Determining or Mediating Media? Model Comparisons on the Role of Media and Risk in Shaping Preventive Behaviors During Crises

Abstract

This study investigates the role of media in risk communication during crises by comparing two theoretical perspectives. The first paradigm views media as a determining factor, while the second sees media as a mediator. The study analyzes two survey data sets: cross-sectional data on nuclear waste issues from eight regions (Mainland China, Hong Kong, Taiwan, Singapore, Japan, South Korea, the United States, and the United Kingdom; N = 8,591) and longitudinal panel data on the COVID-19 pandemic in Hong Kong (N = 920). Results consistently support the the greater explanatory power of the first paradigm. Individuals’ willingness to adopt preventive measures highly depends on the media-constructed risks.

Keywords

media effects risk perception cross-culture crisis communication nuclear waste COVID-19

The role of media during times of crisis or when facing risks has garnered significant attention (Hill, 2001). Media have been recognized for providing platforms that raise the awareness of these risks and promote the adoption of preventive measures (Bekalu et al., 2019; R. Liu et al., 2022). Social scientists have long been attentive to the role of media in influencing behavioral change among individuals. Within academia, two different viewpoints exist on this topic. The first paradigm highlights the determining role of media in shaping people’s attitudes and beliefs, as reflected in major theories throughout the history of communication development, such as the “Magic Bullet” theory, the cultivation theory, and the framing theory (Entman, 1993; Gerbner et al., 1980). Following the emergence of film, sound recording, and broadcasting technology in the late 19th century, communication studies primarily focused on assessing the significant influence of media content on individuals’ attitudes and behaviors (Drew, 2016).

The other paradigm, distinct from the determining role in the previous perspective, emphasizes the mediating role of media in shaping individuals’ behaviors. Many theories in media studies and information studies adopt this viewpoint, such as the uses and gratifications theory and the risk information seeking and processing (RISP) model (Griffin et al., 1999; Katz et al., 1973), suggesting that risk perceptions are mediated by media to subsequently influence public behaviors. That is, risk perceptions cultivate people’s needs for media messages. Individuals first perceive risks and then acquire or seek messages from media platforms. Unlike the determining role attributed in the first viewpoint, media are recognized as mediators in enhancing individuals’ intention to take actions. In this perspective, individuals actively shape their own perceptions of risks, which we refer to as “existing risks.” Conversely, in the first paradigm, individuals passively perceive risks constructed by the media, which we term “constructing risks.”

This study aims to test the two viewpoints by comparing the fundamental rationales considering the interplay of media and risk on individual behaviors. It seeks to empirically test the two fundamental models in the context of two public crises (i.e., the nuclear waste issue and the COVID pandemic) for cross-validation, utilizing cross-sectional survey data from eight regions across Eastern and Western cultures, along with longitudinal survey data from Chinese culture.

This study selected two significant crisis issues for testing, given their widespread influence on the public and the substantial health risks individuals may face during these crises, as well as the considerable impact media have on shaping people’s risk perceptions and subsequent behaviors in both issues. The essential information regarding the nuclear waste issue is that on August 22, 2023, Japan announced it would begin releasing treated radioactive water from the Fukushima nuclear plant into the Pacific Ocean within 48 hours, despite opposition from neighboring regions, such as South Korea and China (Wong, 2023). People, particularly those in regions neighboring Japan, perceived high levels of risk due to this discharge of nuclear waste.

The other issue, the COVID-19 pandemic, represents a prolonged global public health crisis that has resulted in over 7 million deaths as of September 2024 (World Health Organization [WHO], 2024). During this health crisis, beliefs among people in Asia, particularly in Chinese societies, exhibit certain distinctions compared with those in the West. Research has shown that in Western countries, there is a notable polarization in perceptions and misperceptions of COVID-19, driven by varying levels of trust in different information sources (Armstrong & VanDyke, 2025; Pennycook et al., 2022). In contrast, perceptions among Chinese people regarding COVID-19 and related preventive behaviors tend to be more uniform, largely influenced by their exposure to information from various media platforms (Huang et al., 2022; R. Liu et al., 2022).

Thus, this study will adopt these two crises as contexts to test the two models and to address several key questions: (a) What is the role of media in risk communication—determining or mediating role? (b) How do individuals perceive risks—constructing or existing? (c) What is the relationship between media attention and individuals’ risk perceptions? (d) Which model is more effective in predicting individuals’ behavioral outcomes in crisis contexts? The findings of this empirical study will enhance the theoretical understanding of the role of media in risk communication and offer practical insights for crisis management, particularly concerning the utilization of media platforms.

Literature Review

This section will begin by reviewing existing theories on the direct effects of media and risk perceptions on health outcomes, respectively, outlining corresponding hypotheses. We will then review two distinct paradigms for analyzing the interplay between media and risks on individual health outcomes. This review will lead to the formulation of research questions regarding the comparative utility of the two models.

The Role of Media in Health Behaviors

Social scientists have long been attentive to the role of media in studies on media effects. Following the advent of film, sound recording, and broadcasting technology in the late 19th century, the primary focus of communication studies was on evaluating the influence of media content on individuals’ attitudes and behaviors (Geiger & Newhagen, 1993). During the 1930s, the Payne Fund studies investigated how movies impacted children’s emotions, behaviors, and ethical principles (Lowery & Defleur, 1995). Another notable study, conducted by Lazarsfeld et al. (1968), delved into the relationship between radio and newspaper consumption and voting behavior, adding to the body of research in this field.

Specifically, the impact of media on people’s health-related behaviors has been highly emphasized (Dai et al., 2022; Moorhead et al., 2013; Wang et al., 2023). Existing research has shown that attention to media messages can positively influence individuals’ motivation to pursue beneficial health outcomes (Balatsoukas et al., 2015) as well as avoid undesirable health consequences (Santoro et al., 2015). Particularly, when individuals are confronted with specific diseases, media have emerged as a significant channel for disseminating disease-related information to motivate them to take certain preventive measures, such as those related to cancer prevention (Sarkar et al., 2018).

During the COVID-19 pandemic, social scientists revealed the significant impact of media on preventive behaviors (Fenta et al., 2024; Huang, Liu, & Sun, 2025; Huang, Liu, & Zhang, 2025). An empirical study focusing on COVID-19 has shown that social media serve as critical information channels that can potentially exert a positive influence on people’s preventive health behaviors (P. L. Liu, 2021). Furthermore, a cross-sectional survey of 1,536 adults found that greater attention to COVID-19 news in the media was positively associated with people’s adherence to preventive measures (Melki et al., 2022). Therefore, in light of previous studies, we first propose the direct impacts of media attention on people’s preventive behaviors:

H1: Individuals’ media attention is associated with their preventive behaviors.

The Role of Two Types of Risk Perceptions in Health Behaviors

Besides the direct effects of media, risk perception has long been emphasized to influence health behaviors. The dual process model of risk judgment outlines two distinct modes of processing that are fundamental to how people comprehend risk—the affective and cognitive processes, which are governed by independent systems that can influence individual responses to health risks (Chae & Lee, 2019; Huang et al., 2021). These parallel and interacting modes of information processing are believed to work together to guide reasoning and decision-making (Loewenstein et al., 2001).

The cognition-based process posits that individuals rationally assess the cost and benefits of an issue to shape their attitude toward it (Huang, Cai, et al., 2025; Zanna & Rempel, 2008). This process is rational, deliberate, and analytical, but it is also slow and effortful, as it responds to normative rules of probability and logic. In contrast, the affective risk response addresses the emotional changes that occur in reaction to certain stimuli (Niedenthal et al., 1999). This process is automatic, experiential, and intuitive, as well as fast and efficient (Slovic & Peters, 2006).

Both types of risk perceptions may independently influence individuals’ health behaviors. Research has indicated that both affective and cognitive risk perceptions can have positive effects on preventive behaviors. For instance, Ferrer et al. (2013) suggested that worry, a common form of affective risk response, along with cognitive risk perceptions, was associated with increased health-protective behaviors. Similarly, during the COVID-19 pandemic, a study conducted in China found that both cognitive and affective factors were crucial predictors of behavioral responses to the pandemic. Interestingly, the affective factors demonstrated stronger predictive power than the cognitive factors in anticipating people’s responses to COVID-19 (Shen et al., 2021). Therefore, this study will propose the direct effects of the dual processes of risks on people’s preventive behaviors:

H2: Individuals’ (a) affective risk perceptions and (b) cognitive risk perceptions are associated with their preventive behaviors.

Model Comparison—Determining Versus Mediating

Previous research draws attention to the role of both media and risk perceptions in shaping individuals’ ultimate behaviors. Within academia, two distinct paradigms exist—one where media plays a determining role, suggesting that media influences how people perceive risks, and another where media serves a mediating role, indicating that risk perceptions cultivate people’s needs for media messages.

Determining Media Role

This paradigm highlights the determining role of media in shaping people’s attitudes and behaviors. This perspective has been significant since the early days of communication studies. Research on media effects began in the 1920s and 1930s, and gained considerable attention by the late 1950s with the advent of television (Katz & Lazarsfeld, 1955). These developments resulted in a surge of research concerning media effects (Valkenburg et al., 2016). By the 1980s, thousands of empirical studies had been conducted, exploring the cognitive, emotional, attitudinal, and behavioral impacts of media on individuals (Potter & Riddle, 2007).

At the same time, there has been a significant surge in the theoretical development of this topic. Notably, the “Magic Bullet” theory illustrates media messages as a “bullet” fired from a “media gun” that directly penetrates the audience’s mind, emphasizing the assumption of media’s powerful and immediate influence (Berger, 1995). Similar theories, such as cultivation theory and framing theory, also stress the determining impacts of media content on human attitudes and behaviors. Cultivation theory (Gerbner et al., 1980), developed alongside the widespread use of television, suggests that the more time individuals spend immersed in the television world, the more likely they are to accept the social reality depicted on screen. Likewise, framing theory (Entman, 1993) explains how media highlight specific topics and contextualize them within a framework of meaning (i.e., the process of framing), which subsequently affects audience perceptions.

As noted by Valkenburg et al. (2016), these early theories focus on the unidirectional impact of media on people’s behaviors. Indeed, within this paradigm, the determining role of media in shaping human perceptions is emphasized, suggesting that media exert a direct influence on individuals. In this viewpoint, media construct reality for the audience, ultimately shaping their attitudes and behaviors. Consequently, individuals are seen as passive consumers who rely heavily on media for information to perceive reality.

Mediating Media Role

Unlike the determining role of media emphasized in the first paradigm, this paradigm suggests that media play a mediating role in shaping individuals’ attitudes and behaviors. It includes theories, such as the uses and gratifications theory (Katz et al., 1973) and various theoretical frameworks in information behavior studies. These theories argue that individuals do not engage with media randomly; instead, they focus on specific messages that align with their needs or beliefs (Katz & Lazarsfeld, 1955). Specifically, the uses and gratifications theory posits that people actively seek out particular media to fulfill specific needs (Katz et al., 1973). This theory emphasizes need-based media usage, highlighting how social and psychological factors can influence individuals’ intentions to seek information (Valkenburg et al., 2016).

Similarly, many theoretical frameworks analyzing information behaviors underscore the crucial role of individuals’ beliefs and perceptions, particularly risk perceptions, in their intention to obtain information from media. The RISP model and the comprehensive model of information seeking (i.e., CMIS) exemplify this mechanism. The RISP (Griffin et al., 1999) identifies affective responses, such as fear, and perceived hazard characteristics, such as risk perceptions, as significant factors influencing individuals’ information behaviors. It implies that individuals first form an initial perception of an issue, leading to a subsequent need for related information, which prompts media engagement. In the case of the CMIS (Johnson & Meischke, 1993), it incorporates the ideas from Katz et al.’s (1973) uses and gratifications theory, and suggests that perceived threats—termed salience in the model—affect people’s intentions to seek information from specific sources. Thus, these theories share the rationale that individuals form their perceptions first and then turn to media platforms for messages, acknowledging the mediating role of media in risk communication.

In contrast to the first paradigm, which attributes a more dominant role to media, this perspective recognizes media as a crucial intermediary for disseminating risk information to the public, especially in crisis situations. In this paradigm, risks are already established, and media serve as one of the channels for disseminating these existing risks. Moreover, from this viewpoint, individuals take a more active role in forming their own perceptions of risks, while in the first paradigm, they are seen as passive recipients of risks constructed by the media.

Therefore, given the distinct perspectives on the relationship between media and risks in the context of risk communication, this study examines the comparative utility of these two paradigms—the determining media and the mediating media (Figure 1). Accordingly, the following research questions are raised:

RQ1: What is the relationship between media attention and risk perception?

RQ2: Which is the most robust among the following two models: (A) media attention → affective and cognitive risk perceptions → preventive behaviors (Model 1) and (B) affective and cognitive risk perceptions → media attention → preventive behaviors (Model 2)?

Method

Participants and Procedure

To examine the proposed hypotheses and empirically compare the two theoretical models, two separate surveys were conducted, including a cross-sectional survey on the nuclear waste issue from eight regions (i.e., Mainland China, Hong Kong, Taiwan, Singapore, Japan, South Korea, the United States, and the United Kingdom; N = 8,591) and a three-wave longitudinal panel survey data on the COVID pandemic in Hong Kong (N = 920). The societies selected for the nuclear issue included Asia, particularly those affected by Japan, due to the widespread influence of this issue on the population. The United States and the United Kingdom were included for comparison with cultural variations and to validate findings across different cultures. In addition, the survey conducted during the COVID pandemic in Hong Kong aimed to examine potential causality with several time lags and to cross-validate the proposed framework with different issues.

To recruit participants, three independent agencies were engaged. Participants from Mainland China, Singapore, Japan, South Korea, the United States, and the United Kingdom were sourced by Dynata Hong Kong Limited. Participants from Hong Kong were recruited through Rakuten, an online survey agency overseeing a panel exceeding 50,000 individuals from the region. Furthermore, the Chungliu Education Foundation was selected to recruit participants in Taiwan, as it had the capability to reach a more diverse range of areas in the region, including rural areas, thereby enhancing the representativeness of the sample.

For both surveys in all eight regions, we adopted a combination of probability proportional to size (PPS) and quota sampling techniques. These methods were calibrated against demographic factors, such as gender, age, and residence, using population census data in each respective region as a reference. For instance, in Mainland China, the reference used was the 2010 national population census. The rationale for this integrated approach is due to two reasons. Primarily, quota sampling helps facilitate intentional engagement with specific subpopulations under structured oversight. Also, the combination of PPS and quota sampling ensures the sample’s representativeness, especially with respect to established demographic benchmarks (Wartberg et al., 2015).

Pretest

After receiving ethical approval for engaging human subjects in the study, we recruited 10 student participants in each region to complete pretests of the survey. These participants filled out the questionnaire and provided feedback and comments on it. We carefully considered their input and made necessary revisions to the wording of the questionnaire to improve the validity of the research instrument.

Main Study

Data set 1: Nuclear Issue

Participants were invited to complete a web-based questionnaire using the Qualtrics platform for Taiwan and the QuestionPro platform for the other seven regions. Surveys were carried out within specific timeframes with a total of 8,591 valid responses collected (Mainland China: November 6–December 20, 2023, N = 1,011; Hong Kong: October 30–November 30, 2023, N = 1,418; Taiwan: November 21–November 22, 2023, N = 1,087; Singapore: November 6–December 22, 2023, N = 1,032; Japan: November 10–November 24, 2023, N = 1,020; South Korea: November 10–November 30, 2023, N = 1,009; the United States: November 6–December 22, 2023, N = 1,006; the United Kingdom: November 6–December 22, 2023, N = 1,010).

Data set 2: COVID Issue

Given that the nature of cross-sectional data presents certain limitations in examining the causality of the study variables, we have added longitudinal survey data to cross-validate the results and enhance the rigor of our findings. A three-wave longitudinal panel survey was conducted in Hong Kong between November 2021 and September 2022 (N = 940) using the Qualtrics platform for data collection. Data were collected at several-month intervals, with each wave’s data meticulously cleaned to ensure accuracy by excluding participants who provided inconsistent information regarding their age, gender, and education levels across waves. In the first wave (November 2021), 1,501 participants submitted valid responses. For the second wave (March–April 2022), participants from the first wave were recontacted, achieving a response rate of 76.8% (N = 1,153). In the third wave, 79.8% of participants from the previous two waves completed the questionnaires, resulting in a final data set of 920 participants from Hong Kong (August–September 2022).

Data quality was rigorously maintained in both surveys through strict quality checks. This included excluding participants who failed attention checks or completed the survey in less than a third of the median response time. Participants in both surveys had a balanced gender distribution. The distribution of age, gender, education level and household income in eight regions is shown in Table 1.

Table 1.

The Demographics of Two Data sets.

Demographics	Data set 1: nuclear								Data set 2: COVID
Demographics	Mainland China (N = 1,010)	Hong Kong (N = 1,418)	Taiwan (N = 1,087)	Singapore (N = 1,032)	Japan (N = 1,020)	South Korea (N = 1,009)	United States (N = 1,006)	United Kingdom (N = 1,010)	Hong Kong (N = 920)
Age (40 or above) (%)	58.4	65.9	58.9	66.9	76.1	68.3	64.7	67.5	65.9
Gender (male) (%)	53.4	44.4	52	50.6	48.2	50.2	49.5	49.9	45.4
Education (bachelor or above) (%)	61.5	56	85.6	63.1	64.7	83	63.7	53.1	54.3
Income (medium)	RMB 15,001–25,000 (US$2,100–3,500)	HKD 40,001–50,000 (US$$5,200–6,500)	NTD 60,001–90,000 (US$1,920–2,880 USD)	SGD 8,000–8,999 (US$5,920–6,659)	JPY 300,001–400,000 (US$2,100–2,800)	KRW 501–600 (US$3,807–4,560)	US$6,000-6,999	GBP 2,999–3,999 (US$3,808–5,078)	HKD 40,001-50,000 (US$5,200–6,500)

Measures

Media attention was measured by the tendency of individuals to pay attention to nuclear waste and COVID information through various media platforms. The media included television, newspapers, radio, websites (e.g., Sina website), media-sharing websites (e.g., Sina Weibo), social network sites (e.g., WeChat), and short video platforms (e.g., Douyin) in Mainland China. For all other regions, the media included television, newspapers, radio, websites (e.g., Yahoo, CNN, BBC), social media (e.g., Facebook, WhatsApp), and short video platforms (e.g., TikTok). Participants were asked to rate on a 7-point Likert-type scale (1 = very little to 7 = very much). Particularly, for data set 2, COVID issue, this variable was measured in two waves (Waves 1 and 2). Our measurement of media attention aggregates various types of media, including traditional and social media platforms, into a single index. While it is acknowledged that different media channels may exert distinct effects on risk perception and preventive behaviors, this study emphasizes general media attention to the two issues rather than the heterogeneity of media effects. The descriptive statistics of the variables in all eight regions and both issues are shown in Tables 2 and 3.

Table 2.

The Descriptive Statistics of the Study Variables in Data set 1 (Nuclear Issue).

	Mainland China (N = 1,010)		Hong Kong (N = 1,418)		Taiwan (N = 1,087)		Singapore (N = 1,032)		Japan (N = 1,020)		South Korea (N = 1,009)		United States (N = 1,006)		United Kingdom (N = 1,010)
Variables	M (SD)	α	M (SD)	α	M (SD)	α	M (SD)	α	M (SD)	α	M (SD)	α	M (SD	α	M (SD)	α
Media attention	3.74 (1.08)	.80	3.24 (1.08)	.80	3.21 (.9)	.74	3.82 (1.35)	.88	2.72 (1.15)	.84	3.38 (1.23)	.84	3.76 (1.82)	.94	3.21 (1.49)	.90
Affective risk perception	5.46 (1.38)	.95	4.37 (1.64)	.97	4.48 (1.39)	.96	4.66 (1.39)	.95	3.50 (1.49)	.96	5.68 (1.33)	.95	4.27 (1.50)	.93	4.27 (1.30)	.92
Cognitive risk perception	4.57 (.95)	.90	4.42 (1.42)	.94	4.73 (1.27)	.93	4.71 (1.27)	.91	3.40 (1.48)	.94	5.74 (1.28)	.96	4.22 (1.56)	.89	3.79 (1.43)	.89
Preventive behavior	4.81 (1.30)	.86	3.14 (1.42)	.91	3.36 (1.33)	.90	3.66 (1.52)	.92	1.69 (1.02)	.93	4.63 (1.45)	.89	3.76 (1.78)	.93	3.37 (1.73)	.92

Table 3.

The Descriptive Statistics of the Study Variables in Data set 2 (COVID Issue).

	Wave 1		Wave 2		Wave 3
Variables	M (SD)	α	M (SD)	α	M (SD)	α
Media attention	3.65 (.94)	.79	3.70 (.95)	.77	–	–
Affective risk perception	4.33 (1.16)	.91	4.63 (1.19)	.91	–	–
Cognitive risk perception	4.24 (.98)	.81	4.90 (.93)	.74	–	–
Preventive behavior	–	–	–	–	5.46 (.75)	.75

Affective risk perception was measured using the positive and negative affect schedule (PANAS) scale developed by Watson et al. (1988). Participants responded to seven items on a 7-point Likert-type scale (1 = strongly disagree and 7 = strongly agree) regarding how they felt about the risks caused by the nuclear waste issue and COVID. The affect measured included fear, sadness, helplessness, worry, dissatisfaction, anger, and disgust. For the COVID issue, this variable was measured in two waves (Waves 1 and 2, see Tables 2 and 3).

Cognitive risk perception was measured using four items on a 7-point Likert-type scale (1 = strongly disagree and 7 = strongly agree) adapted from Champion et al. (2008). Participants rated their rational assessment of the risk from the two issues. For the nuclear waste issue, the question items included: “The impact on my daily life is pervasive,” “The impact on my health is substantial,” “The probability of contracting diseases has increased,” “The probability of contracting diseases from consuming contaminated food has increased,” and “I have no confidence in avoiding the negative impacts of this risk.” For the COVID issue, the questions included: “The virus is almost everywhere,” “It is easy to get infected,” “The chance of getting infected is high,” “I am not confident in avoiding the negative impacts of this risk,” and “I cannot avoid this risk.” It was measured in both waves (Waves 1 and 2, see Tables 2 and 3).

Preventive behaviors were measured using seven items on a 7-point Likert-type scale (1 = strongly disagree to 7 = strongly agree). Participants rated their tendency to engage in preventive actions for both issues. For the nuclear waste issue, the question items included: “Avoiding consuming food products from the Fukushima region,” “Avoiding consuming food products from Japan,” “Avoiding purchasing and using household/lifestyle products from Japan,” “Avoiding buying and using any products from Japan,” “Reducing the number of trips to Japan for travel,” “Stocking up on seafood products,” and “Stocking up on table salt.” For the COVID issue, participants rated their inclination to engage in COVID-19 preventive actions, including wearing masks, washing hands, maintaining social distance, and getting vaccinated.

Control variables included the participants’ gender, age, educational level, and monthly household income.

Results

First, we compared the study variables in both data sets to provide an overview of the nuclear waste and COVID issue across regions. For the nuclear waste issue, Tables 2, 3,and Figure 2 show that except for Japan where the issue originated, the other regions exhibited relatively high levels of both affective and cognitive risk perceptions. In addition, participants from these regions reported taking preventive actions to avoid potential health risks associated with the issue. Among the seven non-Japan regions, people from South Korea and Mainland China demonstrated the highest levels of risk perceptions and the strongest intentions to take preventive measures. In contrast, the United States and the United Kingdom showed the lowest risk perceptions and intentions.

For the COVID issue, Table 3 indicates that participants in Hong Kong demonstrated relatively high levels of risk perceptions, both affectively and cognitively. They also tended to engage in preventive behaviors to protect themselves from the disease. It is also noteworthy that both media attention and risk perceptions showed a slight increase from the first wave to the second. This may be due to the growing severity of the pandemic between the times when the first and second waves of the survey were conducted.

We then conducted correlation analyses of the study variables with SPSS for both databases. The results demonstrated that all the study variables, including media attention, affective risk perception, cognitive risk perception, and preventive behavior, were significantly correlated.

Next, we used hierarchical regression analysis in SPSS to test our first two hypotheses in both data sets. The results in Table 4 indicate that controlling for demographics, people’s media attention to the nuclear waste issue was significantly correlated with their preventive behaviors in all eight regions (Mainland China: β = .09, p < .01; Hong Kong: β = .17, p < .001; Taiwan: β = .23, p < .001; Singapore: β = .19, p < .001; Japan: β = .20, p < .001; South Korea: β = .05, p < .05; the United States: β = .25, p < .001; and the United Kingdom: β = .11, p < .001). Likewise, people’s media attention to the COVID issue was significantly correlated with their preventive actions (β = .08, p < .01). Thus, H1 was fully supported. Respondents who paid more attention to the issue via media were more likely to take preventive actions to avoid potential health risks.

Table 4.

Hierarchical Regression Analysis in Two Data sets.

	Data set 1: nuclear								Data set 2: COVID
Study variables	Mainland China	Hong Kong	Taiwan	Singapore	Japan	South Korea	United States	United Kingdom	Hong Kong
Control variables
Gender	–.10**	–.10***	–.06	–.02	–.07*	–.10**	–.04	–.14***	–.12***
Age	–.09*	.21***	.09**	–.10**	–.10**	–.08*	–.33***	–.20***	.18***
Education level	.06	.00	.01	.07*	.06*	.00	.38	–.05	–.01
Household income	.01	–.03	–.11***	–.01	.03	–.01	.06	–.02	.10
Independent variables
Media attention	.09**	.17***	.23***	.19***	.20***	.05*	.25***	.11***	.08**
Affective risk perception	.09**	.41***	.28***	.21***	.16***	.23***	.22***	.21***	.14***
Cognitive risk perception	.40***	.26***	.29***	.40***	.40***	.46***	.32***	.29**	.17***
N	1,010	1,418	1,087	1,032	1,020	1,009	1,006	1,010	920
R ²	.27	.53	.39	.44	.34	.45	.41	.29	.14

p < .05. **p < .01. ***p < .001.

Considering the second hypothesis, the results in Table 4 indicate that individuals’ affective risk perceptions to the nuclear waste issue were significantly correlated with their preventive behaviors in all eight regions (Mainland China: β = .09, p < .01; Hong Kong: β = .41, p < .001; Taiwan: β = .28, p < .001; Singapore: β = .21, p < .001; Japan: β = .16, p < .001; South Korea: β = .23, p < .001; the United States: β = .22, p < .001; and the United Kingdom: β = .21, p < .001). Likewise, this significant correlation was also found between cognitive risk perceptions and preventive behaviors in all regions (Mainland China: β = .40, p < .001; Hong Kong: β = .26, p < .001; Taiwan: β = .29, p < .001; Singapore: β = .40, p < .001; Japan: β = .40, p < .001; South Korea: β = .46, p < .001; the United States: β = .32, p < .001; and the United Kingdom: β = .29, p < .001). In addition, the significant correlations between risk perceptions and preventive behaviors were also found for the COVID issue (affective: β = .14, p < .001; cognitive: β = .17, p < .001). Thus, both H2a and H2b were supported. For both issues, people who perceived a higher level of risks, both emotionally and cognitively, tended to take preventive measures.

We then conducted model comparisons of the two conceptual models in two data sets using path analysis in Amos. Table 5 shows the path statistics and the model fit. First, we tested the respective paths in Model 1 (i.e., determining media). The results in Table 5 indicate that for the nuclear waste issue, media attention had direct effects on individuals’ preventive behaviors in most regions, except for Mainland China (β = .04, p = .31), South Korea (β = –.02, p = .50), and the United Kingdom (β = .02, p = .78). The direct effect was not found for the COVID issue (β = .11, p = .07). In addition, indirect effects were found in both data sets. As shown in Table 5, media attention was significantly associated with both affective and cognitive risk perceptions, which in turn significantly influenced preventive behaviors in all regions for the nuclear waste issue. A similar pattern was found for the COVID issue. Therefore, the direct and indirect effects of media attention on preventive behavior through risk perception were found for both issues across all regions, empirically supporting the utility of the determining media perspective.

Next, we examined the paths in Model 2 (i.e., mediating media). The results in Table 6 indicate that for the nuclear waste issue, affective risk perception had direct effects on individuals’ preventive behaviors in all regions. For the COVID issue, the direct effect of affective risk perception was also significant. In addition, cognitive risk perception had direct effects on individuals’ preventive behaviors in both data sets and across all regions. However, indirect effects of both types of risk perceptions on preventive behaviors were not generally found. The impact of affective risk perception on media attention was only found in the United States (β = –.20, p < .001), while the impact of cognitive risk perception on media attention was found in most regions, except for Taiwan (β = –.07, p = .16), Japan (β = .05, p = .52), and South Korea (β = .02, p = .57). In addition, the impact of media attention on preventive behaviors was not found in Mainland China (β = .04, p = .31), South Korea (β = –.01, p = .65), and the United Kingdom (β = .04, p = .68). For the COVID issue, although the significant impact of affective (β = .11, p < .05) and cognitive risk perception (β = .17, p < .01) on media attention was found, media attention did not significantly affect preventive behavior (β = .02, p = .61). The indirect effects were still not supported. Therefore, the direct and indirect effects of risk perception on preventive behavior through media attention were not generally found across issues and regions, nor were the utility of the mediating media perspective.

Table 5.

Path Analysis of Model 1 in Two Datasets.

	Data set 1 (nuclear)								Data set 2 (COVID)
	Mainland China	Hong Kong	Taiwan	Singapore	Japan	South Korea	United States	United Kingdom	Hong Kong
Direct path
Media attention—preventive behavior	.04	.13***	.17***	.18***	.25***	–.02	.16***	.02	.11
Indirect paths
Media attention—affective risk perception	.14***	.16***	.13***	.31***	.07*	.09*	.11**	.17***	.22**
Media attention—cognitive risk perception	.28***	.18***	.14***	.36***	.07*	.08*	.55***	.39***	.15**
Affective risk perception—preventive behavior	.15***	.47***	.31***	.27***	.16***	.17**	.19***	.20***	.15**
Cognitive risk perception—preventive behavior	.45***	.28***	.21***	.36***	.37***	.49***	.40***	.35***	.46***
Model fit
RMSEA	.052	.058	.052	.053	.055	.051	.059	.054	.063
CFI	.949	.956	.957	.956	.962	.965	.947	.950	.901
GFI	.903	.907	.914	.908	.915	.921	.900	.909	.910
BIC	2,169.190	2,863.824	2,170.718	2,084.739	2,037.117	1,869.462	2,319.854	2,135.472	1,843.048

Note. Controlling for gender, age, education level, and household income. For data set 2, the variable “media attention” was measured in Wave 1, while “affective and cognitive risk perception” was assessed in Wave 2, and “preventive behavior” was evaluated in Wave 3. RMSEA = root mean square error of approximation; CFI = comparative fit index; GFI = goodness-of-fit index; BIC = Bayesian information criterion.

p < .05. **p < .01. ***p < .001.

Table 6.

Path Analysis of Model 2 in Two Datasets.

	Data set 1 (nuclear)								Data set 2 (COVID)
	Mainland China	Hong Kong	Taiwan	Singapore	Japan	South Korea	United States	United Kingdom	Hong Kong
Direct path
Affective risk perception—preventive behavior	.15***	.48***	.23***	.23***	.13*	.17**	.23***	.23***	–.21***
Cognitive risk perception—preventive behavior	.45***	.25***	.29***	.40***	.37***	.49***	.35***	.32***	.41***
Indirect paths
Affective risk perception—media attention	–.09	.00	.12	.04	.02	.06	–.20***	–.07	.11*
Cognitive risk perception—media attention	.32***	.20**	–.07	.34***	.05	.02	.48***	.37***	.17**
Media attention—preventive behavior	.04	.12***	.31***	.17***	.25***	–.01	.19***	.04	.02
Model fit
RMSEA	.052	.064	.057	.055	.060	.054	.058	.056	.066
CFI	.949	.945	.949	.952	.954	.960	.949	.948	.900
GFI	.903	.936	.906	.902	.902	.914	.903	.905	.911
BIC	2,169.190	3,214.222	2,328.707	2,154.765	2,239.191	1,971.749	2,266.035	2,183.175	1,786.480

Note. Controlling for gender, age, education level, and household income. For data set 2, the variable “affective and cognitive risk perception” was measured in Wave 1, while “media attention” was assessed in Wave 2, and “preventive behavior” was evaluated in Wave 3. RMSEA = root mean square error of approximation; CFI = comparative fit index; GFI = goodness-of-fit index; BIC = Bayesian information criterion.

p < .05. **p < .01. ***p < .001.

In addition, the model fit statistics for both models are shown in Tables 5 and 6. Compared with Model 2, Model 1 exhibited generally better comparative fit index (CFI) values in both data sets, and more favorable Bayesian information criterion (BIC) values. Therefore, from a statistical perspective, the determining media perspective also demonstrated greater utility for both issues.

Discussion

Research on the role of media in risk communication presents two distinct paradigms: determining media and mediating media. This study utilized survey data related to two public crises across eight regions, from East to West, to comprehensively test the utility of these two models in explaining media effects on individuals’ behavioral intentions.

The findings from the survey data validate the superiority of the first model across various cultures, emphasizing the determining role of media in risk communication, which highlights the significant impact of media on individuals’ risk perceptions and preventive behaviors. This suggests that media attention influences individuals’ perceptions of risks, which in turn shapes their intentions to take preventive measures. Specifically, it indicates that individuals’ willingness to adopt preventive measures is largely determined by media-constructed risks rather than their own existing risk perceptions.

Direct Effects of Media Attention and Risk Perception on Preventive Behavior

This study suggested that media attention to the discharge of nuclear waste issue as well as the COVID pandemic issue directly affected people’s preventive behaviors; likewise, two types of risk perceptions (i.e., affective and cognitive risk perception) also influenced preventive behaviors. People who paid more attention to the issue on media platforms were more likely to adopt the preventive measures promoted by the media. Similarly, those who perceived risks at higher emotional and cognitive levels were more likely to express a strong intention to take related actions.

Our findings align with several studies in terms of the direct effects of media attention on preventive behaviors (Balatsoukas et al., 2015). For instance, R. Liu et al. (2022) suggested that attention to health-related information through mass media can directly promote preventive behavior (specifically vaccination intentions in this study). Likewise, our findings are consistent with previous studies that reveal the direct impact of affective and cognitive risk perceptions on individuals’ preventive behaviors (Ferrer et al., 2013). For example, one study done by Shen et al. (2021) suggested that both cognitive and emotional perceptions of risks were key predictors of behavioral responses to the pandemic. Thus, this finding indicates the direct impacts of media attention and risk perception on preventive behaviors, highlighting the need for further analysis of the relationship between media attention and risk perception, as well as their joint effects on behaviors.

Determining Media Prevailing Over Mediating Media

Moving beyond the direct effects, this study further explored the relationship between media attention and risk perception, as well as their combined effects on behavior. Thus, this compared the determining media perspective with the mediating media perspective. Specifically, it examined which comes first—media attention or risk perception—and how they influence behavior.

First, this study, adopting longitudinal panel data for COVID issue and cross-sectional data in eight regions for the nuclear issue, validates the superiority of the determining media model across various cultures. Although both data sets exhibit similar patterns, the causal predictions from the cross-sectional data on the nuclear issue may introduce potential bias. Therefore, we rely more on the longitudinal panel data for model comparisons and to address the question of causality.

According to the results of the panel survey, media play a pivotal role in shaping individuals’ risk perceptions and subsequent behaviors. Media attention precedes risk perception. The indirect effects of media attention on preventive behaviors operate as follows: media attention first influences individuals’ affective and cognitive risk perceptions, which in turn affects their preventive behaviors. Thus, the key implication of this finding is that it is the media-constructed risks, rather than individuals’ existing risk perceptions, that significantly influence their willingness to adopt preventive behaviors. People’s existing perception of risk is less important compared with what they perceive through media. Possibly, the superiority of the determining media perspective may stem from the all-pervasive nature of media platforms. Media, particularly social media that can be easily accessible via mobile phones, relentlessly demand individuals’ attention. People often have no choice but to engage with this information and as a result are intangibly influenced by it. In addition, regarding the COVID-19 issue, individuals’ perceptions of the virus and vaccines may help explain this finding. It is possible that underrepresentation of vaccination issues in media coverage, or labeling them as hoaxes (Gostin et al., 2019), contributes to the perception that the virus poses less risk than it actually does.

Second, this finding aligns with the core idea of the determining media perspective, which emphasizes the significant effects of media. It somehow supports the direct and powerful effects of media proposed by the magic bullet theory, cultivation theory, and framing theory. Importantly, this direct effect on behavior follows a flow: From media to individuals’ perceptions and beliefs about the issue—specifically, risk perception in this study—and then to their intentions to adopt certain behaviors. In addition, this study supports the fundamental rationale of the determining media paradigm, which posits that media constructs reality for the audience, and that audiences rely on media for information, ultimately shaping their attitudes and behaviors.

Third, this finding also aligns with previous empirical studies on media effects, particularly those based on theories like the health belief model (i.e., HBM), which suggests that media influence individuals’ threat perceptions, specifically perceived susceptibility and perceived severity (Rosenstock, 1974). Empirical studies have provided evidence for these pathways (R. Liu et al., 2022). For instance, by integrating the HBM, Hita et al. (2023) found that media messages during the COVID pandemic affected individuals’ fear of the disease, as well as their perceived susceptibility to and severity of COVID, which in turn influenced their social distancing practices.

In addition, our study clarifies the relationship between media attention and risk perception, indicating that media attention precedes individuals’ risk perceptions. Consequently, people depend on media for information, which further shapes their understanding of risks. This direction of influence aligns with numerous previous empirical studies (Niu et al., 2022). For instance, Zhao et al. (2019) conducted a natural experiment demonstrating that media reports about disasters had significant effects on public risk perceptions.

Practically, when faced with public crises, public institutions and health advocates should strategically utilize media platforms to effectively shape individuals’ risk perceptions and promote preventive behaviors. By creating engaging and informative content that resonates with diverse audiences, they can raise awareness about potential health risks and the importance of taking preventive measures. Specifically, they should provide clear, accurate, and actionable information, including clear explanations of specific health risks individuals may face with statistics and detailed guidance on effective preventive behaviors.

Limitations

The study has several limitations that future studies should address. First, the cross-sectional design of the data gathered from eight regions for the nuclear waste issue restricts the analysis of causal relationships within the theoretical framework. Second, to compare these two models, we ideally need to control certain variables. For example, when testing the mediating media perspective, we should ensure that individuals do not receive any media messages before forming their perceptions of risk. However, achieving this in practice is not feasible. Future studies should consider adopting lab experiments to rigorously manipulate this scenario. Third, the measurement of media attention combines various media types into a single index, which may overlook the potential heterogeneity of media effects. Different media platforms can have unique influences on behaviors. Future studies should specify the media platforms to examine their respective effects. Fourth, although we examined two different crisis issues, they were somewhat limited to the contexts of health and environmental concerns, which may introduce bias. Future studies should explore other scientific issues to validate these findings.

Conclusion

Generally, this study explores the role of media in risk communication during crises by examining two distinct theoretical models (i.e., determining media and mediating media). Survey data of two different issues (i.e., the nuclear waste issue and COVID pandemic) in eight regions from West to East validate the generally better utility of the determining media perspective in both issues and across all regions. This underscores the considerable influence of media on individuals’ risk perceptions and preventive behaviors. Specifically, it suggests that individuals’ readiness to adopt preventive measures is primarily driven by risks constructed by the media rather than their existing assessments of risk.

Our study contributes to both theoretical understanding and practical implications that public institutions and health advocates can use to address a crisis. Theoretically, this study reviewed two schools of thought that explain the role of media in risk communication and empirically compared these two paradigms. The evidence demonstrates the superiority of the determining media paradigm in elucidating the significant influence of media on human behaviors. Specifically, our study provides a key theoretical contribution by showing that individuals’ intentions to adopt preventive measures are largely shaped by media-constructed risks rather than their own risk perceptions, which challenges many previous analyses of risk.

Practically, public institutions should strategically leverage media platforms to effectively influence individuals’ risk perceptions and encourage preventive behaviors. To achieve this, they should focus on delivering clear, accurate, and actionable information. This includes offering straightforward explanations of specific health risks that individuals might encounter, supported by relevant statistics and data. In addition, providing detailed guidance on effective preventive behaviors will empower individuals to make informed decisions about their health. By fostering a well-informed public, these efforts can lead to higher rates of compliance with health recommendations and ultimately improve community health outcomes.

Footnotes

Data Availability Statement

Due to privacy concerns, the supporting data are not available now.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the City University of Hong Kong (grant nos.: 9380119, 7005703, 9610573, 7006022, 7200712, and 9610662) and the Hong Kong Research Grants Council (grant nos.: 9063007 and 9043780).

Ethical Approval

The study was conducted in accordance with the Declaration of Helsinki and approved by the Human Subjects Ethics Sub-Committee of City University of Hong Kong (reference no.: HU-STA-00000570; approval date: September 29, 2023.

Informed Consent Statement

Written consent was obtained from all the subjects involved in this study.

ORCID iDs

Ruoheng Liu

Haodong Liu

Xiaohui Wang

Author Biographies

Yi-Hui Christine Huang (PhD, University of Maryland), ICA Fellow, is a chair professor of Communication and Media at the Department of Media of Communication, City University of Hong Kong. Her research interests include strategic communication management in Chinese societies, risk communication in health and technology, crisis communication and management, relationship management and conflict resolution, and cross-cultural communication.

Ruoheng Liu is a PhD candidate at the Department of Media and Communication at City University of Hong Kong. Her research interests include risk communication, health communication, and media effects studies.

Leyi Zhang is a PhD student at the Department of Media and Communication at City University of Hong Kong. Her research focuses on international communication, media psychology, and misinformation.

Qinxian Cai is a PhD candidate at the Department of Media and Communication at City University of Hong Kong. His research interests include strategic communication, crisis management, and risk communication.

Haodong Liu is a PhD student at the Department of Media and Communication at City University of Hong Kong. His research focuses on strategic communication, health communication, computational communication, and media effects.

Xiaohui Wang (PhD, Nanyang Technological University) is an assistant professor at the Department of Media of Communication, City University of Hong Kong. His research converges at the intersection of information, communication, and new technologies, with special focuses on social media, computational social science, consumer health informatics, and informetrics.

References

Armstrong

C. L.

VanDyke

M. S.

(2025). Verify, accept, or ignore: The complexities of media trust along Southeastern U.S. coastal communities during severe weather events. Journalism & Mass Communication Quarterly. https://doi.org/10.1177/10776990251347659

Balatsoukas

Kennedy

C. M.

Buchan

Powell

Ainsworth

(2015). The role of social network technologies in online health promotion: A narrative review of theoretical and empirical factors influencing intervention effectiveness. Journal of Medical Internet Research, 17(6), e3662. https://doi.org/10.2196/jmir.3662

Bekalu

M. A.

McCloud

R. F.

Viswanath

(2019). Association of social media use with social well-being, positive mental health, and self-rated health: Disentangling routine use from emotional connection to use. Health Education & Behavior, 46, 69S–80S. https://www.jstor.org/stable/48625502

Berger

(1995). Essentials of mass communication theory. Sage.

Chae

Lee

C.-J.

(2019). The psychological mechanism underlying communication effects on behavioral intention: Focusing on affect and cognition in the cancer context. Communication Research, 46(5), 597–618. https://doi.org/10.1177/0093650216644021

Champion

V. L.

Monahan

P. O.

Springston

J. K.

Russell

Zollinger

T. W.

Saywell

R. M.

Jr. Maraj

(2008). Measuring mammography and breast cancer beliefs in African American women. Journal of Health Psychology, 13(6), 827–837. https://doi.org/10.1177/1359105308093867

Dai

Huang

Y.-H. C.

Jia

Cai

(2022). The paradoxical effects of institutional trust on risk perception and risk management in the Covid-19 pandemic: Evidence from three societies. Journal of Risk Research, 25(11-12), 1337–1355. https://doi.org/10.1080/13669877.2022.2108122

Drew

(2016). Technological determinism. In Burns

(Eds.), A companion to popular culture (pp. 165–183). Wiley Publications. https://doi.org/10.1002/9781118883341.ch10

Entman

R. M.

(1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51–58.

10.

Fenta

E. T.

Bogale

E. K.

Anagaw

T. F.

(2024). The role of social media on COVID-19 preventive behaviors worldwide, systematic review. PLOS ONE, 19(7), e0306284. https://doi.org/10.1371/journal.pone.0306284

11.

Ferrer

R. A.

Portnoy

D. B.

Klein

W. M. P.

(2013). Worry and risk perceptions as independent and interacting predictors of health protective behaviors. Journal of Health Communication, 18(4), 397–409. https://doi.org/10.1080/10810730.2012.727954

12.

Geiger

Newhagen

(1993). Revealing the black box: Information processing and media effects. Journal of Communication, 43(4), 42–50. https://doi.org/10.1111/j.1460-2466.1993.tb01303.x

13.

Gerbner

Gross

Morgan

Signorielli

(1980). The “mainstreaming” of America: Violence profile number 11. Journal of Communication, 30(3), 10–29.

14.

Gostin

L. O.

Ratzan

S. C.

Bloom

B. R.

(2019). Safe vaccinations for a healthy nation: Increasing US vaccine coverage through law, science, and communication. Journal of the American Medical Association, 321(20), 1969–1970.

15.

Griffin

R. J.

Dunwoody

Neuwirth

(1999). Proposed model of the relationship of risk information seeking and processing to the development of preventive behaviors. Environmental Research, 80(2), S230–S245. https://doi.org/10.1006/enrs.1998.3940

16.

Hill

(2001). Media risks: The social amplification of risk and the media violence debate. Journal of Risk Research, 4(3), 209–225. https://doi.org/10.1080/13669870152023773

17.

Hita

M. L. R.

Grégoire

Lussier

Boissonneault

Vandenberghe

Sénécal

(2023). An extended health belief model for COVID-19: Understanding the media-based processes leading to social distancing and panic buying. Journal of the Academy of Marketing Science, 51(1), 132–152. https://doi.org/10.1007/s11747-022-00865-8

18.

Huang

Y.-H. C.

Cai

Wang

Sun

(2025). What drives public support for health policies? The protection-motivated mediating model of institutional trust and risk paradox. Social Science and Medicine. https://doi.org/10.1016/j.socscimed.2025.118310

19.

Huang

Y. H. C.

Liu

(2022). Go for zero tolerance: Cultural values, trust, and acceptance of zero-COVID policy in two Chinese societies. Frontiers in Psychology, 13, 1047486. https://doi.org/10.3389/fpsyg.2022.104748

20.

Huang

Y. H. C.

Liu

Sun

(2025). Understanding the conditional media effects: Trust, new media exposure, and preventive behaviors during the COVID-19 pandemic. Chinese Journal of Communication, 18(1), 96–113. https://doi.org/10.1080/17544750.2024.2374269

21.

Huang

Y. H. C.

Liu

Zhang

(2025). Does age matter? The impact of social media exposure on people’s mental health. Health Policy and Technology, 14(3), 100997. https://doi.org/10.1016/j.hlpt.2025.100997

22.

Huang

Y.-H. C.

Wang

Fong

I. W.-Y.

(2021). Examining the role of trust in regulators in food safety risk assessment: A cross-regional analysis of three Chinese societies using an integrative framework. SAGE Open. https://doi.org/10.1177/21582440211061579

23.

Johnson

J. D.

Meischke

(1993). A comprehensive model of cancer-related information seeking applied to magazines. Human Communication Research, 19(3), 343–367. https://doi.org/10.1111/j.1468-2958.1993.tb00305.x

24.

Katz

Blumler

J. G.

Gurevitch

(1973). Uses and gratifications research. The Public Opinion Quarterly, 37(4), 509–523.

25.

Katz

Lazarsfeld

P. F.

(1955). Personal influence: The part played by people in the flow of mass communications. Transaction Publisher.

26.

Lazarsfeld

P. F.

Berelson

Gaudet

(1968). The people’s choice: How the voter makes up his mind in a presidential campaign. Columbia University Press. https://doi.org/10.7312/laza93930

27.

Liu

P. L.

(2021). COVID-19 information on social media and preventive behaviors: Managing the pandemic through personal responsibility. Social Science & Medicine, 277, Article 113928. https://doi.org/10.1016/j.socscimed.2021.113928

28.

Liu

Huang

Y.-H. C.

Sun

Lau

Cai

(2022). A shot in the arm for vaccination intention: The media and the health belief model in three Chinese societies. International Journal of Environmental Research and Public Health, 19(6), 3705. https://doi.org/10.3390/ijerph19063705

29.

Loewenstein

Weber

E. U.

Hsee

C. K.

(2001). Risk as feelings. Psychological Bulletin, 127(2), 267–286.

30.

Lowery

S. A.

Defleur

M. L.

(1995). Milestones in mass communication research: Media effects (3rd ed.). Longman.

31.

Melki

Tamim

Hadid

Farhat

Makki

Ghandour

Hitti

(2022). Media exposure and health behavior during pandemics: The mediating effect of perceived knowledge and fear on compliance with COVID-19 prevention measures. Health Communication, 37(5), 586–596. https://doi.org/10.1080/10410236.2020.1858564

32.

Moorhead

S. A.

Hazlett

D. E.

Harrison

Carroll

J. K.

Irwin

Hoving

(2013). A new dimension of health care: Systematic review of the uses, benefits, and limitations of social media for health communication. Journal of Medical Internet Research, 15(4), e1933. https://doi.org/10.2196/jmir.1933

33.

Niedenthal

P. M.

Halberstadt

J. B.

Innes-Ker

Å. H.

(1999). Emotional response categorization. Psychological Review, 106(2), 337. https://doi.org/10.1037/0033-295X.106.2.337

34.

Niu

Jiang

Liu

Yang

Song

(2022). The influence of media consumption on public risk perception: A meta-analysis. Journal of Risk Research, 25(1), 21–47. https://doi.org/10.1080/13669877.2020.1819385

35.

Pennycook

McPhetres

Bago

Rand

D. G.

(2022). Beliefs about COVID-19 in Canada, the United Kingdom, and the United States: A novel test of political polarization and motivated reasoning. Personality and Social Psychology Bulletin, 48(5), 750–765. https://doi.org/10.1177/01461672211023652

36.

Potter

W. J.

Riddle

(2007). A content analysis of the media effects literature. Journalism & Mass Communication Quarterly, 84(1), 90–104. https://doi.org/10.1177/107769900708400107

37.

Rosenstock

I. M.

(1974). The health belief model and preventive health behavior. Health Education Monographs, 2(4), 354–386.

38.

Santoro

Castelnuovo

Zoppis

Mauri

Sicurello

(2015). Social media and mobile applications in chronic disease prevention and management. Frontiers in Psychology, 6, 567. https://doi.org/10.3389/fpsyg.2015.00567

39.

Sarkar

G. M.

Lyles

C. R.

Ramo

Linos

Bibbins-Domingo

(2018). Using social media to target cancer prevention in young adults: Viewpoint. Journal of Medical Internet Research, 20(6), e203. https://doi.org/10.2196/jmir.8882

40.

Shen

Min

Chu

Y. J.

(2021). The effect of cognition and affect on preventive behaviors during the COVID-19 pandemic: A cross-sectional study in China. BMC Public Health, 21(1), Article 722. https://doi.org/10.1186/s12889-021-10784-y

41.

Slovic

Peters

(2006). Risk perception and affect. Current Directions in Psychological Science, 15(6), 322–325. https://doi.org/10.1111/j.1467-8721.2006.00461.x

42.

Valkenburg

P. M.

Peter

Walther

J. B.

(2016). Media effects: Theory and research. Annual Review of Psychology, 67(1), 315–338. https://doi.org/10.1146/annurev-psych-122414-033608

43.

Wang

Zheng

Huang

Y. H. C.

(2023). Examining the pathways from information acquisition to preventive behaviors: The information-response model. Science Communication, 45(6), 724–750. https://doi.org/10.1177/10755470231204317

44.

Wartberg

Kammerl

Bröning

Hauenschild

Petersen

K.-U.

Thomasius

(2015). Gender-related consequences of Internet use perceived by parents in a representative quota sample of adolescents. Behaviour & Information Technology, 34(4), 341–348. https://doi.org/10.1080/0144929X.2014.928746

45.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063. https://doi.org/10.1037/0022-3514.54.6.1063

46.

Wong

(2023, August 24). Fukushima: China retaliates as Japan releases treated nuclear water. BBC News. https://www.bbc.com/news/world-asia-66577769

47.

World Health Organization. (2024). WHO Coronavirus Disease (COVID-19) Dashboard. https://covid19.who.int/

48.

Zanna

M. P.

Rempel

J. K.

(2008). Attitudes: A new look at an old concept. In Fazio

R. H.

Petty

R. E.

(Eds.), Attitudes: Their structure, function, and consequences (pp. 7–15). Psychology Press.

49.

Zhao

Rosoff

John

R. S.

(2019). Media disaster reporting effects on public risk perception and response to escalating tornado warnings: A natural experiment. Risk Analysis, 39(3), 535–552. https://doi.org/https://doi.org/10.1111/risa.13205