Role of Supervisory Support and Management Commitment to Service Quality

There is no room for debate over the rising importance of services in today’s world. This has been established well in the previous chapter on review of literature by various researchers over the time. The service sector dominates the economies world over. India is no exception. We as a nation are developing and being active members in the world trade, we need to know the rules of the game. Presently, the Indian service sector is in a developing state. The competition at such stages ought to be intense. To achieve a competitive advantage companies, are striving hard to develop competencies. The capabilities in the form of human resources are a must for every firm focusing on providing quality services to secure a competitive edge. The human resource practices of a firm and employees behaviour may affect the quality of services. Moreover, it is imperative for any organization to provide quality services to secure customer loyalty which is important for the firm performance.

A research design is the plan, scheme, blueprint, series of procedures, conceptual framework to describe how the defined problem will be catered to. The entire structure and strategy of the use of data to information i.e. from its collection, through its editing, to its analysis is a research design. The present study was exploratory and causal in nature. Exploration was needed for this study as the variables for humanresource management practices and service behaviorantecedents of employees were large in number.Moreover, these variables have not been standardizedand formalized in researches so far. Exploratoryresearch design was used for chapters 4 and 5. Afterestablishing the human resource management andservice behaviour variables and factors, we needed toknow how these variables will affect service qualityperformance variable. In causal research design, theconcern is: how one variable affects another or isresponsible for change in another (Cooper andSchindler, 2008). Thus, a causal research design wasused to see the effect/impact of human resourcemanagement variables and employees’ servicebehaviour antecedent variables on service qualityperformance and further, the effect/impact of servicequality performance on intentions to resign and jobsatisfaction.

The relevant data for the present study has beenobtained from primary sources. A well-structured andpre-tested questionnaire was used for data collection.The questionnaire was drafted consulting relevantliterature (Yavas et al., 2003; Schneider et al., 1980;Peccei and Rosenthal, 2001; Hogan et al., 1984;Schneider and Bowen, 1985). We also interacted withsome employees in the service sector before finalizingthe questionnaire. Before collecting data from respondents, primary drafts of the questionnaire were pre-tested. Based on thewritten and verbal comments, some items were re-worded to eliminate ambiguity, some were deleted,and, some were reframed. Few sections/variables ofthe questionnaire were stated negatively to avoid theresponse bias. The questionnaire consisted of threesections. The first section contained 16 background questions and second section contained 18 statements about the human resource management practices. The third section contained 39 statements related to employees’ service behaviour (Appendix-2). The background questions in the first section were related to gender, marital status, age, work experience of the employee in that organization, the overall work experience of the employee, and the number of organizations changed by the employee. Questions were also related to number of employees in the organization, employee union for the organization, percentage union membership, whether HRM department exists in the organization etc. (for details see chapter 4, Table 4.1). Statements related to human resource management practices were 18 in number including statements on human resource practices like job analysis, human resource planning, recruitment policy, socialization, training and development, performance appraisal, service performance, pay for performance, compensation, benefits, workforce diversity, HR information system, career management, attitude surveys, and adoption of new HR practices. The third section contained 39 statements on service behaviour antecedents of employees focusing on employee rewards, staff training, empowerment, teamwork, role ambiguity, organizational commitment, service quality performance, intentions to resign, job satisfaction, co-worker support, role modeling and leadership style items, internalization of service excellence, job competence, and job autonomy. The scaling technique used was non-comparative, itemized rating scale i.e. Likert scale. The respondents were asked to rate statements on a five point rating scale where one indicated that respondents strongly disagree, two meant for disagree, three for neutral, four indicated agreeing and five meant strongly agree about what was described in the statement. This technique was used owing to its easier construction and administration, as it is easier for respondents to understand its usage. Each questionnaire was accompanied by a covering letter (request letter) and authorization letter from the Haryana School of Business, which stated the purpose of research, encouraged voluntary participation by employees, ensured the anonymity of their responses, and thanked the respondents for their cooperation.

The universe comprised of male and female employees of banks and insurance companies in the National Capital Region. We used the universe/populationmethod initially for the survey. For this purpose an exhaustive list of banks andinsurance companies in India was prepared, with thehelp of newspapers, magazines and internet. Forgetting permission for survey we wrote letters to thehead offices of banks and insurance companies. Afternot getting favourable response and keeping the time,distance and cost involved in mind, we decided themethod of convenience sampling to approach thebanks and insurance companies personally and seekpermission. A sample was taken from the universe in amanner that would foster both the quality andrepresentativeness of data to facilitating better analysis and interpretation. The commercial banking industry in India ischaracterized by the coexistence of three distinctownership groups: public banks, private banks, andforeign banks (Kumbhakar and Sarkar, 2003). Hence, we decided to visit banks that would lie in each of these categories. Finally, we got permission from six banks and four insurance companies. Out of the sixbanks, two public banks i.e. State Bank of India (SBI)and Punjab National Bank (PNB), two private banks i.e.The Industrial Credit and Investment Corporation ofIndia Limited (ICICI) and Housing DevelopmentFinance Corporation Bank Ltd (HDFC Bank Ltd), andtwo foreign banks i.e. Citi Bank and StandardChartered Bank, became the part of the study. The Insurance Industry has a number of players too.We personally visited the insurance companies alsoand got permission from four insurance companies.Out of the four insurance companies two companieswere from the public sector and two were from privatesector. Public companies were Life InsuranceCorporation (LIC) and Oriental Insurance Corporation(OIC). Private/multinational companies were TATA-AIGand AVIVA Life insurance. On the basis of the permission for survey we got fromthe head offices, the questionnaire was administered to800 respondents of six banks (two governments andfour private / multinational and four insurancecompanies (two governments and two private /multinational). Eighty questionnaires were decided to be distributed toeach organization, which were further divided amongst the branches of that organization randomly dependingon the number of employees present at that time of the survey (who were present at the date and time of thesurvey which was decided for a particularorganization). Thus, random sampling was used to administer the questionnaires to the respondents. As soon as we finished distributing 80 questionnaires to a particular organization and its various branches, we stopped the distribution and moved to the next organization. The questionnaires were distributed personally with the permission letter by visiting the institutions and their branches. This system of distributing questionnaires facilitated and enhanced the researcher’s knowledge regarding the human resource management practices and employees’ service behaviour antecedents through discussion and interview. On the same day, few filled up questionnaires were obtained back. For the rest of the questionnaires two more visits were paid to each branch. Few questionnaires from each branch were also received through mail. Finally, out of the 800 distributed questionnaires, we could collect 615 questionnaires. Only 530 of the returned questionnaires by whatever methods were found suitable for final analysis and without discrepancies. Remaining 85 questionnaires were rejected as they were incomplete in one respect or the other. The detailed description of the sample can be seen through Table 3.1. Out of the sample of 530 respondents, 312 (150 government sector and 162 private/multinational) respondents were from banks and 218 (123 government sector and 95 private/multinational) were from insurance companies. All respondent employees/ managers were of Indian origin. The overall response rate was 66.25%. The brief distribution of the sample is presented in Table 3.2.

The data collected is converted to meaningful informationby its proper classification, tabulation, analysis,interpretation, and presentation. For this purposeappropriate statistical tools were used. The collected datawas properly classified and tabulated first, and then it wasanalyzed and interpreted with the help of various statisticaltools such as means, grand means, standard deviations,percentages, factor analysis, correlations, analysis ofvariance (ANOVA), and multiple regression. The data wasalso checked for reliability.

Reliability analysis was used to refine the human resourcepractices and service behaviour antecedents’ scale.Reliability of the scale shows the extent to which a scaleproduces consistent results, if measurements are maderepeatedly. It can be defined as the extent to whichmeasures are free from random error (Malhotra, 2004).Reliability is the extent to which a measure is consistent;this can be demonstrated by either stability within themeasure (consistency) or stability over time (Rubio, 2005).Numerous marketing and organizational behaviour basedscale development articles illustrated the use of reliability analysis for trimming and retaining items (Netemeyer et al., 2003). There are various approaches to assess reliability, which include the test-retest, alternative forms, split-half coefficient alpha, Kuder-Richardson 20 and Cronbach’s alpha methods. We have used Cronbach’s alpha method for reliability. If the correlation matrix resembles an identity then all correlation coefficients would be zero. However, for factor analysis there must be significant correlation among at least some of the variables. However, increasing the sample size could cause the Bartlett test to become more sensitive to detecting correlations among the variables. Choice of the method used to extract factors is an important step in factor analysis. Since the objective of the study has been to summarize most of the original information (variance) in a minimum number of factors, Principal Component analysis was used. This method considers the total variance and derives factors that contain small proportion of unique variance and in some instances error variance. Here linear combinations of variables are used to account for variation of each dimension in a multivariate space. The variance of the factors is called eigen values. Eigen values represent the amount of variance accounted for by a factor. These are also referred to as the latent root or characteristic root. Communality is the total amount of variance and original variable shares with all other variables. Factor loadings are the correlation between the original variables and the factors. Guidelines exist for identifying significant factor loadings based on sample size. When the set of variables is large the analysis first extracts the largest and best combinations of variables and then proceeds to smaller less understandable combinations. Hence, the number of factors to be extracted becomes an important issue in the absence of any set criterion. There are several criteria to decide. They are: A Priori Method: In which the researcher knows how many factors to expect which enable him/her to specify the number of factors to be extracted beforehand; Determination based on Eigen values: In this approach only the factors with eigen values greater than 1.0 are retained, the other factors are not included in the model;