Trust in health-related settings is often conceptualized as “the optimistic acceptance of a vulnerable situation in which the truster believes that the trustee will care for the truster’s interests” [(27), p. 615]. Following Hendriks et al.’s (28) work, the present study conceptualizes the trustworthiness of experts as the totality of expertise (e.g., knowledge), integrity (e.g., honesty), and benevolence (e.g., good intention).

The Trust, Confidence, and Cooperation (TCC) model (29) suggests that trust in authorities and experts is a key determinant of risk perception and acceptance in risk communication (11, 30). Specifically, lower trust in information sources may heighten risk perception and reduce perceived benefits, leading to unfavorable judgments toward recommended behaviors (31–33).

Trust in experts is not merely a function of perceived competence but also institutional credibility and alignment with societal values (34, 35). During the COVID-19 pandemic, institutional trust became a cornerstone of compliance with public health measures, particularly in state-aligned systems like China (36, 37). However, trust can be paradoxical: while it reduces skepticism toward institutional directives (35), it may also heighten scrutiny of expert warnings about risks (38). For example, transparent communication about vaccine uncertainties could inadvertently elevate risk perceptions while fostering trust (36). This duality underscores the need to disentangle trust’s role in shaping risk–benefit perceptions and affective responses within centralized health systems.

Studies have shown that people who trust experts may also be more aware of the uncertainties and risks associated with boosters, which leads to heightened perceptions of risk (39, 40). This may be due to experts being transparent about the evolving scientific knowledge or the potential for adverse outcomes (41, 42). Alternatively, trust in experts may also make individuals more critical or cautious, thus perceiving fewer benefits from the booster shots when new scientific data or advice creates uncertainty (28, 43).

Consequently, while trust in experts plays a crucial role in shaping one’s health decisions, it may also increase individuals’ awareness and skepticism about the benefits of specific health interventions. From these findings, we hypothesize that:

In various health-related contexts, trust in experts functions as a protective mechanism against uncertainty to alleviate negative emotions such as anxiety. When individuals perceive experts as competent, honest, and acting in the public’s best interests, they are less prone to feelings of fear or worry (44). Studies on public health crises have shown that higher trust in experts reduces worry or anxiety (45, 46). Based on the evidence:

In addition, trust in experts may influence how individuals respond to social pressure or expectations. Ho et al. (47) demonstrated that trust in scientific authorities strengthens the influence of social norms on science-related behaviors, while more recent studies (48, 49) showed trust in experts amplifies how people respond to social expectations edge. In other words, when people trust experts more, they are more likely to behave according to social expectations. This interaction between trust and subjective norms leads us to ask:

RQ2: Will trust in experts moderate the relationship between informational subjective norms and information insufficiency, such that higher trust in experts will strengthen the positive association between informational subjective norms and information insufficiency?

Uncertainty arises when “situations are ambiguous, complex, unpredictable or probabilistic; when information is unavailable or inconsistent; and when people feel insecure in their state of knowledge or the state of knowledge in general” [(50), p. 478]. In healthcare, there are several forms of uncertainty, such as uncertainty relating to the effect of treatments or prevention measures, personal vulnerability, and possible side effects that one can suffer (51). Such continuously changing preventive measures against the variants of SARS-CoV-2 overloaded the public and thus accelerated the development of the so-called ‘infodemic,’ making information searching even harder (52–56).

In contexts with persistent uncertainty, like the COVID-19 pandemic, individuals may adapt by accepting uncertainty rather than seeking to eliminate it. Consequently, higher uncertainty might paradoxically reduce perceived information insufficiency as individuals accept that complete knowledge is unattainable.

While RISP traditionally assumes information seeking reduces uncertainty, uncertainty management theory provides a complementary perspective where individuals may strategically manage rather than eliminate uncertainty (50, 57–59). When faced with high levels of uncertainty, individuals may experience heightened affective responses, such as worry or anxiety, that can motivate them to seek more information (50, 61).

Meanwhile, high levels of uncertainty often make individuals more attuned to social norms and environmental cues to determine the appropriate course of action. In situations with high uncertainty, individuals tend to rely more on the expectations of society or societal pressures, especially in a collectivist culture like China. In these situations, collective norms can influence personal conduct such that social expectations by peers, communities, or health authorities can significantly impact individuals’ motivation toward information.

Under uncertainty, individuals may feel less control due to ambiguity or become more proactive in seeking trusted sources. This adaptation may increase their perceived control. The ability to access, understand, and utilize information, despite ambiguities, can help bolster one’s self-efficacy regarding decision-making. Therefore, rather than feeling overwhelmed, they might consider themselves better equipped to handle health-related decisions at times of great uncertainty. Thus, we hypothesize:

Furthermore, levels of uncertainty may influence one’s perceptions of the benefits and risks associated with COVID-19 booster shots. For instance, a heightened level of perceived uncertainty may make people more sensitive to potential risks while diminishing potential benefits, which may impact their confidence in decision-making. This begets the following hypotheses:

Due to model complexity and sample constraints, hypotheses H7 through H9 are treated as exploratory. These hypotheses aim to offer preliminary insights into the broader influence of uncertainty on informational subjective norms, perceived control, and risk–benefit appraisals. Confirmatory testing of these pathways would require larger and more diverse samples in future studies.

While the hypotheses (see Figure 1) address direct relationships between key variables, the complex nature of risk information-seeking behavior suggests that meaningful indirect relationships may be at play. The current study aims to capture these complex relationships, providing a comprehensive framework for understanding how various factors, particularly trust in experts and perceived uncertainty, influence information-seeking intention about COVID-19 booster shots.

Although our literature review centers on the perceived trustworthiness of experts, we operationalize the concept using Hendriks et al.’s (28) validated scale of the perceived untrustworthiness of scientists (e.g., whether they are irresponsible) on a 7-point Likert scale. To further clarify, respondents were asked to indicate their level of trust in several types of experts, including public health officials, scientists, and governmental health agencies (e.g., CDC China). While the term ‘experts’ is used generically in the analysis, the items reflect institutional affiliations relevant to the Chinese context. All items were reverse-coded so that higher scores reflect higher trust in experts. This approach ensured that higher composite scores reflect greater trust rather than untrustworthiness (M = 5.94, SD = 1.06, Cronbach’s α = 0.94).

The Uncertainty Management Theory suggests that severity, susceptibility, and self-efficacy are associated with one’s perceived medical uncertainty (63). Severity and susceptibility also shape one’s health-related behavior, such as information seeking (64). Therefore, we measured perceived uncertainty by asking participants to what extent they were uncertain about the severity of and susceptibility to the risks of COVID-19 boosters on a 7-point Likert scale (M = 4.81, SD = 0.86, Cronbach’s α = 0.70).

Participants rated their agreement (1 = strongly disagree, 5 = strongly agree) on four items (e.g., “Most of my family whose opinions I value expect me to seek information about the risks posed by receiving COVID-19 booster doses”). The items were averaged to create a composite, with higher scores representing higher normative expectancy for booster-related information seeking (M = 3.88, SD = 0.684, Cronbach’s α = 0.80).

To measure the construct, participants rated their agreement (1 = strongly disagree, 5 = strongly agree) on four items (e.g., “I know where to look for information about the risks of receiving COVID-19 booster shots”). We created a composite by averaging the four items; a higher composite score reflects better control (M = 3.98, SD = 0.62, Cronbach’s α = 0.74).

Risk perceptions are often captured as the perceived likelihood of the risk occurring and the perceived seriousness of the risk if it were to happen. We measured risk likelihood with four questions, “how likely is it that you will be impacted by the potential risks posed by COVID-19 booster shots,” “If you were impacted by the potential risks posed by COVID-19 booster shots, how serious would that impact be,” “how likely is it that society will be impacted by the potential risks posed by COVID-19 booster doses,” and “if society were impacted by the potential risks posed by COVID-19 booster doses, how serious would that impact be?.” Response options were on a 5-point Likert scale ranging from “1 = not at all likely/serious” to “5 = extremely likely/serious.” The items were averaged to create a composite, with higher scores representing higher risk perceptions (M = 3.12, SD = 1.26, Cronbach’s α = 0.90).

In addition, measurements for benefit perception were adapted from the work of Lin et al. (54) and Sun et al. (65) to better align with the context of COVID-19 booster shots. Respondents were asked, “How beneficial will COVID-19 booster shots be for you personally,” and “How beneficial will COVID-19 booster shots be for society?” on a 5-point Likert scale ranging from “1 = not at all beneficial” to “5 = extremely beneficial (M = 4.28, SD = 0.65, Cronbach’s α = 0.72).

We intended to stay consistent with previous negative and positive affective response measurements. For negative affective responses, we based our measures on Witte’s (66) fear appeals but modified them to five items: “When I think about COVID booster shots, I get… frightened,” “…tense,” “…nervous,” “…anxious,” and “…uncomfortable.” Participants rated their agreement with these statements on a 7-point scale (1 = strongly disagree, 7 = strongly agree). Items were averaged to create a composite, with higher scores representing more negative responses toward the booster shots (M = 2.7, SD = 1.28, Cronbach’s α = 0.91). When assessing positive affective responses on a 7-point scale (1 = strongly disagree, 7 = strongly agree), we modified Watson et al.’s (67) work and asked respondents, “When I think about COVID booster shots, I feel… hopeful,” “…enthusiasm,” “…relieved,” and “…confident.” We then averaged the scores to create a composite (M = 5.5, SD = 0.94, Cronbach’s α = 0.86).

This concept describes the discrepancy between a person’s current knowledge level and the perceived level they think (arithmetic difference between current and needed knowledge) to align with prior RISP studies (2). While latent variable modeling was considered, this approach ensured consistency with the model’s original formulation. To measure this difference, we first asked, “How much do you currently know about the risks/benefits posed by COVID-19 booster doses?” and “How much do you need to know about the risks/benefits posed by COVID-19 booster doses?” on a scale of 0 to 100 (“0 = nothing” to “100 = all there is to know.”). We then calculated perceived knowledge insufficiency by subtracting the current knowledge score (M = 68.35, SD = 17.38) from perceived need (M = 67.02, SD = 19.62) to identify the extent to which participants’ current knowledge meets or falls short of perceived needs (M = −1.33, SD = 23.41).

Information-seeking intent was measured with four items on a Likert-type scale ranging from “1 = strongly disagree” to “5 = strongly agree” (e.g., “I will try to seek information about the risks posed by COVID-19 booster doses in the next 6 months),” “I will look for information about the risks posed by COVID-19 booster doses in the next 6 months,” (Kahlor, 2010). Items were averaged to create a composite, with higher values reflecting stronger intentions to seek information about COVID-19 booster shots (M = 3.98, SD = 0.7, Cronbach’s α = 0.85).

Our most striking finding contradicts core RISP assumptions: emotional reactions did not motivate information seeking about boosters, unlike in Western contexts. In contrast to the RISP model, affective responses did not drive information seeking. Instead, social expectations dominated decision-making. This reflects China’s cultural prioritization of communal harmony over individual concerns. In other words, social norms typically outweigh affective responses (10). For another, the dominance of social norms over affective responses aligns with institutionalized health communication in centralized systems, where collective compliance often supersedes individual emotions (76).

Additionally, China’s strong government guidance during the COVID-19 pandemic may have minimized the relevance of individual emotions in information-seeking decisions (38, 77). The consistency of subjective norms across geographic and socioeconomic subgroups underscores their primacy in collectivist health communication. While trust in experts is often theorized to drive compliance, its non-significance in our analyses suggests that institutional trust in China may function indirectly by reinforcing social conformity rather than directly motivating information-seeking behavior (38). This interpretation aligns with cultural frameworks where communal expectations supersede individual trust in authorities (78).

Results suggest that health communication strategies in collectivist contexts should focus on creating social consensus and institutional guidance about information-seeking behavior rather than appeals to emotions. In collectivist contexts, such as China, where trust in governmental and institutional expertise is generally high, individuals may perceive that their current knowledge, derived from official sources, is sufficient for making informed health decisions. This cultural context may reduce the likelihood of perceived knowledge gaps, as individuals may rely more heavily on institutional guidance rather than personal exploration to feel adequately informed (38, 78).

Moreover, the first RQ examined how positive and negative affective reactions influenced information-search intentions. The results showed that both types of affects failed to predict information search behavior significantly; however, positive affect was more closely related to this factor (β = −0.754, p = 0.346) than negative ones (β = −0.17, p = 0.17). This finding raises questions about traditional assumptions in risk communication that negative emotions are primary drivers of information seeking (12). The non-significant results for both affect types might suggest that in long-term health crises like COVID-19, emotional responses become less influential in driving information-seeking behaviors (77). In China, where state-aligned messaging consistently emphasizes collective resilience, individuals may have habituated to fear appeals, redirecting their focus toward normative expectations. Additionally, cultural norms prioritizing emotional restraint in collectivist societies (10) may suppress the role of personal affect in health decisions, further explaining the dominance of subjective norms over emotional drivers. Therefore, these results suggest that affective pathways in the RISP model might indeed be culturally bounded, particularly in collectivist contexts characterized by prolonged and institutionalized risk communication. These same cultural dynamics that muted emotional expression also appeared to influence how individuals processed uncertainty and managed information-seeking decisions.

The significant role of uncertainty in shaping both informational subjective norms and perceived behavioral control extends our understanding of how uncertainty influences information-seeking behavior. Our findings show that uncertainty positively predicts subjective norms (β = 0.475, p < 0.001) and perceived behavioral control (β = 0.412, p < 0.001), suggesting that under conditions of uncertainty, individuals may become more attuned to social cues while also feeling more capable of managing information-seeking behavior. This pattern aligns with uncertainty management theory (50), which suggests that individuals employ multiple strategies to cope with uncertainty. However, the positive relationship between uncertainty and perceived behavioral control is somewhat surprising, as previous research has often found uncertainty to decrease perceived control. This unexpected finding may be explained by the unique context of COVID-19 in China, where a robust information infrastructure and clear institutional guidance might have enhanced individuals’ confidence in their ability to find information despite underlying uncertainties (53). These findings indicate that public health organizations should not shy away from acknowledging uncertainties but rather combine such acknowledgments with clear guidance on information-seeking channels and resources, particularly in contexts with strong institutional support systems.

To further clarify how uncertainty influences information-seeking behavior, we investigated whether perceived behavioral control mediates the relationship between uncertainty and the intention to seek information. Results revealed a significant direct effect of uncertainty on control (β = 0.412, p < 0.001) and a marginally significant indirect effect on intentions through perceived behavioral control (β = 0.30, p = 0.078). Although marginally significant, this indirect effect (β = 0.30) suggests a practically meaningful mechanism, that is, institutional guidance appears capable of transforming uncertainty into actionable agency, enhancing perceived control by nearly one-third of a standard deviation, thereby promoting greater information-seeking. Though exploratory, this finding extends uncertainty management theory by illustrating how institutional infrastructure can transform ambiguity into agency, without altering our primary conclusion about normative dominance.

RQ2 aimed to investigate whether trust in experts moderated the relationship between informational subjective norms and information insufficiency. Our analysis revealed that neither the main effects (trust: β = −0.502, p = 0.231; subjective norms: β = −0.328, p = 0.408) nor their interaction (β = 0.297, p = 0.171) significantly predicted information insufficiency, which suggests that the relationship between informational social norms and perceived knowledge insufficiency remains consistent regardless of trust levels. This result is particularly interesting in the Chinese context, where trust in authorities and social norms typically strongly influence behavior (78). One possible explanation is that trust and norms function as parallel rather than synergistic forces in institutionalized contexts. Since both derive from alignment with official narratives, their joint influence may be redundant. Alternatively, the high baseline trust in health authorities may have restricted variability, which limits the detection of interactive effects.

The lack of moderation reflects the institutionalized communication environment, where trust and norms operate independently rather than interactively (53). In state-aligned systems, institutional trust is deeply ingrained and functions as a stable heuristic for accepting official guidance, while social norms derive from collective compliance with centralized directives. This separation suggests that the expertise of experts does not amplify the effect of norms on knowledge gaps, as both are already aligned with institutional authority. Thus, health communication strategies should treat trust-building and social norm cultivation as separate but complementary approaches rather than assuming they will enhance each other’s effects.

Finally, the strong relationship between informational subjective norms and information-seeking intentions (β = 0.979, p = 0.001), together with only marginal effects of perceived behavioral control (β = 0.727, p = 0.093), suggests that social influences may be of significance in collectivist cultures facing health crises. This result aligns with recent research on COVID-19 communication in East Asian contexts, where social norms often play a dominant role in shaping health behaviors [e.g., (21)]. Thus, the implication is that health communication strategies in collectivist contexts should stress community participation and leverage social networks to promote information-seeking behaviors.

Firstly, our study relies on behavioral intentions rather than actual behaviors. This gap suggests that social-cognitive models are generally more accurate in predicting one’s intentions than actual behaviors [e.g., (78–82)]. In essence, it is essential to exercise caution when extending the study’s findings to explain individuals’ actual seeking behavior regarding COVID-19 booster shots. Therefore, future studies should consider closing the gap by incorporating additional variables, such as Ajzen (3).

Secondly, participants reported low perceived knowledge insufficiency, likely reflecting strong trust in institutional communication sources in China’s COVID-19 context. This unexpected finding suggests that in contexts with strong institutional messaging, people may feel adequately informed despite objective knowledge gaps. Future research could examine contexts where perceived information gaps are more pronounced, such as when conflicting information exists, or trust in institutions is lower.

Thirdly, while providing valuable insights, the generalizability of our study is limited due to the relatively homogeneous sample, which predominantly comprises young, educated individuals residing in urban areas. Although we conducted subgroup analyses across geographic and socioeconomic groups and found consistent patterns, these results should still be interpreted cautiously. For example, this demographic profile likely amplifies the observed role of subjective norms and institutional trust, given that urban and educated groups typically have greater digital access, higher institutional trust, and more exposure to official public health messaging in China. Consequently, our results might overstate the influence of norms and institutional trust while potentially underestimating barriers such as lower health literacy and reduced media access that rural or lower socioeconomic populations face. Future research should employ stratified sampling or weighting methods to ensure representation of rural and socioeconomically diverse populations, thereby improving the external validity of the findings. Additionally, health communication campaigns should explicitly test the effectiveness of norm-based and trust-oriented messages in more diverse demographic subgroups to determine the generalizability and efficacy of these strategies across different population segments.

Fourthly, China’s zero-COVID policy and state-controlled media created a unique risk communication environment characterized by institutional trust and social norms, which may limit the generalizability of our findings to other cultural contexts or different types of health risks. However, other collectivist or state-influenced health systems may exhibit similar dynamics. Future comparative studies (e.g., East Asian or Middle Eastern contexts) can explore how institutional trust interacts with social conformity in shaping public health behavior, enhancing the model’s cross-cultural applicability.

Finally, our cross-sectional design captures these relationships at only a single point in time, despite the likely evolution of perceptions and behaviors during a prolonged crisis. Therefore, it may limit our ability to infer causality or capture the temporal dynamics of risk perception and information-seeking during the evolving stages of the COVID-19 pandemic. Future longitudinal studies could better illustrate how these processes unfold over time, particularly as information needs and institutional messages evolve.