Structural Model of Community Social Capital for Enhancing Rural Communities Adaptation against the COVID-19 Pandemic: Empirical Evidence from Pujon Kidul Tourism Village, Malang Regency, Indonesia

: Pujon Kidul Village, Pujon District, Malang Regency, is an area with tourism potential that has been developed since 2017 with the concept of agricultural tourism. Throughout the development of tourism villages, Pujon Kidul Village has succeeded in accelerating economic growth and providing jobs for the community. However, during the COVID-19 pandemic, tourism villages have been severely affected, leading to the temporary closure of tourist attractions and community businesses. This research aimed to identify what indicators form social capital variables and the relationship between social capital variables and community adaptation patterns in dealing with pandemics in the study location. This was quantitative research with conﬁrmatory factor analysis to determine the indicators of forming social capital and structural equation modeling analysis to determine the relationship between the variables. Based on the ﬁndings, it is known that trust in forming a social network is 0.468. Furthermore, the social network forms community actions of 0.046 and inﬂuences community resilience by 0.007. Therefore, good social capital will make it easier for the community to participate in collective action as a form of caring for each other during the pandemic. This action also inﬂuences the community to survive in a pandemic crisis, thus creating an adaptation pattern for the Pujon Kidul Tourism Village community in facing a pandemic.


Introduction
Since December 2019, the world has been shocked by the outbreak of coronavirus COVID-19; this virus is endemic in almost all parts of the world and has become a hot topic of discussion because of the many deaths it causes worldwide. Since then, the number of COVID-19 cases has increased rapidly and caused a pandemic. Several countries, including Indonesia, have established policies such as social distancing to reduce the high rate of disease transmission [1]. This policy is carried out on a large scale and impacts people's lives in small rural areas. Efforts to deal with pandemics in rural areas can be carried out effectively because of community solidarity [2]. Individual or community solidarity is one of the concepts of social capital used to increase individuals' or communities' capacity for social development [3].
The COVID-19 pandemic impacts not only the economic aspect but also the environment and all aspects of people's lives, especially the social aspects of society [4][5][6]. Likewise, in the face of the current pandemic crisis, the existence of norms, trust, and social networks to support social capital allows individuals to more easily access various sources such as information, assistance, and other shared resources as a form of support between

Data Collection Methods
The primary and secondary data collection methods were used to obtain information and data in this study. The secondary data were obtained from literature and agency studies, the primary data-through questionnaires, interviews, and observations. The sampling method in this study used a population of 1250 households living in Pujon Kidul Village. This study used proportional stratified random sampling, taking samples from the subpopulations in the study population considering their size [30]. The sample in this study was determined based on the table of Isaac and Michael (1981). The sample was determined with a 5% margin of error, yielding a sample of 275 households from the population of 1250 households. The distribution of this sample was spread across three hamlets, namely Maron Hamlet, Tulungrejo Hamlet, and Krajan Hamlet.

Methods of Analysis
This research was conducted using a quantitative approach based on the two research objectives. The first objective was to determine the factors that form social capital using confirmatory factor analysis (CFA). The second objective was to describe the relationship of social capital with collective action and community resilience using structural equation modeling (SEM) analysis.

Confirmatory Factor Analysis (CFA)
Confirmatory factor analysis (CFA) is part of SEM (structural equation modeling) analysis. In the CFA test, validity and reliability tests are carried out on the indicators forming the latent variables of the study, wherein one variable is measured by one or more indicators [31]. The CFA analysis has the following stages [32]: Performance of the theoretical model development. b.
Drawing of a path diagram of the relationship between variables and indicators. c.
Change of the path diagram in the model into equations. d. Obtaining model estimate values. e. Assessment of the model identification. f.
Assessment of the goodness-of-fit criteria.
The CFA in this study was conducted to determine whether the indicators for forming community social capital could adequately form the latent variables of social capital (social networks, norms, and trust).

Structural Equation Modelling (SEM)
This study used structural equation modeling (SEM) analysis, which can analyze the relationship between constructs in research, including indicators and latent variables, as well as measure errors in direct measurement [33]. SEM analysis can be performed after confirmatory factor analysis (CFA). In addition, the following steps can be taken for comprehensive modeling in SEM analysis [ This study used CFA and SEM analysis with AMOS 24 on latent variables and indicators. In addition, this study used SEM to analyze the relationship of the community social capital with collective action during the pandemic in Pujon Kidul Tourism Village.

Geographical Conditions in Pujon Kidul Tourism Village
Pujon Kidul Village is administratively located in Pujon District, Malang Regency [34] ( Figure 1). According to the Pujon Kidul Village Profile (2021), Pujon Kidul Village has an area of 486.40 hectares and is divided into three hamlets, nine community units (RW), and twenty neighboring units (RT). Pujon Kidul Village is located between 7 • 21 N and 7 • 31 LS and 110 • 10 W and 111 • 40 E with the following regional boundaries: g. Modification of the model. h. Development of the discussion, research suggestions, policy implications, and conclusions.
This study used CFA and SEM analysis with AMOS 24 on latent variables and indicators. In addition, this study used SEM to analyze the relationship of the community social capital with collective action during the pandemic in Pujon Kidul Tourism Village.

Economic Conditions in Pujon Kidul Tourism Village
The main potential of Pujon Kidul Village is in agriculture and plantations; with this potential, the majority of community work is in the agricultural and livestock sectors with a total of 1730 people. Therefore, this is an opportunity used by the community to become an agriculture-based tourism village with the main tourist object being cafe sawahs. The existence of these tourism activities makes Pujon Kidul Village an agrotourism village and provides an increase in the economy for the community and the village original income (PADes). Therefore, the economy of Pujon Kidul Village is engaged in the agricultural and tourism sectors.
However, the condition of Pujon Kidul Tourism Village during the COVID-19 pandemic is undoubtedly different from before the outbreak. According to the Pujon Kidul Bumdes data (2022), the most visible impact of changes due to the pandemic is a decrease in the number of tourists, which impacts a decrease in income. For example, the number of tourists visiting tourism villages decreased in 2019 by 601,858, decreasing to 418,272 in 2020, 224,162 in 2021, and as many as 99,254 in April 2022. In addition, due to the pandemic, in Pujon Kidul Village, tourist attractions were also closed temporarily due to the policy of implementing community activity restrictions (PPKM) (Figure 2a

Economic Conditions in Pujon Kidul Tourism Village
The main potential of Pujon Kidul Village is in agriculture and plantations; with t potential, the majority of community work is in the agricultural and livestock sectors w a total of 1730 people. Therefore, this is an opportunity used by the community to beco an agriculture-based tourism village with the main tourist object being cafe sawahs. T existence of these tourism activities makes Pujon Kidul Village an agrotourism village a provides an increase in the economy for the community and the village original inco (PADes). Therefore, the economy of Pujon Kidul Village is engaged in the agricultural a tourism sectors.
However, the condition of Pujon Kidul Tourism Village during the COVID-19 p demic is undoubtedly different from before the outbreak. According to the Pujon Kid Bumdes data (2022), the most visible impact of changes due to the pandemic is a decre in the number of tourists, which impacts a decrease in income. For example, the num of tourists visiting tourism villages decreased in 2019 by 601,858, decreasing to 418,272 2020, 224,162 in 2021, and as many as 99,254 in April 2022. In addition, due to the p demic, in Pujon Kidul Village, tourist attractions were also closed temporarily due to policy of implementing community activity restrictions (PPKM) (Figure 2a

Explanation of Social Capital Characteristics during the Pandemic in Pujon Kidul Tourism Village
Characteristics of the Pujon Kidul Tourism Village community social capital can determined based on the choice of answers to the questions addressed to the responden The answers are explained and illustrated with descriptive statistics. There were five swer choices for each question given: strongly disagree (SD), disagree (D), moderat agree (MA), agree (A), and strongly agree (SA). The following are indicators that meas social capital:

Trust (T). This variable is divided into seven indicators, including trust in neighb
(T1), trust in immigrants (T2), trust in the government (T3), trust in traditional lead (T4), trust in religious leaders (T5), trust in tourism institutions (T6), and commu cation between people (T7). 2. Norms (N). This variable is divided into two indicators, including obedience to c toms (N1) and attendance at traditional events (N2).

Social networks (NW). This variable consists of five indicators, including willingn
to build cooperation (NW1), participation in religious activities (NW2), participat in social activities (NW3), willingness to give opinions during meetings (NW4), a participation in community groups (NW5).
Based on the data on social capital characteristics of Pujon Kidul Village in Table 1 can be seen that the answers from 275 respondents were dominated by choice five,

Explanation of Social Capital Characteristics during the Pandemic in Pujon Kidul Tourism Village
Characteristics of the Pujon Kidul Tourism Village community social capital can be determined based on the choice of answers to the questions addressed to the respondents. The answers are explained and illustrated with descriptive statistics. There were five answer choices for each question given: strongly disagree (SD), disagree (D), moderately agree (MA), agree (A), and strongly agree (SA). The following are indicators that measure social capital:

2.
Norms (N). This variable is divided into two indicators, including obedience to customs (N1) and attendance at traditional events (N2).

3.
Social networks (NW). This variable consists of five indicators, including willingness to build cooperation (NW1), participation in religious activities (NW2), participation in social activities (NW3), willingness to give opinions during meetings (NW4), and participation in community groups (NW5).
Based on the data on social capital characteristics of Pujon Kidul Village in Table 1, it can be seen that the answers from 275 respondents were dominated by choice five, or strongly agree (SA), for the 14 indicators, which means that the people of Pujon Kidul Village strongly agree with the indicators of trust (K), norms (N), and social networks (J). The detail of the data of respondents is in the supplementary material.

Explanation of Collective Action Characteristics during the Pandemic in Pujon Kidul Tourism Village
Characteristics of the collective action of the Pujon Kidul Tourism Village community can be determined based on the choice of answers to the questions addressed to the respondents. The answers are explained and illustrated with descriptive statistics. There were five answer choices for each question given: strongly disagree (SD), disagree (D), moderately agree (MA), agree (A), and strongly agree (SA). In addition, the following are indicators that measure collective action: a.
Collective action from village governments consists of three indicators, including decision-making in taking collective action from the village government (CG1), participation in collective action from the village government (CG2), and frequency of collective action from the village government (CG3) b.
Collective action from community initiatives consists of three indicators, including decision-making in carrying out collective action from community initiatives (CC1), participation in collective action from community initiatives (CC2), and frequency of taking collective action from community initiatives (CC3).
Based on the data on the characteristics of collective action in Pujon Kidul Village in Table 2, it can be seen that the answers from 275 respondents were dominated by choices three and five. Several types of collective action involving the community dealt with the pandemic in Pujon Kidul Tourism Village. The respondents' answer choices are based on the questionnaire results described in Figure 3, illustrating that most respondents were involved in collective action based on the type of action: from the government or from the community. In that case, an action from the community is an action with most respondents often involved. The main difference is that in decision-making, the community more often follows the decision-making of the community's collective action than the decision-making of the collective action of the government. However, even though the community is involved in making decisions for collective action from the community, it is not necessarily the community that participates in taking action from the community, and it is not necessarily the community that participates that takes collective action from the community. Likewise, with collective action from the government (Figure 3), it can be seen that the graph is decreasing for the two types of collective action. The respondents' answer choices are based on the questionnaire results described in Figure 3, illustrating that most respondents were involved in collective action based on the type of action: from the government or from the community. In that case, an action from the community is an action with most respondents often involved. The main difference is that in decision-making, the community more often follows the decision-making of the community's collective action than the decision-making of the collective action of the government. However, even though the community is involved in making decisions for collective action from the community, it is not necessarily the community that participates in taking action from the community, and it is not necessarily the community that participates that takes collective action from the community. Likewise, with collective action from the government (Figure 3), it can be seen that the graph is decreasing for the two types of collective action. The types of activities carried out by the community dealing with COVID-19 are described in Figure 4 in the types of collective action from the community and the government. The community's collective action consists of four activities, while the collective action of the government consists of five activities. Collective action from the community, including distributing necessities and other basic needs, mentoring youth prayer groups, providing health protocol facilities, and cleaning villages, come from community initiatives, while collective action of the village government includes the socialization during the COVID-19 pandemic, social assistance from the village government, distribution of masks, hand sanitizer, and vitamins, procurement of health protocol facilities in public places, and training on tourism products and creative economy. The types of activities carried out by the community dealing with COVID-19 are described in Figure 4 in the types of collective action from the community and the government. The community's collective action consists of four activities, while the collective action of the government consists of five activities. Collective action from the community, including distributing necessities and other basic needs, mentoring youth prayer groups, providing health protocol facilities, and cleaning villages, come from community initiatives, while collective action of the village government includes the socialization during the COVID-19 pandemic, social assistance from the village government, distribution of masks, hand sanitizer, and vitamins, procurement of health protocol facilities in public places, and training on tourism products and creative economy.

Explanation of Community Resilience Characteristics during the Pandemic in Pujon Kidul Tourism Village
Characteristics of community resilience of the Pujon Kidul Tourism Village community can be determined based on the choice of answers to the questions addressed to the respondents. The answers are explained and illustrated with descriptive statistics. The answer choices consisted of five answer choices for each question given: strongly disagree (SD), disagree (D), moderately agree (MA), agree (A), and strongly agree (SA). Additionally, the following are indicators that measure community resilience: community resilience, knowledge of COVID-19 (CR1), community security (CR2), availability of health protocol facilities in the village (CR3), and community perceptions of assistance and contributions (CR4).
Based on the data on the characteristics of community resilience in Pujon Kidul Village in Table 3, it can be seen that the answers from 275 respondents were dominated by choices four and five. Some respondents have a very understanding of COVID-19 and feel very safe in the community. This is supported by the availability of health protocol facilities in good condition that could be used in the village ( Figure 5). However, some respondents feel that the community and the government assist the community, but the assistance is only enjoyed by certain groups (Figure 4).

Explanation of Community Resilience Characteristics during the Pandemic in Pujon Kidul Tourism Village
Characteristics of community resilience of the Pujon Kidul Tourism Village community can be determined based on the choice of answers to the questions addressed to the respondents. The answers are explained and illustrated with descriptive statistics. The answer choices consisted of five answer choices for each question given: strongly disagree (SD), disagree (D), moderately agree (MA), agree (A), and strongly agree (SA). Additionally, the following are indicators that measure community resilience: community resilience, knowledge of COVID-19 (CR1), community security (CR2), availability of health protocol facilities in the village (CR3), and community perceptions of assistance and contributions (CR4).
Based on the data on the characteristics of community resilience in Pujon Kidul Village in Table 3, it can be seen that the answers from 275 respondents were dominated by choices four and five. Some respondents have a very understanding of COVID-19 and feel very safe in the community. This is supported by the availability of health protocol facilities in good condition that could be used in the village ( Figure 5). However, some respondents feel that the community and the government assist the community, but the assistance is only enjoyed by certain groups (Figure 4).

Confirmatory Factor Analysis (CFA), the Results of the Analysis for the First Purpose of the Research
Confirmatory factor analysis was conducted to analyze the indicators of social capital formation in the study locations. Social networks, beliefs, and norms are the dimensions that make up social capital. The CFA analysis in this study is a second-order CFA. First, the latent variable of trust (T) has indicators including trust in neighbors (T1), trust in immigrants (T2), trust in the government (T3), trust in traditional leaders (T4), trust in religious leaders (T5), trust in tourism institutions (T6), and communication between people (T7). Second, the norms variable (N) consists of obedience to customs (N1) and attendance at traditional events (N2). The social networks variable (NW) consists of willingness to build cooperation (NW1), participation in religious activities (NW2), participation in social activities (NW3), willingness to give opinions during meetings (NW4), and participation in community groups (NW5). The CFA analysis was carried out in two stages with the following estimation results and goodness of fit (Figure 6a,b).

Confirmatory Factor Analysis (CFA), the Results of the Analysis for the First Purpose of the Research
Confirmatory factor analysis was conducted to analyze the indicators of social capital formation in the study locations. Social networks, beliefs, and norms are the dimensions that make up social capital. The CFA analysis in this study is a second-order CFA. First, the latent variable of trust (T) has indicators including trust in neighbors (T1), trust in immigrants (T2), trust in the government (T3), trust in traditional leaders (T4), trust in religious leaders (T5), trust in tourism institutions (T6), and communication between people (T7). Second, the norms variable (N) consists of obedience to customs (N1) and attendance at traditional events (N2). The social networks variable (NW) consists of willingness to build cooperation (NW1), participation in religious activities (NW2), participation in social activities (NW3), willingness to give opinions during meetings (NW4), and participation in community groups (NW5). The CFA analysis was carried out in two stages with the following estimation results and goodness of fit (Figure 6a,b).
Based on the estimated value and feasibility of goodness of fit ( Figure 6), it can be seen that in stage 1, there were two invalid confidence variables because they had a loading factor of 0.5, including T6 (0.097), T7 (0.061), and NW3 (0.469). Two indicators needed to be discarded to proceed to the next stage. In the next stage (second stage), the CFA model fit was recalculated, and the indicators that make up the social capital variable were obtained. The trust variable was formed by trust in neighbors (T1) (0.689), trust in immigrants (T2) (0.533), trust in the government (T3) (0.527), trust in traditional leaders (T4) (0.631), and trust in religious leaders (T5) (0.572). The social networks variable was formed by willingness to build cooperation (NW1) (0.548), participation in religious activities (NW2) (0.526), willingness to give opinions during meetings (NW4) (0.629), and participation in community groups (NW5) (0.719). The norms variable was formed by obedience to customs (N1) (0.747) and attendance at traditional events (N2) (1.099).
The loading factor value describes how much influence the indicator has on the formation of the latent variable. For example, this indicates that the indicator of trust in neighbors was the most influential in forming trust. On the other hand, attendance at traditional events was the most influential in forming norms, and participation in community groups was the most influential indicator in forming social networks in Pujon Kidul Village. religious leaders (T5), trust in tourism institutions (T6), and communication between people (T7). Second, the norms variable (N) consists of obedience to customs (N1) and attendance at traditional events (N2). The social networks variable (NW) consists of willingness to build cooperation (NW1), participation in religious activities (NW2), participation in social activities (NW3), willingness to give opinions during meetings (NW4), and participation in community groups (NW5). The CFA analysis was carried out in two stages with the following estimation results and goodness of fit (Figure 6a,b).

Structural Equation Modelling (SEM), the Results of the Analysis for the Second Purpose of the Research
Structural equation modeling was conducted to identify the relationship between social capital and community adaptation patterns depicted through collective action and community resilience in Pujon Kidul Village. SEM analysis was carried out with three alternative models, which described the relationship between the social capital, collective action, and community resilience variables. These three models had different paths. Model 1 described the direct relationship of norms and networks to trust. Then, trust was directly related to the collective action of the government and the community which was considered to affect the community resilience variable. Model 2 described the direct relationship of the social networks and trust variables to the norms variable. The norms were directly related to the collective action of the government and community, which affected the community resilience variables. Finally, model 3 described the direct relationship between the trust, norms, and social networks variables. The social networks were directly related to collective action that was considered to be able to affect community resilience.
In this study, the three models were compared with the estimated value and goodness of fit ( Figure 6). Then, one of the fittest models was selected based on the chi-squared, probability, GFI, CMIN/DF, TLI, AGFI, RMSEA, and CFI criteria.
Based on the estimated value and goodness of fit (Figure 7a-c) of the three SEM models, model 3 was the fittest model among the three alternative models. This is because in the path diagram's depiction, the model's eligibility requirements were at least 4-5 GOFI criteria, and the three models were fit. However, model 3 was the fittest because it met five criteria. There is a lower chi-squared limit value of 299.954 (lower than the df limit value), the CMIN/DF value was 1.648, the GFI value was 0.906, the AGFI value was 0.881, the TLI value was 0.962, the CFI value was 0.967, and the RMSEA value was 0.049. Meanwhile, the relationship between the variables described in model 3 is a direct relationship between the two variables of trust and norms. The relationship between the two is directly related to the network, and the network is directly related to the community's collective action that affects community resilience. The influence of the relationship between variables in Model 3 can be seen from the values of the Squared Multiple Correlation (R2). First, trust positively influences social networks, which is 0.468. Then the network variable has a positive effect on community action with a value of 0.046, and the community action variable has a positive effect on community resilience with a value of 0.007. The positive effect of community action means that the higher the trust, the better the network formed, and a well-formed network increases the community's initiative to take collective action. Then the higher the community's initiative to take collective action, the better the community's resilience; thus, the relationship between these three variables can form an adaptation pattern during the Pujon Kidul Village community pandemic.
Social capital is the main capital in moving individuals or groups of people to live their daily lives. Strong social capital will raise the public's desire to be in-volved. The Pujon Kidul community, which has good social capital conditions, will make it easier for the community to participate in collective action as a form of caring for each other during the pandemic. Then this action also influences the community to survive in a pandemic crisis. Therefore, social capital can support an adaptation pattern of the Pujon Kidul Village community facing a pandemic.

Conclusions
The indicators that describe the latent variable of social capital in Pujon Kidul Tourism Village are the latent variable of trust formed by five indicators: trust in neighbors, trust in immigrants, trust in the government, trust in traditional leaders, and trust in religious leaders. Likewise, the latent variable of norms is formed by two indicators consisting of adherence to customs and attendance at traditional events. Finally, the social network variable consists of 4 indicators formed Willingness to build cooperation, Participation in religious activities, Willingness to give opinions during meetings, and Participation in community groups. The influence of the relationship between variables in Model 3 can be seen from the values of the Squared Multiple Correlation (R2). First, trust positively influences social networks, which is 0.468. Then the network variable has a positive effect on community action with a value of 0.046, and the community action variable has a positive effect on community resilience with a value of 0.007. The positive effect of community action means that the higher the trust, the better the network formed, and a well-formed network increases the community's initiative to take collective action. Then the higher the community's initiative to take collective action, the better the community's resilience; thus, the relationship between these three variables can form an adaptation pattern during the Pujon Kidul Village community pandemic.
Social capital is the main capital in moving individuals or groups of people to live their daily lives. Strong social capital will raise the public's desire to be in-volved. The Pujon Kidul community, which has good social capital conditions, will make it easier for the community to participate in collective action as a form of caring for each other during the pandemic. Then this action also influences the community to survive in a pandemic crisis. Therefore, social capital can support an adaptation pattern of the Pujon Kidul Village community facing a pandemic.

Conclusions
The indicators that describe the latent variable of social capital in Pujon Kidul Tour-ism Village are the latent variable of trust formed by five indicators: trust in neighbors, trust in immigrants, trust in the government, trust in traditional leaders, and trust in religious leaders. Likewise, the latent variable of norms is formed by two indicators consisting of adherence to customs and attendance at traditional events. Finally, the social network variable consists of 4 indicators formed Willingness to build cooperation, Participation in religious activities, Willingness to give opinions during meetings, and Participation in community groups. SEM's findings link social capital variables, collective action, and community resilience interrelationships. For example, model 3 SEM describes the relationship between trust variables that affect network variables by 0.468. The network variables affect community actions by 0.046, and community action variables affect community resilience by 0.007. Therefore, better social capital will form a good pattern of community adaptation through collective action and community resilience during the Pujon Kidul Village pandemic.