Mediating role of socio-religious behavior on consumption pattern changes of middle-class Muslims: An analysis using SEM

Abstract This study aims to investigate the mediating effect of socio-religious behavior on consumption pattern changes of middle-class Muslims in the new normal era. In terms of methodology, this study is a correlational survey research. Research data were collected through questionnaires as many as 553 middle-class Muslim respondents who live in South Sulawesi Province, Indonesia. Reliability was accepted with cronbach’s alpha coefficient, and validity was accepted by confirmatory factor analysis. Data analysis and hypothesis testing were carried out through structural equation modeling and multiple regression. Two softwares (SPSS and LISREL) were used for statistical data processing. The results of the study indicate that the full mediation of socio-religious behavior was found in the relationship between religiosity, attitude, and consumption pattern changes, while in the relationship of subjective norms, perceived behavioral control and consumption pattern changes, socio-religious behavior was found as a partial mediation. In other words, the model proposed to form a fundamental change in consumption patterns of the middle-class Muslims does not work optimally. However, this study still finds an important thing that socio-religious behavior is able to mediate the relationship between religiosity, attitude, and consumption pattern changes.


PUBLIC INTEREST STATEMENT
The purpose of this study is to investigate the mediating effect of socio-religious behavior on consumption pattern changes of middle class-Muslims in the new normal era. This study uses the framework of the Theory of Planned Behavior (TPB) in investigating the mediating effect of the socio-religious behavior. The results of the study show that the full mediation of socio-religious behavior was found in the relationship between religiosity, attitude and consumption pattern changes, while in the relationship between subjective norms, perceived behavioral control and consumption pattern changes, socio-religious behavior was found as a partial mediation. Thus, the new insight that can be revealed in this study is that socio-religious behavior can be used as a means to form changes in consumption patterns of the middle-class Muslims in the new normal era, namely the socio-religious-based consumption pattern model, if it is based on a positive attitude and strong teachings of Islam.

Introduction
Socio-religious behavior has played an important role in fundamentally forming the consumption patterns of middle-class Muslims in Indonesia. This behavior refers to participation in socioreligious activities and involvement in religious networks (Bloom & Arikan, 2013) and is the behavior of most middle-class Muslims who dominate the Muslim population in Indonesia (World Population Review, 2022) with a total of 64.5 million or 27.5% of the total Muslim population (Riska Rahman, 2020). After the Covid-19 pandemic, their socio-religious behavior can be seen on the consumption patterns of those who care more about halal products. The consumption of halal products in Indonesia has grown by an average of 14.96% per year since 2020 and is expected to grow to 2025 (Indonesia Halal Lifestyle Center, 2022). Their socio-religious behavior has been integrated into their local wisdom which has become a way of life for around 4,000 to 5,000 people in Indonesia (IWGIA, 2021) as a daily rule for them (Syaparuddin, 2018). Thus, socioreligious behavior has a mediating effect on the consumption pattern changes of middle-class Muslims in the new normal era.
This study used the theory of planned behavior (TPB) in predicting the mediating effect of socioreligious behavior on the consumption pattern changes of middle-class Muslims. TPB is based on a generalized assumption of mainstream neoclassical economics that human beings are generally rational and infinitely self-centered or selfish (Ajzen, 1991). This assumption was under attack by the critics of behavioral economics who assumed that human beings are subjective and boundedly rational (Herbert et al., 2004). Capitalizing this assumption, this study retained attitude, subjective norms (SNs), and perceived behavioral control (PBC) as the predictors of socio-religious behavior, which replaced generalized rational intention of TPB by the actual behavior of consumption pattern changes. This study also used an additional predictor, it is religiosity. Religiosity is a power that is internally able to influence its followers to perform a behavior (Purwanto et al., 2022;Suleman et al., 2021). On this basis, religiosity was added to TPB.
From the previous studies, the use of TPB in the context of mediating effect of socio-religious behavior on consumption pattern changes has not gotten a sufficient attention from experts. However, there are several previous studies related to this study, and they are the ones conducted by Suleman et al. (2021), Dinh et al. (2022), Ramazani and Kermani (2022), and Shimul et al. (2022), and Rehman et al. (2022). These studies tend to keep socio-religious belief as belief and knowledge only because each one stands alone, so that it is only used as a predictor not as subjective behavior which is used as an intervening variable. This study didn't but kept the socioreligious behavior as a subjective behavior by replacing generalized rational intention of TPB, which was used as an intervening variable that had a mediating effect on consumption pattern changes of middle-class Muslims.
In general, this study aims to complete previous studies that have not given a sufficient attention to the investigation of the mediating effect of socio-religious behavior on consumption pattern changes of middle-class Muslims, while specifically this study aims to predict the effect of religiosity, attitude, subjective norms (SNs), and perceived behavioral control (PBC) on consumption pattern changes of middle-class Muslims through the mediation of socio-religious behavior as an intervening variable. Thus, this study is considered as the first study designed to investigate the mediating role of socio-religious behavior on consumption pattern changes of the Indonesian middle-class Muslims in the new normal era.

Theory of planned behaviour (TPB)
TPB was first developed by Ajzen in 1985, and it is the most widely used theory in understanding and predicting behavior. TPB has been used and tested under various circumstances to predict rational intention and actual behavior (Kautonen et al., 2015). In TPB, intention is predicted by individual beliefs about attitude, SNs, and PBC toward actual behavior. Attitude is a psychological deviation that is shown by assessing the likes and dislikes of certain entities (Eagly & Chaiken, 1995). SNs are social pressures on people to perform or not perform a behavior. PBC describes the perceived behavior that is considered controllable (Liou & Contento, 2001). However, Ajzen then dichotomized the construction of PBC into separate constructs such as internal and external controls (Ajzen, 2002). The internal control is usually equated with one's self-efficacy, while the external control refers to one's beliefs about the support or opposition he/she will find in the environment (Hockerts, 2017). In this study, TPB is used to investigate the mediating effect of socio-religious behavior on consumption pattern changes of the Indonesian middle-class Muslims in the new normal era.

Socio-religious behavior
Religious social behavior refers to participation in social religious activities and involvement in religious networks (Bloom & Arikan, 2013). In the context of Indonesia, the socio-religious behavior is the act of carrying out religious teachings with full faith and sincerity and implemented in the social area of the community (Syaparuddin & Nuzul, 2021). This act is a form of appreciation of the teachings of Islam that is learned and practiced. Not just doing daily worship routines but also doing the activities that have a strong motive in carrying out religious teachings which are interpreted as worship in the form of concrete and meaningful social action decisions for others and the environment (Dermawan, 2014). This is what religious social behavior means in this study and is the behavior of most middle-class Muslims who dominate the Muslim population in Indonesia (World Population Review, 2022) with a total of 64.5 million or 27.5% of the total Muslim population (Riska Rahman, 2020).

Middle-class Muslims
Middle-class communities, according to Riki Rahman and Hazis (2018), are those whose positions are between the working class and the bourgeoisie or between the upper class and the lower class, while middle-class Muslims are a group of Muslims who are placed between the working class and the bourgeoisie or between upper and lower class communities whose livelihoods are lucrative and well educated. In the context of Indonesia, middle-class Muslims are middle-class groups that apply Islamic principles, norms, and values as the identity of individuals and groups that developed according to political events that occur. These middle-class Muslims emerged as a form of discrimination and alienation against Muslims which then triggered the activities of the bourgeoisie, education, and philanthropy as a symbol of political revival. The context of the middle-class santrinization and resantrization has a significant effect on the growth of the group of middle-class Muslims in Indonesia (Fisipol-UGM, 2017). This is what middle-class Muslims mean in this study.

Religiosity and socio-religious behavior
Religiosity is considered a quality of religion and devotion to its followers' beliefs about a religion (Hassan Fathelrahman Mansour & Mohammed Elzubier Diab, 2016) and is one of the main constructs used in research on religion (Abou-Youssef et al., 2015). In this study, the religion in question is Islam, so religiosity is how much someone is committed to the teachings of Islam. Religiosity is belief in God with a commitment to follow predetermined instructions (Nurhayati & Hendar, 2020). Meanwhile, Johnson et al. (2001) define religiosity as how a person is committed to his religion and its teachings which is reflected in his attitude and behavior. From this understanding, a person's behavior can reflect his religious beliefs. While socio-religious behavior is the act of carrying out religious teachings with full faith and sincerity and implemented in the social area of the community (Syaparuddin & Nuzul, 2021), this act is a form of appreciation of the teachings of Islam that is learned and practiced. Not just doing daily worship routines but also doing the activities that have a strong motive in carrying out religious teachings which are interpreted as worship in the form of concrete and meaningful social action decisions for others and the environment (Dermawan, 2014). The discussion above shows that there is synergy between the implementation of religious teachings and social action. It means that religiosity is strongly related to socio-religious behavior in forming consumption pattern changes. As discussed above, we thus propose the following hypotheses: H1a: Religiosity is positively related to socio-religious behavior H1b: Religiosity is positively related to consumption pattern changes

Attitude and socio-religious behavior
Attitude is a judgment that a person likes or dislikes (Ajzen, 1991). Attitude influences behavior through a careful and reasoned decision-making process (Wang et al., 2021). Attitude obtained from experience will have a direct influence on behavior (Duits et al., 2002). This direct influence will be realized if conditions and situations allow (Yazdanpanah & Forouzani, 2015). If the individual is in a situation that is completely free from various forms of pressure or obstacles that interfere with the expression of his/her attitude, it can be expected that the forms of behavior that appear are expressions of the actual attitude (Liao et al., 2007). The formation of a behavior, starting from a good understanding of the information (stimulus) then the attitude shown will be in accordance with the information. Then, the attitude will cause a response in the form of behavior or action against the stimulus or object earlier. If the acceptance of new behavior through a process based on a positive attitude, the behavior will last a long time (Levine, 2007). In this study, socio-religious behavior is an expression of attitude towards the synergy between the implementation of religious teachings and social actions in forming consumption pattern changes. This means that attitude is strongly related to socio-religious behavior in forming consumption pattern changes. As discussed above, we thus propose the following hypotheses: H2a: Attitude is positively related to socio-religious behavior H2b: Attitude is positively related to consumption pattern changes

Subjective norms (SNs) and socio-religious behavior
SNs are defined as individual perceptions of the possibility of a reference group in the form of groups or individuals who agree or disagree with a behavior (Baber, 2018). SNs are part of the belief that certain people are unlikely to approve or approve certain behaviors (Hudi et al., 2019). Meanwhile, according to Ajzen (1991) SNs are a person's consideration with respect to whether other people think that he or she needs to do this or not. SNs refer to the individual's perception of the surrounding social pressure in doing or not doing a certain behavior. In TPB, norms are determined by normative beliefs, which assess certain social pressures from individuals (Ibrahim et al., 2017). In the context of this study, norms are related to support from other parties for socioreligious behavior in forming consumption pattern changes in the new normal era. The higher the support is, the higher the socio-religious behavior in forming consumption pattern changes is. The above understanding shows that SNs have a strong relationship to socio-religious behavior in forming consumption pattern changes. As discussed above, we thus propose the following hypotheses: H3a: Subjective norm is positively related to socio-religious behavior H3b: Subjective norm is positively related to consumption pattern changes

Perceived behavioral control (PBC) and socio-religious behavior
PBC is the extent to which individuals perceive attractive behavior (Liou & Contento, 2001). PCB refers to the perceived ease or difficulty in performing a behavior. In other words, individuals' selfconfidence in performing a certain task significantly effects their intention and behavior (Ajzen, 1991). It consists of two parts: they are how much jurisdiction the individual has over behavior and how confident the individual feels about acting or not performing a certain behavior. If an individual feels more likely to control behavior about a decision to act, the more likely that individual will do so (Ajzen, 1991). Many scientists have also confirmed this relationship (Bonne et al., 2007;Hui Liu Liuning Zhou et al., 2012;Liou & Contento, 2001). The above understanding shows that if an individual feels more likely to control behavior in realizing socio-religious behavior to form consumption pattern changes, the more likely the individual is to change his/her consumption patterns. It means that PBC is strongly related to socio-religious behavior and also to consumption pattern changes. As discussed above, we thus propose the following hypotheses: H4a: Perceived behavioral control is positively related to socio-religious behavior H4b: Perceived behavioral is positively related to consumption pattern changes

Socio-religious behavior and consumption pattern changes
Socio-religious behavior is the act of carrying out religious teachings with full faith and sincerity and implemented in the social area of the community (Syaparuddin & Nuzul, 2021). This act is a form of appreciation of the teachings of Islam that is learned and practiced (Dermawan, 2014). Not just doing daily worship routines but also doing the activities that have a strong motive in carrying out religious teachings which are interpreted as worship in the form of concrete and meaningful social action decisions for others and the environment (Bloom & Arikan, 2013). The form of religious social behavior is the synergy between the implementation of religious teaching and social action. While consumption patterns are the use of goods or services by individuals or groups to meet their needs (Kuusipalo et al., 1989), this consumption pattern can change if it is influenced by several things, one of which is social and cultural matters. Social and cultural factors of the community play a strong role in the attitude of choosing goods to be consumed. Social and cultural factors like this usually develop in the community according to environmental conditions, religion, customs, and local customs (Amirudin & Sabiq, 2022). The above understanding shows that socio-religious behavior can form consumption pattern changes and can also mediate religiosity, attitude, SNs, and PBC in forming in consumption pattern changes. As discussed above, we thus propose the following hypotheses: H5: Socio-religious behavior is a positively related to consumption pattern changes

Mediating effect of socio-religious behavior on consumption pattern changes
Religiosity, attitude, SNs, and PBC can form behavior, including socio-religious behavior, because it is a manifestation of the synergy between the implementation of religious teachings and social actions (Syaparuddin & Nuzul, 2021). The above understanding shows that the synergy between the implementation of religious teachings and actions in changing consumption pattern is formed by attitude, SNs, and PBC. If so, then socio-religious behavior can mediate religiosity, attitude, SNs, and PBC in forming consumption pattern changes. As discussed above, we thus propose the following hypotheses:  Figure 1 shows the proposed structural model, comprising three types of variables, including four predictors (i.e., religiosity, attitude, subjective norms, and perceived behavioural control), one mediator variable (socio-religious behavior), and one endogenous variable (consumption pattern changes). Each component of the model was based on the literature review. Overall, this study thereby contributes to the literature by focusing on identifying the mediating effect of socioreligious behavior on consumption pattern changes.

Measurement
To confirm the conceptual model proposed in this study, the questionnaire that was created included two main parts. The first section briefly describes the purpose of the study and guidelines for filling out the questionnaire and its relation to socio-demographic information. The questions include the respondent's age, marital status, gender, occupation, education, and income/month. The second part, for the development of the model structure, consists of a multiple-choice item scale by adopting a five-point Likert Scale ranging from "Strongly Disagree" (1) to "Strongly Agree" (5). This section has 39 questions; 6 questions for consumption pattern changes (Y2), 7 questions for socio-religious behavior (Y1), 8 questions for religiosity (X1), 6 questions for attitude (X2), 6 questions for subjective norms (X3), and 6 questions for perceived behavioral control (X4).

Data collection
Data collection was carried out in the last 3 months (mid-April to late July) in 2022 by applying the convenience sampling technique in distributing questionnaires to middle-class Muslim respondents who live in South Sulawesi, Indonesia. The questionnaires were distributed online in Google Form format to them via social media, WhatsApp. The complete answers from respondents who have been successfully recorded on Google Forms that have been made as many as 553 records. This means that the sample obtained is 553 respondents from the middle-class Muslim population in South Sulawesi Province, Indonesia. Hair et al. (2010) suggested that the minimum sample requirement for using multivariate analysis is ten times the number of research instruments, and this study has fulfilled it (Boomsma & Hoogland, 2001;Shore et al., 2000).

Data analysis
The data that has been collected is classified into two parts: they are the identity of respondents, and respondents' answers. The identity of the respondents from the questionnaire distributed to them is described in tabulation form, while respondents' answers are first analyzed before being described, through three steps, they are (1) Factor analysis, this step uses the IBM SPSS Statistics 25 application to state the significant correlation between indicators and variables used through the value of Bartlett's Test of Sphericity (≤0.05) and the certainty of sufficient sample support through the KMO value (KMO 0.5). In addition, to state the value of the factor loading (≥0.70) and MSE (≥0.50) for each indicator, as well as the Eigenvalue (>1) and its Variance (%) for each variable (B. Williams et al., 2010), (2) Measurement model assessment, this step uses the Lisrel 8.70 application (Joreskog & Sorbom, 1989) to state that the proposed model is fit to use through the Chi Square (≤ value [df; Sig. 5%]), Probability (≥0.05), CMIN/DF (≤2.00), and RMSEA (≥0.05). In addition, the fitting test of the model can be seen from the values of the following indices (≥0.90): NFI, NNFI, CFI, IFI, GFI, and AGFI (Hair et al., 2010), and also to state the indicators that used is valid, as well as the variables used are reliable (Yang & Montgomery, 2011), and (3) Structural model assessment, this step uses Weighted Least Square (WLS) method with the Lisrel 8.70 application (Joreskog & Sorbom, 1989) to state R-Square value and the significance of the direct and indirect (mediating) effect (≥1.96) of the exogenous variables on the endogenous variable (Rigdon & Hoyle, 1997) for hypothesis test. Table 1 represents demographic information, including gender, age, marital status, education level, occupation, and monthly income. The respondents are largely male (55%), with the overall dominance of farmers (33%) and teachers/lecturers (27%) and 39% of unmarried respondents. The respondent with the highest education level holds a bachelor's degree (34%). Regarding age, our respondents are mainly in the "20-29" category (23%) and "40-49" category (33%). While the income aspect is in the "uncertain/other" category (63%) and the "Rp 5 million to 7 million" category (25%). Table 2 shows the factor analysis of the variables coded Y1 which consists of seven items, Y2 which consists of six items, X1 which consists of eight items, X2 which consists of six items, X3 which consists of six items, and X4 six items. The value of Kaiser-Meyer-Olkin (KMO, 0.50) measure of sample adequacy is >0.8, and the value of Bartlett's Test of Sphericity measure of factor analysis to be continued is <0.05. The value of loading factor (≥0.70) and MSE (≥0.50) measure of requirements for further processing is >0.80. The value of eigenvalue measure of requirements to be used as factors is >1, and the value of total variance explained is ≥72.848% which is significant and acceptable. Table 3 shows that the value of Chi-Square is 4789.61, the probability is 0.000, and the value of RMSEA is 0.104. All results were not found to be good for the model fit values (L. J. Williams et al., 2009). Therefore, it is necessary to modify the measurement model. Table 4 shows that the value of Chi-Square (CMIN) is 52.34, the probability is 0.075, and the value of RMSEA is 0.021. These results were found to be good for the model fit values after modification. In addition, the fitting test of the model can also be seen from the values of CMIN/DF, NFI, NNFI, CFI, IFI, GFI, and AGFI (Agus Widarjono, 2010; Hair et al., 2010) and the results were found to be good for the model fit values after modification. Table 5 shows that the value of standardised loadings is above 0.50, and statistically significant t-values are above 1.96. All indicators (after modification) were found valid and significant. Table 6 shows that the values of CR and AVE are above 0.70 and 0.50 Hair et al. (2010). All variables were found reliable.

R-Square
Two values of R-Square obtained, are 0.75 (75%) and 0.63% (63%). The first one is that the contribution of religiosity, attitude, subjective norm, and perceived behavioral control to socio-    (Hair et al., 2010). The second one is that the contribution of socio-religious behavior, religiosity, attitudes, subjective norms, and perceived behavioral control to consumption pattern changes is also quite robust for its proposed model (Hair et al., 2010).

Structural model assessment and hypothesis test
The results of the hypothesis test are presented in Table 7. Based on the structural model assessment, only four hypotheses (H1a, H2a, H2b, and H5) of the direct effects and two hypotheses (H6a, H6b) of the indirect (mediating) effects were statistically proved to be significant, as they followed the t-value of more than 1.96 (≥1.96; Joreskog & Sorbom, 1989).

Findings and discussion
This study generally aims to investigate the mediating effect of socio-religious behavior on consumption pattern changes of Indonesian middle-class Muslims in the new normal era. Meanwhile, this study specifically aims to examine the effect of religiosity, attitude, subjective norms, and perceived behavioural control on consumption pattern changes of middle-class Muslims in the new normal era through the mediation of socio-religious behavior. Therefore, this study offers a new insight into the socio-religious-based consumption pattern model, namely the consumption pattern of caring for food security and prosperity for middle-class Muslims in the new normal era.
In realizing this, this study proposes 13 hypotheses to be tested. However, the test results show that only five of the thirteen hypotheses were accepted, while eight of another ones were not. The direct effects on religious social behavior have four hypotheses proposed (H1a, H2a, H3a, and H4a). The results of hypotheses H1a and H2a show a positive and significant relationship between socioreligious behavior and religiosity and attitude. This finding is in line with the prior studies by Suleman et al. (2021), Sutrisno and Dularif (2020), and Dinh et al. (2022). As a result, religiosity and attitude are important factors that are considered unidirectional in forming socio-religious behavior of Indonesian middle-class Muslims in the new normal era. However, hypotheses H3a and H4a only have a marginal relationship between socio-religious behavior and subjective norms and perceived behavioural control, but they have a positive relationship with each other. As a result, subjective norms and perceived behavioural control have the potential to be important factors that are considered unidirectional in forming socio-religious behavior of Indonesian middle-class Muslims in the new normal era.
Meanwhile, the direct effects on consumption pattern changes have five hypotheses proposed (H1b, H2b, H3b, H4b, and H5). The results of hypotheses H1b, H3b, and H5 show a positive and significant relationship between consumption pattern changes and socio-religious behavior, attitude, and perceived behavioural control. This finding is in line with the prior studies by Abrar et al. (2021), Nguyen et al. (2022), and Dilotsotlhe (2021). As a result, socio-religious behavior, attitude, and perceived behavioral control are important factors that are considered unidirectional in forming consumption pattern the changes of Indonesian middle-class Muslims in the new normal  Meanwhile, the mediating effects have four hypotheses (H6a, H6b, H6c, and H6d). The results show that hypotheses H6a and H6b are full mediations, while hypotheses H6c and H6d are only partial mediations. As a result, the mediating role of socio-religious behavior is only found in the relationship between religiosity and attitude and consumption pattern changes of Indonesian middle-class Muslims in the new normal era. The results of test of Goodness of Fit in relation to full mediations show that all variables that indicate full mediations only leave two indicators each, they are Y2: Y2.2 and Y2.5, Y1: Y1.4 and Y1.7, X1: X1.6 and X1.8, and X2: X2.4 and X2.6 (See , Table 3,  Table 4 and Table 5). If these indicators elaborated each other then interpreted, it can be stated that their unsavory behavior and friendship in social life can form their consumption patterns to care more about the future necessities of their household and also care more about the budget constraint they have in meeting the necessities of life and the future of their household in the new normal era. Their unsavory behavior and friendship were formed because they tried to sincerely follow the teachings of Islam in all things in their lives and stay away from earning through unlawful means in Islamic law, as well as their decision to make some change in their consumption patterns to care more about Maslahah (benefit) in the new normal era because they are considered as good decisions and in accordance with Sharia (Islamic law) Reviewing at the theoretical level, a person's involvement in a particular religion may be measured with his obedience to perform his religious orders (Benk et al., 2016;Boone et al., 2013). Religion contains commands, prohibitions, rewards, and sins. Besides offering its followers hope and heaven, religion also imposes threats in the form of sin and hell on offenders. This is an internal motivator that is very important for individual behavior (Sutrisno & Dularif, 2020). If religion is implemented in social life, religion will give a noble identity to its followers (Akah et al., 2020;Weaver & Agle, 2002). People who believe and do good deeds will get happiness in the form of entering heaven. Meanwhile, sinners will go to hell. This doctrine can make a Muslim obedient to Islamic law, including consumption patterns that are in accordance with sharia (Islamic law). Ideally, the higher a person's level of religiosity is, the greater the influence of religion in forming his character and behavior is. This means that a religion that teaches sharing with others, lots of friendship, staying away from the unsavory behavior and earning through

Conclusion and managerial implications
The model proposed to map the fundamental changes in consumption patterns of middle-class Muslims in the new normal era turned out to be unable to work optimally. As a result, TPB, which is used to map the formation of changes in consumption patterns, is not able to accommodate it perfectly, so that another theory is needed to support it. However, the present study found an important thing that socio-religious behavior can mediate the relationship between religiosity, attitude, and consumption pattern changes. However, it cannot mediate the relationship between SNs, PBC, and consumption pattern changes. Therefore, the new insight that can be revealed in the present study is that socio-religious behavior can be used as a means to form the consumption pattern changes of Indonesian middle-class Muslims in the new normal era, namely the socioreligious-based consumption pattern model, if it is based on a positive attitude and strong teachings of Islam.
This study also offers some practical implications. First, the government should make socioreligious behavior as the basis in an effort to restore the welfare level of the middle-class Muslims in the post Covid-19. Second, middle-class Muslims should form a socio-religious-based consumption pattern in an effort to restore their welfare level in the post Covid-19. Third, this study is useful for the government and the general public in determining development policies, evaluation, and control systems for restoring the welfare level of middle-class Muslims in the post covid 19.

Limitations and future study directions
Although this study has succeeded in showing a limited mediating effect of socio-religious behavior on fundamental changes in consumption patterns of Muslims in the new normal era, it seems that there are still some limitations in this study and can be considered for improvement in the future. One of them is the geographical approach, we only surveyed in general the middle class-Muslims in the province of South Sulawesi, Indonesia. Therefore, it should be noted that the result of study in this area depends on the geographical characteristics of the study population, and perhaps it may provide similar or different results for other study population.