Sustainable Community Resilience: Linking Perceived and Actual Household Disaster Preparedness

Abstract

From a public management perspective, actual household disaster preparedness is an important component of sustainable community resilience because it strengthens households’ capacity to withstand, adapt to, and recover from disruptions. However, despite it being an important community-level dimension, the mechanisms linking perceived and actual household disaster preparedness are not well understood. Although prior research has demonstrated the close association between preparedness and socio-attitudinal factors, the mechanisms through which these factors influence preparedness-related behavior are unclear. This study examined whether perceived disaster preparedness is associated with actual household disaster preparedness, and whether this relationship is mediated by institutional trust, civic participation, and perceived well-being and moderated by gender and education. The analysis was based on data from an omnibus survey of 550 adults in Lithuania and employed moderated serial mediation models using PROCESS Models 6 and 58. The findings showed no direct association between perceived disaster preparedness and actual household disaster preparedness; instead, the relationship operated indirectly. Civic participation was the only significant simple mediator: higher perceived disaster preparedness was associated with lower civic participation, which in turn was associated with lower actual household disaster preparedness. Institutional trust and perceived well-being did not mediate this relationship. The negative association between institutional trust and actual household disaster preparedness was stronger among more highly educated respondents, whereas the positive association between civic participation and actual household disaster preparedness was stronger among men than among women. These findings suggest that differentiated preparedness strategies are needed to strengthen actual household disaster preparedness as a foundation of sustainable community resilience, and, more specifically, of the long-term sustainability of communities facing recurring risks and disruptions.

1. Introduction

Disasters continue to test the capacity of households, communities, and public institutions to cope with disruption. In this context, actual household disaster preparedness is an important component of sustainable community resilience, because prepared households are better able to withstand, adapt to, and recover from shocks, while also reducing pressure on public response systems [1,2]. In this paper, sustainable community resilience is understood as a community-level expression of community sustainability, referring to the capacity of communities to maintain essential functions and adapt over time in the face of disruption. It is used in a narrower sense than sustainable development, which encompasses broader environmental, economic, and social development goals. Research has identified many factors associated with disaster preparedness, including prior experience, risk perception, preparedness knowledge, social capital, institutional and community support [1,3,4,5]. At the same time, studies increasingly distinguish between perceived preparedness and actual preparedness, showing that people may feel prepared without necessarily having taken concrete preparedness actions [6,7]. This distinction is important because interventions that raise confidence or awareness may not automatically strengthen actual preparedness.

An important gap remains, however, in understanding how perceived preparedness becomes actual household disaster preparedness. Existing studies have often examined either subjective perceptions or objective preparedness measures, but have paid less attention to the social and attitudinal mechanisms that connect them. Yet sustainable community resilience depends not only on whether people believe they are prepared, but also on whether they trust institutions appropriately, participate in civic networks that support preparedness, and possess the social and motivational capacities needed to translate preparedness beliefs into action [3,8,9]. The literature also suggests that these pathways may not operate uniformly across population groups, because preparedness is shaped by demographic and socioeconomic heterogeneity [10,11].

This study addresses that gap by examining whether perceived disaster preparedness is associated with actual household disaster preparedness and whether this relationship is mediated by institutional trust, civic participation, and perceived well-being and moderated by demographic variables. By doing so, the study contributes specifically to research on community sustainability at the local level.

The remainder of the article is structured as follows: In Section 2, the relevant literature is reviewed and the hypotheses guiding the study are developed. In Section 3, the data source, study participants, measurement approach, and statistical analysis methods are described. In Section 4, the empirical findings are presented, beginning with the preliminary analysis and proceeding to hypothesis testing, which shows that perceived disaster preparedness has no direct association with actual household disaster preparedness, but operates through indirect pathways. In Section 5, these findings are discussed, with particular attention to the indirect linkages identified and the role of social-attitudinal and demographic factors in shaping them. In Section 6, the main conclusions are presented.

2. Literature Review and Hypothesis Development

Household disaster preparedness is widely recognized as a core component of community resilience, and more specifically of sustainable community resilience. The latter is understood as the collective, long-term capacity of communities to anticipate, adapt to, and recover from hazards while strengthening transformative capacities and well-being without compromising core functions [12]. Households are the first unit expected to cope with disruption when formal response systems are delayed, overwhelmed, or not accessible. In this sense, preparedness is a matter of possessing material supplies, but also of having the cognitive, social, and institutional conditions that enable households to respond to risk effectively [4,7]. This link is important because resilience is built through interconnected social, economic, institutional, infrastructural, and environmental capacities rather than through material resources alone [12]. Recent research has shown that household preparedness is multidimensional and should not be reduced to a single indicator, since people may express confidence in their readiness while lacking the practical resources needed to cope with disaster [6,13,14].

A key distinction in the preparedness literature is between perceived disaster preparedness and actual household disaster preparedness. Perceived disaster preparedness refers to a subjective evaluation of readiness, that is, the extent to which individuals believe they are prepared for a future emergency. It is therefore closely related to confidence, self-efficacy, risk appraisal, and perceived ability to manage disruption. Actual household disaster preparedness, in contrast, refers to observable and reportable preparedness actions and resources, such as emergency kits, stored water and food, evacuation plans, communication arrangements, or prior training [6,7]. Studies increasingly emphasize that these two dimensions are related but not equivalent. Brown et al. [6], for example, show that perceived and actual household disaster preparedness overlap, but are shaped by different factors. Similarly, Hasan et al. [13] found that university students assessed themselves as moderately prepared, although their actual preparedness was lower, while Hong et al. [14] documented high self-reported confidence alongside weaker practical capacity to cope with loss of essential services. Tkachuck [15] also reported a positive but incomplete correspondence between perceived and actual preparedness, and Velásquez-Espinoza and Alcántara-Ayala [16] further show in educational settings that risk awareness and confidence in preparedness do not necessarily imply strong operational readiness. Together, these findings suggest that feeling prepared does not automatically translate into household readiness, and that a gap may exist between cognitive confidence and enacted preparedness behavior.

This distinction is analytically important because an exclusive focus on either perceived or actual preparedness risks producing an incomplete understanding of household resilience. From a resilience perspective, such misalignment matters because households that overestimate their readiness may delay additional action, whereas households with some objective preparedness may still feel unready and respond hesitantly during a disaster. From the perspective of sustainable community resilience, such a gap is consequential because long-term resilience depends not simply on subjective confidence, but on whether households can mobilize concrete capacities across preparedness, response, recovery, and adaptation [12,17]. This suggests that the relationship between perceived disaster preparedness and actual household disaster preparedness is unlikely to be exhausted by a simple direct association alone and instead may operate through broader social and psychosocial variables. On this basis, the present study proposes the following hypothesis:

H1. 

Perceived disaster preparedness is associated with actual household disaster preparedness through indirect social-attitudinal pathways rather than through a direct effect.

The literature further suggests that the relationship between perceived disaster preparedness and actual household disaster preparedness is not purely individual, but socially embedded. One important factor in this regard is institutional trust. Trust in government, public authorities, and emergency management institutions is often assumed to facilitate preparedness by increasing acceptance of official information and guidance. However, the available evidence points to a more complex relationship. Nikkanen et al. [4] show that trust in authorities does not necessarily increase preparedness directly and may have only a limited effect on preparedness measures in some contexts. More broadly, the literature suggests two competing pathways: trust may encourage preparedness when institutions are seen as credible and supportive, but high trust may also reduce households’ sense of personal responsibility if people assume that authorities will manage the crisis on their behalf [4,18]. Related work also points to the importance of institutional support for preparedness capacity, especially when institutions provide coordination, communication, and credible structures for action [19]. In addition, Qiu et al. [2] show that institutional communication itself can shape citizens’ willingness to co-produce preparedness, although the effect depends on how emergency information is framed. In the context of sustainable community resilience, this institutional dimension matters because institutional resilience depends on adaptive governance, coordinated stakeholder mechanisms, transparent decision-making, and inclusive policy implementation [12]. From this, we propose the following hypothesis:

H1a. 

Institutional trust mediates the relationship between perceived disaster preparedness and actual household disaster preparedness.

A second relevant factor is civic participation. Recent disaster risk reduction research increasingly shows that preparedness is not only a private household matter, but also a socially relational process shaped by participation in networks, community groups, and collective preparedness activities. Das et al. [8] define civic participation as activities undertaken beyond immediate personal responsibilities to support community well-being and show that such participation contributes to collective action for disaster risk reduction through communication, planning, resource maintenance, and social-capital formation. Satizábal et al. [20] similarly emphasize that meaningful preparedness depends on relationships among publics, emergency agencies, local government, and community “connectors,” rather than on one-way information delivery alone. This is consistent with Chang et al. [21], who show that merely distributing preparedness materials is insufficient and that effective preparedness support requires active facilitation, discussion, and capacity-building tailored to individual needs. This is also central to sustainable community resilience, because resilient communities depend on community capital, social networks, trust, cooperation, participation, and collective action [22,23]. From this, we propose the following hypothesis:

H1b. 

Civic participation mediates the relationship between perceived disaster preparedness and actual household disaster preparedness.

Within this socially embedded understanding of preparedness, perceived well-being also deserves particular attention. Research has shown that higher well-being, as reflected in quality of life, psychological well-being, and spiritual well-being, can support protective behavior, preparedness intentions, and resilience-related action [24,25,26]. At the same time, some studies suggest that well-being may shape preparedness indirectly through risk perception and coping orientation [27]. In this way, perceived well-being may influence both how prepared households feel and whether they mobilize concrete preparedness practices. Individuals with higher perceived well-being may feel more efficacious, future-oriented, and capable of planning, while lower perceived well-being may reduce motivation, emotional capacity, or willingness to engage in preparedness activities. From a sustainable resilience perspective, perceived well-being is therefore not only an outcome of resilience but also a potential driver of household readiness. From this, we propose the following hypothesis:

H1c. 

Perceived well-being mediates the relationship between perceived disaster preparedness and actual household disaster preparedness.

Institutional trust, civic participation, and perceived well-being represent interrelated social and psychosocial dimensions of preparedness. One plausible pathway is that perceived disaster preparedness influences actual household disaster preparedness through institutional trust and subsequent civic participation. Institutional trust may shape whether households view public authorities as credible sources of guidance, while civic participation may provide the networks and practical opportunities through which preparedness is enacted [2,19]. Accordingly, the following serial pathway is proposed:

H2. 

Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through institutional trust and then civic participation.

Prior research suggests that institutional trust is positively associated with perceived well-being, partly through greater perceived security and confidence in surrounding support structures, while perceived well-being has been linked to stronger disaster preparedness behaviors and protective action [24,28,29]. This reasoning leads to the following serial pathway hypothesis:

H3. 

Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through institutional trust and then perceived well-being.

A similar logic applies to the relationship between civic participation and perceived well-being. Research on community resilience, inclusive participation, and community empowerment suggests that civic participation not only provides information and practical opportunities for preparedness, but also strengthens social connectedness, collective efficacy, and shared problem-solving. In the broader perspective of sustainable community resilience, these participatory and relational capacities matter because resilient communities depend not only on material resources, but also on community capital, social networks, and collective action that can be sustained over time [22,23]. These qualities are closely related to perceived well-being. In turn, individuals with higher perceived well-being may be better able to sustain preparedness-related action because they possess stronger motivational and emotional resources for planning and coping. Thus, civic participation may reinforce actual household disaster preparedness both directly and indirectly through its contribution to perceived well-being. This provides the basis for the following serial hypothesis:

H4. 

Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through civic participation and then perceived well-being.

Taken together, the reviewed literature suggests that institutional trust, civic participation, and perceived well-being should not be treated as isolated correlates of preparedness. Rather, they appear to form an interconnected set of institutional, social, and psychosocial conditions through which perceptions of readiness may be translated into concrete preparedness. From the perspective of sustainable community resilience, this interdependence is important because long-term resilience is shaped by the interaction of institutional capacity, participatory governance, social connectedness, and adaptive community resources rather than by any single factor alone [12,22]. Recent work on co-production, disaster literacy, and participatory risk communication further supports this integrated perspective by showing that preparedness depends not only on what households have, but also on how they understand risk, how they are engaged by institutions, and how they are embedded in social relationships and communication environments [2,30]. This broader logic supports a full serial pathway linking all three mediating factors in sequence:

H5. 

Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through institutional trust, civic participation, and perceived well-being.

The relationship between perceived disaster preparedness and actual household disaster preparedness also needs to be considered in light of sociodemographic characteristics. Existing studies repeatedly show that preparedness varies across gender, age, education, income, homeownership, household composition, and place of residence, although the direction and strength of these effects are not always consistent. Brown et al. [6] found that some factors were more strongly associated with actual than perceived preparedness, while Hasan et al. [13] reported no overall gender differences in preparedness levels but did find gender-specific patterns in predictors. Nikkanen et al. [4] show that age and rural residence were associated with greater preparedness measures, and Cong et al. [31] demonstrate that preparedness is also shaped by wider social vulnerability at the community level. Gülsoy et al. [30] further strengthen this line of evidence by showing that lower education, lower income, and older age were associated with lower disaster literacy and lower disaster preparedness, while disaster experience was associated with higher levels of both. These findings imply that sociodemographic variables should not be treated merely as controls; rather, they may condition the strength of the relationship between perceived and actual disaster preparedness. From the perspective of sustainable community resilience, this is important because resilience capacities are unevenly distributed across social groups, and the effectiveness of institutional, civic, and psychosocial mechanisms may therefore differ according to households’ social positions and resources. Among demographic variables, education and gender were selected as moderators because prior disaster preparedness research suggests that both variables are consistently associated with differences in preparedness-related perceptions, risk interpretation, civic engagement, and preparedness behaviors [4,6,10,13,30]. Education may shape how individuals interpret institutional information and personal responsibility for preparedness, whereas gender differences have been linked to variation in preparedness perceptions and reporting patterns. Other demographic variables, including age, income, occupation, and marital status, were examined in preliminary analyses; however, they did not demonstrate theoretically consistent and statistically meaningful moderation effects in the present study and were therefore not retained in the final moderated mediation models. Accordingly, we propose the following hypothesis:

H6. 

Demographic variables will moderate the effect of perceived disaster preparedness on (a) institutional trust, (b) civic participation, and (c) perceived well-being.

The following hypotheses follow from the expectation that demographic characteristics shape not only levels of preparedness, but also the way institutional, civic, and psychosocial factors are translated into actual preparedness behavior. In this sense, the effects of institutional trust, civic participation, and perceived well-being may vary across social groups:

H7. 

Demographic variables will moderate the effect of institutional trust on actual household disaster preparedness.

H8. 

Demographic variables will moderate the effect of civic participation on actual household disaster preparedness.

H9. 

Demographic variables will moderate the effect of perceived well-being on actual household disaster preparedness.

The final set of hypotheses concerns moderated mediation effects, namely whether the indirect pathways differ across demographic groups.

H10. 

The indirect effect of perceived disaster preparedness on actual household disaster preparedness through institutional trust will differ across demographic variables.

H11. 

The indirect effect of perceived disaster preparedness on actual household disaster preparedness through civic participation will differ across demographic variables.

H12. 

The indirect effect of perceived disaster preparedness on actual household disaster preparedness through perceived well-being will differ across demographic variables.

3. Materials and Methods

3.1. Data Source

This study analyzed cross-sectional data collected as part of a representative omnibus survey of the Lithuanian population. The survey was carried out between 22 February and 5 March 2024 by a public opinion research company, based on a questionnaire coordinated with the researchers. The survey was implemented in compliance with the ICC/ESOMAR International Code on Market, Opinion and Social Research and Data Analytics, which sets recognized ethical and professional standards for survey research, including respondent protection, data confidentiality, and research transparency. Lithuanian residents aged 18 years and older participated in the survey. Respondents were selected through a multistage stratified random sampling procedure.

3.2. Study Participants

The final sample consisted of 550 adults, all of whom reported their gender, with 46.3% male and 53.7% female. Participants ranged in age from 18 to 86 years (M = 48.19, SD = 17.08; median = 48). Education levels were reported by respondents and spanned from primary education to university master’s/doctoral degrees, with professional education (23.2%) being the most common category. Most participants were employed or retired, with occupation data available for 550 individuals; the largest group identified as specialists (29.6%), followed by workers (27.4%) and pensioners (21.4%). Employment sector information also indicated that 47.7% worked in the private sector, 18.6% in the public sector, 1.1% in NGOs, and 32.7% were not employed (e.g., students, retirees, and those managing the household). Household income was reported by respondents and ranged from 0 to 6000, with a mean of 1923.17 and a median of 1800, indicating substantial variability in economic conditions across respondents. Marital status was provided by participants, most of whom were married (55.7%), followed by those who live together (14.8%) and never married (13.2%). The majority of participants identified as Lithuanian, comprising 95.3% of the sample (n = 525). Smaller proportions identified as Russian (1.8%, n = 10) and Polish (2.3%, n = 12), while 0.5% (n = 2) reported belonging to other nationalities. Overall, the nationality data demonstrated that the sample was predominantly Lithuanian in ethnic composition, which reflects the actual ethnic composition of Lithuania’s population. The achieved sample was treated as representative of the Lithuanian adult population within the limits of the applied multistage stratified sampling design.

3.3. Research Instrument

As the data were collected as part of an omnibus survey, the present study included only those questionnaire items that were relevant to its analytical objectives. The research instrument thus comprised measures of actual household disaster preparedness, perceived disaster preparedness, institutional trust, perceived well-being, civic participation, and selected demographic characteristics. Before inclusion in the analysis, the items were reviewed for conceptual fit and combined into a measure.

Actual household disaster preparedness was measured by the presence of disaster-related supplies and arrangements prepared at home. This operationalization is consistent with research that conceptualizes household preparedness as an action- and resource-based construct reflected in concrete household measures such as emergency supplies, emergency plans, and other practical provisions maintained prior to a disaster [1,6]. This measure was based on dichotomous items asking respondents which preparedness-related items they had in their household. These included, for example, backup copies of important documents, non-perishable food supplies for three days for each family member, an evacuation plan for the respondent and the whole family, etc.

Perceived disaster preparedness was assessed using items reflecting subjective readiness, risk awareness, and perceived preparedness-related capabilities. This approach is in line with studies that distinguish perceived preparedness from actual preparedness and treat it as a subjective assessment of how prepared individuals believe they are for an emergency, including confidence in their capacity to undertake preparedness-related action and to cope effectively with disaster conditions [6,32,33].

Institutional trust was measured with items in which respondents evaluated their trust in different institutions, including the Lithuanian Parliament, the Lithuanian Government, the police, etc., on a four-point Likert-type scale. Although the exact institutional list was adopted from the public survey company’s standard omnibus survey instrument rather than designed specifically for this study, it aligns with disaster preparedness research that highlights confidence in authorities, organizational preparedness, and trust in formal response structures as important institutional dimensions shaping preparedness-related beliefs and actions [1,6].

Perceived well-being was assessed using items capturing respondents’ evaluations of (1) life satisfaction in general, (2) satisfaction with their health, and (3) satisfaction with their material circumstances, rated on a five-point Likert-type scale. Higher values indicated more positive self-evaluations of well-being. The three items were treated as indicators of a broader subjective well-being construct reflecting respondents’ perceived quality of life across general, physical, and material domains. This measure reflects the broader resilience literature in which well-being is treated as a relevant psychosocial dimension of adaptive functioning [34].

Civic participation was measured by dichotomous items reflecting participation in civic and community-related activities, such as involvement in the activities of a public organization or movement. Each item captured whether or not the respondent had engaged in a specific activity, and these items were then aggregated to create an overall indicator of civic participation. This approach was intended to capture the behavioral dimension of civic engagement rather than attitudes alone, e.g., participated in community activities in the neighborhood, participated in a civic or political discussion. The selection of the items was guided by the theory on civic participation. This operationalization was consistent with research that treats participation, engagement, collective action, networking, and involvement in community decision-making and community-based processes as important social dimensions of resilience-building and preparedness-related capacity [23].

The analysis also included selected demographic characteristics: education, ethnicity, main occupation, average monthly household income, marital status, and gender.

3.4. Statistical Analysis

The final analytical sample (N = 550) consisted of respondents with complete data on the study variables. Descriptive statistical analysis was conducted for all study variables, and the results are reported as means (M) and standard deviations (SD). The internal consistency of multi-item scales was assessed using Cronbach’s alpha coefficient based on reliability analysis of the constructed indices (

Table 1. The background of variables included in the study.

Model Variable	Code	Construct Measurement Description	Cronbach’s Alpha
Actual household disaster preparedness	ADP	9 dichotomous items	0.713
Perceived disaster preparedness	PDP	3 items on a 5-point Likert scale	0.711
Institutional trust	INT	10 items on a 4-point Likert scale	0.828
Perceived well-being	PWB	3 items on a 5-point Likert scale	0.746
Civic participation	CVP	6 dichotomous items	0.636
Education	EDU	8 ordinal categories	-
Gender	GEN	categorical variable	-

).

The index measuring institutional trust (INT), composed of 10 items, demonstrated good internal consistency (Cronbach’s α = 0.828). The perceived well-being index (PWB), consisting of 3 items, showed acceptable reliability (Cronbach’s α = 0.746). The civic participation index (CVP), based on 6 items, indicated moderate internal consistency (Cronbach’s α = 0.636). The actual household disaster preparedness variable (K34) was constructed as a composite index from 9 dichotomous items reflecting the availability of specific preparedness supplies and demonstrated acceptable internal consistency (Cronbach’s α = 0.713). Overall, the reliability results indicate that the indices are suitable for further statistical analysis.

Pearson’s product–moment correlation analysis was used to examine bivariate relationships among the study variables. To assess group differences, inferential statistical analyses were performed. Independent samples t-tests were applied for variables with two categories, while one-way analysis of variance (ANOVA) was used for variables with more than two categories. When statistically significant differences were identified, post hoc comparisons were conducted using appropriate multiple comparison procedures.

The proposed theoretical models (Figure A1, Appendix B) were tested using moderated mediation analysis, also referred to as conditional process modelling. The analyses were conducted using the PROCESS macro (version 3.5) for SPSS developed by Hayes [35]. Model 6 was applied to test serial mediation effects, while Model 58 was used to examine moderated mediation effects.

The analytical approach enabled the assessment of (1) the direct and indirect effects of perceived disaster preparedness (PDP) on actual household emergency preparedness (K34) through the mediators—institutional trust (INT), perceived well-being (PWB), and civic participation (CVP) as specified in PROCESS Model 6 (Figure A1, Appendix B); and (2) the conditional indirect effects of perceived disaster preparedness (PDP) on actual household emergency preparedness (ADP), moderated by education (EDU) and gender (GEN). These moderators were retained because preliminary analyses indicated meaningful moderation patterns, whereas other demographic variables did not demonstrate statistically significant conditional effects, as specified in PROCESS Model 58 (Figure A2, Appendix B). In these models, the moderating effects of education (EDU) and gender (GEN) on the relationships within the mediation pathways were examined, allowing the strength of the indirect effects to vary across levels of the moderator. Prior to the analyses, the dataset was screened for missing values. Cases with missing responses on variables included in the mediation and moderation models were handled using listwise deletion within the PROCESS procedure. The proportion of missing data was low and did not materially affect the effective sample size or the stability of the estimated coefficients.

Prior to hypothesis testing, the assumptions underlying the regression-based mediation and moderated mediation models were evaluated. Linearity between predictors and outcome variables was assessed through inspection of scatterplots and partial regression plots. Normality of residuals was examined using standardized residual histograms and normal probability (Q–Q) plots. Homoscedasticity was evaluated by visual inspection of residual-versus-predicted value plots. Multicollinearity was assessed using correlation coefficients and variance inflation factor (VIF) statistics, with no evidence of problematic multicollinearity observed. The diagnostic analyses indicated that the assumptions for regression-based PROCESS analyses were adequately satisfied.

All hypotheses were tested within the specified PROCESS models. In addition to regression coefficients and bootstrap confidence intervals, explanatory power was evaluated using R² statistics obtained from the PROCESS regression outputs. Statistical inference for indirect and conditional effects was based on bias-corrected bootstrap confidence intervals (95% CI) generated from 5000 bootstrap samples, following established recommendations for mediation and moderated mediation analysis. Indirect effects were considered statistically significant when the confidence intervals did not include zero. All analyses were conducted in SPSS 29v using PROCESS v3.5.3 [35].

Before testing the hypothesized mediation and moderated mediation models, preliminary analyses were conducted. These analyses provide the descriptive and correlational context for the multivariate results presented below.

4. Results

4.1. Preliminary Analyses

Descriptive statistics and a series of independent samples t-tests and one-way ANOVAs were conducted to examine whether preparedness-related outcomes differed across key demographic groups, including gender (GEN), education (EDU), age (d2), income (d8), marital status (GEN0), occupation (d5), and employment sector (d6). The results indicate that demographic variation is present, although pattern strength varies by predictor. Also, bivariate correlations were conducted to assess preliminary associations among the key constructs (see

Table 2. Relations between study constructs.

	Mean	Std. Dev.	PDP	INT	CVP	PWB	ADP
PDP	2.79	1.425	1
INT	2.37	0.420	0.081	1
CVP	9.68	5.203	−0.147 **	−0.057	1
PWB	2.51	0.675	0.087 *	0.264 **	−0.131 **	1
ADP	2.20	2.021	−0.104 *	−0.171 **	0.153 **	−0.098 *	1

Notes: N = 550. Correlation is significant at the * p < 0.05, ** p < 0.01 level (2-tailed).

).

Perceived disaster preparedness (PDP) showed significant but modest relationships with several variables. It was positively associated with perceived well-being (PWB; p < 0.05) and negatively associated with civic participation (CVP; p < 0.01), suggesting that individuals reporting greater perceived well-being tend to feel more prepared, whereas those more civically engaged may evaluate their preparedness more critically. Although the correlation with institutional trust (INT) was small and non-significant, the direction supported its inclusion in the subsequent mediation model.

Institutional trust demonstrated moderate positive associations with perceived well-being (p < 0.01) and negative associations with actual household disaster preparedness (ADP; p < 0.01), indicating that higher trust may be accompanied by reduced preparedness behaviors. Civic participation was positively associated with actual household disaster preparedness (p < 0.01), while civic participation and perceived well-being were negatively correlated (p < 0.01), reflecting a pattern in which active civic involvement aligns with greater preparedness but lower subjective well-being.

Finally, actual household disaster preparedness exhibited small but significant correlations with perceived preparedness, institutional trust, civic participation, and perceived well-being, underscoring the interconnectedness of psychological and behavioral aspects of disaster readiness.

Taken together, these preliminary correlations underscore the interdependence among perceived preparedness, institutional trust, civic participation, perceived well-being, and objective preparedness behaviors. Importantly, although several correlations among the study constructs were statistically significant, their magnitudes were generally small to moderate and remained well below commonly accepted thresholds for problematic multicollinearity (e.g., r > 0.70). In particular, the association between institutional trust and perceived well-being was modest in size, suggesting that the constructs are related yet empirically distinct. These findings provided the necessary empirical basis for testing the hypothesized mediation and moderated-mediation pathways in the subsequent PROCESS analyses.

4.2. Hypotheses Testing Results

Serial Mediation Effects

A serial multiple-mediator model, PROCESS Model 6 (see Figure A1, Appendix B), was estimated to examine whether perceived disaster preparedness (PDP) influenced actual household disaster preparedness (ADP) indirectly through institutional trust (INT), civic participation (CVP), and perceived well-being (PWB). The explanatory power of the regression equations varied across mediators. Perceived preparedness explained only a small proportion of the variance in institutional trust (R² = 0.005), whereas the model explained 1.8% of the variance in civic participation (R² = 0.018), 8.0% of the variance in perceived well-being (R² = 0.080), and 4.7% of the variance in actual household disaster preparedness (R² = 0.047). Although the explained variance was modest, this pattern is consistent with disaster preparedness research, where preparedness behaviors are shaped by multiple interacting institutional, psychological, social, and contextual influences.

Direct Effect. The direct pathways from perceived disaster preparedness (PDP) to each mediator and to the final behavioral outcome provide critical insight into the extent to which preparedness perceptions operate through social-attitudinal mechanisms rather than through direct behavioral translation. The estimated direct effects are presented in

Table 3. Direct effects of perceived disaster preparedness.

Direct Effect Pathway	Estimated Effect (β)	SE	Stand. Effect (β)	t	p	LLCI	ULCI
PDP → INT	0.022	0.013	0.071	1.666	0.096	−0.004	0.048
PDP → CVP	−0.472	0.166	−0.121	−2.842	0.005	−0.798	−0.146
PDP → PWB	0.027	0.021	0.053	1.285	0.199	−0.014	0.068
PDP → ADP	−0.098	0.064	−0.065	−1.538	0.125	−0.224	0.027

Notes: X = PDP (Perceived disaster preparedness), Y = ADP (Actual household emergency preparedness). M1 = INT (Trust), M2 = CVP (Civic participation), M3 = PWB (Well-being).

and interpreted below.

The direct effect of perceived preparedness on actual household disaster preparedness was non-significant (PDP, β = −0.098, p = 0.125). Such a pattern aligns with a growing body of research showing that perceived preparedness does not necessarily translate into actual preparedness behaviors, but may depend on intervening psychological, social-attitudinal, and civic factors such as emotional well-being, institutional trust, and civic participation [6,36]. High perceived readiness may even reduce motivation to undertake additional preparedness actions, as individuals may assume that their current level of readiness is sufficient; this is a phenomenon often described as risk normalization or a false sense of security [9].

Thus, the non-significant direct effect (c′ path, PDP → ADP) implies that the influence of perceived preparedness is not direct but operates primarily through indirect social-attitudinal pathways, consistent with the logic of serial mediation models.

Perceived Preparedness → Civic participation (PDP → CVP) was the only significant direct pathway (PDP, β = −0.472, p = 0.005, Table 2). The negative association indicates that individuals who feel more personally prepared tend to exhibit lower levels of civic participation. This is theoretically plausible, because disaster risk reduction literature shows that preparedness can be approached either as an individual-level activity or as a collective process rooted in community participation, shared action, and collaboration with other stakeholders [8]. The same literature also stresses that focusing only on individual preparedness may be insufficient, because stronger resilience depends on communities’ capacity to mobilize resources, participate in decision-making, and act collectively [8,12]. From this perspective, when preparedness is interpreted primarily as a personal responsibility rather than a shared community goal, individuals may feel less compelled to engage in civic or collective preparedness activities. This helps explain why civic participation plays a central mediating role in the model.

Perceived Preparedness → Perceived Well-Being (PDP → PWB) was non-significant (β = 0.027, p = 0.199, Table 2). This finding indicates that feeling prepared does not substantially improve or impair perceived well-being. More broadly, resilience literature suggests that well-being and longer-term adaptive capacity depend on the interaction of social support, institutional inclusion, governance, and community resources rather than on preparedness perceptions alone [12]. Thus, well-being appears more likely to function as a downstream mediator shaped by wider social and institutional processes than by preparedness perception directly.

Indirect Effects. The serial multiple-mediator model (Model 6) evaluated whether perceived disaster preparedness (PDP) affects actual household disaster preparedness (ADP) through multiple social-attitudinal mediators: institutional trust (INT), civic participation (CVP), and perceived well-being (PWB). The results (

Table 4. Indirect effects in serial multiple-mediator model.

Estimated Pathways	Estimated Effects (β)	BootSE	Bootstrapped CI (95%)
			LLCI	ULCI
Total effect	−0.141	0.064	−0.268	−0.015
Direct effect
PDP →ADP	−0.098	0.064	−0.224	0.027
Indirect effects
H1a: PDP → INT → ADP	−0.013	0.009	−0.034	0.002
H1b: PDP → CVP → ADP	−0.022	0.012	−0.049	−0.003
H1c: PDP → PWB → ADP	−0.005	0.006	−0.018	0.003
H2: PDP → INT → CVP → ADP	−0.001	0.001	−0.003	0.000
H3: PDP → INT → PWB → ADP	−0.001	0.002	−0.006	0.001
H4: PDP → CVP → PWB → ADP	−0.001	0.001	−0.004	0.000
H5: PDP → INT → CVP → PWB → ADP	0.000	0.000	0.000	0.000
Total indirect effect	−0.043	0.017	−0.081	−0.013

Notes: N = 550 is the total number of the dataset. ADP is the dependent variable in the model using the Process v3.5 macro Model 6.

) provide strong support for H1, confirming that the influence of perceived preparedness is primarily indirect and operates through social-attitudinal pathways rather than a direct behavioral mechanism.

The results (see Table 3) of the serial multiple-mediator model provide strong empirical confirmation that the association between perceived disaster preparedness (PDP) and actual household disaster preparedness (ADP) is transmitted primarily through indirect social-attitudinal mechanisms, thereby supporting H1 (see

Table A1. Hypotheses tested in Model 6.

Hypothesis ID	Hypothesis
H1	Perceived disaster preparedness (PDP) is associated with actual household disaster preparedness (ADP) through indirect social-attitudinal pathways rather than through a direct effect.
H1a	Indirect effect through M1 only: Institutional trust (INT) mediates the relationship between perceived disaster preparedness and actual household preparedness (PDP → INT → ADP).
H1b	Indirect effect through M2 only: Civic participation (CVP) mediates the relationship between perceived disaster preparedness and actual household preparedness (PDP → CVP → ADP).
H1c	Indirect effect through M3 only: Perceived well-being (PWB) mediates the relationship between perceived disaster preparedness and actual household preparedness (PDP → PWB → ADP).
H2	Serial mediation M1 → M2: Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through institutional trust and then civic participation (PDP → INT → CVP → ADP).
H3	Serial mediation M1 → M3: Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through institutional trust and then perceived well-being (PDP → INT → PWB → ADP).
H4	Serial mediation M2 → M3: Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through civic participation and then perceived well-being (PDP → CVP → PWB → ADP).
H5	Full serial pathway M1 → M2 → M3: Perceived disaster preparedness is indirectly associated with actual household disaster preparedness through the full chain of institutional trust → civic participation → perceived well-being (PDP → INT → CVP → PWB → ADP).

, Appendix A). Although the total effect of perceived preparedness on actual household disaster preparedness was statistically significant (β = −0.141, 95% CI [−0.268, −0.015]), the direct effect became non-significant once the mediators were included (β = −0.098, p = 0.125), indicating that the link between subjective preparedness and actual behavior is fully mediated. The significant total indirect effect (β = −0.043, 95% CI [−0.081, −0.013]) demonstrates that changes in actual preparedness behavior occur through social-attitudinal pathways rather than through direct action, reaffirming the central proposition of H1 that preparedness perceptions alone are insufficient for behavioral readiness. Importantly, the explanatory power of the full mediation model remained modest (R² = 0.047), indicating that although the identified mediators contribute meaningfully to preparedness behavior, substantial variance remains attributable to additional contextual, social, and situational factors not captured within the present model.

Examination of the individual mediation paths shows mixed support for H1a–H1c. The indirect path through institutional trust (H1a) was small and non-significant (β = −0.013, CI [−0.034, 0.002]), suggesting that perceived preparedness neither meaningfully enhances nor diminishes institutional trust in a way that affects actual preparedness (see Table 3). Similarly, the indirect path via well-being (H1c) was not significant (β = −0.005, CI [−0.018, 0.003], Table 3), indicating that perceived well-being does not mediate the link between perceived and actual preparedness.

In contrast, H1b was the only simple mediation hypothesis that received clear empirical support, with the indirect effect through civic participation emerging as significant (β = −0.022, 95% CI [−0.049, −0.003], Table 3). The finding that civic participation emerged as the only significant mediator suggests that the translation of perceived preparedness into actual preparedness depends more strongly on active social engagement and participation-related mechanisms than on institutional trust or subjective well-being. Because civic participation involves concrete actions such as information exchange and engagement in preparedness-related activities, it acts as a direct link between perceptions and household preparedness behavior. This finding suggests that individuals who feel more tend to exhibit lower civic participation, which in turn predicts lower actual preparedness, an effect consistent with theoretical arguments that personal preparedness can reduce participation in community-based preparedness actions. Overall, the findings supported H1 and H1b, whereas H1a and H1c hypotheses were not supported.

The serial mediation hypotheses (H2–H5) were not supported: the indirect paths involving sequential combinations of institutional trust, civic participation, and perceived well-being were all small and statistically non-significant. These results indicate that while the mediators each play distinct roles in the broader preparedness process, the psychological transmission mechanism linking perceived preparedness to behavioral preparedness is driven predominantly by the single mediating effect of civic participation, rather than by serial chains of attitudinal factors. Collectively, the findings underscore that subjective readiness influences actual preparedness only indirectly, primarily through social participation, highlighting the importance of community-level processes in shaping individual preparedness outcomes. This provided the conceptual foundation for testing whether these pathways differed across demographic groups.

4.3. Moderation and Moderated Mediation by Demographic Variables

To examine whether the mediation pathways varied as a function of demographic variables, a moderated serial mediation model, PROCESS Model 58, was estimated. Model 58 tests moderation on both the X → M and M → Y components and evaluates whether demographic variables have effects. The conditional process analyses showed that most demographic variables, including age, occupation, employment sector, income, and marital status, did not significantly moderate the indirect effects of perceived to actual disaster preparedness. For these demographic factors, all conditional indirect effects included zero and all indices of moderated mediation were non-significant, indicating that the underlying psychosocial mechanisms were largely invariant across groups. In contrast, the analysis conducted for education and gender groups demonstrated selective and statistically meaningful moderation effects.

4.3.1. Role of Education as a Moderator

The role of education (EDU) as a moderator of the indirect effects was assessed using the PROCESS v3.5 macro-Model 58 developed for the moderated mediation analysis. This regression model was chosen to test whether education altered the strength of the relationships between the socio-attitudinal mediators (institutional trust (INT), civic participation (CVP), and perceived well-being (PWB)) and actual household emergency preparedness (ADP). The analysis results are presented below (see

Table 5. Regression Results for Predicting Actual Household Disaster Preparedness (ADP) with moderation effects of education (EDU).

Estimated Pathways	Predictor	B	SE	t	p	LLCI	ULCI
	constant	2.188	0.088	24.842	0.001	2.015	2.361
PDP → ADP	PDP	−0.096	0.064	−1.495	0.135	−0.222	0.030
INT → ADP	INT	–0.592	0.215	−2.751	0.006	−1.014	−0.169
CVP → ADP	CVP	0.046	0.016	2.790	0.005	0.014	0.078
PWB → ADP	PWB	−0.152	0.134	−1.137	0.256	−0.416	0.111
EDU → ADP	EDU	0.009	0.052	0.173	0.863	−0.093	0.111
H7: INT × EDU → ADP	INT × EDU	−0.248	0.130	−1.901	0.048	−0.504	−0.008
H8: CVP × EDU → ADP	CVP × EDU	0.009	0.009	0.965	0.335	−0.009	0.028
H9: PWB × EDU → ADP	PWB × EDU	0.085	0.079	1.077	0.282	−0.070	0.239
Model Summary	R	R-sq	MSE	F	df1	df2	p
	0.236	0.056	3.966	3.999	8	542	0.001

Notes: N = 550 is the total number of the study dataset. PDP (Perceived preparedness), INT (Institutional trust), CVP (Civic participation), PWB (Perceived well-being), EDU (Education). The following variables were mean-centered prior to analysis: EDU, PDP, INT, CVP, PWB.

).

The overall model predicting actual household disaster preparedness was statistically significant (F(8, 541) = 3.999, p < 0.001), indicating that the predictors collectively provide explanatory power beyond chance. Although the model explained a modest proportion of variance in actual preparedness (R² = 0.056), such effect sizes are common in disaster preparedness research because preparedness behaviors are influenced by numerous institutional, social, psychological, and situational determinants operating simultaneously [1,12].

Education did not exert a direct influence on actual preparedness (β = 0.009, p = 0.863), nor did it significantly moderate the effects of civic participation or perceived well-being, providing no support for hypothesis H8 or H9 (see

Table A2. Hypotheses for Model 58 (Moderated Serial Mediation).

Hypothesis ID	Hypothesis
H6a	Demographic variables will moderate the effect of perceived disaster preparedness on institutional trust (PDP × d_i → INT).
H6b	Demographic variables will moderate the effect of perceived disaster preparedness on civic participation (PDP × d_i → CVP).
H6c	Demographic variables will moderate the effect of perceived disaster preparedness on perceived well-being (PDP × d_i → PWB).
H7	Demographic variables will moderate the effect of institutional trust on actual household disaster preparedness (INT × d_i → ADP).
H8	Demographic variables will moderate the effect of civic participation on actual household disaster preparedness (CVP × d_i → ADP).
H9	Demographic variables will moderate the effect of perceived well-being on actual household disaster preparedness (PWB × d_i → ADP).
H10	The indirect effect of perceived disaster preparedness on actual household disaster preparedness through institutional trust will differ across demographic variables (moderated mediation: PDP → INT → ADP, moderated by d_i).
H11	The indirect effect of perceived disaster preparedness on actual household disaster preparedness through civic participation will differ across demographic variables (PDP → CVP → ADP, moderated by d_i).
H12	The indirect effect of perceived disaster preparedness on actual household disaster preparedness through perceived well-being will differ across demographic variables (PDP → PWB → ADP, moderated by d_i).

, Appendix A). Specifically, the interaction between civic participation and education (CVP × EDU) was not significant (β = 0.009, p = 0.335), indicating that the positive effect of civic participation on actual preparedness is similar across all education levels. Likewise, the perceived well-being × education interaction (PWB × EDU) was not significant (β = 0.085, p = 0.282), suggesting that well-being predicts preparedness uniformly across the educational spectrum.

The interaction between institutional trust and education (INT × EDU), corresponding to hypothesis H7 (see Table A2, Appendix A), was statistically significant (β = –0.248, p = 0.048), indicating that education moderated the relationship between institutional trust and actual household disaster preparedness. The conditional effects analysis demonstrated that the relationship between institutional trust and actual preparedness differed across educational levels. At low levels of education (−1 SD), institutional trust exhibited the strongest negative association with preparedness (β = −0.493, SE = 0.152). At the mean level of education, the negative relationship became weaker (B = −0.210, SE = 0.118), while at high levels of education (+1 SD), the association remained negative but was comparatively attenuated (β = −0.073, SE = 0.119). These findings indicate that the negative association between institutional trust and preparedness behaviors weakened as education increased (Figure 1).

These results suggest that higher educational attainment reduces the strength of the negative relationship between institutional trust and preparedness behaviors. One possible explanation is that highly educated individuals may place greater confidence in institutional competence and emergency-management systems, thereby perceiving less need for additional household-level preparedness. In this sense, education may strengthen the institutional substitution effect, whereby confidence in public capacity partially replaces individual preparedness initiative.

Figure 1 visually demonstrates the moderating role of education in the institutional trust pathway. At lower levels of education, institutional trust exhibited only a weak association with preparedness behaviors. In contrast, among respondents with higher educational attainment, greater institutional trust was associated with substantially lower levels of actual preparedness. Although the interaction effect was modest in magnitude, the pattern remained statistically significant and theoretically meaningful.

No other indirect pathways were moderated by education. These findings suggest that hypothesis H6a was supported for the institutional trust mediator (M1), whereas no evidence of moderation emerged for mediators M2 or M3. Consequently, hypotheses H6b and H6c must be rejected, indicating that the moderating influence of education is narrow and mechanism-specific rather than broad, shaping only how institutional trust operates within the actual preparedness. The conditional indirect analysis results for statistically significant causal relationships are presented in

Table 6. Conditional indirect effects of education evaluated by using the PROCESS v3.5 macro-Model 58.

Moderator Education (EDU)		Effect	BootSE	Boot LLCI	BootULCI
Conditional indirect effects (PDP → INT → ADP)
M−1SD	−1.724	−0.003	0.009	−0.021	0.016
M	0.000	−0.015	0.011	−0.040	0.000
M+1SD	1.724	−0.035	0.025	−0.104	−0.003
Moderator Education (EDU)		Effect	BootSE	Boot LLCI	BootULCI
Conditional indirect effects (PDP → CVP → ADP)
M−1SD	−1.724	−0.012	0.015	−0.049	0.010
M	0.000	−0.020	0.012	−0.048	−0.001
M+1SD	1.724	−0.030	0.022	−0.077	0.006
Moderator Education (EDU)		Effect	BootSE	Boot LLCI	BootULCI
Conditional indirect effects (PDP → PWB → ADP)
M−1SD	−1.724	−0.005	0.012	−0.031	0.017
M	0.000	−0.005	0.006	−0.019	0.004
M+1SD	1.724	0.000	0.011	−0.026	0.021
Conditional effects of the focal predictor (INT → ADP)
Moderator Education (EDU)		Effect (S.E.)	t	Boot LLCI	BootULCI
(IV → ME3)
M−1SD	−1.724	−0.493 (0.152)	−0.531	−0.773	0.444
M	0.000	−0.210 (0.118)	−2.751 ***	−1.014	−0.169
M+1SD	1.724	0.073 (0.119)	−3.261 ***	−1.632	−0.405
Moderated mediation index (MMI)		BootSE		Boot LLCI	BootULCI
−0.079		0.039		−0.167	−0.019

Note: Values for moderator education (EDU) are presented for ±one SD from the mean; Bootstrap sample size = 5000. Statistical significance: *** p < 0.001. Boot LLCI, boot lower level of confidence interval; BootULCI, upper level of confidence interval.

.

Taken together, these results suggest that education selectively moderates the preparedness process, influencing the role of trust but not the roles of civic participation or perceived well-being. Thus, only hypothesis H7 receives empirical support, whereas H8 and H9 are not supported. Consequently, the moderated mediation results indicate that hypothesis H10, predicting demographic variation in the indirect effect through trust, was supported for education as a moderator. In contrast, hypotheses H11 and H12, which predicted moderating effects on the civic participation and perceived well-being pathways, were not supported, as neither pathway showed statistically significant conditional effects across educational levels. These findings reveal that the moderating role of education is mechanism-specific rather than general, influencing only how institutional trust functions within the preparedness process, while leaving the civic participation and perceived well-being pathways unaffected. Overall, the pattern demonstrates that education shapes actual preparedness indirectly by altering the dynamics of institutional trust, but does not modify the socio-attitudinal effects of civic participation or perceived well-being. Thus, H7 received support for the institutional trust pathway, whereas H8 and H9 were not supported for education-based moderation.

4.3.2. Role of Gender as a Moderator

In addition, the PROCESS v3.5 macro-Model 58 was used to test the role of gender (GEN_Male) as a moderator of the indirect effects. The moderated mediation analysis showed the gender uniquely qualifies the civic participation relationship, highlighting an important demographic nuance in reporting actual household disaster preparedness. The overall model predicting actual preparedness was statistically significant, F(8, 542) = 5.365, p < 0.001, explaining 7.3% of the variance in actual preparedness (ADP). The moderated model including gender explained 7.3% of the variance in actual preparedness (R² = 0.073), indicating modest but meaningful explanatory power for the conditional process model. Among the direct predictors, institutional trust (INT) negatively predicted actual preparedness (β = −0.594, p = 0.005), whereas civic participation (CVP) showed a positive effect (β = 0.052, p = 0.002). Gender itself was a significant predictor (β = 0.526, p = 0.002), indicating that males reported higher levels of actual preparedness. Remarkably, the interaction between civic participation and gender (CVP × GEN_Male) was statistically significant (B = 0.082, p = 0.013), revealing that the effect of civic participation on actual preparedness differs by gender. The interactions between institutional trust and gender and between well-being and gender did not reach significance. Consequently, only hypothesis H8 receives empirical support, whereas H7 and H9 are not supported. Moreover, these findings suggest that respondent gender shapes how some factors, particularly civic participation, are associated with reported household disaster preparedness. This points to the importance of demographic heterogeneity not only in preparedness itself, but also in how household preparedness is reported (

Table 7. Regression Results for Predicting Actual Household Disaster Preparedness (ADP) with moderation effects of gender.

Estimated Pathways	Predictor	B	SE	t	p	LLCI	ULCI
	constant	2.163	0.084	25.685	0.000	1.997	2.328
PDP → ADP	PDP	−0.056	0.064	−0.876	0.382	−0.183	070
INT → ADP	INT	−0.594	0.213	−2.793	0.005	−1.012	−0.176
CVP → ADP	CVP	0.052	0.016	3.184	0.002	0.020	0.084
PWB → ADP	PWB	−0.174	0.131	−1.328	0.185	−0.431	0.083
GEN → ADP	GEN	0.526	0.173	3.041	0.002	0.186	0.865
H7: INT × GEN → ADP	INT × GEN	0.213	0.423	0.503	0.615	−0.618	1.044
H8: CVP × GEN → ADP	CVP × GEN	0.082	0.033	2.492	0.013	0.017	0.146
H9: PWB × GEN → ADP	PWB × GEN	−0.092	0.271	−0.342	0.733	−0.624	0.439
Model Summary	R	R-sq	MSE	F	df1	df2	p
	0.271	0.073	3.886	5.365	8	542	0.000

Notes: N = 550 is the total number of the study dataset. PDP (Perceived preparedness), INT (Institutional trust), CVP (Civic participation), PWB (Perceived well-being), GEN (gender). The following variables were mean-centered prior to analysis: GEN_Male, PDP, INT, CVP, PWB.

).

Tests of the highest order conditional interactions indicated that gender significantly moderated only the relationship between civic participation (M2) and the outcome (ADP, F(1, 542) = 6.211, p = 0.013), whereas the interactions for institutional trust (M1, F(1, 542) = 0.253, p = 0.615), and perceived well-being (M3, F(1, 542) = 0.117, p = 0.733) were not significant. These findings suggest that hypothesis H6b was supported for the civic participation mediator (M2), whereas no evidence of moderation emerged for mediators M1 or M3. Consequently, hypotheses H6a and H6c must be rejected.

Furthermore, the conditional indirect effects analysis revealed that the mediating role of civic participation (CVP) differed meaningfully by gender, and for men the indirect effect was considerably stronger. Additionally, gender did not moderate the IV → M1 or IV → M3 pathways (M1 × X: F(1, 541) = 2.218, p = 0.137; M3 × X: F(1, 541) = 1.116, p = 0.291), gender did significantly moderate the IV → M2 (civic participation) path (F(1, 547) = 2.620, p = 0.010). Moreover, the significant moderation of the M2 → Y relationship produced a meaningful conditional indirect effect through the civic participation pathway. This finding shows that gender does influence the strength of the preparedness → civic participation relationship, although the direction and magnitude of the moderation varied across gender groups. Thus, H6b was supported for the civic participation mediator (M2), while no evidence of moderation appeared for the M1 or M3 mediators (see

Table 8. Conditional Indirect Effects by gender (PROCESS Model 58).

Moderator Gender (GEN_Male)	Effect	Boot SE	Boot LLCI	BootULCI
Conditional indirect effects (PDP → CVP → ADP)
Female (GEN = −0.425)	−0.005	0.009	−0.025	−0.003
Male (GEN = +0.575)	−0.084	0.037	−0.171	−0.025
Conditional effects of the focal predictor
Moderator Gender (GEN_Male)	Effect (S.E.)	t	Boot LLCI	BootULCI
(M2 → CVP)
Female (GEN = −0.425)	0.017 (0.021)	2.808	0.024	0.058
Male (GEN = +0.575)	0.099 (0.025)	3.893	0.049	0.149
Moderated mediation index (MMI)	BootSE		Boot LLCI	BootULCI
−0.079	0.038		−0.169	−0.019

Note. Values represent conditional indirect effects estimated using 5000 bootstrap samples. Gender was effect-coded and mean-centered for analysis (Female = −0.425; Male = +0.575). LLCI = lower-limit confidence interval; ULCI = upper-limit confidence interval.

).

Turning to hypothesis H8, which proposed that demographic variables would moderate the effect of civic participation on actual preparedness, the results provided clear support. The interaction between civic participation and gender predicting actual preparedness (CVP × GEN_Male → ADP) was statistically significant, indicating that the positive association between civic participation and actual preparedness differs for men and women. Specifically, the conditional effects analysis further demonstrated clear gender differences in the civic participation pathway. Among women, civic participation showed only a modest positive association with actual preparedness (β = 0.017, SE = 0.021), whereas among men the association was substantially stronger (β = 0.099, SE = 0.025). These findings indicate that increases in civic participation were associated with considerably greater increases in preparedness behaviors among male respondents (see Figure 2). The findings further suggest that civic participation may be most behaviorally effective among male respondents, for whom community engagement translated more strongly into concrete preparedness actions.

Finally, among the three hypothesized moderated mediation pathways (H10–H12), only the civic participation pathway demonstrated evidence of demographic moderation. Neither institutional trust (H10) nor perceived well-being (H12) showed significant moderated indirect effects, as indicated by non-significant indices of moderated mediation. In contrast, the civic participation pathway (H11) exhibited a significant index of moderated mediation, reflecting that gender meaningfully alters the strength of the indirect effect from perceived to actual preparedness through civic participation. Accordingly, gender significantly moderated the civic participation pathway, providing support for H8 in the gender-specific model, whereas no support emerged for H7 or H9 in relation to gender. To improve clarity,

Table A3. Summary of hypothesis testing results.

Hypothesis	Description	Result
H1	PDP indirectly affects ADP through mediators	Partially supported
H1a	Mediation through institutional trust	Not supported
H1b	Mediation through civic participation	Supported
H1c	Mediation through perceived well-being	Not supported
H2	Serial mediation via INT → CVP	Not supported
H3	Serial mediation via INT → PWB	Not supported
H4	Serial mediation via CVP → PWB	Not supported
H5	Full serial mediation via INT → CVP → PWB	Not supported
H6	Demographic moderation of PDP → mediators	Not supported
H7	Demographic moderation of INT → ADP	Supported (education)
H8	Demographic moderation of CVP → ADP	Supported (gender)
H9	Demographic moderation of PWB → ADP	Not supported
H10	Moderated mediation via institutional trust	Supported (education)
H11	Moderated mediation via civic participation	Supported (gender)
H12	Moderated mediation via perceived well-being	Not supported

(see Appendix A) summarizes the outcomes of all hypothesis tests conducted in the present study. As shown in Table A3, support was strongest for the civic participation pathway and for the conditional effects involving education and gender, whereas institutional trust and perceived well-being demonstrated more limited mediation effects.

5. Discussion

This study’s results are important not only for explaining household disaster preparedness, but also for understanding preparedness from the perspective of sustainable community development and public management. Community resilience depends on the interaction of household capacities, community relationships, and institutional support, while preparedness is increasingly understood as a co-produced public good rather than solely a matter of private responsibility [1,2]. The findings offer several theoretical and empirical contributions, particularly regarding the role of civic participation, institutional trust, perceived well-being, gender, and education in shaping actual preparedness.

Consistent with the first hypothesis, perceived preparedness did not directly predict actual household disaster preparedness. This aligns with research showing that perceived and actual household disaster preparedness are often only weakly aligned, and that feeling prepared does not necessarily translate into concrete action [6]. Prior studies suggest that preparedness perceptions may fail to generate action when psychological, social, and contextual mechanisms intervene [3,37,38]. The present results reinforce this interpretation by demonstrating that preparedness operates mainly through indirect pathways rather than through a direct conversion of perceived preparedness into actual preparedness. This distinction is important for sustainable community resilience, which depends not only on perceived readiness but on the capacity to convert beliefs into sustained individual and collective action [5,12].

Among the three mediators examined, civic participation emerged as the only significant mediator linking perceived to actual household disaster preparedness. This finding is consistent with scholarship documenting the central role of civic participation and collective action in motivating protective behaviors and strengthening disaster readiness [8,9]. Civic participation appears to function as a social mobilizer that converts abstract preparedness beliefs into concrete actions by embedding individuals in collective norms of responsibility, information exchange, and community readiness. In resilience terms, such participation helps build social capital, solidarity, and locally available capacities that are essential for sustainable community resilience before disasters occur [5,8]. In contrast, higher perceived preparedness was associated with lower civic participation, suggesting that individuals who consider themselves ready may disengage from community preparedness activities. This is theoretically consistent with community engagement theory, which posits that preparedness depends on reciprocal relationships among citizens, social networks, and civic agencies rather than on individual judgment alone [9]. The absence of mediation effects for institutional trust and perceived well-being suggests that these constructs operate more distally from actual preparedness behavior. Although institutional trust and well-being may shape general attitudes, confidence, or resilience perceptions, they do not necessarily produce concrete preparedness actions at the household level. In contrast, civic participation reflects active involvement in social networks, collective problem-solving, and community-based activities that are directly connected to preparedness-related practices. This distinction may explain why civic participation alone emerged as a statistically meaningful mediator in the present study.

Institutional trust did not mediate the preparedness-behavior relationship, although previous studies show that trust plays an important role in disaster risk reduction and can support preparedness, risk awareness, and resilience-building activities [3]. In the present study, the indirect path through institutional trust was small and non-significant, suggesting that institutional trust may matter conditionally rather than uniformly. At the same time, the association between institutional trust and actual household disaster preparedness was not uniform, suggesting that its role may be more complex than a consistently positive interpretation would suggest.

One possible explanation is that high institutional trust may, under some conditions, create a sense of security, potentially leading households to rely more on public authorities for disaster management rather than taking individual preparedness measures. Although institutional trust is commonly viewed as beneficial for public cooperation and policy legitimacy, very high confidence in institutional response capacity may unintentionally reduce individuals’ motivation to undertake personal preparedness actions. Individuals who strongly trust governmental and emergency-management institutions may in some contexts assume that authorities will provide sufficient protection and support during crises, thereby transferring responsibility for preparedness from the household to public institutions. In this sense, institutional trust may operate not only as a resilience-supporting resource, but also as a mechanism that reduces perceived personal responsibility for preparedness.

This interpretation is consistent with studies showing that trust in institutions can influence compliance with public guidelines while also reducing the perceived need for personal action [39,40]. Thus, the present findings do not necessarily contradict previous research but rather suggest that the effects of trust may depend on whether it reinforces household initiative or substitutes for it. This is also consistent with prior research suggesting that preparedness may decline when citizens perceive emergency management primarily as a governmental responsibility rather than a shared responsibility between institutions and households [3,9]. From a sustainable community resilience perspective, trust remains important, but resilience is likely to be stronger when institutional trust is complemented by participation, empowerment, and community capacity rather than substituted for them.

Similarly, perceived well-being did not serve as a significant mediator, indicating that well-being may play a more distal role in actual household preparedness. Research on emergencies and civic participation more often links well-being to social cohesion, solidarity, coping, and resilience during emergencies than to pre-event household preparedness behaviors themselves [37]. The present findings therefore underscore the need to distinguish between drivers of preparedness intentions and drivers of preparedness actions, as these may be led by different psychological and social mechanisms. This distinction also matters for sustainable community resilience, because well-being may be better understood as an outcome of supportive social and civic conditions than as a direct trigger of household preparedness action.

The moderated mediation analyses expand theoretical understanding by demonstrating that indirect pathways differ across demographic groups. Education moderated the institutional-trust pathway: individuals with higher education showed a stronger negative association between institutional trust and actual household disaster preparedness. This pattern aligns with research showing that the relationship between education and trust is neither linear nor uniform. Hakhverdian and Mayne [11] show that education interacts with institutional context to shape trust in different directions depending on governance quality and performance. This interpretation is broadly consistent with our finding that highly educated individuals may place greater weight on institutional capacity, thereby reducing their own preparedness behaviors.

Gender moderated the civic participation pathway, with civic participation showing a stronger positive association with reported household preparedness among male respondents than among female respondents. This is consistent with evidence that preparedness-related perceptions and reported behaviors may vary by gender [10]. In this context, the observed difference may reflect gendered patterns in how respondents perceive and report actual household preparedness, rather than differences in the household’s preparedness per se.

Overall, the results demonstrate that household disaster preparedness is embedded in a broader social, institutional, and demographic context. Institutional trust, civic participation, perceived well-being, gender, and education shape how perceived preparedness is linked to actual preparedness. These findings resonate with research emphasizing that preparedness depends on both individual and community-based resources, including social capital, participation, and shared capacities [5]. More broadly, they suggest that sustainable community resilience is strengthened not simply by increasing perceived readiness, but by cultivating the social relationships, civic participation, and institutional conditions that enable households to act on that readiness.

Implications. These findings have several implications for public management. First, disaster preparedness should be managed not only as a matter of household action, but also as a process that is strengthened through civic participation. For public agencies, this means investing in community networks, local organizations, and participatory arrangements that can support participation and preparedness action [9,38]. Second, communication with highly educated groups should be designed meaningfully so that confidence in institutional capacity does not reduce household preparedness motivation. Public risk communication should frame preparedness as a shared responsibility between households and institutions, while avoiding excessive reliance on institutional response capacity alone. The findings further indicate that preparedness strategies may require differentiation across social groups, particularly where civic participation more strongly translates into preparedness behaviour. From a public management perspective, these results highlight the importance of balancing institutional trust with the promotion of active household preparedness and recognizing that preparedness perceptions and behaviours may vary across respondent groups.

Limitation. One limitation of this study is its cross-sectional design, which means that firm conclusions about causal direction cannot be drawn. While the analyses identified statistically significant associations and indirect pathways, they could not establish the temporal ordering of the variables. Further longitudinal or experimental studies are needed to assess causality more directly.

Second, although the study included several important variables related to household resilience, its analytical scope remained relatively broad, which may have limited the detection of more nuanced patterns across contextual conditions. Several preparedness-related factors, including geographic location, housing conditions, insurance status, household composition, prior disaster experience, and community-level social capital, were not included. Future research would benefit from integrating a wider range of contextual and household-level variables in order to better capture the heterogeneity of preparedness-related behaviors.

A third limitation concerns the use of self-reported measures, which may be affected by social desirability, recall error or differences in how respondents interpret preparedness-related questions. A fourth limitation is that the study was conducted in one national context (Lithuania), which may restrict the generalisability of the findings to countries with different institutional, cultural and safety environments.

Future research. Longitudinal and cross-cultural studies should examine how preparedness perceptions evolve over time and under different institutional conditions, and how these changes contribute to sustainable community resilience. Future work might also incorporate additional moderators such as political ideology, hazard experience, social networks, or place-based identity, as these factors may shape not only household preparedness but also the broader social and institutional capacities. Qualitative studies may further deepen understanding of how individuals interpret civic roles, institutional obligations, and preparedness responsibilities, which may remain partly hidden in quantitative models.

6. Conclusions

This study suggests that household disaster preparedness is best understood as a socially embedded process rather than as feeling prepared. Perceived preparedness was associated with actual preparedness through indirect pathways, especially through civic participation, while institutional trust and perceived well-being did not emerge as significant mediators. The findings show that these pathways vary across demographic groups, particularly by education and gender, indicating that preparedness does not develop uniformly across respondent groups. Overall, the study indicates that household disaster preparedness is associated less with perceived preparedness itself than with the social pathways through which it is translated into action, particularly civic participation, and with demographic differences in how these pathways operate. For governing bodies, this means that preparedness efforts should raise household awareness and encourage civic participation through targeted communication with different social groups. For NGOs and community organizations, the findings emphasise the importance of fostering local engagement and community-based initiatives, in order to translate preparedness awareness into actual household action. This is important for improving household preparedness and linking household action with civic participation.