This section outlines the methodology used in the study. Specifically, it covers research framework, procedure of conducting the research, participants or target sample, measurement of scales, data collection and analysis. The study applied the quantitative research method to examine the relationships among Job Demands, Emotional Exhaustion and Psychological Contract Breach.
Research Framework
The framework of this study was built based on research objectives and literature review part which is presented in figure 3.1 as below. In this model, Job Demands with three dimensions namely Workload, cognitive and Emotional Demand were regarded as independent variables and Emotional Exhaustion is the dependent variable. Meanwhile, Psychological Contract Breach served as moderator to test its influence on the relationship between Job Demands and Emotional Exhaustion.
Figure 3.1. Research framework
Psychological Contract Breach
Job Demands (1) Workload
(2) Cognitive Demand (3) Emotional Demand
Emotional Exhaustion
Control Variables (1) Age
(2) Gender (3) Job Tenure H2; H2a -b-c
H1; H1 a-b-c
24
From the illustration of research framwork, the hypotheses formulated in this study include:
Hypothesis 1: Job Demands have a positive relationship with Emotional Exhaustion.
Hypothesis 1a. Workload has a positive association with Emotional Exhaustion.
Hypothesis 1b. Cognitive Demand has a positive association with Emotional Exhaustion.
Hypothesis 1c. Emotional Demand has a positive association with Emotional Exhaustion.
Hypothesis 2: Psychological Contract Breach moderates the relationship between Job Demands and Emotional Exhaustion, such that higher Psychological Contract Breach strengthens the relationship between Job Demands and Emotional Exhaustion.
Hypothesis 2a. Psychological Contract Breach strengthens the relationship between Workload and Emotional Exhaustion, such that higher Psychological Contract Breach strengthens the relationship between Workload and Emotional Exhaustion.
Hypothesis 2b. Psychological Contract Breach strengthens the relationship between Cognitive Demand and Emotional Exhaustion, such that higher Psychological Contract Breach strengthens the relationship between Cognitive Demand and Emotional Exhaustion.
Hypothesis 2c. Psychological Contract Breach strengthens the relationship between Emotional Demand and Emotional Exhaustion, such that higher Psychological Contract Breach strengthens the relationship between Emotional Demand and Emotional Exhaustion.
25
Research Procedure
The process of conducting the research was presented in this segment. First detailed explanations for each step was introduced and then Figure 3.2. below would summary the whole procedure.
Research Background and Framework
First, the interested topic was chosen after reading various past articles and then discussion with the thesis advisor. Also the background of the study was analyzed through the online information available about service industry in Vietnam. Afterwards, the framework model of the study was formulated with specific sample target and Psychological Contract Breach as moderator in investigating the correlation of Job Demands to Emotional Exhaustion.
Literature Review
Researcher focused on empirical study to learn about the connection of all variables proven in the past besides linking the theory to the proposed relationship among all constructs. In this research, literature about Job Demands, Emotional Exhaustion and Psychological Contract Breach is the main interest.
Hypotheses Formation
Based on research purpose and literature review part, hypotheses about the relationship of all variables was generated, which also stated specifically the moderating effect of Psychological Contract Breach.
Questionnaire Design
Researcher reviewed all measurements related to each variable in the study with the focus on the validity and reliability of each scale, besides the well-known of its application. Subsequently, the scale of Job Demands (three dimensions), Emotional Exhaustion and Psychological Contract Breach would be determined and arranged in a scientific way together with control variables in demographic part.
26
Data Collection
The survey was then be distributed to service workers in Vietnam via convenience and snowball sampling methods, which took around two months to complete. The majority of the survey was uploaded online via social media for better approach to participants while the rest was printed to give to friends.
Data Analysis
The software SPSS version 23 was chosen to conduct analysis for the data with step-by-step analysis including descriptive analysis, reliability and correlation analysis, confirmatory factor analysis (CFA) and hierarchical regression analysis.
Research Result and Implication
From the analysis, results were interpreted with hypothesis confirmation or rejection. Besides, implications and limitations of the research would also be presented based on the results and the whole research process.
Proofread and References
After the main parts of the study are completed, proofread from research and also the comments from thesis advisor was necessary to revise any errors in time. Besides, references would be stated and combined in the last part of the research so that it is easier for readers to refer to the original source.
27
Figure 3.2. Research procedure
Research Background and Framework
Literature Review
Hypotheses Formation
Questionnaire Design
Data Collection
Data Analysis
Research Results and Implications
Proofread and References implication
28
Research Designs
This study applied quantitative research method to find out the relationships between Job Demands and Emotional Exhaustion, with the moderating impact of Psychological Contract Breach.
Specifically, self-reported survey was conducted. The questionnaire in this study comprises of four major sections including Job Demands (Workload, Cognitive Demand and Emotional Demand), Emotional Exhaustion, Psychological Contract Breach and demographic details. Before distributing to participants, back translation method was applied for convenience of readers. All the scales were originally in English, while target sample is fulltime workers in Vietnam service industry, therefore first, the survey was translated into Vietnamese by an English teacher in Vietnam. Subsequently, two Vietnamese graduate students with at least 7.0 IELTS level and also live in English-speaking countries were invited to translate back to English so that the questionnaire reaches the consistency in terms of translation.
In designing the questionnaire, the researcher also paid attention to common method variance (CMV) bias. For example, some participants may intentionally want to provide all positive and same answers to all questions. Therefore, in the questionnaire there are three items belonging to Psychological Contract Breach were reversed score to avoid random and systematic answers affect the result. Furthermore, the variables were not explained in detail in the questionnaire together with the proposed hypothesis according to the suggestion of Spector and Brannick (2009) is also another solution for this problem.
Research Sample and Data Collection
The sample for this study were fulltime workers in Vietnam service industry, and the division of this sector was based on the Services Sectoral Classification developed by World Trade Organization (WTO). Regarding Vietnam already joined WTO for more than a decade, this list is considered suitable for doing research purpose and globally accepted. Specifically, there are twelve main types of business in the service sector including: “Business services, communication services, constructions and related engineering services, distribution services, educational services, environmental services, financial services, health related and social services, tourism and travel related services, recreational, cultural and sporting services, transport services and others” (World Trade Organization, 1991, pp.1-7). In total, there were 245 valid respondents who are fulltime
29
Vietnamese service workers in this survey after two months of collecting data, from 15th January 2020 to 15th March 2020.
Data was collected via convenience and snowball sampling using online social media and also paper to reach more participants. Specifically, convenience sampling, which belongs to non-probability sampling method, refers to collecting target sample that is accessible and approachable by the researcher. Therefore, this survey was sent to the researcher’s Vietnamese friends and relatives who worked in service industry. Besides, snowball sampling was also applied in this research. Once the survey reached one participant, it would be introduced to their networks including friends, colleagues, or relatives who are also the sample target of the research. In this way the time for collecting data was reduced and the coverage density of the questionnaire increased substantially. The description of target sample with specific criterion was also provided in the research introduction so that only people who worked in Vietnam service industry joined. The main social media platforms used in distributing the questionnaire include Facebook, Gmail, and Zalo.
Measurement
The survey for this study comprises of question items for three main variables and demographic information, which is stated in Appendix A. Specifically, there are 15 items for Job Demands, 5 items for Emotional Exhaustion and 5 items for Psychological Contract Breach. The questionnaire is originally in English but was translated into Vietnamese for convenience of the participants, together with back translation method to English for accuracy and validity check, as stated in the research design section.
Job Demands
There are three core dimensions in Job Demands variable, namely Workload, Cognitive Demand and Emotional Demand.
Workload (Quantitative demand) was measured with the scale developed by Spector and Jex (1998), with 5 questions focusing on the amount and pace of work. Examples of Workload inventory include “My job requires me to work very fast” and “I often have to do more than I can do well”. In the original research, the Cronbach’s Alpha for this scale was 0.82. Answers were stated according to 5-point Likert frequency scale, from 1 (never) to 5 (always), with higher score representing higher Workload experienced.
30
Cognitive Demand scale was applied from the Copenhaghen Psychological Questionnaire – version two (COPSOQ II) by Pejtersen et al. (2010) as this measurement was designed specifically for jobs that have high psychologically-related factors. It is also a sandardized and validated scale with Cronbach’s alpha of 0.74 conducted among 3517 samples in the original paper. Example items include “Does your work require you to remember a lot of things” and “Does your work require you to make difficult decisions?”. Answers were also stated according to 5-point Likert frequency scale, from 1 (never) to 5 (always).
Emotional Demand indicator wias measured with the set of six question items developed by Bakker et al. (2003), reaching 0.74 for reliability (Cronbach’s Alpha value). Some examples can be stated including “Is your work emotionally demanding” or “Do you have to deal with clients who do not treat you with the appropriate respect and politeness? Answers were also stated according to 5-point Likert frequency scale, from 1 (never) to 5 (always).
Emotional Exhaustion
The measurement of this variable was adopted from Maslach Burnout Inventory- General Survey (MBI-GS). This questionnaire was developed by Schaufeli et al. (1996) with 0.87 internal consistency score. Two other dimensions also stated in this scale are Cynicism and Professional Efficacy. However, this study only focused on Emotional Exhaustion which was measured by five questions. Items for this scale were assessed with a 5-point frequency scale ranging from 1 (never) to 5 (always). Examples of this construct are “I feel emotionally drained from my work” and “I feel used up at the end of the workday”.
Psychological Contract Breach
Psychological Contract Breach was measured by the scale based on Robinson and Morrison (2000) which contained five questions and recorded 0.92 score for its Cronbach’s Alpha. This is also considered global assessment for employee’s perception of the degree to which organizations fulfilled its promises. There are three items were reversed such as “Almost all the promises made by my employer during recruitment have been kept so far” and “So far my employer has done an excellent job of fulfilling its promises to me”. Responses were designed according to 5-point agreement scale ranging from 1 (strongly disagree) to 5 (strongly agree), higher score indicates higher Psychological Contract Breach, except for the reversed questions.
31
Control Variables
Empirical studies showed that some demographic characteristics could have substantial influence on Emotional Exhaustion. Therefore, in this study age, gender and job tenure were selected as control variables so that in testing the hypotheses formulated, the effect of these control variables would not distort the result.
Age. Past literature showed that age was inversely correlated to Emotional Exhaustion (Bekker, Croon, & Bressers, 2005). The older people become, the less Emotional Exhaustion they suffer.
Another research of Morgan, Van Haveren and Pearson (2002) also agreed with this finding, explaining this relationship by using coping mechanisms as the method of dealing with burnout when people ages. Recent scholars such as Seery and Corrigall (2009) also found the same the connection with Emotional Exhaustion.
Gender. Meanwhile, gender differences also were recorded to affect Emotional Exhaustion (Brewer and Shapard, 2004). This gender difference was found to be greatest among young and new employees who just had few years of working experience, women often suffer higher Emotional Exhaustion (Bakker, Demerouti, & Schaufeli, 2002; Markus et al., 2018). Maslach (2001) also found that women tend to suffer more Emotional Exhaustion than men. This relationship could be explained via gender role theory by Eagly and Wood (1982), which stated that women tend to reveal their own emotions so more likely to get emotional depletion after times, while men prefer to suppress emotions.
Therefore, gender is considered control variable in this study.
Job tenure. Job tenure in the organization is the period of time any employees work in a company. Long time ago, Maslach and Jackson (1981) already found besides age, tenure also has negative impact on Emotional Exhaustion. Even recent study conducted by Topa, Pozo and José (2018) among 364 public employees in service sector also believed that the newer the staff members was, the higher Emotional Exhaustion recorded.
32
Pilot Test
Before collecting main data for this study, a pilot test was conducted with the aim of testing the reliability and face validity of all scales, and also the appropriateness of the questionnaire after comments from participants. After two weeks of distributing the survey online for pilot test, from December 27th 2019 to January 10th 2020, there were 85 respondents, all were fulltime workers in Vietnam service industry. The result of this pilot test provided general picture for researcher to understand the sample, while showing the reliability of each scale, represented by Cronbach’s Alpha value. According to Wim, Patrick and Patrick (2008), Cronbach’s alpha value in the range of 0.6 and 0.8 is considered as acceptable. From the pilot test, the majority of variables had good reliablity (over 0.7) and Cognitive Demand’s scale also had acceptable reliability. All the translation and contents of questions in the survey got good feedbacks from participants without confusing or misleading terms.
Table 3.1.
Reliability of Each Scale (N= 85)
Variable Total item number Cronbach’s Alpha
Workload 5 .784
Cognitive Demand 4 .617
Emotional Demand 6 .855
Job Demands 15 .826
Emotional Exhaustion 5 .898
Psychological Contract Breach 5 .761
Data Analysis
After valid responses were collected, data was imported to Statistical Package for Social Sciences (SPSS) 23.0 version. The analysis techniques were conducted step-by-step as below:
Reliability Analysis
The Cronbach’s Alpha score represents the reliability and internal consistency of each scale. In other words, it shows whether the measurements selected in the questionnaire would return the stable responses or not. This coefficient ranges from 0 to 1 with higher number indicating higher reliability.
Specifically, if Cronbach’s alpha score reached 0.7 upwards, the scale would be considered appropriate and acceptable (Nunnally & Berstein, 1994).
33
Descriptive Analysis
General information including means, standard deviation, and frequency distribution of all variables was then analyzed by descriptive analysis. This method also analyzed all demographic characteristics of participants such as age, gender, education, specific sub-industry, marital status, educational level in number and percentages.
Pearson Correlation Analysis
Then the degree of correlation among all variables including Job Demands (Workload, Cognitive Demand and Emotional Demand), Emotional Exhaustion and Psychological Contract Breach was tested using Pearson correlation analysis. The result would range from -1 (negative linear correlation – same with other negative value) to 1 (positive linear association – same with other positive value) for the variables included. However, if the value is 0, there would be no relationship among all constructs.
Confirmatory Factor Analysis (CFA)
This study will apply Confirmatory Factor Analysis (CFA), one of the structural equation modeling (SEM) techniques which are normally used in quantitative research. The purpose of CFA is to evaluate whether the observed data fits well with proposed model, using Mplus software to conduct.
In running CFA, attention was focused on Chi-square (c2), Comparative Fit Index (CFI), Tucker Lewis Index (TLI), The root mean square error of approximation (RMSEA) and finally The standardized root mean-square residual (SRMR). For Chi-square value, the probability greater than chosen alpha score (0.05) indicates good fit (Tabachnick & Fidell, 2007). Moreover, CFI and TLI all range from 0 to 1 in value, if the score reach over 0.9 means reasonable fit. Last but not least, RMSEA and SRMR below 0.08 is considered reasonable fit of the whole model (Browne & Cudeck, 1989).
Hierarchical Regression Analysis
Finally, all hypotheses were verified using hierarchical regression analysis. In this study this analysis was adopted to test the correlation among Job Demands (include three dimensions), Emotional Exhaustion and Psychological Contract Breach. Control variables was inserted in the first model, then Job Demands (independent variables) and Emotional Exhaustion (dependent variables) were also entered in the second one. The interaction term of psychological contract and Job Demands (each dimension was entered according to each hypothesis) was put in the third model.
34
35