The intention to use QR code payment in an emerging market – the role of “Attitude” as mediator

As one of the applications of mobile banking, payment by scanning a QR code not only reduces the risks when paying via cards but also brings about more convenience than cash payment and more safety than other forms of payment, especially in the context of the current strong and complicated Covid-19 epidemic. However, the adoption of this payment technology seems to be left open in emerging markets, especially Vietnam. Within this context, this study conducted a survey on 441 customers about their intention to use QR code payment when shopping in Vietnam market. Research results show that personal innovation, perceived security and usefulness, perceived ease of use and facilitating conditions affect the attitudes towards intention to use the QR code payment service. In particular, perceived ease of use, perceived security and usefulness are the two factors that most strongly influence the intention to accept the use of QR code payment service though attitude mediator. The results of this research can assist managers with considerations in constructing development plan for new technological products and services in such an emerging market as Vietnam


I. Introduction
It can be said that science and technology is one of the fastest growing areas. Currently, the 4th scientific and technological revolution is taking place vigorously affecting all aspects of life in the society. More and more scientific and technological achievements are applied to life, bringing great benefits to people in many fields. Increased production values, improved quality of life, changing awareness and habits, etc. are the most visible impacts in today's society. Along with the development of e-commerce and information networks, it has facilitated the design of technology products and services to meet the needs of bringing convenience to consumers. The increasingly popular online shopping habit has led to increasingly diverse forms of payment. When making online shopping, consumers have a lot of payment options such as payment via card, payment by e-wallet, cash payment upon delivery, payment via mobile devices. Credit card fraud or credit cards with stolen personal information are problematic for customers and banks, especially in the case of money transfers or online shopping. Moreover, with the development of the field of mobile information, payment via mobile devices is gradually becoming a new trend, and technologies such as QR codes, contact payments, contactless and digitized card information are on the increase. Payment by scanning a QR code is one of the applications of mobile banking. QR technology reduces the risk of card payments, is much more convenient than cash payments, and safer than other forms of payment. With such tremendous benefits; however, not any customer can easily accept the use of this new technology utility, especially when the shopping habit of paying by cash or credit card of consumers is still  Hossain et al., (2018) in China… These studies show the factors that influence customers' use of QR codes such as information quality, system quality, interoperability and subjective norms, usefulness, and ease of use, acceptability. However, previous studies have focused on the influence of these factors on the traditional use of QR codes without exploring the role of mediator variables as a new perspective.
Moreover,different countries, different times with different cultures can greatly affect the development of the process of implementing QR code services of banks and the use of QR codes among different users in various ways. Therefore, determining which factors and the extent of influence that such factors have on the intention to accept the use of QR code payment when making purchases is an urgent issue in emerging markets like Vietnam.

II.
Theoretical background 2.1 Concept of QR code and research context: QR code is a matrix code, also known as twodimensional barcode, built in 1994 by the Denso Wave Company (Japan) with an aim of fast matrix decoding at high speed. QR stands for Quick Response. This is a barcode matrix that can be read by barcode readers or smartphones that can take pictures with a specialized application to scan bar codes. (Denso Wave; 2020). Payment method by scanning QR code is done according to the process that customers will open their mobile banking application, use QR Pay payment feature to scan QR code on the invoice via camera of mobile phone and mobile banking application will notify the bank. After checking, the banking system will execute the order to transfer money from the customer's account to the account of the point of sale. The amount of the invoice will be deducted from the customer's account via the customer's mobile banking application. Most of the current retail chain chains such as supermarkets and shopping centres are applying this form of payment. Payment by QR code scanning started from Japan and has spread around the world. In Europe and North America, QR codes have been booming since the 2000s, with marketers and manufacturers putting QR codes in many brands of products and services, from wine brands to soap bottles, candy packages ... (Okazaki et al., 2012). In Asian countries such as India, Indonesia, Singapore, Vietnam ... QR code payment is becoming more and more popular as consumers switch from the habit of using cash to payment using technology applications. In Vietnam, it is not only until the last few years that QR codes have begun to blossom and promote the role of payment activities of consumers. It can be said that the popularity of smartphones in the technological age has brought a "golden opportunity" for QR codes to develop payment functions, when previously this function was still gloomy in Vietnam. The popularity of smartphones, mobile networks 3G, 4G facilitate a favourable condition for suppliers to develop new forms of products and services to meet the diverse needs of customers. According to data from the General Statistics Office (2017),it was during the first 3 months of 2017 whenVND 20,000 billion was spent by the Vietnamese on buying smartphones. Amonga population of more than 90 million people with mostly young people, approximately 50% of Vietnamese people get a lot of exposure to the internet and up to 70% of the population use smartphones. With this rate, Vietnam is hence ranked high in Asia -Pacific and the world in the use of the internet and smart phones.According to Vietnam E-commerce Association (2019), the scale of Vietnam ecommerce market in 2025 will rank third in Southeast Asia, after Indonesia (100 billion USD) and Thailand (43 billion USD). In addition, most of the current customers who use smartphones are young, knowledgeable and interested in experience. They always update new technology utilities, especially the forms of payment integrated in mobile applications, making payment ISSN: 00333077 3442 www.psychologyandeducation.net connection easy and convenient without using cash or different types of bank cards. Therefore, mobile payment is having many development opportunities in Vietnam.

Attitude towards intention to use QR code payment
Attitude reflects favorable or unfavorable feelings that people show through their behaviors (Fishbein, 1963). Loudon and Della Bitta (1993) define attitudes as how people perceive an object as positive or negative, something beneficial or harmful, or something that enhances pros or cons. The definition regards attitudes as a feeling or assessment on objects. Lian et al. (2014) suggest that an attitude is an internal intention of an individual towards a target audience and a source, a preparation for a behavior. Attitude is therefore a complex type of psychological process that includes cognitive, emotional and behavioral trends and has long-term and consistent characteristics (Zheng et al.2018). Attitudes are important because they direct the people's thoughts (rational function), affect emotions (emotional function) and affect behaviors (action function). This is an important basis for marketers to deliver effective marketing strategy that influences decision-making and behavioral changes. Attitudes are important because they direct the people's thoughts (rational function), affect emotions (emotional function) and affect behaviors (action function). Attitudes affect the decision to consume, accept or abandon something -this is the behavior of a person. The different models such as TAM, TRA, TPB show that attitudes play an important role in consumer decision-making as they are the best predictors of behaviors. A positive relationship between attitudeand positive use has been highly supported by many scholars (Meharia, 2012); (Hansen, Risborg and Steen, 2012); (Lee and Ryu, 2013).

Theory of Techology Acceptance Model
(TAM) Based on a theory developed from TRA and TPB,TAM was proposed by Davis and Richard Bagozzi (Bagozzi, 1992;Davis, 1989). Indeed, significant value of TAM in predicting technology acceptance among potential users has been evident (Ajzen 1991). TAM focuses on explaining behavior acceptance and use of consumer technology.The TAM states that perceived usefulness (PU) and perceived ease of use (PEoU), determine a person's decision to accept a new technology (Davis, 1989)PEoU refers to a concep that using a particular system is effortless or simply easy to do (Davis, 1989;Taylor & Todd, 1995) and PU refers to "the degree to which a consumer believes that online shopping will provide access to useful information, facilitate comparison of offers and enable a faster checkout"(Vijayasarathy (2004,748). This model is applied in many fields such as mobile phones

Studies related to payment by QR code
There have been some studies related to the form of QR code payment in the world shown in the following The results indicate that the intention to use is significantly related to the QR code usage. Actual users use QR codes more than users experience and have a positive relationship between ownership of electronic devices and the use of QR codes. An early adopter has a negative relationship with the use of QR codes. In addition, perceived usefulness of QR codes and updating knowledge of electronic devices does not have a significant relationship with the speed of using QR codes.
The authors conduct online research on the internet in Spain.
Attitude, innovativeness and subjective norms, compatibility, security, personal innovativeness and mobility.

surveys
The study of the authors analyzed the user's acceptance of payment systems using QR codes on mobile devices. The outcomes illustrate that opinions, change and subjectivesstandards are the decisive factors for the intention to use this technology in the future. In addition to the basic elements of the technology adoption model (TAM), the authors also identified new variables that affect the acceptability of using the mobile QR payment system as compatibility, security, personal innovativeness and mobility. The research has shown that QR codes have a great influence on purchasing intent and customer satisfaction. The results of the paper also confirmed that QR codes affect the perception flow, the combination of factors affecting shoppers' satisfaction and ultimately the intention to buy.
Through the summary of these studies, it can be seen that there are many empirical studies on accepting new technologies of customers in the field of payment. In Vietnam, there have beenmany studies on accepting the use of new technology in new banking industry such as Mobile banking, E-banking .., but there has not been any official research project on using QR code payment in Vietnam. Therefore, it is highly necessary to study the intention to use QR code www.psychologyandeducation.net payment method when buying goods in emerging markets like in Vietnam and it is a new trend in Vietnam. These studies will be the basis for proposing a research model to serve in the Vietnamese context.

III.
Hypotheses and research model With the popularity of smartphones, laptops, and tablets in payment operations as today, the deployment of QR code payment services is an indispensable trend. There are many different studies around the world in the field of electronic banking (E-banking) as well as mobile banking (mobile banking) or the intention to use new technologies such as studies by Venkatesh et al.  2015). There are differences in the selection of factors to measure the level of technology acceptance in these studies. However, it can be summarized into the following factors: perceived usefulness and perceived ease of use will affect the attitude and intentions that lead to the acceptance of technology use by the research subjects. In addition, a number of factors such as perceived security, subjective norms are factors emphasized by the authors in a number of studies in the field of accepting mobile banking services and payment services by QR code. In short, it can be seen that the group of factors affecting the process of accepting QR code payment services on mobile phones will have appropriate variables such as perceived usefulness, perceived security, perceived ease of use, social norms are the factors that are suitable for the study on the intention to use QR code payment in mobile banking. These are independent variables that directly impacton the acceptance of using new technology services, or indirectly influence intention to use via mediator variable is "attitude". In Vietnam, this is also a new type of service, the customer experience of this service is not substantial. Within the payment via electronic banking, customers are more or less cautious of the safety and security. With the above characteristics, the authors choose perceived ease of use, perceived usefulness, social norms, perceived security as the factors that influence the extent of adoption of QR code payment method, whether it is through mediator variable "attitude" or not. The research hypotheses are as follows: * Perceived ease of use "Perceived ease of use is the degree to which a person believes that using a particular system will save effort" (Davis, 1989, p.320). There are many studies that also use similar concepts as "perceived ease of use" such as "complexity" and  2012), also confirms and demonstrates that ease of use is a factor influencing the intention to adopt new technology and services.The higher customers' perceived ease of use on QR code service is, the easier it is for customers to accept the service. Therefore, the following hypothesis is proposed: Hypothesis H1: Attitude is amediator variable between perceived usefulness and intention to use QR code payment when making purchases.
* Perceived usefulness Previous studies have shown a link between perceived usefulness and perceived ease of use, which influences attitudes and behaviors of adopting technology (David, 1989; Huarng, Yu and Huang 2010, Liébana-Cabanillas et al., 2014). A study in the field of accepting QR code payment technology in the world by Shin et al. (2012) also proved that this is one of the determinants of the customer's adoption of QR code technology. Therefore, the authors hypothesized: Hypothesis H2: Attitude is a mediator variable betweenQR codeintention to use and perceived usefulnesswhen making purchases.
* Perceived security Risk and safety assessment is a major concern in electronic payment systems (Ashrafi, 2008 (Lee, 2009). Therefore, perceived security of QR code payment systems must be under supervision (Schierz et al., (2010); (Meharia, 2012) for the success of using QR code in payment. The form of payment by scanning a QR code is normally connected to the customer's online account. Therefore, this factor plays an extremely important role in choosing the form of payment. Customers must feel the security, information security, accurate transaction process, account information and transactions are highly secure so as to decide to use this service. Therefore, the authors hypothesized: Hypothesis H3: Attitude is amediator variable between perceived security and intention to useQR code payment when making a purchase.
* Social norms Social norms refers to the level of individuals' perception that reliable people think they should use an innovation (Venkatesh et al., 2003). It acts as a combination of subjective norms components in the TRA model (Ajzen, 1991) and social images in the model (Moore and Benbasat, 1991). In the study of Taylor and Todd (1995), the subjective norms was determined to include the effects of the opinions of other talented people known as reference groups, including colleagues, friends, supervisors, and technology experts. Research shows that pressure from reference groups to adopt an effective innovation as it contributes to reducing risk through association with awareness. One component related to "social influence" is the "social factor". "Social element" is the case in which individuals representing a certain group of people are behaving in certain ways by members of society.Venkatesh et al. (2003) believe that "social norms" can significantly influence user intention when users make a required innovation. In these cases, "social norms" may be important in the early stages of acceptance but is unlikely to be used in the long term (Venkatesh and Davis, 2000). In fact, users with mobile banking services often spread to their colleagues, relatives, and partners. Therefore, the authors hypothesized: Hypothesis H4: Attitude is amediator variable between Social Norms and the intention to use QR code payment when making a purchase. Figure 1. The proposed research model

IV.
Methodology The study was mainly conducted through qualitative and quantitative research to collect primary data with two stages as follows: At the first phrase, the authors conduct qualitative research to adjust the research scale and complete the questionnaire for official research. The method used was in-depth group interviews with a small number of 10 random customers and 5 experts knowledgeable about the field of mobile payment at banks and the field of technology. The interview took place with each person on a set of questions, especially openended questions used for people invited to the interview about issues related to the use of QR code payment in Vietnam. The results were brought into stop in 15 people because the information obtained was overlapped in some people and new information was not exploited.
The first objective of the in-depth interview is to examine and screen independent variables in the theory model proposed by the authors and to preliminarily determine the relationship between the independent and dependent variables. This indepth interview will help the author identify the factors that are suitable for the characteristics of the QR code payment method in Vietnam and the factors affecting the acceptance of using QR code payment method when making a purchase. The next objective is to check the appropriateness of  (2008), to be able to analyze the exploratory factor, it is necessary to collect a data set with a sample size of at least 4 or 5 times the number of observed variables. To ensure the reliability of the data, the authors determined the sample size to be 500 people. The sample was selected as students and employees. The authors selected a convenient sampling method for undergraduate and graduate classes as well as handed out tickets directly at the counter at banks in Vietnam. A small number of samples were surveyed online in remote areas such as Hanoi, Vinh Phuc ... Quantitative research was conducted from May 2019 to January 2020 with interviews among customers who have used or have never used QR code payment. As a result, a total of 441 respondents answered the questionnaire completely. The data from 441 votes were encrypted, input, cleaned and conducted analyzed on SPSS 20.0 and 3.2.9 smartPLS software. V.

Sample description statistics
The survey was conducted by convenient samplingin VietNam.

Results of factor analysis of EFA
Discovery factor analysis is a quantitative analysis method used to shorten a set of many interdependent observation variables into a smaller set of variables (called factors) so that they are more meaningful but still contains most informative content of the original set (Hair &ctg 1998). Factor loading of less than 0.5 in EFA will be disqualified. Furthermore, a difference of no more than 0.3 between two factors loading (ie, a question uploading both factors) will be disqualified. The extraction method used isvarimax rotation with principal components and varimax rotation stops when extracting elements with eigenvalue = 1.The scale proves to be reliable when the total variance is> 50% (Gerbing & Anderson, 1988), provided that the KMO index> = 0.5.KMO acts as a factor that determines the appropriateness of EFA, 0.5 ≤ KMO ≤ 1, then factor analysis is suitable. The Bartlett's test considers the correlation between observed variables. If this test is statistically significant (Sig <0.005), the observed variables are correlated with each other on the whole (Hoang Trong & Chu Nguyen Mong Ngoc, 2008).
For independent variables: EFA analysis of 1st prefixes: KMO = .908, with the Signal of Bartlett's Test = 0.000 <0.05 for the data found to parse. In eigenvalue = 1.035> 1 with the bad method is 80.684> 50% (satisfactory), then there are 4 factors extracted from 15 observed variables. This prefix is a decantation of 80,684% the data variable. However, in the rotation match, it shows that some observed variables have a weight (Factor loading) less than 0.5, moreover, there is a difference of not more than 0.3 between the two loading factors, so it will be disqualified, which is variable PU1 . We continue to analyze EFA factor for the second time.However, some observed variables have a factor loadingof less than 0.5and there is a difference of not more than 0.3 between the two factor loading, so they will continue to be eliminated.The observed variablesthat are eliminated after the first and second EFA factor analysisinclude PU1, PU2, PU3, PU4.
In the third EFA factor analysis, KMO coefficient = .893, with significance level of Bartlett's Test = 0.000 <0.05, showed that the data is suitable for factor analysis. At eigenvalue = 1,238, 3 factors were extracted with the total variance explained at 81,003% (>50%). Thus, after the 3rd EFA analysis, these 11 observed variables ensure the EFA analysis standard (satisfactory) and they are included in the analysis of the PLS path model. Details of the analysis results are presented in the following For mediator variables "attitude" and "intention to use QR code payment" variable: By performing a factor analysis with the "Attitude" variable, we have KMO (Kaiser-Meyer -Olkin) value of 0.845 and Bartlett test with Sig of 0.000 ≤ 0.05. Furthermore, there is a single factor with an eigenvalue greater than one with a total variance of 78,837% greater than 50%. As a result, the variable "Attitude" is measured by four observed variables: ATT1, ATT2, ATT3, ATT4.
Similarly for the variable "Intention to use", the KMO value is 0.734 and the Bartlett test has Sig of 0.000 ≤ 0.05. Value of the total variance explained = 80,648%> 50%: satisfactory; then it can be said that this factor explains 80,648% of data variation and one factor is "intention to use" is measured by 3 variables ITU1, ITU2, ITU3.
Through the results of reliability analysis and EFA analysis, the actual research model is different from the theoretical model. The new model consists of three independent variables (Perception of security, Perceived ease of use, Social norms). The research hypothesis is proposed as follows: Hypothesis H1: Attitude is a mediator variable between Perceived ease of use and Intention to accept payment of QR code when making a purchase.
Hypothesis H2: Attitude is a mediator variable between Perceived security and Intention to use of payment QR code when making a purchase.
Hypothesis H3: Attitude is a mediator variable between Social Norms and the Intention to use of QR code payment when making a purchase.

Analysis of PLS path model
The determination of reliability and convergence validity is essential. According to theHair et al. (2017)  The coefficient of determination R 2 = 0.738 for the endogenous hidden variable «Intention to Use», in which factors affecting the attitude towards the intention to use of QR codes have the coefficient R 2 = 0.693, which means that factors explain 69.3% of «Intention to Use» through the variable «Attitude».
The internal model shows that Attitude has a great influence on Intention to Use (beta = 0.693), of which Perceived Ease of Use (beta coefficient = 0.455) and Perceived Security (beta coefficient = 0.332) are two factors that have a strong influence on «Attitude» and thereby Intention to Use. The results show that almost all customers highly appreciate the ease of use that QR code payment method brings. This can be explained by the fact that the most current banks supported the safety and convenience of QR code payment, people who are ready with smartphones and connected to the Internet will easily accept QR code payment. The results also show that social norms also has a significant influence on the attitude towards the intention to accept QR code payment (beta = 0.209), which is also easy to understand due to the fact that Vietnamese people have collective culture. Therefore, their use of products, especially new ones, will also be easily influenced by people around them.

Mediation Effects of "Attitude" in PLS
The impact of multiple mediators is analysed using a mediation processes of a PLS described by Nitzl et al. (2016). This approach will simultaneously examine various external and internal factors (i.e., perceived ease of use, perceived security and social norms, intention to use) to provide a clear and comprehensive panorama of the relationship between external and internal factors, as well as the attitude and intention to use in the research model. This process consists of 2 steps: Step (1) Determining the significance of indirect effects, Step (2) Determining the type of effect and / or of mediation. In step 1, PLS researchers must start by testing the indirect impactaxb of the two paths for significance. Alternatively, researchers should apply bootstrap routines to test the significance of the indirect effect a × b. SmartPLS can test the meaning of both internal and external models, using a process called bootstrapping. In this process, a large number of test samples (egg, 5000) are taken from the original sample with changes to create bootstrap www.psychologyandeducation.net standard errors, which will give relative T-values for checking the significance of structure lines. Bootstrap results show the relative stability of the data. The meaning of the indirect impact (axb) is shown in the In the subsequent step, with an aim to determining the impacts ofmediatorvariable, the significance of the direct effect must be assessed. If the direct effect is not significant, the impact will be "full mediation" and on the contrary, If the direct effect is significant, the impact will be "partial mediation" (Nitzl et al., 2016). Statistical significance of p-value <0.05, shown in the effects of PEoU on ITU (p = 0.011) and PS's impact on ITU (p = 0.000), showing that these are the "Partial mediation" impacts. Particularly, the impact of SN on ITU is not significant (p = 0.907). As a result, the attitude is "full mediation" between Social norms and the intention to use of QR code payment method when making a purchase.

VI. DISCUSSIONS AND IMPLICATIONS
This paper aims to understand the role of mediator variable "attitudes" in identifying factors that may influence Vietnamese consumers' intention to use QR codes, to provide Proposals and valuable information for bank managers as well as entrepreneurs to facilitate customers in the payment process, contributing to make the shopping process quick and efficient. The research results show that factors such as social norms, perceived ease of use, and perceived security all influence the attitude towards the intention to use QR code payment on customers' purchases. In particular, attitude plays a mediator role in the impact of these factors on Intention to Use. This also shows that creating a positive attitude towards the use of QR codes is essential, because it enables the enhancement of this form of consumer payment (beta = 0.693). Specifically, if Vietnamese consumers maintain a positive attitude with the intention to use QR codes, they will be more willing to engage in their QR code usage. This result can be particularly useful for bank managers or business people in general seeking to improve a positive review of QR codes. Moreover, the factors of Perceived Ease of Use (beta factor = 0.455) and Perceived Security (beta factor = 0.332) have the strongest impact. Therefore, in order to promote the form of QR code payment, increase familiarity with customers, firstly, banks need to coordinate with related parties to make the QR code payment process simple, easy to implement. Moreover, at the banks or transaction points, there should be a guide who can give clear explanation of the steps to install QR code payment application on smartphones, creating peace of mind, excitement,familiarity to the customer, helping customers to master new technology services, thereby increasing the intention to use this form of payment. Sometimes, the use of new payment services will make customers feel afraid. Therefore, in order to increase perceived security among customers, banks and transaction points need to perform accurately, clearly, conveniently and answer questions of customers quickly and promptly.
What's more, it is a necessity for banks to ensure absolute safety of customer information when using and have regulations on risk prevention in payment transactions by scanning QR codes. The increase in points of acceptance of QR code payment, the establishment of a wide distribution network in provinces and cities all over the country, the enhancement of discount methods, and customer support can make contributions into facilitating the best conditions for customers in their application of QR code payment.