Consumer Behavior toward Halal Food Using Theory of Planned Behavior and Theory of Interpersonal Behaviors among Muslim Students in Tasikmalaya

The share of halal products in Indonesia is enormous and grows yearly. With this rapid development, the City of Tasikmalaya plans to become the pioneer of the first halal tourism in West Java by prioritizing halal certification in food. However, the facts on the ground state that many Muslim students still do not understand and tend not to care about consuming halal food. This study aims to reveal the magnitude of the influence of attitudes, subjective norms, perceived behavioral control, religiosity, and habits that are formed on purchase intentions and actual purchases of halal food on Tasikmalaya Muslim Students. The method used in this research is descriptive quantitative by distributing it to 404 respondents. The data analysis technique in this study uses Partial Least Structural Equation Modeling (PLS-SEM). The results show that attitude, subjective norm, perceived behavior, and perceived behavior positively affect purchase intention. Purchase intention and habit positively affect the actual purchase, while habit moderates the relationship between purchase intention. The last, religiosity has a positive effect on attitude and purchase intention.


Introduction
The halal label is a significant factor in consumer purchasing decisions because the halal market continues to increase (Lutfie et al., 2016;Anggadwita et al., 2020).One of the cities that wants to move into the halal food industry is the City of Tasikmalaya.This city wants to be the pioneer of the first halal tourism in West Java due to the solid Islamic tradition and its regional background, which is famous for its many Islamic boarding schools (Kompasiana.com, 2021).However, according to Rosidi et al. (2018), students with the most significant potential in the halal food market tend to be more concerned with practical, affordable, and fast food.
Based on this, this study aims to examine the buying behavior of halal food among Tasikmalaya Muslim students.The research will integrate consumer behavior theory, namely, the Theory of Planned Behavior and the Theory of Interpersonal Behavior, as well as the religiosity factor.The method used in this research is to use the survey method and distribution of questionnaires to 400 respondents.

Literature Review
According to Heizer et al. (2017), operations management is a series of activities that create value in goods and services by converting inputs into outputs.Moreover, based on Heizer et al. (2017), ten operations management decisions include: process and capacity strategy (design of goods and services); managing quality; process and capacity strategy; location strategy; layout strategy; human resources and job design; supply chain management; inventory management; scheduling; and maintenance (Ivanov et al., 2021).
Based on what was stated by Heizer et al. (2017), supply chain management describes the coordination of all supply chain activities, starting from raw materials and ending with customer satisfaction.Thus, the supply chain includes suppliers, manufacturers or service providers, distributors, wholesalers, or retailers who deliver products and services to end customers (Latifah et al., 2021).According to Khan et al. (2022), halal supply chain management can be interpreted as the management of the process of handling halal products starting from raw materials from suppliers to finished products in the hands of consumers, whose entire process is based on Sharia law.
Developed a theory called the Theory of Interpersonal Behavior.This theory proposes that behavioral interests are determined by feelings (feelings) that humans have towards behavior which is called affect.Subsequent behavior is influenced by what humans have done, which is called habit, behavioral interests, and facilitating conditions (Triandis, 1980).Emphasized that "habit" is behavior that has been automated in existing situations to shape behavior directly without needing an interest-building process first.
According to Chae et al. (2020), a purchase intention is a form of consumers who wish to buy and choose a product based on their experience, use, and desire for a product.According to Prabowo & Priambodo (2021), an actual purchase is a series of physical and mental actions experienced by consumers.This theory is the result of the development of the Theory of Reasoned Action (TRA) by Icek Ajzen to become the Theory of Planned Behavior (TPB) in 1985.The purpose of this theory is to predict individual behavior precisely.Stated that the foundation of attitudes is formed around ideas or beliefs embedded in the attitude formation model and structure (Ajzen, 1985).Then according to Kania et al. (2020); Susilawati et al. (2022), attitude is a consumer's response to whether he likes or dislikes an object.
Perceived Behavioral Control, also known as behavioral control, is a person's feelings about how easy or difficult it is to manifest a particular behavior.Ajzen & Fishbein (2005); Apsari et al. (2019) stated that religiosity comes from the English "religion," which means religion, then the word "religious" is added, which means religious or pious.Based on this frame of mind, this study adopts research from Amalia et al. (2020) which analyzes factors from the Theory of Planned Behavior, Theory of Interpersonal Behavior, and Religiosity on purchase intention and actual purchase (actual purchase) of halal food.(Amalia et al., 2020) The hypothesis in this study refers to (Amalia et al., 2020)

Methodology of Research
The research method used in this research is a quantitative method with a descriptive research type because this research is directed to provide accurate symptoms and facts regarding the characteristics of the population in a particular area.Quantitative researchers will see the relationship of variables to the object under study, which is cause and effect (causal), so there are independent and dependent variables in their research.The strategy used in this study used a survey method to obtain data from certain natural places by distributing questionnaires via Google form to 404 respondents from Tasikmalaya Muslim Students.The unit of analysis used is the individual obtained directly from the source (primary data) because this research will analyze the answers or information from the various opinions of each individual.Based on the implementation time, this study used a cross-sectional method which only made one observation.The data analysis technique uses Partial Least Squares Structural Equation Modeling (PLS-SEM), which aims to examine the relationship between variables by looking for a relationship or influence between variables.This model is suitable for use when research has several variables and indicators.

Normality Test
Based on Figure 2 above, it is known that the significance value is 0.000 <0.05, so it can be concluded that the residual data is not normally distributed.Therefore, data processing can use Smart-PLS.Convergent validity is carried out to determine the validity of each relationship between indicators and their variables.The convergent validity test can be seen from the loading factor of each construct indicator, which has a value of > 0.7 for confirmatory research and a loading factor value of 0.6-0.7,which is acceptable for exploratory research, as well as a value of Average Variance Extracted (AVE) > 0.5 (Table 1).Based on the Table 1, each indicator on the variable meets the criteria, which is more than 0.7.Then, the outer model of the convergent validity test can also be seen from the AVE values.The recommended AVE value is > 0.5 to indicate good convergent validity.The discriminant validity test has attachments with different construct measurements and is not highly correlated.To test discriminant validity, one must pay attention to the cross-loading value of each variable > 0.70 (Table 2).Using the PLS-SEM data analysis technique with SmartPLS 3.0 software, the reliability test can use the Composite Reliability (CR) value (Table 3) to measure the reliability value of a construct.The rule of thumb for this measurement is Composite Reliability > 0.7 for confirmatory research, and CR values of 0.6-0.7 are still acceptable for exploratory research.Based on these results, the CR value of each construct exceeds 0.7.This is in accordance with the rule of thumb for the CR value, which is greater than 0.7, so all variables can be declared reliable.

Inner Model Measurement
The Table 4 shows the R-square value in the study, namely the attitude variable, with an R-square value of 0.838 or 83.8%; this means that 83.8% of the attitude variable can explain the religiosity variable.While an error variance causes the remaining 16.2% at the time of measurement, there may be other variables that can be explained by the attitude variable, which requires further research.Furthermore, the purchase intention variable with an R-square value of 0.936 or 93.6%; this means that 93.6% of the purchase intention variable can explain the variables of attitude, religiosity, subjective norm, and perceived behavioral control.In contrast, the remaining 6.4% is caused by an error variance at the time of measurement, so there may be other variables that can be explained by the purchase intention variable, which requires further research.Furthermore, the actual purchase variable with an R-square value of 0.710 or 71% means that 71% of the actual purchase variable can explain the variables of attitude, religiosity, subjective norm, perceived behavioral control, and habit.While an error variance causes the remaining 29% at the time of measurement, other variables can be explained by the actual purchase variable, which requires further research.

Hypothesis Testing
Table 5 can answer this research question by looking at the significance value between variables, namely the t-value > 1.64.Attitude, subjective norm, religiosity, and perceived behavioral control variables show a positive and significant influence on purchase intention because they have a tvalue above 1.64, so the hypothesis is accepted.The purchase intention and habit variables show a positive and significant influence on the actual purchase variable because it has a t-value above 1.64, so the hypothesis is accepted.Furthermore, the habit variable can moderate the relationship between the purchase intention variable and the actual purchase.Figure 2 shows the magnitude of the path coefficient value for each relationship between the variables contained in this study using SmartPLS.The large values of the path coefficients in the structural model indicate the path coefficients of the hypothesized causal variables.This value indicates the strength of the influence of the independent variable on the dependent variable.An example is the first hypothesis: attitude influences purchase intention with a regression coefficient of 0.105.This value indicates that the attitude variable affects 10.5% of the purchase intention variable.

Figure 2. Structural Model
The relationship between attitudes towards purchase intention has a t-value of more than 1.64, equal to 1.705, indicating that attitude influences purchase intention, so H1 is rejected with a regression coefficient of 0.105.The relationship between norms and purchase intention has a t-value of more than 1.64, equal to 3.225, indicating that subjective norms affect purchase intention, so H2 is accepted with a regression coefficient of 0.166.
The relationship between perceived behavioral control and purchase intention has a t-value of more than 1.64, equal to 1.967, indicating that perceived behavioral control influences purchase intention, so H3 is accepted with a regression coefficient of 0.072.The relationship between purchase intention and actual purchase has a t-value of more than 1.64, equal to 9.542, indicating purchase intention towards the actual purchase, so H4 is accepted with a regression coefficient of -0.645.
The relationship between habit and actual purchase has a t-value of more than 1.64, equal to 4.256, indicating that perceived behavioral control influences purchase intention, so H5 is accepted with a regression coefficient of 0.329.The habit variable moderates the relationship between the purchase intention and actual purchase variables.The t-value of more than 1.64, equal to 2.383, indicates that habit negatively moderates the relationship between the purchase intention variable and the actual purchase variable, so H6 is accepted with a regression coefficient of 0.140.
The relationship between religiosity and attitude has a t-value of more than 1.64, equal to 47.479, indicating that perceived behavioral control influences purchase intention, so H7 is accepted with a regression coefficient of 0.915.The relationship between religiosity and purchase intention has a t-value of more than 1.64, equal to 9.928.It indicates that religiosity affects intention, so H8 is accepted with a regression coefficient of 0.658.

Conclusion
This study aims to determine the effect of Attitude, Subjective Norm, Purchase Behavioral Control, and Religiosity on Purchase Intention and its direct effect on Actual Purchases with Habit as a moderating variable.Data processing to assist in analyzing the relationship between variables in this study uses Partial Least Square (PLS) with SmartPLS 3.2.9software.Based on the results of the analysis in the previous chapter, it can be concluded that Attitude has a positive and significant effect on purchase intention by having a t-value of more than 1.64, which is equal to 1.705, and a regression coefficient of 0.105.Subjective norm positively and significantly affects purchase intention by having a value.The t-value is more than 1.64, equal to 3.255, and the regression coefficient is 0.166.significant to the actual purchase by having a t-value of more than 1.64, which is 9.542 and a regression coefficient of 0.645.Habit has a positive and significant effect on Actual Purchases by having a t-value of more than 1.64, equal to 4.256, and a regression coefficient of 0.329.Habit moderates the relationship between purchase intention variables and actual purchase variables with a t-value of more than 1.64, which is equal to 2.383, and a regression coefficient of 0.140.Religiosity has a positive and significant effect on Attitude by having a t-value of more than 1.64, equal to 47.479, and a regression coefficient of 0.915.Religiosity has a positive and significant effect on purchase intention, with a t-value of more than 1.64, 9.928, and a regression coefficient of 0.658.

Figure 1 .
Figure 1.Conceptual Framework(Amalia et al., 2020) , namely: H1: Attitude significantly influences Purchase Intention of halal food by Muslim students in Tasikmalaya.H2: Subjective Norm significantly influences the Purchase Intention of halal food by Muslim students in Tasikmalaya.H3: Perceived Behavioral Control significantly influences Purchase Intention of halal food by Muslim students in Tasikmalaya.H4: Purchase Intention significantly influences the Actual purchase of halal food by Muslim students in Tasikmalaya.H5: Habit significantly influences the Actual Purchase of halal food by Muslim students in Tasikmalaya.H6: Habit negatively moderates the impact of Purchase Intention toward Actual Purchase of halal food by Muslim students in Tasikmalaya.H7: The religiosity of Muslim students in Tasikmalaya can impact Attitude toward Purchase Intention of halal food.H8: The religiosity of Muslim students in Tasikmalaya can impact Purchase Intention of halal food.

Figure
Figure 2. Normality Test

Table 4 .
Value of Structural Inner ModelIn the model, fit test analysis is done manually using following GoF formula.Based on the results of the model fit test that has been carried out, the model fit test value is greater than 0.38, equal to 0.86.So it can be concluded that the model used in this study has a good model fit.