Introduction

The concept of resilience is widely used in both academic and policy contexts. Originally derived from physics, resilience has been integrated into various fields of study. In 1999, Mileti (1999) first introduced the term “disaster resilience,” which refers to a region’s ability to withstand extreme disasters that cause systemic destruction, injuries, and declines in productivity and quality of life, without depending on external aid. The Sendai Framework emphasizes the importance of investing in disaster risk reduction to enhance resilience. The United Nations Office for Disaster Risk Reduction (UNDRR) defined resilience as “the ability of a system, community or society exposed to hazards to resist, absorb, accommodate, adapt to, transform and recover from the effects of a hazard in a timely and efficient manner, including through the preservation and restoration of its essential basic structures and functions through risk management” (UNDRR, 2005).

From the perspective of research subjects, resilience studies include individual resilience (Matsukawa et al. 2024), families to communities (Longstaff et al. 2010; Magis, 2010; Kais and Islam, 2016), cities, and even nations and systems (Tierney, 2012). The concept of resilience spans multiple disciplines, and its assessment is also multifaceted. Communities are the primary social units affected by and respond to a disaster. Various disciplines employ different data sources to evaluate community resilience. Geography-related fields and engineering approaches, along with social and health sciences, have established distinct indicators to measure community resilience (Sim et al. 2021). The concept of perceived community resilience has been utilized to understand community resilience from the perspective of survivors of earthquakes in Turkey (Doğulu et al. 2016). The Community Advancing Resilience Toolkit (CART) has investigated its correlations with disaster risk reduction efforts (Cui et al. 2018), household livelihood assets (Wei et al. 2022), disaster preparedness in rural China (Sim et al. 2021) and responses to typhoons in Hong Kong (Guo et al. 2020). These studies employed the concept of perceived community resilience to examine community recovery and preparedness.

When disasters occur, ethnic minorities tend to be more vulnerable (Wisner et al. 2003; Hung et al. 2021). Particularly during the recovery process, minorities groups are more vulnerable to the impacts of disasters (Perry and Green, 1982; Fothergill et al. 1999; Carter-Pokras et al. 2007). Bolin and Stanford (1991) also found that ethnic minorities have greater housing and reconstruction needs. A study conducted in Colombia acknowledged the vulnerability of minority groups and emphasized their participation in disaster risk reduction and management activities (Gomez et al. 2023). In China, disasters have become a significant factor limiting the increase in farmers’ incomes and efforts to combat poverty in the southwestern minority regions (Zhuang et al. 2010). Enhancing the resilience of ethnic minorities is very important in practical terms.

Due to the location between the Pacific Rim Seismic Belt and the Eurasian Seismic Belt, China is highly susceptible to seismic events (Chen et al. 2013). The assessment findings show that earthquake disasters have significantly impacted many areas in the ethnic minority regions of Southwest and Northwest China (Borjigidai et al. 2014a, 2014b). From 2010 to 2018, the average annual direct economic losses caused by natural disasters in ethnic minority areas accounted for 32.9% of the country, and the death toll (including missing persons) accounted for 55.1% (Wei and Liu, 2020).

Over the past two decades, Sichuan and Xinjiang have experienced several significant earthquake disasters that have influenced China’s disaster reduction policy. The first major earthquake occurred in Bachu-Jiashi in Xinjiang with a magnitude of 6.8 on February 24, 2003. The second occurred in Sichuan on May 12, 2008, with a magnitude of 8.0. Following the Bachu-Jiashi Earthquake, the Earthquake Safety Project for Rural Dwellings (ESPRD) was initiated in Xinjiang (Zhang et al. 2013). After the Sichuan Earthquake, the Law of Earthquake Prevention and Disaster Reduction was revised to enhance safety measures. On May 11, 2017, a magnitude 5.5 earthquake struck Tashkurgan Tajik County in the Kashgar Prefecture, an area predominantly inhabited by the Tajik ethnic group.

The areas affected by the Bachu-Jiashi Earthquake and the Tashkurgan County Earthquake are located in three counties in Southern Xinjiang, which has been identified as one of the concentrated contiguous poverty-stricken areas (CCP-areas). In 2011, China identified 14 CCP-areas where ecological fragility and frequent natural disasters have contributed to the persistence of poverty. The Tibetan areas of Sichuan and the Wumeng Mountain Area are also included in this classification. The combination of poverty and disaster impacts has exacerbated a vicious cycle of “hazard–poverty–fragility” (OCNCDR et al. 2017).

This study explores strategies for enhancing community resilience in rural ethnic minority areas in Sichuan and Xinjiang. Specifically, we utilized a modified CART approach to investigate the status of disaster risk reduction in areas with diverse ethnic groups and to identify the influencing factors.

Based on the research purpose, this paper will answer the following questions:

  1. (1)

    What is the current state of perceived community resilience in rural ethnic communities?

  2. (2)

    Are there differences in perceived community resilience among rural ethnic communities?

  3. (3)

    What factors influence perceived community resilience in rural ethnic communities?

Literature review

Community resilience

Policy decisions and initiatives are formulated at broader global, regional, or national levels but are put into practice at the family or community level (Kais and Islam, 2016). Chandra et al. (2011) view it as the enduring capacity of a community to withstand adversities and recover from them. Norris et al. (2008) define community resilience as a process that connects various adaptive capacities to positive trajectories of function and adaptation following disturbances. Longstaff et al. (2010) define it as the ability of a community to absorb disturbances, undergo change, and maintain its fundamental functions, structures, identity, and feedback. Ungar (2011) highlights the roles of social capital, physical infrastructure, and culture explicitly, defining community resilience as interdependent patterns of community social capital, infrastructure, and culture that endow a community with the potential to recover from severe changes, maintain adaptability, and support growth in learning from crises. Ostadtaghizadeh et al. (2015) define community resilience as the ability of a system, community, or society that might be exposed to hazards to effectively resist, absorb, adapt to, and recover from the effects of a hazard to achieve and maintain an acceptable level of functioning. Community resilience can be defined as the ability of a community to transform its environment through the deliberate collective action of recovery, necessitating their effective response to adversity and learning from it (Pfefferbaum et al. 2013b).

Pfefferbaum introduced the Communities Advancing Resilience Toolkit (CART) that assesses community resilience across four dimensions: connection and caring, resources, transformative potential, and disaster management (Pfefferbaum et al. 2013a, 2013b, 2017). The aim is to identify reform measures and enhance community resilience. Subsequently, an expanded version was developed that added five items related to information and communication, forming a fifth dimension of the CART (Pfefferbaum et al. 2015). Perceived community resilience can be a perception of the physical and socioeconomic conditions for resilience (Guo et al. 2020). In CART, perceived community resilience is measured by assessing each individual’s perceptions of their community’s reactions to natural crises (Yang et al. 2020). Compared to other community resilience assessment methods, CART has demonstrated significant effectiveness in disaster preparedness and emergency response, encompassing natural disasters (Pfefferbaum et al. 2016; Cui et al. 2018; Sim et al. 2021; Kim et al. 2023) and public health incidents (Zhang, 2022; Tavares et al. 2023).

Considering the diverse social, economic, and cultural backgrounds, the unique characteristics of Chinese communities, and the current state of Chinese-specific community development, previous studies have validated the cross-cultural adaptability and effectiveness of the CART scale in China (Hu et al. 2017). The Chinese version of the CART has been used extensively in empirical studies in Sichuan, Northwest China, and Hong Kong (Cui et al. 2018; Guo et al. 2020; Sim et al. 2021; Wei et al. 2022). Based on previous studies, we believe CART indicators can effectively investigate rural ethnic minority communities in Sichuan and Xinjiang.

DRR policies and disaster preparedness

Disaster risk reduction (DRR) policies aim to strengthen resilience by improving disaster preparedness. Due to variations in the development processes of disaster emergency responses and public awareness, discrepancies exist in the public’s perception of responsibility allocation for disaster preparedness. According to a study by Wehde and Nowlin (2021), respondents place the largest share of responsibility on individual households rather than governments at any level. However, Zheng and Wu (2020) found that the more the public recognizes that individuals have responsibilities for disaster preparedness, the more likely they are to participate in emergency preparedness activities. Similarly, regarding the preparedness of the Chinese public, even in areas where a series of mitigation activities have been conducted, data shows that only half of the public has prepared emergency supplies at home or received disaster education, with fewer than 20% participating in emergency preparedness training (Cui et al. 2018).

The “rule of relative advantage” proposed by Norris et al. (2008) also applies to countries and communities, as different communities have varying resources and, therefore, receive different levels of support. In this study, the measurement of DRR policy focuses on the Earthquake Safety Project for Rural Dwellings (ESPRD) and Model Communities for Disaster Risk Reduction (MCDRR). ESPRD aims to enhance the seismic resilience of rural self-built houses by offering subsidies, training for construction artisans, and providing earthquake-resistant housing designs. It was initially implemented in Xinjiang as a pilot project in 2004 and was officially promoted throughout China in 2007 (Wu and Wu, 2020). The program enhances residents’ living conditions and reduces safety risks (Wang et al. 2005; Wu and Wu, 2020), and improves the construction of buildings (Wang, 2008; Yang et al. 2024).

The MCDRR aims to enhance community resilience through disaster prevention, mitigation capabilities, and emergency management measures. These efforts ultimately contribute to strengthening community resilience (Xie, 2023). The concept of MCDRR was first introduced in China’s Eleventh Five-Year Plan for Comprehensive National Disaster Reduction in 2007. This study adheres to national standards for MCDRR and identifies five key variables to assess the construction status of disaster-mitigation communities, considering both material and non-material preparedness perspectives (MCAPRC, 2011).

Disaster experience and disaster preparedness

The impact of disaster experience on individual disaster preparedness behaviors is a topic of interest in various studies. Chan and Ho (2018) found a positive correlation between disaster experiences and preparedness; however, this correlation varied at individual and household levels. Matsukawa et al. (2024) discovered that individuals with disaster experience exhibit greater resilience. A study on earthquake preparation in the United States revealed that disaster experiences can help residents better cope with disasters (Heller et al. 2005). Similarly, a Swiss study indicated that flood experiences positively affect individuals’ risk perception of flooding, influencing their preparedness behaviors (Siegrist and Gutscher, 2006). Individuals with flood experiences showed a deeper understanding of floods, higher response efficiency, and a stronger willingness to engage in adaptive actions than those without such experiences (Siegrist and Gutscher, 2008). Furthermore, earthquake experiences have been found to positively impact disaster mitigation behaviors and risk perception (Jackson, 1981). However, it is important to note that different disaster experiences can result in distinct impacts. For example, even in areas with repeated local flood disasters, most respondents did not think about disaster preparedness (Chan et al. 2017). Recurrent post-disaster depression and fear have been shown to inhibit mitigation behavior (Hansson et al. 1982).

This study categorizes the disasters that frequently occur in the area into four types: floods, geological disasters resulting from heavy rain, earthquakes, and avian influenza. The categorization is based on individuals’ experiences with these types of disasters, and the final data reflects the number of different types of disasters that people have encountered.

Methodology

Sampling and data collection

This study utilizes primary data collected through the “Community Vulnerability Assessment Survey” conducted by our team in 2018. Samples were collected from Sichuan and Xinjiang using stratified sampling. The study examines community resilience in ethnic minority areas, considering disaster policies and individual factors.

We selected ethnic minority communities that are at high risk for earthquakes, frequently experience disasters, and are characterized by CCP-areas. In Sichuan, we selected Xide County and Zhaojue County, where the Yi ethnic group resides in Liangshan Yi Autonomous Prefecture, Danba County, and Xiangcheng County, home to the Tibetan ethnic community. In Xinjiang, we focused on Uyghur ethnic communities in Kashgar City and Tajik ethnic communities in Tashkurgan Tajik County.

Among China’s ethnic minorities, the Uyghurs are the second largest group, the Yi are the sixth, and the Tibetans are the eighth. These groups make up significant portions of the 10 most populous ethnic groups in China (DED-NEAC and DCS-NBS, 2019). The Uyghur people primarily reside in the Xinjiang Uygur Autonomous Region. The Yi ethnic group is predominantly found in the provinces of Yunnan, Sichuan, Guizhou, and Guangxi, with the Liangshan Yi Autonomous Prefecture being the most significant area inhabited by the Yi people. Tibetans are mainly located in the Tibet Autonomous Region and in the provinces of Sichuan, Qinghai, Gansu, and Yunnan. Although the Tajik population is relatively tiny, Tashkurgan Tajik County is the only Tajik autonomous county in China. Sixty percent of the Tajik ethnic group lives in Tashkurgan Tajik Autonomous County in Xinjiang, while the remainder is spread across southern Xinjiang.

After determining the region, the questionnaire was developed using a random selection progress based on the population proportions of each community. To overcome language barriers, individuals fluent in both Mandarin and local ethnic languages were hired in each minority area. After completing pre-survey training, face-to-face surveys were conducted using the selected sampling list. A total of 1794 questionnaires were collected, of which 1188 were considered valid, accounting for 66.22% of the total collected (Table 1).

Table 1 Cities/counties Sampled.
Full size table

Measurement and description

Perceived community resilience

The measurement scale used in this study to assess perceived community resilience was adapted from the Chinese version of the CART. The Chinese version of the five-dimensional CART has demonstrated excellent internal consistency, a high model fit, and clear construct validity (Hu et al. 2017). This study utilized a modified version of CART after a trial investigation. The modification involved excluding the dimension with the lowest Cronbach’s coefficient alpha, which was linking and care. The revised CART includes four dimensions: disaster management, information and communication, resources, and transformative potential. Together, these dimensions comprise a total of 21 items. All items are rated on a Likert scale ranging from 1 to 5, where a score of 5 indicates complete alignment between the community’s current state and the description provided, while a score of 1 signifies complete misalignment, with decreasing levels of agreement represented by scores from 5 to 1.

We used Cronbach’s coefficient alpha to assess the reliability of the questionnaire in measuring the items. The overall Cronbach’s alpha for the scale was 0.96, indicating high consistency. Specifically, Cronbach’s alpha values for the dimensions were 0.91 for disaster management, 0.87 for information and communication, 0.89 for resources, and 0.93 for transformative potential. All dimensions and the overall scale showed Cronbach’s alpha values >0.85, indicating good internal consistency. Perceived community resilience is reflected in the mean and standard deviation of 21 core community resilience items across four domains, as well as the overall perceived community resilience scores. The mean scores for the core community resilience items range from 3.66 to 4.14 (Table 2).

Table 2 Perceived Community Resilience (CART) and domains (N = 1188).
Full size table

Preparedness

In our survey, we proposed five preparedness activities for the community. Three of these activities focused on stockpiling materials within the community, while the other two aimed at building capacity for residents. We ask questions that include “In your community/village, is there any equipment for flood control, landslide, and other natural disaster emergency rescue or cleanup?” (emergency rescue equipment), “Are there any public service facilities such as health centers in your community/village?” (public service facilities) and “In the community/village where you live, is it easy to get common medicines if you are ill and need to take them for a long time?” (medical supplies). Non-material preparedness includes distributing preparedness brochures and participating in emergency preparedness training.

There are notable differences between material and non-material disaster preparedness, as shown in Table 3. Approximately 47.52% of respondents reported that their communities lack emergency rescue equipment or are unsure of its availability. In contrast, 36.08% of respondents indicated that multiple types of rescue equipment are available and easily accessible in their community. Furthermore, 89.73% of communities have public service facilities such as health clinics. This highlights that emergency equipment is the most significant gap in disaster preparedness. Regular emergency preparedness training is conducted in 39.31% of communities (villages) with significant participation, while 56.57% provide brochures on disaster preparedness and risk reduction.

Table 3 Descriptive statistics of disaster mitigation community development variables.
Full size table

Natural disasters can lead to structural damage and building collapses, resulting in loss of life and property. In China’s rural areas, houses are often self-constructed and may include brick-timber, wooden, stone, or rammed-earth structures. Each of these building types responds differently to seismic activity (Pan et al. 2023). ESPRD aims to enhance the seismic performance of self-built buildings. The physical safety of housing directly impacts residents’ disaster resilience, which in turn affects the overall resilience of the community. Perceptions of housing safety are an independent variable related to perceived community resilience and ESPDR policies. To assess respondents’ perceptions of housing safety, the questionnaire included the question: “Do you believe that your current home is safe during a natural disaster?” The results showed that 7.24% of respondents feel their homes are very unsafe, 17.26% consider their homes somewhat unsafe, 29.04% indicated that their housing is very safe, and 30.89% stated it is relatively safe. This indicates that nearly 60% of respondents view their housing as relatively safe.

As shown in Table 4, nearly half of the respondents have not experienced any disasters, with approximately 32.32% having experienced one type of disaster and only 1.01% having experienced all four types of disasters.

Table 4 Descriptive statistics of demographic variables.
Full size table

Essential socioeconomic and demographic variables include gender, age, ethnicity, marital status, household size, education, employment status, and socioeconomic status as control variables. The data shows a balanced gender ratio among respondents, with each gender representing 50%. Notably, 70.12% of the respondents have a maximum educational level of elementary school or below, while only 1.52% have attained a university-level education or higher. Though the educational attainment in rural areas tends to be relatively low, most respondents perceive their socioeconomic status as middle-tier, comprising approximately 68.43%. Most respondents reported household sizes ranging from 4 to 9 members, which accounts for 77.02% of the total. Additionally, 31 respondents indicated that their households have 10 or more members, about 2.60% of the total. Large households are a common characteristic of ethnic minorities.

Data analysis

First, we use box plots to display variations in community resilience among ethnic minorities and provinces. Next, using the Spearman correlation coefficient, we quantify the relationship between the independent and perceived community resilience variables. Finally, we establish a multivariate linear regression model to explore further each independent variable’s specific impact on the perceived community resilience indicators. The expression is as follows:

$${\rm{y}}={\rm{w}}_{1}{\rm{x}}_{1}+{\rm{w}}_{2}{\rm{x}}_{2}+\ldots +{\rm{w}}_{\rm{n}}{\rm{x}}_{\rm{n}}+{\rm{b}}$$

In the expression, y represents the perceived community resilience and its domains, \({\rm{x}}_{1},\,{\rm{x}}_{2},\ldots ,\,{\rm{x}}_{\rm{n}}\) represents the dependent variable, and \({\rm{w}}_{\rm{n}}\) represents the regression coefficient corresponding to \({\rm{x}}_{\rm{n}}\), and \({\rm{b}}\) represents the intercept of the model.

The data analysis primarily utilizes Stata 16.0.

Results

Comparison of perceived community resilience

Due to the varying perceptions of community resilience among different ethnic groups, we conducted a further analysis across different regions. Figure 1a presents a boxplot illustrating the median scores and distribution of overall perceived community resilience across different ethnic communities. This reflects the residents’ perceptions, with scores ranging from 1 to 5; higher scores indicate higher levels of community resilience. The results indicate significant differences in community resilience among various ethnic minorities. The Uyghur has the highest score (4.56) in community resilience, followed by the Tajik community (4.25), the Tibetan community (3.78), and the Yi (3.46) community in descending order.

Fig. 1: Comparison of perceived community resilience.
Fig. 1: Comparison of perceived community resilience.
Full size image

a Comparison of community resilience across different ethnic minorities. b Comparison of community resilience across different provinces.

Figure 1b illustrates the median resilience scores for various ethnic communities across different provinces, revealing significant disparities in perceived community resilience. The Uyghur and Tajik communities in the Xinjiang Uyghur Autonomous Region have consistently scored above 4, with a score of 4.50. In contrast, the Yi and Tibetan communities in Sichuan Province have a lower score of 3.59, falling below the 4 mark. As shown in the figure, the findings highlight the notable differences in community resilience between these two provinces.

Correlation analysis

The Spearman correlation coefficient was utilized to examine the relationships between the dependent variables, which include perceived community resilience and its various domains, and the independent variables. The findings (Fig. 2) indicate that perceived community resilience is significantly positively associated with respondents’ education, employment status, and socioeconomic status, although these associations are relatively weak. In contrast, perceptions of housing safety, ethnicity, province, and material and non-material community preparedness exhibit more substantial positive correlations. Furthermore, a negative correlation exists with gender, age, and disaster experience. There are also notable differences between community resilience and its dimensions, with significant positive correlations between household size and perceived community resilience, resources, and development potential. However, factors related to information and communication and disaster management do not show any relationship with household size.

Fig. 2: Correlation results between perceived community resilience and independent variables.
Fig. 2: Correlation results between perceived community resilience and independent variables.
Full size image

a Correlations between perceived community resilience and influencing factors; b correlations between disaster management and influencing factors; c correlations between information and communication and influencing factors; d correlations between resources and influencing factors; e correlations between transformative potential and influencing factors. Some variables are replaced by abbreviations.

Regression analysis

Further, a multiple linear regression model was employed to examine the relationship between various influencing factors and perceived community resilience and its respective domains (Table 5). The regression results indicate that demographic variables partially impact community resilience. Compared to individuals under 25, those aged 30–40 and 40–50 tend to associate with perceived community resilience negatively. Specifically, individuals who self-identify as belonging to the lower-middle, middle, upper-middle, or upper classes perceive their community resilience more positively than those of lower socioeconomic status. Additionally, the number of family members positively affects community resilience, although this impact is relatively limited.

Table 5 Regression results for perceived community resilience.
Full size table

Disaster experience has a moderately positive impact on community resilience. In other words, the more disasters the community’s residents have experienced, the higher their perceived community resilience. This may be related to post-disaster reconstruction efforts, often including subsidies for new housing and supporting facilities. Among the four domains, disaster experience is positively correlated with disaster management and information and communication domains.

In comparison to the Yi ethnicity, Uyghur and Tibetan ethnicity have a significant impact on community resilience, while the Tajik community does not show a significant effect. Among these groups, the influence of the Uyghur community is the most profound, whereas the Tibetan influence is the least significant. Compared to the Yi ethnic group, the Uyghurs significantly positively impact all four dimensions of community resilience. In contrast, the Tibetan and Tajik communities are not significantly associated with the resources or transformative potential domain of community. Additionally, The perception of housing safety significantly contributes to community resilience.

Material and non-material preparedness within a community are positively associated with community resilience and its various domains. Communities that have easily accessible emergency rescue equipment, public service facilities, and medical supplies, as well as readily available preparedness brochures and high participation in emergency preparedness activities, tend to demonstrate stronger community resilience. We assessed DRR policies through indicators of house safety and community preparedness. The analysis indicates that residents’ perceived community resilience is correlated with disaster reduction policies.

Discussions

This study aims to analyze the current state of perceived community resilience in ethnic minority areas, explore the differences among various ethnic minority communities, and investigate the factors that impact perceived community resilience.

First, we found that individual factors can influence perceived community resilience, though this influence is not definitive. Age, self-perceived socioeconomic status, and ethnicity can shape these perceptions. Specifically, compared to individuals under 25, individuals aged 30–50 often view community resilience negatively. This may relate to emergency drills conducted in schools across China over the last decade, which have increased disaster awareness among young people. Those who perceive their socioeconomic status as higher tend to have a more positive view of community resilience. Consistent with previous studies, socioeconomic status generally has a positive impact on perceptions of community resilience (Sim et al. 2021; Zhang et al. 2023). Our study discovered that larger household sizes are associated with a heightened perception of community resilience. This may be linked to the mutual support among family members following a disaster. Notably, in these survey areas, the number of ethnic minority households is relatively high compared to Han nationality families. Our questionnaire survey revealed that more than 60% of respondents regularly participate or sometimes participate in emergency preparedness training. This finding offers valuable insights for disaster reduction policies at the community level. It highlights the importance of engaging various family members in the planning and response process, ultimately enhancing the community’s overall emergency response capacity.

We observed that different ethnic minority communities demonstrate unique forms of resilience. Within this context, we established a hierarchy among four ethnic minority groups based on their community resilience, ranked from strongest to weakest: Uyghur, Tajik, Tibetan, and Yi communities. Overall, minority communities in Xinjiang show greater resilience than those in Sichuan, which may be attributed to the earlier implementation of ESPRD in Xinjiang. In 2004, Xinjiang initiated a pilot project for rural housing, which was later expanded nationwide in 2007. By 2014, more than 3 million impoverished rural households in Xinjiang had benefited from the ESPRD policy (Wen et al. 2016). It has been shown to improve the seismic performance of housing through both capital investment and engineering improvements, thus strengthening disaster prevention and reduction (Zhang et al. 2013). The ESPRD policy enables residents to enhance their housing conditions through subsidies, which boosts perceptions of housing safety and contributes to overall community resilience (Li et al. 2010; Yao et al. 2017).

This study focused on ESPRD and MCDRR. As an indicator of MCDRR, material, and non-material resources positively influenced respondents’ perceptions of community resilience. However, there is a weak correlation between the perception of housing safety and community resilience, and no correlation exists regarding transformative potential. This indicates that disaster reduction policies have greatly impacted residents’ perceptions of community resilience. The key element is that residents can observe or participate in them. Governments can enhance perceived community resilience by investing in infrastructure, increasing preparedness awareness, and promoting disaster education (Maripe, 2011). Huang et al. (2021) emphasize that the policy system is the most significant factor influencing the emergency response process in China. However, the difference between Xinjiang and Sichuan illustrates that implementing disaster reduction policies must consider inter-provincial disparities.

In this study, the perceived community resilience among the Yi people is the lowest. Ethnicity is one of the factors that contribute to differences in disaster preparedness and coping strategies. This has been evidenced in numerous prior studies (Murphy et al. 2009; Smith and Notaro, 2009). Additionally, research by Wu et al. (2020) confirms that China’s preferential policies for ethnic minority areas help bridge ethnic differences. In 2018, the Chinese government implemented significant poverty alleviation efforts in minority areas, such as the Yi ethnic minority (CPCCC and SCPRC, 2018). Disaster risk reduction and poverty alleviation are strongly interconnected (Schmidtlein et al. 2011; Loayza et al. 2012; Yin et al. 2017; Cheng et al. 2018; Zhang et al. 2020). Disasters can lead to increased poverty, so integrating disaster risk reduction strategies into poverty alleviation policies is essential. Relevant measures such as housing initiatives and community preparedness are key indicators in poverty alleviation efforts. By reducing disaster losses, we can enhance the effectiveness of poverty alleviation strategies.

This study has some limitations. Although resilience is dynamic, we only used cross-sectional data to examine community resilience due to data constraints. Following the policy change, this research can serve as a foundation for future investigations to provide a comparative analysis of the situation after development. We plan to conduct follow-up studies in this area to establish a longitudinal study system for resilience. Additionally, we need to pursue further research from the perspectives of different ethnic cultures and customs.

Conclusions

The study employed a modified CART questionnaire to evaluate community resilience in ethnic minority areas of Sichuan and Xinjiang based on survey data collected in 2018. It investigates the community resilience of different ethnic minorities and analyzes the factors that affect perceived community resilience.

We found variations in community resilience among ethnic minorities, with these differences being more noticeable across different provinces. In addition, the results show that both individual factors and DRR policies have a particular impact on perceived community resilience. However, the effect of the DRR policy is more prominent, which means that the focus should be on its implementation. Policies for disaster reduction that are observed or participated in can enhance residents’ perception of community resilience. Additionally, it is crucial to address the differences in policies between provinces.