This chapter outlines the research methodology. It illustrates the research framework, estates the research hypotheses, describes the research procedures, and research design.
Furthermore, it will explain the statistical analysis methods used to evaluate data.
Research Framework
The research framework of this study is based on the research purpose and research questions explained in chapter 1. There are three factors that might affect the satisfaction level of individuals, which are compensation, training and development and Job stress. Job Satisfaction might also have an effect on organizational commitment.
Figure 3.1. Research framework
Figure 0-1
Research Hypotheses
Based on the research questions and purpose of the study, the following hypotheses were formulated:
H1 Training and Development is positively related to employee job satisfaction.
H2 Compensation is positively related to employee job satisfaction.
H3 Job Stress is negatively related to employee job satisfaction.
H4 Job satisfaction is positively related to organizational commitment.
Training and Development
Compensation
Job Stress
Job Satisfaction Organizational
Commitment H1
H2
H3
H4 H5, H6, H7
28
H5 Job satisfaction mediates the relationship between training & development, and organizational commitment.
H6 Job satisfaction mediates the relationship between compensation and organizational commitment.
H7 Job satisfaction mediates the relationship between job stress and organizational commitment.
Research Design
This study followed a quantitative approach to measure the relationship between the independent variables, the mediator and the dependent variable. The data was collected using a survey questionnaire. This method was selected because the researcher intended to collect data in form of numbers, which were then analyzed using statistical methods to test and verify the relationship between the variables. As the literature explains, a quantitative study tries to verify or falsify a relationship or hypothesis, and it focuses on outcomes or effects between variables (Neuman, 2014). Thus, the quantitative approach was deemed to be more effective for the purpose of this research. Due to that the nature of this study was to understand the effects training and development, compensation, job stress has on job satisfaction and organizational commitment. Survey questionnaires were used because they allow the researcher to gather information on the backgrounds, beliefs, or attitudes of a large number of people (Neuman, 2014).
Instrument
The Research instrument used was a self-reported survey questionnaire. The participants choose the best answer according to their feelings and attitudes. This instrument had a total of 94 items, and was divided into six sections, which are:
I. Organizational Commitment II. Job Satisfaction
III. Training and Development IV. Compensation
V. Job Stress
VI. Personal and demographic data.
29
The different scales used in this study were taken from the work of previous researchers, which have been proved to have a good reliability and validity.
From section I to V this survey questionnaire used a 5-point likert scale questionnaire, where respondents were able to evaluate each item from ―strongly agree to strongly disagree‖ or
―strongly satisfied to strongly dissatisfied‖. For the items in part VI, participants were asked to select from the different available options.
Organizational Commitment
This dependent variable was measured using Allen & Meyer‘s (1990) scale to measure organizational commitment. This questionnaire comprises the three components of commitment which are affective commitment, normative commitment and continuous commitment (Jaros, 1997). This section contains a total of 23- items. This survey used a 5-point likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) a sample item is ―I would be very happy to spend the rest of my career in this organization‖. The internal reliability of the three components of commitment has been estimated using coefficient alpha, the median reliabilities are: .85 for affective commitment, .79 for continuous commitment, and .73 for normative commitment (Allen & Meyer, 1996).
Training and Development
This independent variable was measured using an 18-items scale (Bartlett, 2001). This survey comprises four dimensions to reflect the role and influences of Human Resource Development practices. It uses a 5-poin likert scale ranging from 1 (strongly agree) to 5 (strongly disagree) an illustrative item is ―My manager makes sure I get the training needed to remain in the organization and be effective in my job‖.
The 4 dimensions are:
1. Training motivation (Tharenou & Conroy, 1994) 4-items (α=.87)
2. Perceived support of training from colleagues (Noe and Wilk, 1993) 3- items (α= .83) 3. Management training support (Noe & Schmitt, 1989) 6-items. (α=.96)
4. Training benefits (Noe & Wilk, 1993) 5-items. (α=.82)
30
Employee Job Satisfaction
Employee Job Satisfaction was used as a mediator variable in this study. Job Satisfaction was measured using the 15-item shortened version of the Job Diagnostic Survey (JDS) (Hackman & Oldham, 1976).
This survey comprises 6 dimensions: skills variety, task identity, task significance, autonomy and feedback from the job itself and from the agents. It uses a 5-point liker scale ranging from 1 (very satisfied) to 5 (very dissatisfied) a sample item is ―I am satisfied with my job as it provides me with the opportunity to improve my skills‖. The internal reliability for the five dimensions are: 0.71 for skill variety, 0.59 for task identity, 0.66 for tasks significance and task autonomy, and 0.71, 0.78, 0.59 for feedback from the job itself, from agents and dealing with others (Hackman & Oldham, 1975).
Job Stress
Job Stress was used as an independent variable. Job Stress was measure using a 16-items scale. This section comprises three variables, which are role ambiguity (Kerr, McHugh &
McCrory, 2009). Role conflict and role overload (Lin & Rashid, 2010). A 5-point Likert scale is used for this section ranging from 1 (strongly agree) to 5 (strongly disagree) an illustrative item is ―I know how to go about getting my job done‖. The alpha reliability coefficient for role ambiguity is of .80 (Kerr, McHugh & McCrory, 2009). The reliability for role conflict is of .88, and for role overload is .92 (Lin & Rashid, 2010).
Compensation
This independent variable used a pay satisfaction questionnaire (PSQ) developed by Heneman and Schwab (1985), this scale consists of 17-items based on the four independent dimensions, which are: pay level, benefits, raises, and administration (Heneman & Schwab, 1985). A 5-point Likert scale is used for this section ranging from 1 (very satisfied) to 5 (very dissatisfied). The coefficient alpha reliability of the three dimensions of compensation are: 0.95 for pay level, 0.95 for pay benefits, 0.81 for pay raises and 0.88 for administration. (Heneman &
Schwab, 1985).
31
Personal and Demographic Data
This is the last section included in the survey questionnaire; it provides a descriptive analysis of the survey participants. This section contains information such as the participant‘s age, gender, marital status, education, and tenure.
Instrument Validity
The questionnaire had to be translated into Spanish; therefore the researcher carried out two peer reviews, expert review and back translation of the scales, to make sure the scales were accurate and consistent.
Peer Review
The persons in charge of doing peer review had knowledge on the human resource field.
Moreover they were native Spanish speakers with a high level of English. This was being done in order to get feedback regarding the content and the translation of the scales. According to their suggestions the necessary changes were made.
Expert Review
The different scales were also reviewed by two experts; the experts were professionals who work in the bank industry. They read the questionnaire and evaluated whether the questions were effectively related to the topic. They had a high level of education and experience on the Human Resource field. According to their observation about the content and the translations the necessary changes were made.
Back Translation
Back translation was done to ensure the accuracy of the translation. Two bilinguals were in charge of the translation, one translated the original questionnaire to the target language, the second translated back from the target to the original (Brislin, 1970). In this study, the English questionnaires were translated into Spanish by a bilingual person. Secondly, the Spanish questionnaires were translated back to the original language by another person. After having the two English versions of the questionnaires the researcher compared how similar the
32
questionnaires were, if the questionnaires are very similar it will mean that the Spanish version is equivalent to the English version. The two translators are current Human Resource Development master students. One of them is native in both languages. The other is a native Spanish speaker, and all her education has been thru bilingual institutions, she obtained her bachelor‘s degree in the United States, she also has a high TOEFL score.
After the revised questionnaire was ready, the researcher conducted a pilot study. The questionnaires were applied to 40 employees that work in a different financial institution in Honduras. The researcher tested the reliability and validity of the scales to show that the questionnaire was appropriate to use for the formal study.
Pilot Study
A pilot study helps the researcher to detect any problems or deficiencies in the design of a research instrument before proceeding to the formal study (McBurney & White, 2010). Based on the results of the pilot study, the researcher proceeds to make the necessary changes, increasing the credibility and reliability of the research instrument. For this study the pilot test was collected from full time employees of one of the most important banks in Honduras (Banco Credomatic).
Out of 40 questionnaires delivered, 38 were successfully collected. Cronbach Alpha was used in order to test the internal consistency of the questionnaire. The objective was to obtain a Cronbach‘s Alpha of at least .70 which is acceptable for the research. The questionnaire was also reduced from 96 items to 94, based on the reliability results and the feedback of the HR experts.
Item OCN24 (I do not think that wanting to be a ‗company man‘ or ‗company woman‘ is sensible anymore). This item from the organizational commitment scale was eliminated due to the lack of understanding of this question by experts and participants; also it had low reliability score. Item COMP15 (Pay of other jobs in the company) this item was also eliminated based on the feedback of experts; it was considered to be redundant or similar to the other item (Differences in pay among jobs in the company).
Data Collection
The target populations for this study were employees from Banco Atlántida, a commercial bank in Honduras, Central America. Since this is the leading bank in Honduras the researcher considered appropriate to use it for this investigation. Therefore, the researcher used purposive
33
sampling when selecting this organization. Convenience sampling was used to select the participants. The HR manager in charge of the Atlantic area helped with the distribution and collection of the questionnaires. He was in control to the access to the participants within the organization.
This study only covered branches in the Atlantic area; the Atlantic area counts with the third most important city in Honduras, and is the city where the bank was founded. This is also one of the areas most affected by the economic crisis. The sample participants were full time employees. The researcher distributed a total of 200 questionnaires, and collected 150 questionnaires‘ back.
Data Analysis
The researcher used SPSS 22 statistics software for statistical analysis. The following statistics were used to analyze the data:
Descriptive Statistics
This analysis is used to summarize the numerical results and show them in a clear and concise manner (Healey, 2008). The results will be shown in forms of tables to provide the reader a better picture of the data (Neuman, 2014). The descriptive statistics include the means and the standard deviation of the data.
Correlation
This analysis was used to measure the strength of the association between two continuous variables (Gray & Kinnear, 2012). In this case the strength of relationship between the independent variables (Training and Development, Job Stress, Compensation) with mediator (Job Satisfaction), and also strength of the relationship between the mediator (Job Satisfaction) and the dependent variable (Organizational Commitment).
Hierarchical Regression
This analysis is used when there are two or more independent variables; and the purpose of this regression is to test causation (Gray & Kinnear, 2012). Therefore, this study tested the
34
direction and effect the three independent variables (Training and Development, Job Stress, Compensation) have on the mediator (job satisfaction), and the effect the mediator has on the depend variable (organizational commitment).
Research Procedure
First the researcher reviewed the available literature; based on the literature review the researcher identified the research topic. The study was then carried out based on the researcher‘s interest, which is centered on the financial Industry and on key topics such training and development, compensation, job stress, employee job satisfaction and organizational commitment. After have identified the topic the researcher created the theoretical framework.
Then the researcher adopted the different reliable scales to measure each of the variables. Later on the researcher collected and analyzed the data. The researcher also showed the results from the statistical analysis in SPSS and discussed them. Lastly the researcher gave conclusions and suggestions that will be of benefit to the Honduran financial industry and specially Banco Atlántida. The research process is illustrated below:
35 Figure 3.2. Research procedure Figure 0-2
Literature Review
Research Topic Identification
Instrument Development Theoretical Framework
Participants Identification
Data Collection
Data Analysis
Research Results and Discussions
Conclusions and Recommendations
36