“Determinants of intention to continue using internet banking: Indian context”

It is necessary to understand the customers’ perceptions of internet banking because it helps determining the direction and patterns of intention to continue using internet banking. This could also help bank policymakers to develop appropriate strategies to increase internet banking usage. The study aims to examine the determinants of user’s intention to continue using internet banking since there have been no systematic attempts to understand this aspect, especially in the Indian context. This research suggests and tests an extended model to predict the intention to continue using internet banking in India. The suggested study model was examined using survey data from 206 internet banking users. PLS-SEM was employed for data analysis. The findings imply that the most significant determinants of intention to continue using internet banking are service quality, trust, and user satisfaction. On the other hand, the study finds that intention to continue using internet banking is not impacted by system quality and information quality.


INTRODUCTION
Companies are working to gain more consumers by offering the highest comfort and security levels anywhere at 24/7.Comfort applications include, for example, goods delivery, health, and education.Internet banking helps this transformation.This service allows many banking services (e.g., bill payments, money transfers) to be conducted easily and securely at any time and from anywhere through the internet.Similarly, internet banking helps banking specialists to minimize strain in physical branches, reduce operating costs, and improve service quality (Rahi & Ghani, 2019;Foroughi et al., 2019).
In the initial phase, internet banking was used to offer information about banking services (Shankar & Jebarajakirthy, 2019).Nowadays, customers are using internet banking as a means of conducting many transactions, including account inquiry, account statement, order a checkbook, bills payments, money transfers, pay taxes, stock investment, and term insurance payment (Yu et al., 2015; Tam & Oliveira, 2016; Lin et al., 2020).Providing internet banking has become more of a "need to have" than a service that "nice to have" (Kesharwani & Singh Bisht, 2012).
Despite the possible benefits offered to consumers by internet banking, such as accessing services' banking at anytime and anywhere (Tam & Oliveira, 2017), the use of internet banking could have a strong impact on users' perceptions.Such perceptions could also affect internet banking's continued usage.Continued usage is identified as the user's decision to continue using certain IT already used by the individual (Nabavi et al., 2016).Prior research indicates that continued usage is more critical than the initial one, as the developing cost of a fresh customer may reach up to five times of retaining a present customer (Yuan et al., 2019).Therefore, banks must comprehend the key determinants of the continued usage of internet banking.Understanding why users continue to use IT is critical to the long-term survival and ultimate modern IT success (Bhattacherjee, 2001;Nabavi et al., 2016).It is necessary to understand the customers' perceptions of internet banking because it will help determine the direction and patterns of continued internet banking adoption.This could also help bank policymakers develop appropriate strategies to increase internet banking usage.
Only fewer studies in literature concentrate on the determinants of continued usage of internet banking services (Foroughi et Kumar et al., 2020).A systematic analysis by Shaikh and Karjaluoto (2015) showed no single research that explained the continuation of consumer behavior towards using technology in South Asia or Africa (majorly India, Bangladesh, Pakistan, Bhutan, and Nepal).Consequently, further research is required to determine whether essential factors in the adoption cycle impact continuous behavior in the internet banking context.Therefore, the present study aims to bridge the research gap by suggesting an extended model based on DeLone and McLean's model to verify the customer's continuity intention towards using internet banking in India.Firstly, DeLone and McLean's mode; is extended by the trust factor as another critical construct in the internet banking context.Secondly, this research focuses on switching to continued internet banking usage via information, system, and service quality, trust, and user satisfaction with internet banking.Finally, PLS-SEM was utilized to examine the proposed research model.Although the quality of information, system, and service may impact the first-time adoption of IS, the relationship between "satisfaction" and "intention/use" needs to be understood in terms of continued use (Teo et al., 2008).Moreover, continuous use is essential for IS to be truly capable of generating net benefits (Bhattacherjee, 2001;Tella, 2012).Based on this argument and besides adding "trust" construct, the present study uses "intention to continue usage" instead of "intention/use" (see Figure 1).This would activate the "intention/use" construct in the updated model of D&M.
Some studies were reviewed to justify the relationships, in Figure 1, in the internet banking context.

AIMS AND HYPOTHESES
The present study aims to verify the determinants of user's intention to continue using internet banking since there have been no systematic attempts to understand this aspect, especially in the Indian context.Based on the literature, the following hypotheses were developed: H1: Information quality (InfQ) impacts user satisfaction.
H2: Information quality (InfQ) impacts intention to continue using internet banking.
H4: System quality (SyQ) impacts intention to continue using internet banking.
H6: Service quality (SerQ) impacts intention to continue using internet banking.
H8: Trust (Tr) impacts intention to continue using internet banking.
H9: User satisfaction (US) impacts intention to continue using internet banking.

METHODOLOGY
Since

RESULTS
To meet SEM requirements, both measurement and structural models were applied (Gefen et al., 2000).The measurement model was examined by reliability and validity, while the structural model was tested by testing hypotheses, assessing R 2 , Q², effect size (f²), and model fit.For the CMB, recent work stressed the significance of evaluating CMB's impact on statistical analysis outcomes (Chin et al., 2012).According to Kock (2015), the appearance of VIF greater than 3.3 is suggested as an indicator of multicollinearity problem, and that a model might be CMB.As clarified in Table 3, all VIF scores are fewer than 3.3, confirming the absence of multicollinearity problem and CMB.

Hypotheses testing and R 2 assessing
To test the hypotheses (β-value, t-value, and p-value), a bootstrapping with 5,000 samples was con-ducted.In Figure 2, β-values of the relations between model constructs can be seen.The t-and p-values are employed to examine if β-values are statistically significant (i.e., at *, **, or ***) or not.The results of hypotheses testing are presented in Table 4.
R 2 implies the variation proportion in the dependent variable that independent variable/s collectively interpret.The R 2 value should be at least 10% (Falk & Miller, 1992).Cohen (1988) indicates that the value of R 2 above 0.26 is substantial.Accordingly, as displayed in Figure 2, the R 2 values for US and ICU are 0.661 and 0.620, respectively, which are considered substantial.

Effect sizes (f²) and predictive relevance Q²
The p-value can display the existing effect, but it cannot uncover f² (Cohen, 1988).Effect size in each path in SEM can be revealed by Cohen's f² (Urbach & Ahlemann, 2010).For explaining f² by Cohen (1988): the score of > 0.35 refers to a big effect, 0.15-0.35refers to a moderate effect, 0.02-0.15refers to a small effect, and lower than 0.02 refers to no effect.Table 4 displays all f² scores, which appear to be satisfactory.
In SmartPLS, the blindfolding procedure is used to evaluate Q².The interval point for Q² is above zero.Q² scores over zero imply that the model has a predictive relevance (Hair et al., 2011).As reported in Table 4, all Q² scores are over 0.

Model fit
In PLS-SEM, the model fit evaluation is done employing two ways: SRMR and tests of the overall model fit (Schamberger et al., 2020).SRMR The measure of GoF to identify if GoF scores are not fit, small, moderate, or high is provided by Wetzels et al. (2009) as follows: GoF lower than 0.1 (not fit); 0.1 to 0.25 (small); 0.25 to 0.36 (moderate); above 0.36 (high).In this research, the model's GoF, resulting from the formula mentioned earlier, is > 0.36, which is considered high.

DISCUSSION
Based on DeLone and McLean (2003), the present study examined the determinants of the user's continuity intention towards using internet banking.The results summarized in the prior section show that except H2, H3, H4, and H5, all remaining hypotheses were supported and interpret 66.1% and 62.0% of the variance in UF and ICU.Based on results, InfQ is the key determinant affecting customers' satisfaction with internet banking (β = 0.376, p < 0.001).This result reveals that the higher level of InfQ offered by service suppliers increases the level of satisfaction with internet banking.For instance, if the service supplier's information is relevant, accurate, timely, and perfect, this will encourage customers to develop their satisfaction towards internet banking.A similar finding was also found in previous IS studies (S.

Theoretical implications
It should be noted that this research was conducted in India.India is globally ranked as the second-largest market in terms of total internet users (IBEF, 2020) and has seen tremendous growth in e-banking in the past two decades (Shankar & Jebarajakirthy, 2019; Sujeet & Srikrishna, 2014).However, there is insufficient research on internet banking and continuous customer behavior in the Indian context.Thus, the present study aimed to bridge the research gap by suggesting an extended model based on DeLone and McLean to verify the user's intention to continue using internet banking in India.
In addition to including trust as another significant construct in the internet banking context.This research has revealed useful insights into the behavioral factors of intention to continue using internet banking in the case of India.

Practical implications
From a practical viewpoint, the fundamental aim of IS use studies is to guarantee that any IS is more effective for the community.This is only possible when IS is continually used (Bhattacherjee, 2001;Tella, 2012;Veeramootoo et al., 2018).This study provides an obvious view of the key factors influencing intention to continue using internet banking.The findings from this study can be useful for bank policymakers who can use such results to enhance their internet banking channels in favor of continued usage.The findings imply that the most significant drivers of intention to continue using internet banking are service quality (SerQ), trust (Tr), and user satisfaction (US).
Service quality (SerQ) is a key determinant of intention to continuance using internet banking.Therefore, banks should concentrate on customer's needs, support the customer with prompt and dependable services simultaneously, and present sufficient knowledge, personal care, and attention.This, in return, will motivate further continued usage of internet banking services.
Indian banks need to focus on service quality and gain the trust of the customers.It is important that customers feel safe during online transactions.Furthermore, since user satisfaction is the key factor, banks in India must customize services with effective administration to achieve customer satisfaction, motivating the intention to continue using internet banking.

CONCLUSION
The study aimed to examine the determinants of user's intention to continue using internet banking since there have been no systematic attempts to understand this aspect, especially in the Indian context.The PLS-SEM analysis findings imply that service quality, trust, and user satisfaction are the most important determinants of intention to continue using internet banking.In other words, the results obtained suggest that the higher level of service quality, trust, and user satisfaction in internet banking would help to maintain existing customers and attract new and prospective ones.
This study had some limitations; first, it has not considered the moderate impact of demographic variables; detecting the impact of demographic variables on continuance use would supply deeper insights.Second, the study expanded DeLone and McLean's model by adding trust construct only.Therefore, future research should apply the proposed model by adding more important constructs such as interest.Third, the study was conducted in India, a large multicultural and multilingual developing country.Thus, it would be intriguing to conduct a similar study to test and verify the model in other developing countries' contexts.Past research found an enormous variation in the expected effects with various types of users, cultures, and systems.Fourth, a larger sample may provide greater generalization ability of the proposed model tested.Finally, future research is also required to continue evaluating IS success in different contexts and countries to determine the relationship between costs and benefits of internet banking success.

Figure 2 .
Figure 2. PLS algorithm results al., 2019; Yuan et al., 2019).In the Indian context, most researches concentrate on customers' acceptance or adoption of internet banking (e.g., Sujeet & Srikrishna, 2014; Safeena et al., 2011; Bashir & Madhavaiah, 2015; Kaur & Malik, 2019; (Lin et al., 2020)t relieves the fear that the partner in the exchange will act opportunistically.In the internet banking context, trust is described as a customer's guaranteed confidence in an internet banking provider's capability to offer reliable online services(Bashir & Madhavaiah, 2015).Trust was determined as one of the core factors in internet banking literature(Lin et al., 2020).Literature supposes that customers who trust online banks probably carry on utilizing services provided by those banks (Yu et al., 2015; Vatanasombut et al., 2008).Trust will not just affect the intention to continue using internet banking, yet other factors as well (Lin et al., 2020).Customers will have higher satisfaction if they have higher trust in it (Lin et al.Hsiao et al., 2016).At the same time, Hsiao et al. (2016) and Veeramootoo et al. (2018) reveal that satisfaction has a robust effect on continued usage intention of e-filing and mobile social apps, respectively.Similar results were also found by Foroughi et al. (2019) in the Malaysian mobile banking context.
(Kassim et al., 2012)1)008).According toMolla and Licker (2001), trust refers to two important issues: security and privacy.Besides website and content quality, customers can turn towards security and privacy issues.Teo et al. (2008)define trust as Figure 1.The proposed model , 2015; S. Sharma & M. Sharma, 2019b).Yoon (2002)further contended that the result of online trust may be satisfaction and indicated a positive association between trust and satisfaction.Teo et al. (2008)reported that numerous studies on online trust have dealt with behavioral intentions as the ultimate dependent variable.Trust and satisfaction relationship has been explored in relatively fewer researches.Accordingly, the current study focuses on customers' trust through its impact on both satisfaction and intention to continue using internet banking.Satisfaction is recognized as a key factor in the continuity of any IS usage(Rahman et al., 2017), including internet banking.User satisfaction appraisal is the most commonly utilized indicator of IS efficiency because of its high level of validity and ease of verification(Cheok & Wong, 2015).For advocating users to accept an information system use, it is important to ensure a suitable level of compatibility between the user's needs and expectations(Kassim et al., 2012).vices.This, in turn, may encourage or discourage intention to continue using internet banking.In marketing literature, research on satisfaction and continued usage has protruded as a dominant is-sue in IS (

Table 1
(Kelloway, 1998;Hair et al., 2014)ies, a fivepoint Likert scale was employed for all indicators.Data were gathered from internet banking's users with the assistance of an online questionnaire via Google Docs.This questionnaire was sent via different social media networks like mails.In response, the questionnaire was completed by 206 internet banking users, all of which were valid for analysis.A sample that exceeds 200 is proper to provide sufficient statistical power for data analysis(Kelloway, 1998;Hair et al., 2014).Thus, the sample size of 206 is large enough for data analysis.
McLean (2003)s to verify the user's continuance intention towards using internet banking based on DeLone andMcLean (2003), a quantitative approach is employed.To achieve this, a

Table 2 .
(Urbach & Ahlemann, 2010)sually recommendable to guarantee reliability.The threshold of α is ≥ 0.7(Hair et al., 2014).Composite reliability (CR) is also preferred with scores ≥ 0.70(Urbach & Ahlemann, 2010).As shown in Table3, both α and CR have values > 0.70, reflecting the reliability of the measurement instrument used.Factor loadings and cross-loadings (CMB)are also deliberated in this research.A multicollinearity issue is not desirable in any research.To evaluate the multicollinearity issue, a VIF is usually employed (O'brien, 2007).Note: Factor loadings are remarked in bold.

Table 3 .
Measurement model and multicollinearity examination Note: AVE's square root is in bold.