How news consumption in modern media landscapes relates to threats perceptions and fears in public spaces. A scoping review

Abstract

A high percentage of current news media covers topics relevant to the public such as wars, crime, and terrorist attacks, making threat perceptions and fears in public spaces (TPF-PS) more salient. Due to the unique challenges involved with current media landscapes, research examining how news consumption (NC) is related TPF-PS is still attracting scholarly interest. Therefore, we conducted a scoping review of N = 72 empirical studies since the advent of the smartphone era. The review detects cultivation theory as the main theoretical approach as well as a prevalence of quantitative methods and cross-sectional surveys. Furthermore, most empirical studies are carried out in North America and draw on non-probability sampling methods. Moreover, the review detects (region-based) research trends by breaking down N = 307 identified operationalizations into main classifications of TPF-PS and NC. Additionally, the review indicates a predominance of no associations between NC and TPF-PS. Implications for future research are discussed.

Keywords:

• Scoping review
• news consumption
• digital media
• threat perceptions
• public fears

In current times, endless streams of graphic images and media reports about terrorist organizations and committed violent crimes are broadcasted around the world (Wessler et al. 2022). Arguably, this permanent awareness of the latest events, as well as past crime incidents or terrorist attacks, are making society’s fear of threats more salient. Threat perceptions and fears in public spaces (TPF-PS) (including, e.g. generalized fears of victimization and feelings of insecurity within the general public) are typically regarded as a robust indicator of people’s political mood with meaningful links to outgroup prejudice and support for authoritarian policies (e.g. Osborne et al. 2023). Naturally, intrinsic and extrinsic conditions influencing TPF-PS have been investigated extensively over the past decades. In this light, mass media has long been singled out as responsible for exaggerated TPF-PS following cultivation theory (Gerbner 1998) and other prominent theoretical approaches on media effects such as agenda-setting theory, reinforcing spirals theory, or framing theory (for a general theoretical overview, see Valkenburg, Peter, and Walther 2016). Research examining how news consumption (NC) is related to TPF-PS is still attracting substantial scholarly interest (Potter 2022). This academic interest is growing particularly due to the unique challenges involved with current media landscapes (Morgan, Shanahan, and Signorielli 2014).

In participatory democracies, news reporting serves the crucial function to inform the public about significant events in order to facilitate their involvement in political discourses. Institutionalized actors (e.g. editorial staffs, journalists, news anchors) across traditional media like newspapers, radio, and TV have fulfilled this function for many decades. However, social media, together with the widespread prevalence of smartphones, has changed people’s NC fundamentally (Kümpel 2022). That is, the widespread adoption of smartphones has facilitated unprecedented access to news content, enabling users to stay continuously connected to information sources throughout the day. This following shift in NC behavior, characterized by constant connectivity and on-the-go access, has significant implications for how individuals perceive and respond to threats in their environment.

A scoping review (see Munn et al. 2018) can be a suitable way to gain a systematic overview of certain research efforts in order to identify current mainstream trends and, more importantly, diagnose deficits that may bias the field and impair its societal utility. There are systematic reviews that address the role of (social) media for risk and crisis communication (Rasmussen and Ihlen 2017; Wahlberg and Sjoberg 2000) and also meta-analyses that test links between overall media consumption and public risk perception (Niu et al. 2022)—a particularly challenging task given that the body of literature exhibits many heterogeneous phenomena. What is currently lacking, however, is a methodologically sound systematization concerning the features of empirical studies that deal with NC and TPF-PS.

For this purpose, we conducted a scoping review focusing on empirical literature about links between people’s NC and TPF-PS since the advent of the smartphone era (starting with the introduction of the iPhone in 2007). The advent of the smartphone era serves as a significant milestone for investigating NC in modern media landscapes and its link to TPF-PS for two reasons. First, it represents a significant inflection point in the history of media technology, marking the beginning of a mobile era characterized by ubiquitous connectivity. Second, through a clear temporal boundary around the smartphone era, we can ensure that our review focuses on the most relevant and contemporary empirical literature.

Our review contributes to existing research by pursuing a threefold objective: First, the review identifies both prevalent theoretical and methodological approaches, as well as sample characteristics and sampling methods in this research field. Second, it outlines previous operational approaches of the two primary concepts (i.e. TPF-PS and NC). Third, in line with methodological approaches of recent scoping reviews on other topics [e.g. incidental news exposure (Schäfer 2023) or social media effects on well-being (Valkenburg, van Driel, and Beyens 2022)], the present study aims to categorize the identified operationalizations of NC and TPF-PS into two main classifications and examine their types of association while also considering region-based differences.

Objectives of the present study

Theories, methodologies, and samples

Since its beginnings, media effects research has developed several micro-level theories that withstood the test of time (or rather evolved during it) and remained highly relevant today (see Valkenburg, Peter, and Walther 2016). For NC effects in particular, and arguably even more so concerning people’s TPF-PS, some of these well-established theories have also been applied heavily. Most notably, explorative adaptations of cultivation theory (Gerbner 1998) that argue for a socializing impact of repeated exposure to sensationalist (local) news reporting, alongside framing theory (Entman 1993), which posits that any news content can influence audience perceptions by presenting information in a certain way, have been commonly used as a theoretical background (e.g. Chadee, Smith, and Ferguson 2019; De Coninck et al. 2021). Other popular approaches like information processing theories [e.g. adaptions of the elaboration likelihood model (Petty and Cacioppo 1986) or the limited capacity model (Lang 2000)], uses-and-gratification theory (Ruggiero 2000), or the reinforcing spiral model (Slater 2007) also appear promising to predict and explain nuanced effects on threat perceptions (see, e.g. Matthes, Schmuck, and von Sikorski 2019; Williamson, Fay, and Miles-Johnson 2019). Beyond these major theories, domain-specific approaches which may have originated in related disciplines [e.g. social identity theory (Tajfel and Turner 2004) and intergroup threat theory (Stephan and Stephan 2000)], may play an important role in current theoretical thinking (e.g. Lissitsa, Kushnirovich, and Steinfeld 2023). By systematically determining the theoretical state-of-the-art, it may be possible to identify dominant lenses through which NC is empirically approached as well as peripheral viewpoints that might allow for unique (maybe even disruptive) predictions to expand our understanding of the phenomenon in question. Accordingly, we asked:

RQ1: What theoretical approaches are prevalent in the sampled literature?

When it comes to methodologies, researchers periodically bemoan several patterns across diverse phenomena in communication science (see, e.g. Parry et al. 2022 for research on social media use and well-being or Wei, Fan, and Leo-Liu (2023) for mobile communication research). Specifically, it has been pointed out that many problems in communication research are predominantly approached through a critical-realist lens via quantitative methods, mainly using cross-sectional survey methods and self-report scales. Prior reviews emphasized three shortcomings here: first, a relative lack of social-constructivist and interpretative perspectives that make use of different qualitative methods; second, a more specific dearth of longitudinal and experimental (quantitative) methods that include behavioral indicators; and third, a profound shortage of mixed-method approaches combining paradigms to maximize their respective strengths and minimize their weaknesses. As a consequence of this methodological imbalance, many research areas suffer from relying primarily on fairly decontextualized, inherently standardized, and necessarily researcher-driven findings that tend to neglect how context-sensitive and idiosyncratic communication experiences may be across different people. Additionally, cross-sectional quantitative research does typically (see Grosz, Rohrer, and Thoemmes 2020) not allow for solid insights into causal mechanisms without making strong (i.e. potentially unviable) assumptions (Rohrer 2018). It is unclear whether research on NC and TPF-PS shows similar imbalances. Therefore, we asked:

RQ2: What methodological approaches are prevalent in the sampled literature?

Another frequently highlighted methodological issue concerns sampling methods and sample populations in quantitative research. Regarding the former, reviews have documented that non-probability sampling (i.e. selection of individuals from a population of interest through non-random means, e.g. quota-based, purposive, convenient sampling methods) is prevailing over probability sampling (e.g. simple random, stratified random, clustered sampling methods) in both survey and experimental communication research (Erba et al. 2018). This prevalence is typically considered problematic because population-level inferences from non-probability samples rest on untested assumptions that may at least be partially violated and, thus, introduce additional error (see Cornesse et al. 2020 for a detailed review). Prototypical issues of non-probability sampling methods are, for instance, an overrepresentation of more easily reachable or topically motivated social groups, a method-inherent uncertainty about the distribution of potentially confounding parameters (Lehdonvirta et al. 2021), or, especially when using students of one’s own program, varying expertise in the research topic (Meltzer, Naab, and Daschmann 2012). Accordingly, the prevalence of different sampling methods can serve as an indicator for the quality of available quantitative evidence.

Concerning sample populations, communication scholars have noted a predominance of samples including people from so-called WEIRD (i.e. western, educated, industrialized, rich, democratic) societies that do not represent the vast majority of the world population and thus put the generalizability and transferability of evidence into question (Bates 2021). What is more (and closely related to some non-probability sampling methods) is that samples often do not resemble existing heterogeneity within WEIRD societies but instead tend to draw from a higher educated, racially less diverse, and more socio-economically advantaged subpopulation that is university students.

Given their potential relevance, these well-known challenges concerning sampling methods and sample populations may arguably be even more pressing for empirical inquiries into people’s TPF-PS. It is therefore imperative to obtain an overview of the state of the field, which is why we asked:

RQ3: What a) sample characteristics and b) sampling methods are prevalent in the sampled literature?

Operationalizations of TPF-PS and NC

Threat perception is often considered an ambiguous construct in terms of available operationalizations because its phenomenological and conceptual vicinity to fear responses (toward perceived threats) have troubled researchers across different disciplines (see, e.g. LeDoux 2014; LaTour and Rotfeld 1997). Concerning research on TPF-PS, this ambiguity can be observed for common construct pairings (e.g. crime perception and fear of crime, perceived victimization risk and fear of victimization, or terrorism threat perceptions, and fear of terrorism) where the former represent people’s varying awareness about undesirable circumstances and the latter describe their emotional response to these circumstances. Although this distinction might appear to be conceptually reasonable, it is not self-evident that it is also strictly followed through in standard self-report items, such that threat perceptions and self-perceived fears may occasionally be used (close to) synonymously. To account for this, we opted for accepting operationalizations of both threat perceptions and related fears in this scoping review.

When it comes to public threat domains, researchers typically have to decide between holistic and domain-specific operationalizations (e.g. crime perception vs. violent crime perception) that can differ in terms of their reference, for instance, to broad or more narrow geographical locations (e.g. national vs. regional crime perception; e.g. Ramírez-Álvarez 2021) or different groups of actors (e.g. perceived threat of distinct immigrant populations; e.g. Theorin 2022) and patients (e.g. fear for oneself vs. fear for one’s family or friends; e.g. Nellis and Savage 2012). It is therefore essential to obtain a comprehensive overview on what operational approaches may be predominant (or even overrepresented, perhaps so much so that additional work might only incrementally advance our knowledge) and which ones might be promising but neglected (such that a stronger emphasis might be worthwhile). Thus, we asked:

RQ4a: How is TPF-PS operationalized in the sampled literature?

News can be consumed in myriad ways in modern media landscapes. Traditionally, news reporting has primarily been carried out offline by media institutions (e.g. local and national newspapers, magazines, radio stations, TV channels; in the following referred to as traditional/offline NC). While these traditional news actors are still relevant today—particularly given that most of them have rapidly also gone digital using websites, apps and social media profiles—peoples’ current news diets now not only include (non-institutionalized) digital news reporting via social media, podcasts, or alternative news sources (Kümpel 2022) but also a combined use of several media channels (in the following referred to as digital/online & multichannel NC including online NC, social media NC, and the combined consumption of offline and online news).

Accordingly, research is forced to adapt to these modern (and further dynamically changing) landscapes in order to still be able to address people’s lived NC routines without leaving significant portions in the dark. Besides that (and similar to TPF-PS), NC operationalizations may also differ substantially with respect to their granularity (e.g. private vs. public TV channels, or local vs. national vs. international newspapers, or internet vs. social media vs. Twitter/X) or their topical specificity (e.g. general news coverage vs. news about terrorism vs. news about particular events). Furthermore, the nature of modern media landscapes illustrates that properties of media that are otherwise separated into individual devices are combined into one medium (e.g. NC on smartphones). As discussed in other scoping reviews (e.g. on incidental news exposure; Schäfer 2023), this further emphasizes the role of end-user devices.

Against this background, a scoping review affords an opportunity to explore whether current research is keeping up with current developments and uncover latest trends as well as crucial blind spots in how NC is operationalized when it comes to studying its link to TPF-PS.

RQ4b: How is NC operationalized in the sampled literature?

Association trends and region-based differences

Scoping reviews are viewed as a useful approach to systematically analyzing quantitative empirical evidence on a topic. As outlined above, it is very likely that previous research on NC and TPF-PS includes different, more or less overlapping, or otherwise vaguely defined construct operationalizations. Hence, in line with the methodological approaches of prior scoping reviews (Schäfer 2023; Valkenburg, van Driel, and Beyens 2022), we intend to test whether their associations differ across traditional and digital/online & multichannel NC as well as across fear-related and not fear-related concepts of TPF-PS. In addition, this review aims to identify region-based differences by categorizing the identified operationalizations of NC and TPF-PS into two primary classifications.

RQ5: What a) types of association and b) region-based differences can be identified in the sampled literature in relation to the two main concepts?

Method

Eligibility criteria

All empirical papers must fulfill several criteria in order to be included. First, papers must involve data collections with human participants, regardless of whether they use quantitative, qualitative, or mixed methods (referred to as empirical study). Hence, the present scoping review does not cover other publication forms such as reviews, meta-analyses, or only theory-based studies. Second, at least one term related to media consumption/exposure and at least one term related to TPF-PS must appear in the respective title, abstract, or keywords. Third, papers either have to be published in academic journals or book chapters (including online-first publishing formats). Fourth, they must be written in English or German (i.e. languages that the authors understand proficiently). Fifth, records must be published between 2007 and 2023. As the first iPhone generation was available in 2007, we decided to set this time frame as it best marks the beginning of the smartphone era. Sixth, papers must address the relation between NC and TPF-PS. Accordingly, we exclude threat perceptions that are not triggered by humans (e.g. natural emergencies; Gutteling, Terpstra, and Kerstholt 2018) and personal safety evaluations in non-public (e.g. violence against women or misogyny in private areas) or virtual (e.g. cybercrime) settings, as well as fear-related concepts related to individual concerns about sanitary or health safety (e.g. individual fear of Covid-19 risk; Dryhurst et al. 2020). Figure 1 visualizes our inclusion criteria by differentiating and partially summarizing both formal and content-related criteria.

Figure 1. Flow chart illustrating inclusion criteria for the scoping review.

Search procedure

To identify relevant papers, we utilized Web of Science and Scopus as two well-established academic databases (see Appendix A for search items). We preset our search by only including a) papers from 2007 to 2023, b) articles, book chapters, early access, and conference papers, c) relevant research areas, and d) articles in English and German. On May 18, 2023, a total of N = 719 papers was identified (Web of Science: 315; Scopus: 404). Based on subsequent (iterative) screening of titles, abstracts, and full texts, N = 49 papers were included (see Appendix B).

Coding procedure

Coding was performed by the first and second authors. During initial coding, which in total consisted of 15 randomly selected papers, we went through three coding rounds to arrive at acceptable inter-coder reliability across each category (i.e. Krippendorf’s α > 0.80). Afterward, we randomly assigned the 34 remaining papers to one of both coders. Each operational definition of NC and TPF-PS constituted a separate case. Since we are interested in sample characteristics and sampling methods, additional coding units represented a) subsamples for different countries and b) each empirical study conducted within a single paper (e.g. two-study papers). Overall, this led to N = 72 empirical studies as well as N = 307 operational definitions of NC and TPF-PS (see Figure 2 for an overview).

Figure 2. Flow chart illustrating the scoping review process.

Collected data items

Measures for meta information

Unrelated to any RQs, we first retrieved titles, author names, publication year, and the name of the journal/book. For answering RQ1 and RQ2, we determined the main theory related to our variables of interest (α = 0.84) by examining whether specific theories or theoretical keywords were referred to (see Appendix C) using both an open category and an a priori categorization. Based on previous scoping reviews (e.g. Schäfer 2023), we also coded methodology (quantitative, qualitative, or mixed; α = 1.00), method (cross-sectional survey, longitudinal survey, experiment, eye tracking, focus group interviews, individual interviews, observation, or mixed methods by applying two or more different methods; α = 1.00), and study setting (online, offline; α = 1.00).

For answering RQ3, we first indicated whether a probability or non-probability sampling method was used (α = 1.00). Based on Rasmussen and Ihlen (2017), we also coded the sample region (α = 1.00). Moreover, information on studied populations (α = 0.97) and sample gender distribution (α = 1.00) was investigated. Finally, we recorded if ethnicity or race (α = 1.00) was taken into account for analyses and what kind of socioeconomic factors (including education, financial situation, employment status, and area of residence) were reported (α = 0.95).

Measures for subject matter information

For answering RQ4, we first determined what kind of operational definitions of TPF-PS a paper included (RQ4a). Based on previous research, coding categories included a) fear of (violent) crime, b) threat/crime (trend) perceptions, c) fear of terror, and d) fear of victimization. Additionally, we included a category that indicated the topical specificity of the given operationalizations (α = 1.00). Regarding RQ4b, we coded whether operational definitions of NC of a paper assessed a) media use in general or, more specifically, b) NC [online/digital (online news articles, internet NC, and NC on websites) vs. offline/traditional (e.g. TV, newspaper, radio)], c) NC on social media, or d) NC with regard to a specific topic (e.g. terrorist attacks). We further coded for mixed measurements and thus multichannel NC (i.e. mergers of offline and online NC, for instance, combined operationalizations of NC via TV, Facebook, print and online news) (α = 0.95).

In two open categories, we recorded the given names of all operational definitions of NC and TPF-PS. Furthermore, for getting a detailed picture, we assessed the type of end-user device (smartphones, tablets/laptops/PCs, television, radio, or print; α = 0.89). Additionally, once the role of social media was investigated, we also recorded whether studies focused on specific networks or generic social media use.

Finally, concerning RQ5, we assessed the type of association (i.e. none, positive, negative, or mixed) between NC and TPF-PS (α = 0.99). For exploratory purposes (not reported here), we also wrote a brief report of the qualitative studies’ main findings.

Results

Out of the total of 49 papers, only two were published as book chapters. The remaining papers were journal articles. Figure 3 illustrates a timeline for publication years, revealing that empirical research on the relationship between NC and TPF-PS peaked after 2020.

Figure 3. Timeline illustrating the publication year of the included papers.

Note: N = 49.

Theoretical and methodological approaches

RQ1 asked for the prevalence of theories across sampled papers. While 22.4% (n = 11) did not refer to any specific theory, a majority applied cultivation theory (44.9%, n = 22). Apart from that, several investigations referred to the theoretical premises of framing and priming (14.3%, n = 7) or involved other media-related theories with a focus on the individual and the topic [e.g. intergroup threat theory, hostile media effect; 14.3% (n = 7)]. None of the included studies referred to news value theory and only one study each referred to information processing theories or uses and gratification approaches (2%, n = 1 for each).

Concerning the methodological approaches (RQ2), most empirical studies (N = 72) applied a quantitative methodology (93.1%, n = 67), with cross-sectional surveys (63.9%, n = 46), experiments (18.1%, n = 13), and longitudinal surveys (11.1%, n = 8) being the most prevalent methods. The remaining studies with qualitative methodology (6.9%, n = 5) either engaged in individual interviews (4.2%, n = 3) or focus group discussions (2.8%, n = 2). Regarding the setting, more than half of the sampled empirical studies took place online (56.3%, n = 40, 1 missing value). While most of the remaining studies preferred an offline setting (40.8%, n = 29), two studies (2.8%) combined online and offline settings.

Sample characteristics and sampling methods

Concerning sample characteristics (RQ3a), most empirical studies were carried out in North America (34.7%, n = 25), Northern/Western Europe (26.4%, n = 19), South/Eastern Europe (13.9%, n = 10), and Central America (11.1%, n = 8). Samples from Western Asia (5.6%, n = 4), Eastern/Southeastern Asia (2.8%, n = 2), South America (1.4%, n = 1), Oceania (1.4%, n = 1), or multinational samples (2.8%, n = 2) were rare.

A large majority of studies (70%, n = 49, 2 missing values) only included adults (i.e. starting from 18 years), whereas some studies combined adults and participants younger than 18 years (18.6%, n = 13). Only a small proportion of studies focused specifically on children and teenagers (2.9%, n = 2) and 8.6% (n = 6) exclusively examined students.

Most empirical studies achieved a balanced gender distribution for their samples (maximum 40/60; 84.1%, n = 58, 3 missing values). Unbalanced gender distributions only included a higher proportion of female (versus male) participants (15.9%, n = 11). Concerning other sociodemographic indicators, many studies considered both participants’ education levels (80.6%, n = 58) and indicators for participants’ socioeconomic status based on their area of residence (44.4%, n = 32). Participants’ financial situation (37.5%, n = 27) and their current employment status (25%, n = 18) were rarely included in the analysis. A total of 41.7% (n = 30) studies addressed participants’ ethnicity or race in their analyses.

Regarding sampling methods (RQ4b), n = 8 empirical studies provided no further information. The remaining 64 cases predominantly showcased non-probability (67.2%, n = 43) compared to probability sampling (32.8%, n = 21).

Operational definitions for main concepts

To obtain a comprehensive overview of the operational approaches of TPF-PS (RQ4a) and NC (RQ4b), we take a closer look at all coded operationalizations of NC and TPF-PS (N = 307). Concerning TPF-PS (RQ4a), many empirical studies conceptualized TPF-PS as a fear-related concept (i.e. 46.6%, n = 143). Other studies measured general threat perceptions related to specific actor groups (40.4%, n = 124) or determined people’s general crime (trend) perceptions (6.8%, n = 21). Few exceptions (6.2%, n = 19) addressed concepts related to people’s coping or, more specifically, avoidance behavior (e.g. perceived likelihood of future self-protective action).

Three fears were predominant: a) fear of (violent) crimes (25.7%, n = 79), b) fear of terror/terrorism (13.4%, n = 41), and c) fear of victimization (7.5%, n = 23). Interestingly, a large majority of fear measures were related to people’s fear for oneself (88.1%, n = 126), hardly asking for peoples’ fears for others (2.8%, n = 4), except for some composite measures of fear for another’s and one’s own safety (9.1%, n = 13). Studies also rarely differentiated narrow geographical locations (18.9%, n = 58). That is, only 7.2% (n = 22) of operationalizations referred to broad geographical areas (e.g. national, country, state) and another 9.4% (n = 29) to more narrow locations (e.g. urban, community, neighborhood, rural, municipality). However, some studies made use of combined measurement of national and regional fears (2.3%, n = 7).

Concerning the operational definition of NC (RQ4b), sampled operationalizations manifested a preference for traditional NC/exposure in general (41.0%, n = 126) as well as NC with respect to specific topics (29.3%, n = 90). Among these topics, three emerged as most notable: a) news about terrorism/terrorists (45.6%, n = 41), b) news about immigrants, refugees, or asylum seekers (36.7%, n = 33), and c) news involving (violent) crime scenarios (17.8%, n = 16). Less prevalent but still noteworthy, several studies operationalized NC by merging offline and online exposures (15%, n = 46). Empirical studies that measured people’s NC on social media (without referring to specific topics and merging offline and online NC/exposure) constitute a rarity (5.5%, n = 17), just as it is the case for overall media consumption/exposure (3.9%, n = 12), online-only NC (3.3%, n = 10), and general internet use (2%, n = 6). Of the empirical studies that somehow included NC on social media (i.e. additionally including those empirical studies that either referred to social media within a merged operationalization of NC or to social media news on specific topics) , most did not specify specific platforms (n = 25). If platforms were specified, Twitter n = 8) and Facebook n = 10) were most commonly investigated, alongside a weaker emphasis on WhatsApp n = 2), WeChat n = 1), and other platforms (n = 2).

Looking at end-user devices, roughly 1 out of 4 operationalizations did not mention the specific end-user device at all (25.4%, n = 78). When end-user devices were specified for traditional NC, TV (39.4%, n = 121) and print media emerged as most relevant (10.7%, n = 33), dominating radio (3.9%, n = 12). When it comes to online media, not a single study clearly indicated the specific end-user device (smartphones, tablets/laptops/PCs). Most commonly, operationalizations combined several end-user devices in a composite measure (20.5%, n = 63).

Type of association differentiated by main concepts

To determine general association trends between our variables of interests (RQ5a), we coded quantitative investigations (per operationalization) for whether they identified no significant (57%, n = 171), a positive (28.0%, n = 84), or a negative association (9.7%, n = 29). Mixed findings (i.e. where non-significant, positive, and negative associations came up for different analytical techniques or when including additional control variables; 5.3%, n = 16) were disregarded here.

We first recoded TPF-PS operationalizations, creating a nominal variable (concept type) that distinguishes between fear-related concepts (i.e. fear of crime, fear of terror, and fear of victimization) and general threat and crime (trend) perceptions as the primary research foci (excluding seldom examined variables like coping/avoidance behaviors). A Pearson chi-square test showed a small-sized significant relationship between association type and concept type (χ²(2, n = 265) = 9.41, p < 0.01, Cramer’s V = 0.19). As Table 1 displays, empirical studies most often found no significant associations between NC and TPF-PS. When only counting significant associations, positive links (i.e. higher NC comes with increased TPF-PS, and vice versa) occurred more often than negative ones (i.e. higher NC comes with decreased TPF-PS, and vice versa). Compared to general threat and crime (trend) perceptions, negative links were less common for fear-related concepts.

Table 1. Crosstabulation for the link between type of association and concept type.

			Concept type
			Fear-related concepts	Threat and crime (trend) perceptions	Total
Type of association	No association (n.s.)	Count	72	85	157
		% within association	45.9%	54.1%	100%
		% within concept	59.0%	59.4%	59.2%
	Positive association	Count	44	36	80
		% within association	55.0%	45.0%	100%
		% within concept	36.1%	25.2%	30.2%
	Negative association	Count	6	22	28
		% within association	21.4%	78.6%	100%
		% within concept	4.9%	15.4%	10.6%

Note: N = 265; 0 cells have expected count less than 5; the minimum expected count is 12.89.

Secondly, we also recoded NC operationalizations, again creating a nominal variable named NC type where we differentiated between traditional/offline NC (i.e. receiving news from traditional media institutions such as newspapers) and digital/online & multichannel NC (including online NC, social media NC, and the combined consumption of offline and online news) as theoretically most interesting focus points (excluding generic measures like overall media consumption and internet use not news-related). Another Pearson chi-square test pointed toward a medium-sized significant relationship between association type and NC type (χ²(2; n = 243) = 14.53, p < 0.01, V = 0.25). As can be observed in Table 2, significant differences across NC types again occurred only for negative associations, which were significantly less often identified for digital/online & multichannel NC.

Table 2. Crosstabulation for the link between type of association and news consumption type.

			News consumption (NC) type
			Traditional	Digital & Multichannel	Total
Type of association	No association (n.s.)	Count	89	61	150
		% within association	59.3%	40.7%	100%
		% within NC	58.2%	67.8%	61.7%
	Positive association	Count	38	28	66
		% within association	57.6%	42.4%	100%
		% within NC	24.8%	31.1%	27.2%
	Negative association	Count	26	1	27
		% within association	96.3%	3.7%	100%
		% within NC	17.0%	1.1%	11.1%

Note: N = 243; 0 cells have expected count less than 5; the minimum expected count is 10.00.

Region-based differences between main concepts

For investigating region-based differences (RQ5b), we created a nominal variable differentiating between North America (44.0%, n = 135) and other regions (56.0%, n = 172). Pearson chi-square testing demonstrated a medium-sized significant relationship between region and concept type (χ²(1; n = 288) = 20.41, p < 0.001, V = 0.27). More specifically, general threat and crime (trend) perceptions are more commonly researched outside North America (see Table 3). Pearson chi-square testing also showed a small-sized significant relationship between region and NC type (χ²(1; n = 263) = 4.52, p < 0.05, V = 0.13), revealing that digital/online & multichannel NC is also more commonly researched in other regions than North America (see Table 4).

Table 3. Crosstabulation for the link between region and concept type.

			Concept type
			Fear-related concepts	Threat and crime (trend) perceptions	Total
Region	North America	Count	79	42	121
		% within region	65.3%	34.7%	100%
		% within concept	55.2%	29.2%	42.0%
	Other	Count	64	103	167
		% within region	38.3%	61.7%	100%
		% within concept	44.8%	71.0%	58.0%

Note: N = 288; 0 cells have expected count less than 5; the minimum expected count is 60.08.

Table 4. Crosstabulation for the link between region and news consumption type.

			News consumption (NC) type
			Traditional	Digital & Multichannel	Total
Region	North America	Count	71	31	102
		% within region	69.6%	30.4%	100%
		% within NC	43.8%	30.7%	38.8%
	Other	Count	91	70	161
		% within region	56.5%	43.5%	100%
		% within NC	56.2%	69.3%	61.2%

Note: N = 263; 0 cells have expected count less than 5; the minimum expected count is 39.17.

Discussion

This scoping review investigates the empirical literature about links between people’s NC and TPF-PS since the advent of the smartphone era. Our findings provide several valuable insights. First, when looking at the theoretical state-of-the-art, cultivation theory (Gerbner 1998) and media-related theories with a focus on the individual are the dominant lenses through which NC is empirically approached. Besides these main theoretical approaches, it is also noteworthy that many studies lack any kind of clear theoretical foundation, indicating that exploratory perspectives are co-existing next to theory-driven research (Levine and Markowitz 2023).

While research most commonly applied explorative adaptations of cultivation theory (Gerbner 1998) and framing theory (Entman 1993) as their theoretical background (e.g. Chadee, Smith, and Ferguson 2019; De Coninck et al. 2021), theoretical reconsiderations are visible following the proposition of more deeply involved media users (e.g. Lissitsa, Kushnirovich, and Steinfeld 2023). As a prime example of this trend, Williamson, Fay, and Miles-Johnson (2019) revealed that receiving information about terrorism from multiple media sources increases fear of terrorism, but media sources accessed passively (e.g. TV, radio) are not as influential as media sources accessed more actively (e.g. internet). Indeed, particularly when focusing on digital/online & multichannel NC, studies only found partial support for cultivation theory (e.g. Intravia et al. 2017), suggesting that it may be somewhat outdated and cannot be transferred straightforwardly to contemporary media and NC. However, given that similar issues with lacking empirical support are also occasionally raised for traditional NC, it might also be that cultivation theory in itself is in need of a conceptual update (see Potter 2022).

Furthermore, our results further indicate that researchers’ theoretical choices are not one-dimensional, but instead often contingent upon whatever topic is investigated. For instance, some studies examined NC and TPF-PS with regard to specific actors, such as refugees. Due to this topical focus, these studies were more likely to draw on more concrete theories that are specifically suited for it (e.g. intergroup threat theory or hostile media effects). Additionally, in the absence of prolonged exposures, one-off-exposure experiments typically do not argue with cultivation effects (but rather rely on domain-specific approaches which may have originated in related disciplines (see, e.g. Matthes, Schmuck, and von Sikorski 2019).

Second, our scoping review emphasizes several methodical shortcomings. Most obviously, qualitative methods (and, by that, social-constructivist and interpretative perspectives) constitute a rarity in this research field. However, quantitative methods are not without issues. Here, we identified a dearth of longitudinal and experimental methods, especially those involving behavioral indicators. In fact, long-term studies and experiments only made up a small part of our sampled empirical studies, resulting in a lack of insights into temporality and causality. Proper experimental designs in particular might provide us with much-needed insights into idiosyncratic communication experiences triggered by news articles about public safety topics. On an operational level, it also needs to be noted that most research focuses primarily on people’s awareness about undesirable circumstances (e.g. crime perception) and/or the extent of their emotional response to these circumstances (e.g. fear of crime), thus neglecting behavioral responses (e.g. whether and how NC affects coping and avoidance strategies). Crucially, most of these issues may be best addressed via mixed-method approaches; however, here we have to point toward yet another, perhaps even more profound shortage as our sampled studies were either quantitative or qualitative without attempting to embrace strengths and overcome weaknesses by combining both methodologies.

Third, we systematically determined prevalences across different sample characteristics and sampling methods. Regarding sample populations, communication scholars (similar to other disciplines) have put into question the generalizability of evidence given samples predominantly include people from so-called WEIRD societies (Bates 2021). In a majority of cases, the same can be said to be true for empirical inquiries into people’s NC and TPF-PS where non-probability samples from North America or Northern/Western Europe are predominant. This corresponds with the results of other communication sciences reviews (Erba et al. 2018) and questions the generalizability of available quantitative evidence. Interestingly, many studies control for educational differences among their participants as well as, somewhat exceptionally, various residential indicators (e.g. urban vs. rule, region, place of origin, neighborhood). However, a stronger focus on other socioeconomic factors (e.g. participants’ current employment status or financial situation) might increase adjustment to existing heterogeneity within WEIRD societies.

Fourth, our review provides a comprehensive overview on what operational approaches of TPF-PS and NC are predominant in the literature. For the former, it became clear that previous operationalizations almost equally in numbers represent either people’s varying awareness about undesirable circumstances (i.e. sense-of operationalizations) or their emotional response to these circumstances (i.e. fear-of operationalizations). This finding therefore underlines the assumptions of TPF-PS as an ambiguous construct. Notably, in the case of fear-of operationalizations, fear for (significant) others’ safety is hardly researched, even though other-directed fears may be as important as (or perhaps sometimes even more important) than fear for one’s own safety.

Concerning NC operationalizations, it can be considered alarming that only a small proportion of the empirical studies explicitly considered digital/online & multichannel NC. That is, despite the worldwide advent of smartphones during the past 15 years, digital/online NC and social media NC in particular is hardy researched in relation to TPF-PS. It is further surprising that examinations of online or/and social media NC often makes no reference to end-user devices, meaning that existing research in this field largely neglects specific characteristics of online and social media (Intravia et al. 2017).

Lastly, when breaking down operationalizations into a) fear-related concepts and general threat and crime (trend) perceptions and b) traditional/offline and digital/online & multichannel NC, it appears that significant differences between these main operationalization types are primarily found for negative associations. Moreover, both concepts related to general threat and crime (trend) perceptions as well as digital/online & multichannel NC are understudied in North America (i.e. where the majority of empirical research has been conducted).

Limitations and implications for future research

The findings of this scoping review are subject to several limitations, which come with clear implications for future work. First, each scoping review addresses their topics by selecting specific search terms. Even though this review has indeed defined a large number of word combinations in order to cover available research (see Appendix A), these search terms can never be complete, such that empirical studies that could be relevant may not be included in our sample. A second limitation concerns the use of significance levels instead of effect sizes (following Valkenburg, van Driel, and Beyens 2022) when exploring association types. Third, it needs to be mentioned that we only considered empirical papers written in English and German which might have led us to miss relevant findings written in other languages.

Based on the findings of our scoping review (and the continuing scholarly interest in links between NC and TPF-PS), we also have to discuss implications for future research. From a methodological point of view, future research needs to make better use of mixed-method designs as combining paradigms is the way to maximize validity. Furthermore, as cross-sectional quantitative research does not allow for solid insights into causal mechanisms (see Grosz, Rohrer, and Thoemmes 2020), researchers should be encouraged to conduct longitudinal and experimental studies. In line with other authors (e.g. Intravia et al. 2017), we also argue the need for more specific and nuanced operationalizations, particularly for digital/online and social media NC that take into account idiosyncrasies of different platforms and end devices, that meet current NC practices. For that our knowledge in this research field gets expanded, examining dependent variables related to coping or avoidance behaviors might be a promising focus that is distinct from fear-related and threat perception constructs. For instance, novel insights could be obtained when adverse feelings such as fear that are directly related to a particular crime are investigated alongside individuals’ behavioral actions (e.g. avoiding going out at night).

Our results for association types beg the question whether greater digital/online & multichannel NC can also reduce TPF-PS. In line with cultivation theory, previous research has focused primarily on the extent to which NC increases TPF-PS, that is, negative outcomes. A reconsideration of possible positive outcomes of NC might be promising here to understand how NC relates to TPF-PS in modern media landscapes. Finally, it is clear that future research needs to focus on other regions than the United States and Northern/Western Europe.

Conclusion

This scoping review of empirical research examining how people’s NC is related to their TPF-PS offers a systematic overview of current mainstream trends and, most importantly, diagnoses deficits that may bias the field and impair its societal utility. First and foremost, future research is nothing less than forced to adapt to dynamically changing landscapes in order to stay relevant. Against this background, the findings of the present scoping review suggest a pressing need for more qualitative studies, experiments, and longitudinal studies, as well as a better approach to digital/online & multichannel NC. Only when creating more specific and nuanced operationalizations for contemporary online and social media NC, researchers will be able to understand people’s lived NC routines, so that they can better inform public discourse and political decision-making.

Funding details

Funded by the Austrian Research Promotion Agency (Grant number: FO999895140)

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Excluding Transplanation, Physiology, Physical Geography, Obstretrics Gynecology, Neurosciences, Microbiology, Mathematics, Mathematical Methods in Social Sciences, Materials Science, Literature, Linguistics, History Philosophy Of Science, History, Evolutionary Biology, Development Studies, Arts Humanities Other Topics, Art, Water Resources, Surgery, Religion, Pediatrics, Meterology Atmospheric Sciences, Geology, Geography, General Internal Medicine, Food Science Technology, Energy Fuels, Emergency Medicine, Biodiversity Conservation, Agriculture, Social Work, Research Experimental Medicine, Oncology, Infectious Diseases, Zoology, Transportation, Substance Abuse, Rehabilitation, Immunology, Biomedical Social Sciences, International Relations, Geriatrics Gerontology, Nursing, Medical Informatics, Engineering, Environmental Sciences Ecology.

2 Excluding Chemical Enginerring, Chemistry, Veterinary, Neuroscience, Materials Science, Health Professions. Earth and Planetary Sciences, Immunology and Microbiology, Physics and Astronomy, Biochemistry, Genetics, and Molecular Bilology, Pharmacology, Toxiology and Pharmaceutics, Energy, Agricultural and Biological Sciences, Mathematics, Medicine, Nursing.

Appendix A

Web of Science: in total, 526 results.

488 results if only including studies from 2007 to 2023.

476 results if only including articles, book chapters, early access, and proceeding papers.

459 results if only including articles in English.

315 results if only including relevant research areas.¹

((TI=(‘media use’ OR ‘media consumption’ OR ‘news consumption’ OR ‘media exposure’ OR ‘news exposure’ OR ‘news media’ OR ‘news report*’ OR ‘media platform’ OR ‘social media ‘OR ‘social network*’ OR ‘online media ‘OR ‘internet’ OR ‘mobile phone*’ OR ‘social sharing’)) OR (AB=(‘media use’ OR ‘media consumption ‘OR ‘news consumption’ OR ‘media exposure’ OR ‘news exposure’ OR ‘news media’ OR ‘news report*’ OR ‘media platform’ OR ‘social media ‘OR ‘social network*’ OR ‘online media ‘OR ‘internet’ OR ‘mobile phone*’ OR ‘social sharing’)) OR (AK=(‘media use’ OR ‘media consumption ‘OR ‘news consumption’ OR ‘media exposure’ OR ‘news exposure’ OR ‘news media’ OR ‘news report*’ OR ‘social media ‘OR ‘social network*’ OR ‘online media ‘OR ‘internet’ OR ‘mobile phone*’ OR ‘social sharing’))) AND ((TI=(‘fear of crime’ OR ‘crime perception*’ OR ‘crime trend perception*’ OR ‘fear of violent crime* ‘OR ‘fear of war ‘OR ‘fear of terror’ OR ‘fear of attack*’ OR ‘fear of violence’ OR ‘fear of victimization’ OR ‘reaction* of fear’ OR ‘perceived vulnerability’ OR ‘perceived threat*’ OR ‘sense of security’ OR ‘feelings of security’)) OR (AB=(‘fear of crime’ OR ‘crime perception*’ OR ‘crime trend perception*’ OR ‘fear of violent crime* ‘OR ‘fear of war ‘OR ‘fear of terror’ OR ‘fear of attack*’ OR ‘fear of violence’ OR ‘fear of victimization’ OR ‘reaction* of fear’ OR ‘perceived vulnerability’ OR ‘perceived threat*’ OR ‘sense of security’ OR ‘feelings of security’)) OR (AK=(‘fear of crime’ OR ‘crime perception*’ OR ‘crime trend perception*’ OR ‘fear of violent crime* ‘OR ‘fear of war ‘OR ‘fear of terror’ OR ‘fear of attack*’ OR ‘fear of violence’ OR ‘fear of victimization’ OR ‘reaction* of fear’ OR ‘perceived vulnerability’ OR ‘perceived threat*’ OR ‘sense of security’ OR ‘feelings of security’)))

Scopus: in total, 784 results.

713 results if only including studies from 2007 to 2023.

684 results if only including articles, book chapters, and conference papers.

414 results if only including relevant research areas.²

404 results if only including articles in English.

TITLE-ABS-KEY(‘media consumption’ OR ‘news consumption’ OR ‘media exposure’ OR ‘news exposure’ OR ‘news media’ OR ‘news report*’ OR ‘media use’ OR ‘social media’ OR ‘social network*’ OR ‘online media’ OR ‘internet’ OR ‘mobile phone*’ OR ‘social sharing’) AND TITLE-ABS-KEY(‘fear of crime’ OR ‘crime perception*’ OR ‘crime trend perception*’ OR ‘fear of violent crime*’ OR ‘fear of war’ OR ‘fear of terror’ OR ‘fear of attack*’ OR ‘fear of violence’ OR ‘fear of victimization’ OR ‘reaction* of fear’ OR ‘perceived vulnerability’ OR ‘perceived threat*’ OR ‘sense of security’ OR ‘feelings of security’) AND PUBYEAR < 2024 AND PUBYEAR > 2006 AND PUBYEAR < 2024 AND (LIMIT-TO (DOCTYPE, ‘ar’) OR LIMIT-TO (DOCTYPE, ‘cp’) OR LIMIT-TO (DOCTYPE, ‘ch’)) AND (EXCLUDE (SUBJAREA, ‘MEDI’) OR EXCLUDE (SUBJAREA, ‘NURS’) OR EXCLUDE (SUBJAREA, ‘MATH’) OR EXCLUDE (SUBJAREA, ‘AGRI’) OR EXCLUDE (SUBJAREA, ‘ENER’) OR EXCLUDE (SUBJAREA, ‘PHAR’) OR EXCLUDE (SUBJAREA, ‘BIOC’) OR EXCLUDE (SUBJAREA, ‘PHYS’) OR EXCLUDE (SUBJAREA, ‘IMMU’) OR EXCLUDE (SUBJAREA, ‘EART’) OR EXCLUDE (SUBJAREA, ‘HEAL’) OR EXCLUDE (SUBJAREA, ‘MATE’) OR EXCLUDE (SUBJAREA, ‘NEUR’) OR EXCLUDE (SUBJAREA, ‘VETE’) OR EXCLUDE (SUBJAREA, ‘CHEM’) OR EXCLUDE (SUBJAREA, ‘CENG’)) AND (LIMIT-TO (LANGUAGE, ‘English’)

Appendix B

Table

	Title	Year	Author(s)	Journal/Book
1	(Un)Veiling Our Biases: Activating Religious, Emotional, and Contextual Cues in News Media Representations of Syrian Refugees	2022	Partain, L.P.B. & Weaver, A.J.	International Journal of Communication
2	A Neglected Population: Media Consumption, Perceived Risk, and Fear of Crime Among International Students	2021	Shi, L.	Journal of Interpersonal Violence
3	A Vicious Cycle? Threat of Terror, Perceived Media Bias, and Support for Surveillance Policies	2023	Kaskeleviciute, R. & Matthes, J.	Mass Communication and Society
4	An understudied population? Exploring the factors associated with fear of crime in a semi-rural environment	2022	Lytle, D., Intravia, J., & Randa, R.	Deviant Behavior
5	Association between feeling threatened by a terrorist attack and subjective health: a web survey a week after the attacks of 22 March 2016 in Belgium	2018	Deschepper, R., Six, S., Gidron Y., Depoorter A.-M., Vandekerckhove, M., Gheysens N., Van Overmeire, R., & Bilsen, J.	European Journal of Psychotraumatology
6	Carry-over effect of single media exposure and mass-mediated contact with remote outgroups: From asylum seekers in Europe to an Israeli local outgroup	2023	Lissitsa, S., Kushnirovich, N., & Steinfeld, N.	Journal of Applied Social Psychology
7	Crime in Your Area: Use of Neighborhood Apps Is Associated With Inaccurate Perceptions of Higher Local Crime Rates	2023	Fetterman, A. K., Baker, C. D., & Meier, B. P.	Psychology of Popular Media
8	Crime News Consumption and Fear of Violence: The Role of Traditional Media, Social Media, and Alternative Information Sources	2021	Näsi, M., Tanskanen, M., Kivivuori, J., Haara, P., & Reunanen, E.	Crime & Delinquency
9	Crime, fear and subjective well-being: How victimization and street crime affect fear and life satisfaction	2013	Hanslmaier, M.	European Journal of Criminology
10	Cultivating Support for Punitive Criminal Justice Policies: News Sectors and the Moderating Effects of Audience Characteristics	2017	Simmons, A. D.	Social Forces
11	Different Effects on Different Immigrant Groups: Testing the Media’s Role in Triggering Perceptions of Economic, Cultural, and Security Threats from Immigration	2022	Theorin, N.	International Journal of Communication
12	Direct and indirect effects of crime-related media consumption on public confidence in the police	2020	Choi, J., Yim, H. & Hicks, R.D.	International Journal of Police Science & Management
13	Does Watching the News Affect Fear of Terrorism? The Importance of Media Exposure on Terrorism Fear	2012	Nellis, A.M. & Savage J.	Crime & Delinquency
14	Examining the Relationship Between Media Consumption, Fear of Crime, and Support for Controversial Criminal Justice Policies Using a Nationally Representative Sample	2018	Dolliver, M. J., Kenney, J. L., Williams Reid, L., & Prohaska, A.	Journal of Contemporary Criminal Justice
15	Experiencing violence in a cross-media environment: an interdisciplinary focus group study	2019	Haara, P., Reunanen, E., Kivivuori, J., & Näsi, M.	Nordic Journal of Criminology
16	Exposure to politicized media and prejudice against immigrants in Italy: Identifying its impact and psychological mediators	2016	Vaes, J., Latrofa, M., Vieno, A., & Pastora, M.	Psicologia Sociale
17	Exposure to Terrorism-related News: Associations with Beliefs about Terrorism and Self-protective Action	2021	Wallace, L.N.	Communication Studies
18	Fear of Terrorism and Attitudes Toward Refugees: An Empirical Test of Group Threat Theory	2022	De Coninck, D.	Crime & Delinquency
19	Fear of Terrorism and Fear of Violent Crimes in the United States: A Comparative Analysis	2021	Onat, I., Guler, A., Kula, S., & Bastug, M.F.	Crime & Delinquency
20	Fear of terrorism: media exposure and subjective fear of attack	2019	Williamson, H., Fay, S. & Miles-Johnson, T.	Global Crime
21	Gender and fear of terrorism in Turkey	2009	Wilkox, P., Ozer, M. M., Gunbeyi, M, & Gundogdu, T.	Journal of Contemporary Criminal Justice
22	Impacts of Urban Terror Attacks on Turkish Mothers’ Daily Experiences	2022	Gökyar, M. & Erdur-Baker, Ö.	Journal of Interpersonal Violence
23	Investigating the relationship between social media consumption and fear of crime: A partial analysis of mostly young adults	2017	Intravia, J., Wolff, K.T., Paez, R., & Gibbs, B.R.	Computers in Human Behavior
24	Media and Crime Perceptions: Evidence from Mexico	2021	Ramírez-Álvarez, A. A.	The Journal of Law, Economics, and Organization
25	Media consumption and crime trend perceptions: a longitudinal analysis	2019	Shi, L., Roche, S. P., & McKenna, R. M.	Deviant Behavior
26	Media Consumption and Fear of Crime in a Large Chinese City	2019	Wu, J., Li, F., Triplett, R.A., & Sun, I.Y.	Social Science Quarterly
27	Media Consumption and Fear of Crime: Evidence of the Need for an Intersectional Approach	2021	Rosenberger, J.S., Dierenfeldt, R., & Ingle, H.	Victims & Offenders
28	Media consumption, perceptions of crime risk and fear of crime: Examining race/ethnic differences	2012	Callanan, V.	Sociological Perspectives
29	Media Exposure Predicts Children’s Reactions to Crime and Terrorism	2008	Becker-Blease, K. A., Finkelhor, D., & Turner, H.	Journal of Trauma & Dissociation
30	Media Use and Affective Political Polarization: What Shapes Public Perceptions of Immigrants’ Deservingness?	2013	Willnat, L., Ogan, C., & Shi, J.	Journal of Broadcasting & Electronic Media
31	Media, Fear of Crime, and Authority: Focus Groups in Lima (Perú)	2022	Kanashiro, L., Dammert, L., & Hernández, W.	Smart Innovation, Systems and Technologies book series
32	Media, Gender, and Fear of Crime	2015	Callanan, V., & Rosenberger, J. S.	Criminal Justice Review
33	Mediatized Realities of Migrants in a Comparative Perspective: Media Use, Deservingness, and Threat Perceptions in the United States and Western Europe	2021	De Coninck, D., Ogan, C., Willnat, L., & D’Haenens, L.	International Journal of Communication
34	Mediators of the association between television viewing and fear of crime: Perceived personal risk and perceived ability to cope	2011	Custers, K., & Van den Bulck, J	Poetics
35	Murder She Watched: Does Watching News or Fictional Media Cultivate Fear of Crime?	2017	Chadee, D., Smith, S., & Ferguson, C. J.	Psychology of Popular Media Culture
36	News media and crime perceptions: Evidence from a natural experiment	2018	Mastrorocco, N. & Minale, L.	Journal of Public Economics
37	Perceived development of crime, need for punishment and fear of crime: What influence does media consumption have in Switzerland?	2016	Philipp, C., Koch, K., Borchard, B., Regli, D. & Znoj, H.	Kriminalistik
38	Perceived Threat, Immigration Policy Support, and Media Coverage: Hostile Media and Presumed Influence	2013	Watson, B.R. & Riffe, D.	International Journal of Public Opinion Research
39	Right-Wing Authoritarianism and Social Dominance Orientation as Mediators Between News Media Consumption and Perceived Migrant Threat	2022	De Coninck, D., van Assche, J., & D’Haenens, L.	Mass Communication and Society
40	See something, say something? The role of online self-disclosure on fear of terror among young social media users	2023	Kaskeleviciute, R., Knupfer, H., & Matthes, J.	News Media & Society
41	Supporting the Iron Fist: Crime News, Public Opinion, and Authoritarian Crime Control in Guatemala	2014	Krause, K.	Latin American Politics and Society
42	Terror, Terror Everywhere? How Terrorism News Shape Support for Anti-Muslim Policies as a Function of Perceived Threat Severity and Controllability	2019	Matthes, J., Schmuck, D., & von Sikorski, C.	Political Psychology
43	The Impact of Media on Fear of Crime among University Students: A Cross-National Comparison	2012	Kohm, S.A., Waid-Lindberg, C.A., Weinrath, M. Shelley, T.O., & Dobbs, R.R.	Canadian Journal of Criminology and Criminal Justice
44	The Islamic State in the News: Journalistic Differentiation of Islamist Terrorism from Islam, Terror News Proximity, and Islamophobic Attitudes	2021	von Sikorski, C., Matthes, J., & Schmuck, D.	Communication Research
45	The relationship between media portrayals and crime: perceptions of fear of crime among citizens	2017	Hollis, M.E., Downey, S., Carmen, A., & Dobbs, R.R.	Crime Prevention and Community Safety
46	The Role of the Media in the Fear of Crime: A Qualitative Study in the Portuguese Context	2022	Silva, C. & Guedes, I.	Criminal Justice Review
47	Trial by Media?: Media Use, Fear of Crime, and Attitudes Toward Police	2022	Shin, S. Y. & Watson, B. R.	Journalism Practice
48	Values and Fear of Crime	2009	Russo, S. & Roccato, M.	Social Psychology: New Research
49	Violence Trending: How Socially Transmitted Content of Police Misconduct Impacts Reactions toward Police Among American Youth	2021	Novich, M., & Zduniak, A.	The Emerald International Handbook of Technology-Facilitated Violence and Abuse

Appendix C

Table

1	Coders
2	Title
3	Author(s)
4	Year of publication
5	Name of journal/ book section
6a	Main theory	0 = no specific theory (the study does not include theory-related keywords, e.g. cultivation, frames) 1 = cultivation theory 2 = news value theory 3 = media priming/framing 4 = information processing theories (ELM, HSM) 5 = uses and gratifications theory 6 = other theories (with focus on the individual that are somehow related to media, e.g. social comparison, interpersonal diffusion)
6b	Name of Main Theory
7	Methodology	1 = Quantitative 2 = Qualitative 3 = Mixed method
8	Method	1 = Cross-sectional survey 2 = Longitudinal survey 3 = Experiment 4 = Eye-tracking 5 = Focus group interviews 6 = Individual interviews 7 = Observation 8 = Mixed methods (two or more different methods are applied)
9	Setting	77 = no exact information about setting 1 = online (internet-based) 2 = offline (everything else not online, e.g. at schools, but also telephone interviews) 3 = mixed (combined use of online and offline setting)
10	N Indicate the exact number of included human participants
11	Sampling	77 = No exact information about sampling 1 = Probability sampling (simple random, systematic, stratified random, cluster) 2 = Non-probability sampling (quota, purposive, snowball, self-selection, convenience)
12	Region of Sample	77 = no exact information about sample region 1 = North America 2 = Central America 3 = South America 4 = Northern/Western Europe 5 = South/Eastern Europe 6 = Northern Asia 7 = Central Asia 8 = Western Asia 9 = Southern Asia 10 = Eastern/Southeastern Asia 11 = Africa 12 = Oceania 13 = Multinational
13	Studied Populations	77 = no exact information about age 1 = refers to adult population (starting with 18 years) 2 = refers only to children and teenager populations (until 17 years) 3 = refers to student sample (colleague students) 4 = refers to both adult and younger population
14	Gender Distribution	77 = no exact information about gender distribution 1 = balanced sample (50/50 or maximum 40/60) 2 = unbalanced sample (more female) 3 = unbalanced sample (more male)
15	Socioeconomic Status (SES) Multiple answers possible	77 = reports no socioeconomic factors 1 = education 2 = financial situation (e.g. household income, financial difficulties) 3 = employment status (e.g. not vs. employed, full-time vs. part-time, professional occupancy) 4 = based on area of residence (e.g. urban vs. rule, region, place of origin, neighborhood; also including residential postal codes which are used either for sampling or analysis)
16a	Variables of Interest: DVs/Mediators	1 = fear of (violent) crimes 2 = crime (trend) perceptions 3 = fear of terror/terrorism 4 = fear of victimization/perceived victimization risk (the concept must be explicitly named as such!) 5 = other concepts (fear of threats also against particular groups)
16b	Names of DVs/Mediators
17a	Variables of Interest: IVs/Moderators	1 = media consumption/exposure in general (concerning news vs. not news) 2 = internet use in general 3 = offline/traditional news consumption/exposure 4 = online/digital news consumption/exposure (other than social media -> category 5) 5 = social media use/news consumption on social media 6 = multichannel news consumption/news media use (mix of offline and online news exposure/use) also including unclear descriptions 7 = news referring to a specific topic
17b	Names of IV/Moderators
18	End-User Device	77 = no focus on a specific device / also including unclear descriptions 1 = smartphones 2 = tablet, laptop, PC 3 = television 4 = radio 5 = print (newspaper, magazines) 6 = mixed (if at least one end-user device is identifiable, but the other devices are not; if all end-user devices are not clear -> 77)
19	Investigated Social Media Networks Multiple answers possible	77 = no focus on social media networks (e.g. only media consumption in general; traditional media; internet use in general) 1 = social media in general with no focus on specific networks or with focus on several networks together 2 = Twitter 3 = Facebook 4 = Instagram 5 = WhatsApp 6 = TikTok 7 = Reddit 8 = Snapchat 9 = YouTube 10 = WeChat 11 = QQ 12 = Tumblr 13 = other social media networks
20	Ethnicity	0 = no 1 = yes (including main effects, interaction/conditional effects)
21	Main Findings of Quantitative Studies	77 = qualitative results 0 = no significant relationship reported 1 = significant positive relationship (not including marginal) 2 = significant negative relationship (not including marginal) 3 = report of mixed results (even if there are multiple calculations/models for the same variable)
22	Main Findings of Qualitative Studies Short report of main findings (based on result chapter)