Effects of Gender and Personality Traits of Front-Desk Employees on Customers' Assessment of Service Quality: Evidence from Islamic Banks in the UAE

This paper examined the effects of front-desk employees’ personality traits and their gender on customers’ assessment of Islamic banks' service quality in the United Arab Emirates. We used a sample of 104 front-desk employees and 454 customers. The Mini-Markers instrument was used to assess personality traits and the SERVPERF instrument to collect data on service quality for a pair-matching sample design. We used regression analysis to determine the relationship between the five factors of personality of front-desk employees, their gender, and the overall bank service quality as well as each of its dimensions. The paper provided evidence indicating that personality traits do not have their own independent effects on customers’ perceptions of Islamic banks' service quality. Although Lin, Chui and Hsieh (2001) found supporting results for such linkages when they studied service quality in four service sectors (not including banks) in Taiwan, we concluded that these effects may be dependent on other factors such as culture, research design, or industry characteristics. However, employees’ gender was found to have significant effects on the empathy dimension of service quality. The originality of this paper lies in its attempt to link personality traits of front-desk employees of banks to customers’ perceptions of service quality in the UAE. It is the first study that uses the Mini-Markers and SERVPERF together in the UAE to achieve its purpose. Its value lies in its contribution to the settlement of the debate about possible linkages between employees’ personality traits and service quality perceptions.


Introduction
In today's competitive market, quality attributes are considered critical elements for banks as well as other organizations.Kumar, Kee, and Manshor (2009) reported that long-term growth and profitability of banks rely on banks' ability to attract and retain loyal customers.Although early work on service quality tended to address the conceptualization of service quality, research in the last three decades tended to focus on quality assessment.

Introduction
Much research effort in the last three decades indicates that service quality is determined by technical aspects as well as the delivery of the service.This notion also extends to government services (Foley 2008).The literature of the banking industry also suggests that both instruments (SERVQUAL and SERVPERF) are suitable for assessing service quality (e.g., Cui, Lewis, & Park, 2003;Zhou, 2004;and Vanniarajan & Anbazhagan, 2007).On the other hand, much research effort in the psychology literature has been exerted to identify personality traits (e.g., Norman, 1963;Eysenck & Eysenck, 1963;Cattell, Eber, & Tatsouka, 1970;Comrey, 1970;Costa & McCrae, 1985;McCrae &Costa, 1997, andFunder, 2001).Although different factors and models have been suggested, the five-factor personality traits model (FFM) is widely considered as the most comprehensive model within the field of selection and assessment that is capable of adequately predicting behavior across different settings (Gill & Hodgkinson, 2007).Digman (1990) and Salgado (2003) provided evidence that personality measures developed within the FFM yielded higher predictive validity coefficients with respect to job performance and other organizational behavior relatives to measures based on alternative models of personality.
The five factors of the model are Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism (Costa & McCrae, 1985).These big-five personality traits or factors, as widely described in the literature, represent a sort of grand unified theory of personality since personality measures, as well as trait adjectives, could all be loaded in a factor analysis on one of the big-five personality traits (e.g., DeRaad, 2000).Furthermore, the big-five have been validated in many different settings and in many different languages and cultures (Costa & McCrae, 1995;and DeRaad, 2000).
Comprehensive reviews of the literature on the five factor model (e.g., Funder 2001, Gill & Hodgkinson, 2007) reached similar conclusions on the nature of personality and the model's validity.From a theoretical viewpoint, the five factors are abstractions that represent individuals' experience and behavior.In addition, there is strong evidence to support the validity of the five factor model and the independence of the factors.Lewis, Orledge, and Mitchell (1994) examined university students' attitudes towards bank services.They concluded that students had higher expectations than their perceptions.However, the study did not address service delivery or effect of service providers' personality traits on perceived service quality.Lin, Chui and Hsieh (2001) addressed the relation between personal characteristics of service providers and service quality in four service sectors in Taiwan (life insurance, real-estate agencies, information services, and securities).They also explored possible effects of gender on the relationship between personality traits of employees and customers' perceptions of service quality.The results indicated the existence of a relationship between personality traits and service quality and that gender moderates such a relationship.In addition, the study results indicated that openness and agreeableness were valid predictors of the dimension of assurance.Furthermore, conscientiousness was found to be related to reliability while extraversion was related to responsiveness.Cui, Lewis, and Park (2003) studied service quality in the banking sector in South Korea.Their results indicated that the SERVQUAL and weighted SERVQUAL models did not appear to be valid for use with Koreans.On the other hand, the results of testing the non-weighted SERVPERF and the weighted SERVPERF models showed good fit only when the original 21 items were reduced to five items by partial regression.

Service Quality and Personality in the Banking Industry
Zhou (2004) studied service quality, satisfaction, and customers' behavioral intentions in China's retail banking.The study used factor analysis and a structural equation analysis to relate identified service quality context-dependent dimensions to overall satisfaction and customers' behavioral intentions (SERVPERF).The results indicated that the standard five dimensions of service quality are not good determinants of satisfaction and behavioral intentions regardless of the scale (SERVPERF or SERVQUAL).Bhat (2005) studied service quality among a random sample of 800 bank customers from five banks in Northern India using SERVQUAL.The mean differences between expectations and perceptions were calculated separately for each bank for comparison purposes.The results showed that service quality of foreign banks was higher than that of Indian banks.However, all banks fell below their customers' expectations.Lee and Hwan (2005) studied the relationships among service quality, customer satisfaction, and profitability in the Taiwanese banking industry.The results indicated that service quality influenced customers' attitudes and satisfaction, which influenced purchase intentions.However, customer satisfaction did not influence service quality perceptions.In addition, the results indicated that customer satisfaction with bank service quality significantly influenced bank profitability but not bank market share.Vanniarajan and Anbazhagan (2007) studied the SERVPERF scale in the Indian retail banking.They used factor analysis and multiple regression techniques.In addition, a two-group discriminant analysis was used to assess the discriminatory power of the service quality factors among the three groups of banks.The overall results of the regression analysis indicated significant effect of reliability and assurance on customers' attitudes towards retail banking services.However, there were no effects for customers of cooperative banks.In addition, responsiveness and tangibles were insignificant for the pooled data and for each group of the banks.

Research Instruments
We used in this paper two questionnaires.The first is the Mini-Markers to assess service providers' personality traits.The second is SERVPERF to assess customers' perceptions of service quality provided by the UAE Islamic banks.A copy of each questionnaire appears in Appendixes A and B.
Modern personality research (e.g., Costa & McCrae, 1995;Goldberg, 1992;Wiggins and Trapnell, 1997;and Gill & Hodgkinson, 2007) indicated that the five-factor model of personality is the most popular model.It recognizes Openness to experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism (OCEAN for short) as the fundamental dimensions of human personality.Furthermore, Salgado (2003) showed that personality measures developed within the five-factor model yielded higher prediction validity coefficients relative to measures based on alternative models of personality.This popularity of the five-factor model of personality motivated us to select one of the instruments with item format that was originally designed for the five-factor model.According to Gill and Hodgkinson (2007), the big five traits, like other self-report trait-based measures of personality, are being currently assessed using two major types of item format, namely, statement-based and single adjectival-based descriptors.Statement-based measures are like the ones developed by Costa and McCrae (1992).Single adjectival descriptors are like the ones developed by Goldberg (1990).Gill and Hodgkinson (2007) also pointed out that a considerable body of growing evidence has been accumulated to suggest that the single adjectival description-based approach is methodological superior to the statement-based approach in a number of important respects.First, the nature of the statement-based items is relatively complex when compared to adjectival descriptors.This property of complexity is known to reduce response validity of such statement-based items (Gaskell et al., 1995) and the clarity of any factors extracted from the responses (Dunbar, Ford, Hind, & Der, 2000).Second, the contexts in which statement-items are responded to affect the criterion-related validity of personality scale.Such effects are particularly marked for conscientiousness measures of the five factor model (Schmit, Ryan, Stierwall, & Powel, 1995).Goldberg (1992) indicated that the use of adjectival descriptors not only meet the requirements of reliability and validity but is also congruent with the lexical hypothesis underlying much of the original five factor model.Gill and Hodgkinson (2007) pointed out that currently there are three prominent big five instruments that use adjectival items.They are Goldberg's (1992) adjective Markers, Saucier's (1994) Mini-Markers, and Trapnell and Wiggins's (1990) Interpersonal Adjective Scales-Big Five.Goldberg (1992) developed and tested a set of 100 adjective markers of the big-five personality traits.Each of the big-five factors was represented by 20 markers.He reported reliability coefficient ranging from 0.82 to 0.97 in various data sets.
Based on Goldberg's (1992) adjective markets, Saucier (1994) argued that the original 100 items could still be reduced and briefed to make it desirable and more convenient to the users.Saucier used the set of the 100 markers as a sample of variables from which to derive a smaller brief set that adequately compare to the full set of markers, while maintaining reliability and validity of the original set.Through various samples of moderate size, Saucier extracted through different rotations five principal components of personality with their markers.These markers were chosen from among the original 100 markers on the basis of their highest rotated factor loadings in the combined samples.Saucier termed this set of markers, the Mini-Markers.
We selected the Mini-Markers as the appropriate instrument to assess the personality traits of selected Islamic bank employees.This selection was based on two factors.First, prior research (e.g., Saucier, 1994, andSalgado, 2003) demonstrated the instrument's reliability and validity.Second, the instrument has the advantage of being short and could be administered to employees quickly as it takes only few minutes to complete (Saucier, 1994).
The second instrument is SERVPERF.We preferred to use SERVPERF for two main reasons.First, although prior research has documented evidence on the reliability and validity of both instruments, recent research work in the banking industry (e.g., Cui et al., 2003;Zhou, 2004;Lee & Hwan, 2005;Vanniarajan & Anbazhagan, 2007) provided evidence in favor of SERVPERF over the SEVQUAL.Second, the SERVPERF is shorter than the SERVQUAL as the first contains only one set of items related to service performance.It does not contain the other set dealing with expectations.This issue of length could be viewed as an operational advantage for the SERVPERF instrument.

Piloting the Questionnaires
Piloting the questionnaires is a standard procedure to remove any ambiguity from the questionnaire and correct words or phrases that may cause misunderstanding.Such a procedure is a necessary one to improve the questionnaire reliability and validity.
Most employees and customers of Islamic banks are Arabs and speak primarily the Arabic language.Accordingly, the authors used translated versions (in the Arabic language) of both the Mini-Markers and the SERVPERF instruments to enable those who are not fluent in English language to express their perceptions and ratings in their native language.The English language versions of the two instruments were translated using the process of "back translation" as suggested by Brislin (1970) and Hui and Triandis (1985).First, we employed a translator to express the English version of each instrument into Arabic versions.A second independent translator was then employed to translate the resultant instruments in Arabic versions back to English.This process allowed us to compare the original English version of each instrument to the Arabic equivalent version in order to guard against loss or distortion of content.The process resulted in the correction of few words with the help of a third individual who is fluent in both English and Arabic languages.This process was employed to have high levels of content validity for the Arabic versions of the instruments that would be comparable to the English versions.
After completing the process of back translation, we distributed the Arabic version of the Mini-Markers questionnaire to eight bank employees and two human resource managers who are familiar with personality tests.Some of the expressions used to describe personality traits in the translated Arabic version were revised for better understanding.
The Arabic version of the quality assessment questionnaire was also distributed to twenty four Arab customers of Islamic banks.Based on the feedback received from the customers, few words on the questionnaire were modified.
The English versions of the two questionnaires were also piloted using other five front-desk employees and ten customers of Islamic banks.There was no feedback requiring modifications or changes in these versions of the questionnaires.All data obtained through the pilot process of both questionnaires were not used in the final data analyses.

The Study Population
The population of the study consists of two distinct groups.The first group consists of all employees of the different outlets of four Islamic banks whose management expressed willingness to participate in the study.The second group consists of the customers of the four Islamic banks who are registered as clients/customers at any of the banks' outlets.Thus, the population of the study consists of all customers and employees of four UAE Islamic banks.These banks operate through a network of branches/outlets and carry out various banking activities and services all over the UAE.
Through personal contacts and visits to the Islamic banks' main branches, we were able to obtain information on front-desk employees but not about customers.

The Study Samples
As it is difficult to study the whole population of employees and customers who belong to the four Islamic banks because of their large numbers, we selected a sample of front-desk employees and a second sample of customers from the four UAE Islamic banks.The first sample consisted of 160 front-desk employees, where each bank was represented by 40 employees who deal directly (face to face) with customers and clients.Those employees were selected randomly from lists of employees provided by the banks.Those selected employees constituted the first sample intended to respond to the Mini-Markers questionnaire.Those selected front-desk employees did not know about the study of customers' perceptions that took place after their participation.In addition, the sample of bank employees did not include any member of bank management.
The second sample consisted of bank customers who were selected for the administration of the service quality questionnaire using a site-intercept technique.For each employee selected in the first sample to participate in the study, a sample of five bank customers were selected and approached on site immediately after completing their business transactions with a particular employee.Each one of these five customers was invited to participate in the study.Once he agreed to participate, he was invited to a side room and asked to set down and fill out the questionnaire based on his current experience with the bank employee.Thus, a total of 800 customers at the rate of 200 customers per bank constituted the second sample.This approach to sampling ensured that every employee is assessed by several customers so that possible effects of customers' bias for or against a particular employee could be minimized.

Data Collection
The first author held several personal meetings with the chairmen of the four Islamic banks and branch/outlet directors to explain the nature and objectives of the study and to discuss the planned procedures of distributing the Mini-Markers questionnaires to front desk employees and the quality assessment questionnaires to the customers.The administrators of four Islamic banks agreed to have their front-desk employees participate in the study and instructed them to cooperate with the authors.
We distributed the Mini-Markers questionnaires to the selected employees of the four Islamic banks and asked each one of them to complete the questionnaire within one week and deposit the completed questionnaire in a specially marked box in the branch manager's office.Each questionnaire had a special code for the particular employee at the particular bank.This code was used in the process of identifying employee personality data for pair matching with collected data from bank customers about service quality.The process of distributing and collecting the Mini-Markers questionnaires took about two weeks.
However, we employed sixteen business graduates who were trained in mall intercept technique to survey the sample of bank customers.These data collectors worked in pairs and approached a sample of 800 customers during bank business hours over a period of three weeks.
We held two meetings with the data collectors before the actual collection of the data to explain the nature and purpose of the study, their assignments to the different banks, and the data collection plan.We also accompanied those data collectors to the locations of their assigned bank branches to introduce them to the front-desk supervisors without the knowledge of front-desk employees and to make them familiar with the layout of the bank branch where the customers would perform their transactions.They also had a chance to observe the flow of some actual transactions before the starting date.The data collection plan included the following procedures: 1) Two data collectors (a male and a female) were assigned to collect data from customers of one bank over three-week period.Each data collector was provided with enough copies of the instruments in both Arabic and English languages.The particular version of the quality assessment questionnaire was distributed based on the customer's preference.
2) Data collectors were asked to insert the particular code of the participating employee on the quality assessment questionnaire to ensure accuracy and facilitate data processing at a later stage.
3) Each data collector was instructed to keep himself at distance away from the front-desk employee and to approach the customer after completing his transaction with the intended employee.Each data collector was also instructed to kindly request the customer to participate in the study, which aims at improving the quality of bank services.Once a customer agrees to participate in the study, the data collector invited him into a side room and asked him to set down and fill out the questionnaire based on his immediate experience that has taken place with the specific bank employee.Each customer was also assured of the confidentiality of his responses as the data were to be analyzed after aggregation.Finally, data collectors were instructed to present each customer with a brand name pen (worth about sixteen dollars) as a gift.
4) Each data collector was instructed to thank the customer for his cooperation once he has completed the quality assessment questionnaire.
5) Each data collector was instructed not to approach the planned number of five customers per employee on the same day but over few days and at different times during bank working hours.This procedure was intended to reduce the effect of mood on customers' assessments.
6) Data collectors were informed that the authors would visit them from time to time to monitor the process of data collection.
The data collection process was carried out as planned and data collectors did not encounter any significant problems.The process took about five weeks (two weeks for employees and three weeks for customers).

Research Hypotheses
The psychology literature indicates that performance of an employee is affected by his personality traits.Mount and Barrick (1998), Lounsbury andGibson, (2001), andLounsbury (2003) attempted to link customer's perceptions of perceived service quality to employees' personality traits.This line of thinking resulted in two views.The first indicates the existence of such a link between personality traits of front-desk employees and customers' perception of service quality.The other view indicates no relationship between personality traits of front-desk employees and customers' perception of service quality.The question of which of these two views prevails in the banking industry in the UAE remains to be answered empirically.Thus, this study formulates this question in the first null hypotheses as followed.
H 01 : There is no significant positive relationship between personality traits and customers' perception of overall service quality.
H 02 : There is no significant positive relationship between employees' personality traits and customers' perceptions of the different dimensions of service quality.
H 03 : There is no significant difference between male and female perceptions of any of the dimensions of bank service quality.

Data Analysis
We performed two steps to prepare collected data for analysis.First, we calculate an average score of customers' perceived quality of bank services for each employee.Computation of this average was required for the pair-matching design (to match one score of a customer for perceived bank services quality provided by an employee with that employee's personality traits).Second, we made reverse coding of the negatively worded personality trait items on the Mini-Markers questionnaire.For example, a negatively worded item that was originally scored as five was reversed to the score of one.
The statistical analyses included data reliability analysis and multiple regression analyses.We used five regression models.The first model employed overall service quality as the dependent variable.The other four models used each of the dimensions of reliability, responsiveness, assurance, and empathy as the dependent variables (one at a time).The scores of the five personality dimensions, gender and other variables were used as the independent variables.The multiple regression models were of the following form: Where: Y 1-5 = Perceived overall service quality or one of its dimensions,

Sample Responses and Respondents' Profile
As mentioned in the previous section, a sample of 160 front-desk employees from the four Islamic banks was randomly selected for the administration of the Mini-Markers questionnaire.A second independent sample of 800 bank customers was also selected for the service quality questionnaire (SERVPERF).Returned completed questionnaires from front-desk employees (the first sample) reached 120 questionnaires.However, we excluded four questionnaires because of incomplete data Thus, the usable responses were only 104 returned questionnaires.The distribution of these employee responses among the four banks is shown in Table 1.

Data Reliability
A reliability measure indicates the extent to which the data provided by subjects for the different questionnaire items are consistent.Data consistency usually indicates reduced levels of measurement bias (error) across time and the various items of the instrument.We used alpha coefficient as the measure for internal consistency among the different items for each dimension on the two instruments.
The results of the analysis indicate that the overall reliability level for the Mini-Markers and the SERVPERF were 0.758 and 0.968 respectively.These coefficients are higher than the minimum acceptable level indicated by Sekaran (2003).The reliability coefficient for the Mini-Markers was almost similar to the ones reported in prior studies (e.g., Goldberg, 1992;Salgado, 2003;and Robie, Brown, & Bly, 2008).In addition, the reliability coefficient for the SERVPERF is stronger than the results reported in prior studies (e.g., Parasuraman et al., 1988;Cronin & Taylor 1992;and Vanniarajan & Anbazhagan, 2007).This may indicate that the subjects were more familiar with the service quality instrument than the Mini-Markers instrument given that some research work on service quality was done in the UAE (e.g., Alamri, 2003;Al-Hammadi, 2002).

Tests of Collinearity
Table 4 reports the results of testing for collinearity.According to Cooper and Schindler (2011), an index of the Variable Inflation Factor (VIF) that is 10 or higher indicates a serious collinearity that can damage the results of the regression analysis.The table shows that the indices are too low.These results indicate that collinearity is not an issue for this paper.

Personality Traits and Customers' Perceptions of Bank Service Quality
The regression analyses used the five personality trait dimensions, gender, nationality, and training received as independent variables.Gender, nationality, and training received were used in the form of dummy variables (0, 1) The tangibility dimension of service quality was excluded as the bank employees have no direct or immediate effect on this dimension.Table 5 shows that the model R 2 (coefficient of determination) is very low (about 1.4%).This indicates that the model is missing other independent variables.The table also shows that there were no significant statistical relationships between any of the independent variables and overall service quality (maximum t = -0.767,p ≥ 0.445).These results fail to reject the first null hypothesis.

Multiple Regression Results for the Reliability Dimension of Service Quality
Table 6 shows that the model R 2 (coefficient of determination) is low (about 6.4%).In addition, there were no significant statistical relationships between any of the independent variables and customers' assessment of the reliability dimension of quality (maximum t = -1.361,p ≥ 0.177).These results fail to reject the second null hypothesis.

Multiple Regression Results for the Responsiveness Dimension of Service Quality
Table 7 shows that the model R 2 (coefficient of determination) is very low (about 5.4%).This indicates that the model is missing significant independent variables.In addition, there were no significant statistical relationships between any of the independent variables and customers' assessment of responsiveness dimension of service quality (maximum t = -1.536,p ≥ 0.128).These results fail to reject the second null hypothesis.

Multiple Regression Results for the Assurance Dimension of Service Quality
Table 8 shows that the model R 2 (coefficient of determination) is very low (about 3.9%).This is an indication that other important variables affecting customers' assessment of the assurance quality are missing.In addition, there were no significant statistical relationships between any of the independent variables and customers' assessment of the assurance dimension of service quality (maximum t = -1.211,p ≥ 0.229).These results fail to reject the second null hypothesis.

Multiple Regression Results for the Empathy Dimension of Service Quality
Table 9 shows that the model R 2 (coefficient of determination) is low (about 8.4%).In addition, there were no significant statistical relationships between any of the independent variables and customers' assessment of the empathy dimension of quality except the gender variable (t = -2.371,p ≤ 0.020).While the findings regarding personality traits lead us to conclude that the results fail to reject the second null hypothesis, we note the significant effects of gender on customers' perceptions for the dimension of empathy.The negative relationship indicates that dealing with female employees improves customers' impression of service quality.This result is consistent with the reported results of Lin, Cui, and Hsieb (2001).

On the Effects of Gender
Table 9 above shows significant effect of gender on customers' perceptions of the empathy dimension of service quality.To judge the total effect of gender without personality traits, we performed a separate paired t-sample test on each of the dimensions of the service quality.Table 10 shows the results of the tests assuming equality of the variance.Table 10 above shows that there are no significant differences between the scores of males and females on the overall perceived service quality as well as the different dimensions except empathy.For this dimension, there is significant difference in scores between the two categories of gender (t=2.158,p ≤ 0.033).These results confirm the results of the regression and indicate that the gender effect is not related to the five factors of personality traits.

Discussion
The quality literature indicates that perceived service quality is determined by how customers interact with the service provider as well as how the service is delivered to in terms of accessibility of facilities, appearance and behavior of employees, and the overall treatment of customers.Initially, Parasuraman et al. (1985 and1988) proposed that higher levels of perceived service quality led to an increase in consumer satisfaction.However, Oliver (1980) suggested that a quality attitude is initially formed based on expectations and subsequently revised based on present level of satisfaction with a product or service.
On the other hand, the psychology literature indicates that employees' performances are affected by their personality traits which are represented by the big-five factors.Only recently, some authors (Mount & Barrick, 1998;Lounsbury et al. 2001;Lin et al. 2001;and Lounsbury 2003) have attempted to link customer's perceptions of the perceived service quality to employee personality traits.This line of thinking resulted into two views.The first indicates the existence of such a link between personality traits of front desk employees and customers' perception of service quality.The other view indicates no relationship between personality traits of front-desk employees and customers' perception of service quality.The results reported in Tables 5 to 9 are consistent with the second view that linkages between employee personality traits and customers' perceptions of service quality are not empirically supported.
There are different possible explanations for the absence of the relationship between personality traits and perceived service quality.The first explanation is that in reality there is no such relationship.In other words, these relationships are simply a matter of chance.Only repeated studies on the same population or similar one would provide evidence on the validity of such an explanation.These results contradict the reported results of Lin, Chiu, and Hsieh (2001).But perhaps the composition of the sample studied by Lin, Chiu, and Hsieh (four service sectors not including banks) is one major reason for the differences.A second possible explanation is that employees at Islamic banks have shared values and social convictions that mask their personality traits in a way that the personality trait scores lose their discriminating power in the statistical analysis.In other words, the environments where employees of Islamic banks work help in softening the differences in personality among employees so that they become like members of a similar team.This may be supported by the notion that Islamic banks in the UAE have similar organizational cultures to the extent that individual employees' differences will not affect performance or customers' perceived service quality.Studies on organizational culture (e.g., Judge & Cable, 1997;Sarros, Densten & Cooper, 2005) speak of successful organizations when there is a better fit between organizational culture and employees personality traits and values.This explanation would have some implications for the research methodology and sample selection.The issues of organizational culture could be included as a main variable in the regression model to study the interplay of personality traits, organizational culture, and customers' perceptions of service quality.This implication would also require the use of a third instrument for the measurement of the dimensions of organizational culture.In addition, a more complex design for sample selection would be required where a sample of two sets of banks would be selected first (conventional banks and Islamic banks), followed by the design used in this paper.
Our results contribute toward the settlement of the debate about possible linkages between employees' personality traits and service quality perceptions.The evidence provided indicates that such linkages do not exist.This means that personality traits do not have their own independent effects on customers' perceptions of Islamic bank service quality.Although Lin, Chiu, and Hsieh (2001) found supporting results for the linkages when they studied service quality in four service sectors (not including banks) in Taiwan, we conclude that such effects may be dependent on other factors (e.g., organizational culture, research design, country characteristics).
The results of our paper have some implications.The first implication deals with hiring procedures and the characteristics of employees to be hired.Current emphasis on technical training and personality tests need to be re-examined.Perhaps the organization-fit model should be the main focus for hiring procedures.
A second implication of our results deals with the gender issue.Islamic banks may need to have a second look at female employees as a main key for improving corporate image.From that perspective, some Islamic banks started to have some branches exclusively for women customers.In branches where services are offered to customers of both gender, a balanced hiring for both females and males may give favorable perceptions of quality assessments.
A third possible implication would deal with the reputation, use, and development of personality tests.The rush toward the development and use of such tests by banks in their employee selection process for delivery of quality services is questionable and bank management needs to examine this issue carefully.
Our results are also subject to some limitations.First, as any questionnaire-based research, the extent of data validity and reliability affects reported results.Although the used instruments are well documented in the literature in terms of their reliability measures were of satisfactory level, there is always the chance that some subjects might not have responded in a truthful manner.There is also the chance that bank employees got involved in self-efficacy when filling the personality traits instrument.Thus, the results are limited by the extent to which subjects were honest and truthful in providing the data required.
A second possible limitation deals with the effect of non-response bias.Although the research samples were selected in a proper way to ensure representation, there is the possibility that non-respondents might hold different assessment views than those who responded.However, high response rates would reduce the effect of non-response bias, if any.In addition, the results of splitting each sample into two parts and testing for significant differences in scores on randomly selected items indicate that the effect of non-response bias is insignificant.
A third possible limitation deals with the use of two versions of the instruments (one in Arabic and the other in English).Although we used a common approach for back-translation, it is possible that some subjects had different understanding of some of the items, especially on the Mini-Marker instrument, which might affect the reported results.
A fourth possible limitation deals with the form and specifications of the regression models used.We used linear additive models with specified variables.However, there is a chance that other forms of the model (non-linear or multiplicative) may give different results from the ones obtained from the linear additive models.
It is also possible that other unspecified variables (e.g., organizational culture) would affect the reported results.
A fifth possible limitation deals with the averaging process used to come up with the service quality score.This process might have biased the results toward the centre.That is, the data might have lost some of its discriminating ability.
We recommend some possible avenues for further research.First, it is possible to compare personality traits effects on perceived service quality at Islamic banks with conventional banks.

Copyrights
Copyright for this article is retained by the author(s), with first publication rights granted to the journal.
This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).
Disagree (SD), 2 = Disagree (D), 3 = Neither Agree nor Disagree (N), 4 = Agree (A), 5 = Strongly Agree (SA) indicate your visit type to the bank: __1__ Personal banking __2__ Company representative (Company work C: Over all Service QualityHow would you rate the overall quality of the services provided by the bank?Please circle a number on the following scale to indicate your rating.equal weights (importance) for the quality dimensions.However, you may have different view or different levels of importance.You are asked to assign a level of importance for each dimension using a scale of 1 to 20.Number (1) expresses the lowest level of importance and number (20) expresses the highest level of importance.D1: Tangibles:The physical appearance of the bank (e.g., staff, facilities, and communication materials).The ability of the bank staff to perform the promised service dependably and accurately.The willingness of the bank staff to help customers and provide prompt services.The knowledge and courtesy of the bank staff and their ability to convey trust and confidence.Level of care and individualized attention that the bank staffs provide to customers.any comments you would like to share with us, please provide them in the space below:

Table 1 .
Distribution of responses to the mini-markers questionnaire Completed questionnaires by bank customers (the second sample) reached 556 questionnaires.However, only 454 completed questionnaires were useable for data preparation and analyses.The distribution of these customer responses among the four banks is shown in Table2.

Table 2 .
Distribution of responses to the service quality questionnaire

Table 3
presents the general demographic factors for both samples (employees and customers) in terms of numbers and percentages.

Table 3 .
Distribution of the demographics for the two samples

Table 4 .
Results of collinearity test

Table 5 .
Multiple regression results for overall service quality

Table 6 .
Multiple regression results for the reliability dimension of service quality

Table 7 .
Multiple regression results for the responsiveness dimension of service quality

Table 8 .
Multiple regression results for the assurance dimension

Table 9 .
Multiple regression results for empathy dimension

Table 10 .
Results for testing the difference in service quality scores based on gender Second, it is possible to replicate this work with another sample of employees and customers of Islamic banks in UAE or a gulf state.Results obtained from such replications would help judge the extent to which one would be able to generalize obtained results beyond the particular banks used in this study.Zhou, L. (2004).A dimension-specific analysis of performance-only measurement of service quality and satisfaction in China's retail banking.Journal of Services Marketing, 18(7), 534-546.http://dx.doi.org/10.1108/08876040410561866Pleaseuse the list of common human traits below to describe yourself as accurately as possible.Describeyourself as you see yourself at the present time, not as you wish to be in the future.Describe yourself as you are generally or typically, as compared with other persons you know of the same sex and roughly your same age. For each trait, please mark a number in the table that indicates how accurately the trait describes you using the following rating scale: