Enhancing consumers’ self-reported loyalty intentions in Islamic Banks: The relationship between service quality and the mediating role of customer satisfaction

Abstract The main objective of this paper is to investigate how service quality and customer satisfaction are correlated to self-reported loyalty intentions in Islamic banks. The paper presents primary data collected by self-administered questionnaires involving a sample of 655 respondents from all local Islamic Banks in Jordan. The results show that compliance, tangibility, responsiveness, assurance, and empathy positively linked to self-reported loyalty intentions, with compliance and assurance appearing as the greatest correlation with self-reported loyalty intentions compared to other service quality dimensions. Islamic banks should focus on continuous improvement of service quality due to the direct correlation with customer satisfaction and loyalty. It is important for Islamic banks to review and endorse all policies and procedures to ensure that their documents and investments are undertaken in line with the requirements of Islamic law. Managers should pay special attention to the service quality provided by their employees and develop their skill set through training and education programmes based on Islamic principles. As the first study of its kind in Jordan, the paper’s contribution stems from filling the research gap in examining the relationship between the various service quality dimensions and self-reported loyalty intentions in Islamic banks using the CARTER model.


PUBLIC INTEREST STATEMENT
Islamic banks and non-Islamic banks operate in different realms with regard to acceptable revenue stream. For example, Islamic banks cannot finance project(s) which conflict with the moral value system of Islam such as financing alcoholic breweries and nightclubs, casinos and gambling establishments, assassination, consumption of pork, or any other activity prohibited by Islam known to be detrimental to society. Due to the cost-savings associated with maintaining current customers against spending heavily on new ones, Islamic banks have been forced to reevaluate their current business practices. Therefore, as a group of researchers working in different fields of consumer behaviour and services marketing, our social responsibility relies on creating research that helps this type of responsible business to continue to improve its service quality. It is hoped this study provides relevant recommendations and strategies that will improve their relationship marketing programs to attract and retain customers.

Introduction
Customers have more power than ever before due to better education and easy access to banking channels. As such, they demand more from the products and services they require (Angelova & Zekiri, 2011;Schiffman and Wisenblit, 2019). The 21 st century consumer market raises many questions to those businesses that need to respond to this new era of consumerism (Angelova & Zekiri, 2011). Moreover, because of dramatic changes in the business environment, financial services today, are looking for innovative ways to market and achieve competitive advantage, the aim being to increase customer satisfaction and enhance loyalty (Astuti & Nagase, 2014;Buracom, 2002). This is a position which has necessitated marketers to amend their thinking about the significance of customer relationships.
Establishing, building, and maintaining good customer relationships have become the basis for profitable growth (Angelova & Zekiri, 2011;Astuti & Nagase, 2014). In other words, building customer relationship is the backbone for all financial service industries. Consideration is now given to dimensions of service quality, customer satisfaction and customer loyalty, which at present are the main concerns for businesses in the services sector. Increased emphasis on these issues will lead to an enhancement of a business's performance and will translate into higher profits (Angelova & Zekiri, 2011;Shaon & Rahman, 2015). Given that banks and financial services offer relatively similar products and services, the best way for them to create competitive advantage is by the quality of services offered (Royne Stafford, 1996).
Offering high-quality service leads to reduced perceived risk, which has been proven to have a positive impact on customer satisfaction and loyalty (Ganguli & Roy, 2011). Customer satisfaction and loyalty are broadly recognised as significant sources of competitive advantage for all financial services industries. However, reaching these goals can be difficult and how to achieve this is still unclear for marketing practitioners. Research about customer attitudes towards banks, found that 13% of customers chose to close their accounts because of an encounter with a rude or unhelpful employee; 11% closed their account because they felt their bank is emotionless and impersonal; 16% closed their accounts because of a general perception of poor service (Baron et al., 2009).
In most cases, attempting to achieve exclusive customer loyalty has been challenging. A more attainable objective would be to strive for customer loyalty which could lead to increased market share and overall profitability (Keiningham et al., 2005). Consequently, loyalty is considered a main factor in measuring the success of banks (Casaló et al., 2008). For instance, pressure from shareholders to increase profitability has forced banks to switch from a strategy based on quick sales method, to a strategy that prioritizes long-term relationships with customers (Al-hawari, 2015).
Islamic banks and non-Islamic banks operate in different realms with regard to acceptable revenue stream (Naser & Moutinho, 1997). However, Islamic banks operating in Jordan still face high levels of competition because they all offer similar products and services. This makes it imperative for banks to highlight relationship marketing to gain customer loyalty (Fararah & Al-Swidi, 2013;Naser et al., 1999). Such market conditions require Islamic banking institutions to shift focus from simply increasing customer numbers, to building a strong base of loyal customers (Akroush, 2009). In other words, most banks try to maintain large numbers of customers for the longest time possible. This is a strategy that Islamic banks also practice (Ahmed, 2011). The logical pathway for Islamic banks to maintain market share is to build on their relationship with existing customers. Due to the cost-savings associated with maintaining current customers against spending heavily on new ones, Islamic banks have been forced to re-evaluate their current business practices. This position suggests that Islamic banks will need to develop products and services that will satisfy a highly diverse customer base to remain competitive in a global environment (Anouze et al., 2019). These observations give rise to the following research questions: This study has adopted relationship marketing theory. A main objective of relationship marketing theory is the identification of main drivers that impact important outcomes for the businesses/banks and a sound understanding of the causal relationships between these drivers and outcomes (Hennig-Thurau et al., 2002). Loyalty is a key objective of relationship marketing (Abdullah et al., 2014;Fiol et al., 2009;Hennig-Thurau et al., 2002), occasionally even equated with the relationship marketing concept itself (Hennig-Thurau et al., 2002). Therefore, this study aimed to answer research questions about determining the main drivers of customer loyalty, and to investigate the CARTER model of customer loyalty within the Islamic banking sector. The model proposes that compliance, responsiveness, reliability tangibility, empathy, and assurance correlates to customer loyalty, which, in turn, influences Islamic banks' performance. Hence, this study intent to add to existing literature in several ways, namely: • A limited number of studies have been conducted to examine the relationship between different service quality dimensions and self-reported loyalty intentions, particularly those based consumer's decision-making process in emerging countries such as Jordan (Anouze et al., 2019).
• Loyalty research has been based on theoretical frameworks developed in Western countries (Kueh & Ho Voon, 2007). However, few studies have focused on the Islamic, and more specifically, the Arabic culture (Anouze et al., 2019).
• Additionally, the number of service quality dimensions varies from one country to another and from one industry to another (Abu-El Samen et al., 2013). Therefore, the current study has adopted the CARTER model to be a useful measure of quality in Islamic banks. It is of notice that few studies have been carried out using the CARTER model (e.g., Fauzi & Suryani, 2019;Lone et al., 2017). This study will attempt to use the CARTER model in a novel way to better estimate the prospect of repeat dealings by measuring its' correlation with self-reported loyalty intentions, instead of only measuring its' correlation with the less reliable customer satisfaction.
• Although Jordan is a relatively small country, it has a high level of competition in the banking industry. This is due to foreign players entering the Jordanian banking market and increasing the competition to attract and retain customers, because Jordan is considered one of the region's safest countries (Anouze et al., 2019). It is hoped this study provides relevant recommendations and strategies that will improve their relationship marketing programs to attract and retain customers.

Islamic banking in Jordan: brief background
The number of licensed banks operating in Jordan rose from 21 in 2003 to 26 at the end of 2018. Of this total, 16 are Jordanian (three are Islamic banks), and 10 are foreign (Association of banks in Jordan, 2018). The population of Jordan is estimated at 6.5 million, 92% of which are Muslim, a position which proved alluring for the Islamic banking sector (Mahadin & Akroush, 2019). However, the total asset of banks operating in Jordan increased by 3.02 % at the end of 2018, reaching to JD 48.57 billion, distributed as follows: JD 37.64 billion for Jordanian commercial banks (77.49% of the total assets of banks operating in Jordan); JD 7.95 billion for the Islamic banks (16.37% of total assets) and JD 2.99 billion for foreign banks (6.15% of total assets).
Noticeably, market share of Islamic banking assets in Jordan rose from 13.5% in 2008 to 21.8 % at the end of 2014. The total number of employees in banks operating in Jordan increased from to 21,038 employees at the end of 2017 to 21,262 employees at the end of 2018, 4294 employees in Islamic banks (20.2% of the total number of employees) (Association of banks in Jordan, 2018). This demonstrates that Islamic banks can compete against commercial banks. Therefore, Islamic banks in Jordan should develop service quality, banking operations, efficiency, and experience by the banks' employees to enhance customer loyalty (Mahadin & Akroush, 2019).
Islamic banking in the Kingdom of Jordan plays a vital role in its economy, industry, and social life. This is because the key mission of Islamic banks is the achievement of social and economic development, by delivering special financial services to customers which are consistent with standards of Islamic law (Metawa & Almossawi, 1998). For example, Islamic banks cannot finance project(s) which conflict with the moral value system of Islam such as financing alcoholic breweries and nightclubs, casinos and gambling establishments, assassination, consumption of pork, or any other activity prohibited by Islam known to be detrimental to society (Bellalah & Masood, 2013).
During 2018, Islamic banks in Jordan continued to support community and national events, providing donations for these different organisations. They also provide a wide range of products and services as an alternative for interest-based products, the aim being to attract potential customers involvement in new investment opportunities (Abu-Alkheil et al., 2012). Instead of charging interest, Islamic banks use the concept of profit and loss sharing (PLS). This means customers (debtors) and Islamic banks (provider of the funds) share the risks; both benefit when earnings increase and suffer when earnings decrease (Mbawuni & Nimako, 2017). Parasuraman et al. (1988) developed the SERVQUAL model to evaluate service quality, consisting of five elements: (i) responsiveness, (ii) reliability, (iii) tangibility, (iv) empathy, and (v) assurance. The SERVQUAL model also identified 22 items that cause variance between customer expectations and perceptions. According to Pakurár et al. (2019), the modified SERVQUAL model is useful for addressing customer satisfaction in the Jordanian banking sector. Furthermore, Buttle (1996) is opined that SERVQUAL's five evaluative elements have not been endorsed entirely. Investigators include other elements in the original model, as they believe the original five elements of the model are unable to measure service quality. For example, Othman and Owen (2001) in their a paper entitled "Developing an Instrument to measure customer service quality in Islamic banking", concluded that it is significant for Islamic banks to implement quality of service approach. This, they believe, would help create methods to provide financial services to customers, and deal with fierce rivalry between traditional banks and new entrants into the marketplace. Moreover, their study also has improved a tool to evaluate service quality in Islamic banking with six elements called the CARTER model. The CARTER model is a tool that can be utilised to express and evaluate Islamic banking service quality, producing effective valuation instruments. This model considers the five SERVQUAL elements, and an additional element which is appropriate for the Islamic banking business, called "Compliance with Islamic law". Within the "Compliance with Islamic law" are elements such "keep on Sharia' and principles"; "setting up of Islamic goods and services"; "no interest paid or taken on savings and loans"; and "running of profit-sharing investment products".

Self-reported loyalty intentions
From a behavioural viewpoint, loyalty is defined as repeated transactions over a defined period. It is sometimes measured by regular purchase of products and services, purchasing more and another product from the same company (cross-buying), and recommending the brand/company to other consumers (Ball et al., 2004;Day, 1969;Gronholdt et al., 2000). Mechinda et al. (2009) define self-reported loyalty intentions as "Actual conduct, as a sequence of purchase, as proportion of market share, as probability of purchase, as duration, as frequency and as intensity" In the context of this study, according to Ehigie (2006) self-reported loyalty intentions is demonstrated when customers of a bank are willing to buy new services but also recommend the bank's services to others. According to Dick and Basu (1994), having loyal customers help businesses save numerous costs, because acquiring new customers is more expensive than nurturing existing ones. Furthermore, loyal customers tend to be less price-sensitive (Rundle-Thiele & Maio Mackay, 20011;Mao, 2010), are less likely to complain (Taap et al., 2011), and less inclined to switch to competitors (Buchanan & Gillies, 1990). The belief of Story and Hess (2006) is that loyalty behaviours are expressed in two ways: (i) purchase (primary) behaviours and (ii) nonpurchase (secondary) behaviours.
Primary loyalty behaviours take the form of frequency, volume, share, and retention. They are easily measured, can be measured in currency, transforms directly into revenues and profits, but are not necessarily predictors of future behaviour. However, according to Kumar and Shah (2004): "A majority of existing loyalty programs follow these measures to reward self-reported loyalty intentions. That is, the more you spend with the company, the more rewards you earn." Secondary loyalty behaviours take the form of referrals, advocacy, endorsements, and selective exposure to alternative brands. These behaviours are not easily measured, and their correlation with profits and revenues is less direct but greater in magnitude. Oliver (1997) proposes four stages of loyalty, each consecutive stage stronger than the preceding one. This suggests theoretically that loyalty of customers are developed in four stages: (i) cognitively, (ii) affectively, (iii) conatively, and (iv) behavioural, all described as "action loyalty". Estiri et al. (2011) add that two dimensions, customer satisfaction and service quality, have become significant for marketing theory and practice.

Distinctions between service quality and customer satisfaction
Service quality and customer satisfaction are two concepts identified as having similarities and differences between them. Although both involve evaluation of two specific expressions: 1) expectancy-perception as in service quality (Parasuraman et al., 1985(Parasuraman et al., , 1988, and, p. 2) expectancy disconfirmation as in satisfaction (Rust & Oliver, 1994), there have been debates over whether service quality and satisfaction are sufficiently distinct from each another (Johnston, 1995). However, some service researchers (e.g., Brogowicz et al., 1990), position the two as equivalent concepts that could be considered together. Rosen and Surprenant (1998) are opined that several administration experts utilise the term customer satisfaction as a cumulative term of service quality, without significant explained reason about the theoretical differences between both concepts.
Despite debates regarding similarities and differences between service quality and satisfaction, market researchers (e.g., Boulding et al., 1993;Cronin & Taylor, 1992;Taylor & Baker, 1994), agree that both constructs are separate and unique although they share a close relationship. Regarding differences between service quality and satisfaction, expectations in the gap model were originally based on an ideal standard, whereas within the disconfirmation paradigm uses predictive (would be) expectations as the norm (Rust & Oliver, 1994). Similarly, Cronin and Taylor (1992) and Reeves and Bednar (1994), explain that expectancy in the case of service quality reflects what customers' feel a service "should" offer or "something that should be offered by the company". In the case of satisfaction, expectation reflects predictions of what "would" happen during future transactions, or "something that will happen". Boulding et al. (1993) remind researchers that the term "expectations" should be treated differently in satisfaction and service quality studies, since the term "expectations" can be a potential source of confusion.
Another dominant view relating to satisfaction and service quality suggests that service quality is mostly a cognitive judgement, whereas satisfaction includes a number of cognitive and affective judgements such as equity, attributions, cost/benefit analyses that influence satisfaction (Rust & Oliver, 1994;Oliver, 1993). The cognitive status of service quality is strongly implied in the SERVQUAL scale, which assumes that consumers apply a mental calculus to reach an evaluation (Choi et al., 2004). Although early service quality researchers (e.g., Carman, 1990;Parasuraman et al., 1985Parasuraman et al., , 1988 defined satisfaction as an antecedent of service quality, it has generally been accepted that satisfaction is a consequence of service quality (e.g., Han et al., 2008;McDougall & Levesque, 2000;Rust & Oliver, 1994;Oliver, 1999;Cronin & Taylor, 1992).
The reasoning behind the claim that satisfaction leads to service quality perception, is that service quality represents a global judgement formed from transaction-specific assessments which are the content of customer satisfaction (Carman, 1990;Parasuraman et al., 1988). This suggests that conceptualised perceived service quality is a long-run overall evaluation of a service, whereas satisfaction is a transaction-specific evaluation. Based on these conceptualisations, the suggestion is that incidents of satisfaction over time result in perceptions of service quality. Interesting, the contrasting view by Oliver (1993), is that quality existed ahead of satisfaction. Cronin and Taylor (1992) echo this view, stating that service quality is an antecedent of consumer satisfaction which influences purchase intentions; service quality had less impact on purchase intentions than customer satisfaction.
The conclusion that satisfaction and service quality are unique concepts poses the question: "What then is the specific nature of the relationship between these concepts?" Bitner and Hubbert (1994) recommend that both concepts are dependent on each other. For example, Iacobucci et al. (1995) are opined that satisfaction will not be developed unless service quality is based on customers' desires.

Research model and hypotheses
The current study has adopted CARTER model to be useful to measure quality in the Islamic banks. The research model and hypotheses are shown in Figure 1.

The relationship between compliance and self-reported loyalty intentions
Compliance describes the ability of Islamic banks to consistently adhere to principles of Islam in exchanges or transactions (Misbach & Hadiwidjojo, 2013;Othman & Owen, 2001). One early patronage study on Islamic banking by Erol and El-Bdour, 1989, found that religious motivation is not a primary criterion, nor does it play any part in the selection of interest-free bank services in Jordan. Haron et al. (1994) reported similar results of the attitudes of Malaysian customers toward interest-free banks. Their findings have been disputed by later researchers (e.g., Naser et al., 1999;Othman & Owen, 2001), who confirm that compliance with Islamic principles is the most significant contributing factor for customer satisfaction and dealings with Islamic banks in Kuwait, Jordan and Bahrain, respectively.
A recent study by Dandis and Wright (2020), also confirmed that compliance is one of the main significant dimensions of service quality leading to attitudinal loyalty in Islamic banks in Jordan. Asnawi et al. (2018) found that perceived behavioural control and religiosity is the significant predictor of the intention to consume halal products in international chain restaurants. In relation to this and the fulfilment of customers' needs in service provider, marketers believe that the religion factor often impacts customer behaviour and decision-making when purchasing a product or service (Abedniya & Zaeim, 2011). The following hypothesis is offered in our study on service quality and self-reported loyalty intentions in Jordanian Islamic banks: H1: Compliance has a positive correlation with self-reported loyalty intentions.

The relationship between tangibles and self-reported loyalty intentions
Tangible qualities of service refer to the appearance of the physical surroundings of the service, consisting of physical facilities, building designs, personnel, tools or equipment, way of communication of the service, and other customers in the service facility (Parasuraman et al., 1994). Nusairat et al. (2020) have reported a significant positive effect of restaurant design on customers' behavioural intentions. In this context, physical clues can play an important role in helping customers form an impression of the service they are getting (Zeithaml et al., 2018). Chi Cui et al. (2003)  of service quality by customers in such various cultural environments. Chen et al., found that physical environment quality and personal interaction quality have a positive correlation with behavioural intentions relating to customer loyalty. The following hypothesis is offered in Jordanian Islamic banks: H2: Tangibles have a positive correlation with self-reported loyalty intentions.

The relationship between reliability and self-reported loyalty intentions
Reliability refers to the ability to perform the promised service dependably and accurately (Parasuraman et al., 1994). Kandampully (1998) emphasises that fulfilment of the service promise may inspire a long-term relationship, positively affecting long-term customer retention and sustainment, and subsequently reduce the likelihood of customer defection. McDougall and Levesque (1994) claimed that the major strategic focus should be "getting it right the first time", to ensure the performance segment is completely satisfied and remains loyal. Although some studies (e.g., Kashif et al., 2015;Rehman, 2012), found that reliability had an insignificant influence on customer satisfaction and customer loyalty, other empirical researches (e.g., Kheng et al., 2010;Shaikh & Siddiqui, 2018) reported a positive relationship between reliability and customer loyalty. The following hypothesis is offered in Jordanian Islamic banks: H3: Reliability has a positive correlation with self-reported loyalty intentions.

The relationship between responsiveness and self-reported loyalty intentions
A responsiveness service refers to the willingness of banks to provide accelerated service to customers (Parasuraman et al., 1988). Shao Yeh and Li (2009) expressed that responsiveness is essential as a measure of quality, as well as a diagnostic tool to reveal the service quality. Oly n.d. ubisi and Kok Wah (2005) recommend that personnel commitment to provide quality services, skilful handling of complaints, and efficient delivery of services resulted in long-term satisfied customers. Although Rehman (2012) found an insignificant relationship between the responsiveness dimension and customer satisfaction in Islamic banks in UAE, other empirical researches (e.g., Badara et al., 2013;Shaikh & Siddiqui, 2018), reported a positive relationship between responsiveness and customer loyalty, on the contrary. The following hypothesis is offered in Jordanian Islamic banks: H4: Responsiveness has a positive correlation with self-reported loyalty intentions.

The relationship between assurance and self-reported loyalty intentions
Assurance refers to the knowledge and courtesy of employees and their ability to convey trust and confidence (Parasuraman et al., 1988(Parasuraman et al., , 1994. According to Abbas et al. (2003), sufficient knowledge in various Islamic banking instruments compounded with other relevant aspects such as ethics, professionalism, duties and responsibilities towards customers and organizations, will enhance efficiency and smooth-running of Islamic banking. Arasli et al. (2005) suggest that if banks can instil feelings of confidence in their customers regarding the way they handle customers' affairs and the safety of bank transactions and if they can deal with customers in a professional and competent way, then, customers will be satisfied with their bank and be willing to recommend it to others. Pakurár et al. (2019) found a significant relationship between empathy and customer satisfaction in Islamic banks in UK and Pakistan. The following hypothesis is offered in Jordanian Islamic banks: H5: Assurance has a positive correlation with self-reported loyalty intentions.

The relationship between empathy and self-reported loyalty intentions
Empathy refers to the level of caring and personalised attention the bank provides for its customers to make the customers feel valued and special (Parasuraman et al., 1988(Parasuraman et al., , 1994. Sin et al. (2002) describe empathy as dealing with a business relationship that enables two parties to see the situation from the other's perspective i.e. seeking to understand somebody else's desires and goals. For example, if customers feel they receive individualised and quality attention there is the possibility they will do business there again (Parasuraman et al., 1994). Dandis and Wright (2020) found empathy to be the most influential factor leading to attitudinal loyalty. Similarly, Shaikh and Siddiqui (2018), Taleghani et al. (2011), andButcher et al. (2001) report a positive relationship between empathy and customer loyalty. The following hypothesis is offered in Jordanian Islamic bank: H6: Empathy has a positive correlation with self-reported loyalty intentions.

The relationship between overall service quality, satisfaction and self-reported loyalty intentions
Overall quality of service dimensions is the most important aspect for generating customer satisfaction and loyalty in the banking sector (Eskildsen et al., 2004). Jones et al. (2002) indicate a positive relationship between service quality and resistance to better alternatives, recommendation, and repurchase intention. Deng et al. (2010) found that perceived service quality is the strongest correlation with customer satisfaction and customer satisfaction has the greatest correlation with customer loyalty. Zafar et al. (2012) concluded that the customer satisfaction and customer loyalty have the strong positive relationship with each other and the change in one point of customer satisfaction contributes almost 87 present in customer loyalty. Therefore, satisfaction is a necessary precondition for building long-term customer relationships and is likely to increase loyalty Boonlertvanich, 2019). Customer satisfaction is also considered a mediating variable between service quality and customer loyalty (Boonlertvanich, 2019;Iqbal et al., 2018;Makanyeza & Chikazhe, 2017;Kashif et al., 2015;Cronin et al., 2000;Cronin & Taylor, 1992). The following hypotheses are offered in Jordanian Islamic bank: H7: Satisfaction mediates the relationship between service quality and attitudinal loyalty.

Sampling and data collection
A self-administrated survey has been found to be the best method to obtain data for the following reasons: First, it is designed to deal specifically and directly with the respondents' opinions, beliefs, attitudes, and motive (Burns, 2000). Second, it proves to be an effective tool especially when a researcher has little or no control over behavioural activities (Yin, 2014). Third, it provides the means of accurate assessment of information about samples (Creswell & Creswell, 2018); and serious concern about casual research situations (Hair et al., 2010).
This study did not employ random sampling techniques because the researcher has limited knowledge of the number of customers in Islamic banks in Jordan because these banks keep customers' details confidential. Hence, non-probability convenience sampling was employed in this study because to the population size is unknown and it was impossible to make a randomisation. In fact, there may be no other way to gather data in some cases than to sample a group of individuals who are available, as well as providing a captive audience for achieving a response (Schmidt & Hollensen, 2006). Interestingly, many similar studies have used non-probability convenience sampling (e.g., Kashif et al., 2015;Sayani, 2015). However, the policy of Islamic banks management does not allow any researcher to collect data directly from their customers within the bank branches, but they can administer the survey by themselves. The target population of the current study comprises customers who have an account in Islamic banks operating in the Jordanian market, i.e. (Jordan Islamic Bank (JIB); Islamic International Arab Bank PLC (IIAB); Safwa Bank (formerly known as Jordan Dubai Islamic Bank). The final version of the questionnaire was sent to the headquarters of the Islamic banks, then on to individual branches and employees of those branches for these questionnaires to be completed.
The sample size of the pilot study is typically small, varying from 15 to 30 participants for the initial testing, which depended on the heterogeneity of the target population (Malhotra et al., 2017). However, a pilot study was conducted by giving the questionnaire to 25 respondents, who were representative of the respondents in the main study, to assess the validity and reliability of the research instrument. Based on respondents' feedback, no survey items appeared problematic to respondents. Overall, the result analysis of 25 surveys confirmed acceptable scores of reliability. For the current study, the equation sample size was not used to determine the sample size as the size of the true population is unknown.
The size of a sample affects the quality of the research data, that needs to be taken should reflect the basic characteristics of the population (Proctor, 2005). Larger samples are better than smaller samples because larger samples tend to maximise the accuracy of population estimates and increase the generalisability of the results (Proctor, 2005;Sekaran & Bougie, 2013). Malhotra et al. (2017) reported that sample size in marketing research designs is affected and determined by several aspects, which can be summarised as the following: (1) The average size of samples in similar previous studies; (2) The nature of the research (e. g. in qualitative research, the sample size is typically small, meanwhile with quantitative research, larger samples are required) Table 1 provides an overview the sample sizes used in similar previous studies with details of response rates and usable responses received.
It can be seen in Table 1 that there is no certain sample size in such studies. It has been suggested that "the adequacy of sample size might be evaluated very roughly on the following scale: 50-very poor; 100-poor; 200-fair; 300-good; 500-very good; 1000 or more-excellent" (Comfrey & Lee, 1992). Therefore, the actual number of questionnaires of those who completed and usable was 655, yielding a response rate of 72.8%. This yielded a response rate of approximately 73%. The sample size for this study exceeded the requirements suggested by Hair et al. (1999) that "a sample size of 200 may be required to ensure appropriate use of maximum likelihood estimation, to generate valid fit measures and to avoid drawing inaccurate inferences" (as cited in Molina et al., 2007, p. 260). In other word, the response rate about 73% considered sufficiently large for statistical reliability and generalisability (Stevens, 2012).

Questionnaire and measurements
The scales used to measure tangibles, reliability, responsiveness, assurance, and empathy were operationalised with 22 items adopted from Parasuraman et al. (1988Parasuraman et al. ( , 1994. The scale used to measure compliance was operationalised with five items adopted from Othman and Owen (2001). The measured scale of customer satisfaction was operationalised with six items. One item expressing the interaction satisfaction of bank customers has been taken from the study by Leverin and Liljender (2006); items representing pleasure and reselecting the bank have been adopted from the work of Ndubisi and Wah (2005); and Rust and Oliver (1994). The remaining items seeking responses to customers' satisfaction from their experience and perception of the selected bank as the right choice, have been taken from the Gremler and Gwinner (2000); and Rust and Oliver (1994) studies. Finally, seven items expressing the behavioural response of bank customers have been taken from the study by Zeithaml et al. (1996). A total of forty items were adjusted to the Islamic bank's context. Table 2 exhibits the research sample characteristics. As shown, 78.6% of the sample populations were males and 21.4 % were females. The disparity between males and females can be attributed to the smaller percentage of females in the workforce, to the somewhat conservative social nature in Jordan in which males are more responsible for family finances. Accordingly, the sampling method is a convenience sample and not a random sample for the reasons mentioned beforehand.

Descriptive statistics
Table 2 also shows that 63.5% of the samples are young customers within the age category of 26 − 45 years, consistent with Jordanian demographics. This could result in a more fruitful long-term relationship. As can be seen from Table 2, the highest percentage of the sample customers' monthly income was between 500 and 1000 (40.5%) Jordanian Dinars (equivalent to 465-930 GBP). This implies that the Islamic banks' customers may be sensitive to any fluctuations in the perceived service quality and customer satisfaction that may affect their loyalty in the long term. Table 2 shows that the great majority of the sample population are well educated. 56.5% of respondents had completed a university degree and another 18.5% of respondents had received a college degree. This indicates that they are in a good position to judge service quality, customer satisfaction and customer loyalty. Table 2 also indicated that highest percentage of occupation was for the customers employed in the managerial and professional roles, representing a ratio of 34.7% and 32.5% of the sample, respectively. This holds a strategic implication that would help Islamic Banks understand its positioning and loyalty status in the minds of their customers.

Measurement model
Tests of normality were conducted to satisfy the criterion of multivariate normality, namely, skewness and kurtosis. "The absolute value of skewness greater than 3.0 or the absolute value of kurtosis greater than 8.0 may indicate an abnormal distribution (Chen, 2012). Therefore, it was accepted that the value of skewness and kurtosis should not be greater than 3.0 and 8.0 (Chen, 2012). As shown in Table 3, the values of both skewness and kurtosis were within the acceptable levels, indicating the multivariate normality.

Factor analysis
Principal component analysis was used with VARIMAX rotation on 40 items. This resulted in the elimination of one item (depicted low communality, i.e. 0.5) and remained scale items were again subjected to factor analysis. Factor loadings close to −1 or 1 indicate that the factor strongly affects the variable. Loadings close to zero indicate that the factor has a weak correlation with the variable. In this research, loading of ± 0.40 or more considered to be acceptable (Malhotra et al., 2017). A summary of the factor analysis and reliability test results is presented in Table 4.
As can be seen in Table 4, only item 40 "I will not switch to another bank if I experience a problem with this bank" removed to increase the reliability and content validity, as the factor loading below 0.40. The range for factor loadings was 0.617-0.911for all items. The final factor solution is represented in Table 4 along with the items, factor loadings, explained variance and Cronbach's a value. The range for the value of the KMO statistic was .792-.883 indicating the appropriateness of factor analysis. Cronbach's alpha for all measures exceeds the recommended threshold value of 0.70 (Malhotra et al., 2018;Schmidt & Hollensen, 2006). Therefore, all measures are robust in terms of their reliability. Table 5 shows the mean, standard deviation for the summated scales. The mean scores of the variables were between 3.79 and 4.01, indicating that respondents were inclined towards a favourable direction, i.e. towards agreed. Table 5 also informative of the correlations between the investigated constructs, which in turn provides a preliminary way to test the hypotheses. CARTER dimensions have high positive correlations with customer satisfaction and self-reported loyalty intentions, with empathy appearing as the highest positive with satisfaction (0.755) and

2.
No interest paid nor taken on savings and loans in this bank −.642 −.399

3.
The bank provides Islamic products and service −1.086 1.639

4.
The bank provides of free interest loans −.775 −.092

5.
The bank provides profit-sharing investment accounts −.583 .214   self-reported loyalty intentions (0.658). Additionally, a strong positive correlation was found between overall satisfaction and self-reported loyalty intentions (0.810).

Hypotheses testing
Since the measurement model evaluation provided evidence of reliability and validity, the structural model was examined to evaluate the hypothesized relationships among the constructs in the research model (Agag & El-Masry, 2017). The multiple regression tests and hierarchical multiple regressions were conducted by running two steps of analysis separately, as follows: (1) CARTER dimensions as independent variables and self-reported loyalty intentions as dependent variable. The results show that all hypothesized relationships are supported. Tables 6 and 7 summarise research findings summarised in terms of relationships between CARTER dimensions customer and self-reported loyalty intentions.  As shown in Table 6, the predictive power of this model was indicated by (R = 0.769) and (Adjusted R Square = 0.588). As shown in Table 7, the six variables were highly significant in regressing with behavioural customer loyalty at the significant level (0.000), lower than 0.05 in the 95% confidence level. The result indicates that compliance (Beta: 0. 245) is the most important factor influencing behavioural customer loyalty with Islamic banks in Jordan. Followed by "empathy" (Beta: 0.168), "assurance" (Beta: 0.171), "reliability" (Beta: 0.128), "tangibles" (Beta: 0.120) and "responsiveness" (B: 0.111), respectively.
(1) As shown in Table 8, the empirical findings of this study found that student satisfaction functions as a mediating variable between overall CARTER dimensions and self-reported loyalty intentions.

Discussion
The main objective of this paper is to investigate how the CARTER model and customer satisfaction are correlated with self-reported loyalty intentions in Islamic banks. More specifically, this study is designed to answer four key questions, as follows:

RQ1. Which service quality dimensions are correlated to self-reported loyalty intentions that can be used to measure Islamic banks' performance?
SERVQUAL model is designed to capture consumer expectations and perceptions of a service along the five dimensions (reliability, assurance, tangibles, empathy, and responsiveness) that are believed to represent service quality. Although the SERVQUAL instrument has been widely applied in a variety of industry, there are several criticisms of the approach. Therefore, the current study has adopted CARTER model to measure Islamic bank performance in Jordan. It is comprised of the five dimensions used in SERVQUAL model and one additional dimension, compliance, to measure service quality in the Islamic banks. Further, service quality perceptions differ between consumers due cultural differences. Furrer et al. (2000) indicated that the relationships between cultural dimensions and the relative importance of service quality dimensions can provide beneficial insights for how businesses must allocate resources in different cultural groups. In other words, market segmentation is necessary because international service markets are heterogeneous. Such segmentation helps service marketers to have a better understanding of customer purchase decisions by identifying marketing objectives more precisely by cultural segments. In each segment, the resources can be allocated proportionally to the relative importance of each service quality dimension across different cultural segments (Furrer et al., 2000).

RQ2. Which of the service quality dimensions will have the greatest correlation with self-reported loyalty intentions?
Findings of this study revealed that compliance is the greatest correlation with self-reported loyalty intentions compared to other service quality dimensions. This result is consistent with previous studies (e.g., Anouze et al., 2019;Lee & Ullah, 2011;Souiden & Rani, 2015), that identified compliance as a valuable predictor of consumer behaviour and consumer purchasing decisions with Islamic banks because the customers believe that Islamic banks are Shariah compliant. This result indicate that Islamic banks enjoy a relatively strong power base among those customers who stress obedience to Islamic principles in the practices of the banks they select. This is consistent with the findings of Kaakeh et al. (2019), who state that compliance affect attitude directly and intention indirectly, mediated by attitude. Further, Mbawuni and Nimako (2017) report that consumer attitude, readiness to comply with Islamic law, knowledge, perceived innovativeness, and perceived benefits were critical determinants of bank customers' intention to adopt Islamic banking in both Muslim and non-Muslim sub-groups.
Furthermore, this study's findings revealed that assurance has significant correlation with selfreported loyalty intentions. This result is consistent with Kheng et al. (2010), who reported that assurance positively influences customer loyalty. From the results of this study, customers are satisfied by Islamic banks and their loyalty can make the customers more confident by dealing transactions, depending on Sharia (Islamic Law). Further, treating customers courteously, coupled with the knowledge of employees to answer questions from customers also increase customers' satisfaction and loyalty. Moreover, the findings found that empathy has positive significant correlation with self-reported loyalty intentions; this result is consistent with previous literature (Butcher et al., 2001;Kheng et al., 2010). In this regard, it is clear from data analysis that employees who show higher giving customers individual attention, understanding and knowing the customer and making the effort to understand the specific needs of the customers and convenient operating hours.
Given the significant relationship between reliability, the current study's results reveal that customers' perceptions of their bank's ability to perform the service dependably and accurately, tend to become positive once they perceive that banks keep promises, provide services at the promised time, and give them individual attention. This is in line with prior research (e.g., Kheng et al., 2010;Zafar et al., 2012), which pointed out that reliability plays a critical role on selfreported loyalty intentions. Similarly, Osman et al. (2009) found that reliability was perceived as the essential attributes in choosing the Islamic financial institutions. In this regard, higher dependably and accurate service, increases customer satisfaction and enhance loyalty in Islamic banks.
Findings of this study revealed that the tangible has significant correlation with self-reported loyalty intentions, which is in line with previous literatures (e.g., Kashif et al., 2015;Choudhury, 2014). That is because high standards of physical evidence such as modern equipment, professional appearance of staff and nice brochures can improve customers' first impressions. The significance of a relaxed and comfortable Islamic banking atmosphere is now perceived by the theme, "a branch as comfortable as your living room" (Yavas et al., 1997). Moreover, our findings revealed that responsiveness has significant correlation with self-reported loyalty intentions. This result is consistent with Badara et al. (2013) which have reported that responsiveness positively influences customer loyalty. It is evident from data analysis that the current study shows customer satisfaction with loyalty by telling customers exactly when the banks will likely perform the ability of employees to help them and provide prompt services and readiness to reply to customers' query and request.

RQ3. What is the nature of the relationship between service quality and self-reported loyalty intentions taking into consideration the mediating role of customer satisfaction?
This study found that the six independent variables (compliance, tangibles, reliability, responsiveness, assurance, and empathy) have high positive correlations with customer satisfaction and self-reported loyalty intentions. Customer satisfaction was found to be a mediating variable between service quality and self-reported loyalty intentions. This result is consistent with previous empirical studies (e.g., Boonlertvanich, 2019;Kashif et al., 2015) that indicated that service quality and customer loyalty is highly interrelated. This research supports the understanding that customer satisfaction mediates the relationship between quality of service and self-reported loyalty intentions (Boonlertvanich, 2019;Makanyeza & Chikazhe, 2017;Kashif et al., 2015;Caruana, 2002). In addition to the positive correlation between service quality and self-reported loyalty intentions, higher dimensions of service quality lead to higher customer satisfaction and, eventually, to selfreported loyalty intentions.

Theoretical implications
The present study makes several noteworthy theoretical contributions from various perspectives. First, this study attempts to simultaneously introduce the concepts of service quality dimensions, customer satisfaction and self-reported loyalty intentions in the Jordanian Islamic banking industry. It also aims to fill the gap in emerging banking marketing literature in general and Jordanian banking marketing literature, as this is the first study conducted in Jordan's Islamic banks as per the authors' knowledge.
Second, the current study also has adopted the CARTER model as a tool to measure quality in the Islamic banks, whereas most of the previous studies have focused on using SERVQUAL in understanding the customer satisfaction and bank performance. This paper also aims to add to the banking marketing literature knowledge base with respect to the cross-cultural understanding of service marketing by providing additional evidence that the compliance element plays a vital role in the minds of Jordanian customers. The study found that compliance has been the most important factor for self-reported loyalty intentions. Previous research paid little attention to the compliance factor and how other dimensions of service quality is linked customer loyalty, particularly in the field of Islamic financial services marketing in an emerging country. Finally, the present study attempted to shed light on the core differences between services quality and customer satisfaction to provide a better understanding of these concepts.

RQ4. What recommendations can be implemented at an executive level to increase customer loyalty towards Islamic banks?
Based on the findings of the study, the following recommendations are provided for developing the marketing of the Jordan Islamic Bank: First, compliance, empathy and assurance are the top three influential factors on customer satisfaction and self-reported loyalty intentions. It is important for Islamic banks to review and endorse their policies and procedures to ensure that documents and investments are undertaken in line with Shari'a requirements. They should be closely involved with the actual practice and implementation of financial transactions from employees or practitioners to obtain a more comprehensive and deeper understanding of the issues. In addition, it is important for Islamic banks to give individual attention to their customers and understand their specific needs, to make them feel further valued and unique. It is important for Islamic banks also to pay more attention to their employees' knowledge, use their professional knowledge towards making recommendations and reducing such risk for customers as well as helping them to more confident towards future transactions.
Second, although responsiveness, tangibles and reliability are the least correlated, they are still considered crucial factors. Given the importance of tangibility, managing the evidence and the use of environmental psychology can be treated as powerful marketing tools. Therefore, it is important for Islamic banks to allocate part of their annual budget on renovating and refreshing their physical evidence to keep their all customers fulfilled and improve self-reported loyalty intentions. To improve reliability managers are encouraged to include reliability concerns in their mission statements, set reliability standards, teach the significance of reliability in training programs, appoint reliability teams to study specific services, recommend ways to improve reliability, measure error rates and reward error-free service. To improve responsiveness that managers should pay more care to customers' personnel contact, telephone calls and messages when issues happen, as speedy response can enhance customers' fulfilment and support self-reported loyalty intentions. Furthermore, responsiveness requires that the Islamic bank's personnel appear wellinformed and empowered in answering customers' request for information and dealing with their complaints.
Finally, managers can study the market, analyse customer data, regularly take the feedback of the customers, and incorporate the changes desired by the customers in their feedbacks. Such efforts could enable them to provide better, more innovative, and competitive services tailored specifically to targeted segments' needs and wants. Furthermore, they can provide live contact over the Internet to solve the problem and provide proper information when the customer needs some advice.

Limitations and future research directions
Even while contributing to important issues in the marketing banking literature this study suffers from some limitations. One limitation of this study comes from the fact that the field research has been conducted on one single market and focused on one type of service. It was tested in Jordanian Islamic banks, and the results of the study are based on the customers' perception of Islamic banks. The banking system in Jordan is based on a dual banking system, conventional and Islamic banks operating side by side. This is an important issue for future research to compare Islamic with conventional banks in the Jordanian market. In addition, according to cultural context of the participants, satisfaction and loyalty could be different in other countries (Narteh, 2013). Academics should therefore exercise some caution when citing the results to other settings.
It would be interesting for future research to replicate this study in other countries, knowing that customers from different geographic regions do not necessarily behave the same way, while the methods of operation of the Islamic banks are usually similar in the different economic environments and countries. Such replication could contribute to advance the understanding of different cultural features affecting customer satisfaction and self-reported loyalty intentions. Further work is required also to investigate the current study model in business-to-business relationships in a similar research setting. Finally, this study employed a convenience sample. Although being a strong sample in terms of diversity and size, generalizations of the results must be made with caution. Therefore, future studies can use random sampling of general consumers.