, J. G. (2012). Social networking profile correlates of schizotypy. Psychiatry Research, 200(2), 641e646. McCall, J. M. (2007). Viewer discretion advised: Taking control of mass media influences. Lanham, MD: Rowman & Littlefield. Menifield, C. E., Rose, W. H., Homa, J., & Cunningham, A. B. (2001). The media's portrayal of urban and rural school violence: A preliminary analysis. Deviant Behavior, 22(5), 447e464. Mitchell, A., Gottfried, J., Barthel, M., & Shearer, E. (2016). The modern news consumer: News attitudes and practices in the digital era. Washington: Pew Research Center. Morgan, M., & Shanahan, J. (2010). The state of cultivation. Journal of Broadcasting & Electronic Media, 54(2), 337e355. Morgan, M., Shanahan, J., & Signorielli, N. (2014). Cultivation theory in the twentyfirst century. In R. S. Fortner, & P. M. Fackler (Eds.), The handbook of media and mass communication theory (pp. 480e497). Hoboken, NJ: John Wiley and Sons. Nellis, A. M., & Savage, J. (2012). Does watching the news affect fear of terrorism? The importance of media exposure on terrorism fear. Crime & Delinquency, 58(5), 748e768. Nielsen, A. L., & Bonn, S. (2008). Media exposure and attitudes toward drug addiction spending, 1975e2004. Deviant Behavior, 29(8), 726e752. O'Keefe, G. J., & Reid-Nash, K. (1987). Crime news and real-world blues: The effects of the media on social reality. Communication Research, 14(2), 147e163. O'Keeffe, G. S., & Clarke-Pearson, K. (2011). The impact of social media on children, adolescents, and families. Pediatrics, 127(4), 800e804. Perkins, D. D., & Taylor, R. B. (1996). Ecological assessments of community disorder: Their relationship to fear of crime and theoretical implications. American Journal of Community Psychology, 24, 63e107.

Perrin, A. (2015). Social media usage: 2005-2015. Washington: Pew Research Center. Roberts, J. V., & Stalans, L. J. (1997). Public opinion, crime, and criminal justice. Bolder: WestView Press. Roche, S. P., Pickett, J. T., & Gertz, M. (2016). The scary world of online news? Internet news exposure and public attitudes toward crime and justice. Journal of Quantitative Criminology, 32(2), 215e236. Romer, D., Jamieson, K. H., & Aday, S. (2003). Television news and the cultivation of fear of crime. Journal of Communication, 53(1), 88e104. Ross, C. E. (1993). Fear of victimization and health. Journal of Quantitative Criminology, 9(2), 159e175. Rountree, P. W., & Land, K. C. (1996). Perceived risk versus fear of crime: Empirical evidence of conceptually distinct reactions in survey data. Social Forces, 1353e1376. Skogan, W. G., & Maxfield, M. G. (1981). Coping with crime: Individual and neighborhood reactions. Beverly Hills, CA: Sage. Slingerland, W. J., Copes, H., & Sloan, J. J., III (2006). Media construction of whitecollar violence revisited: An examination of two nightclub tragedies. Deviant Behavior, 27(4), 423e455. Spiller, S. A., Fitzsimons, G. J., Lynch, J. G., & McClelland, G. H. (2013). Spotlights, floodlights, and the magic number zero: Simple effects tests in moderated regression. Journal of Marketing Research L, 277e288. Stafford, M., Chandola, T., & Marmot, M. (2007). Association between fear of crime and mental health and physical functioning. American Journal of Public Health, 97, 2076e2081. Surette, R. (2007). Media, crime, and criminal justice: Images, realities, and policies. Belmont, CA: Thomas Wadsworth. Taylor, R. B., & Hale, M. (1986). Testing alternative models of fear of crime. The Journal of Criminal Law and Criminology (1973), 77(1), 151e189. Thompson, A. C. (2010). From sound bites to sound policy: Reclaiming the high ground in criminal justice policy-making. Fordham Urban Law Journal, 38, 775e820. Truman, J. L., & Morgan, R. E. (2016). Criminal victimization, 2015. Washington, DC: U.S. Department of Justice. Waid-Lindberg, C. A., Dobbs, R. R., & Mname, T. (2011). Blame the media? The influence of primary news source, frequency of usage, and perceived media credibility on punitive attitudes. Western Criminology Review, 12(3), 41. White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica: Journal of the Econometric Society, 817e838. Warr, M. (1990). Dangerous situations: Social context and fear of victimization. Social Forces, 68(3), 891e907. Warr, M. (2000). Fear of crime in the United States: Avenues for research and policy. Criminal Justice, 4, 451e489. Weber, P. (2014). Discussions in the comments section: Factors influencing participation and interactivity in online newspapers' reader comments. New Media & Society, 16(6), 941e957. Weitzer, R., & Kubrin, C. E. (2004). Breaking news: How local TV news and realworld conditions affect fear of crime. Justice Quarterly, 21(3), 497e520. Wyant, B. R. (2008). Multilevel impacts of perceived incivilities and perceptions of crime risk on fear of crime: Isolating endogenous impacts. Journal of Research in Crime and Delinquency, 45(1), 39e64. Ziegele, M., Breiner, T., & Quiring, O. (2014). What creates interactivity in online news discussions? An exploratory analysis of discussion factors in user comments on news items. Journal of Communication, 64(6), 1111e1138.