In this chapter, it provides the methodology of the present study including research framework, research hypotheses, research procedure, sample, pilot test, data collection, measurement, control variables, and data analysis.
Research Framework
The research framework in this present study was conducted from the literature review and the research questions. The present study proposed that the independent variable is Leader-member exchange (LMX). The dependent variable is employee creativity. And the intrinsic motivation acts as a mediator between LMX and employee creativity. The research framework which is shown in figure 3.1.
Figure 3.1. Research framework
30
Research Hypotheses
Based on the research framework, research questions, and literature review in this study.
The hypotheses are established as follow:
Hypothesis 1: LMX is positively related to employee creativity.
Hypothesis 2: LMX is positively related to intrinsic motivation.
Hypothesis 3: Intrinsic motivation has a positive effect on employee creativity.
Hypothesis 4: Intrinsic motivation mediated the relationship between LMX and employee creativity.
Research Procedure
The procedure of this study can be divided into seven steps. The first step, the researcher identifies the objective and issue of the study. Next, the researcher prioritizes the previous literature that are relevant to the subject of the study including theories, frameworks, researches, and data.
After the concept of the study is determined, the questionnaire or instruments of those variables are developed based on the objective of the study. And also, the researcher is designing which methods and data analysis are suitable for the study. The final step, the researcher provides the conclusion and suggestion for future researchers. The total steps of research procedure are shown in figure 3.2.
31 Figure 3.2. Research procedure
32
Participant
In this study, the respondents are employees and their immediate managers working in the Center Educational Technology (CET), Ministry of Education in Thailand. One of the most creative organizations in Thailand. Thus, this population from CET organization was selected because they concerned creativity as the main concept to do their job. Moreover, all managers and employees in the organization are Thailand's nationality with full-time jobor part-time job, also from the eight department which used creative skill as a main part, those departments including:
the division of directing, the educational television department, the educational radio department, the educational production department, the division of maintenance and technicians, the division of engineering and developing learning innovation, the division of digital education for learners with disabilities, and human resources department. And the participants are required job tenure at least three month's tenure because a probationary period in CET organization is three months which the qualification of those employees are being evaluated as a suitable fit for the organization.
Data Collection Process
In this study, using the purposive sampling is created to serve a very particular need or target. And CET organization is chosen as a participant in this study because the most abilities that they concerned as main function is creativity skill. The tasks in CET required the creativity, original ideas, innovation, and development. Thus, the reason why the researcher chose CET organization as a case study because they concerned creativity as the strength to operate in their organization. At first, the total of 215 participants from 8 divisions are selected because this number of samples is represented the entire number of an employee in CET and it is suitable for a quantitative study in which the aim to conduct inferences and analyze data. However, after the data was collected, the total questionnaires of 201 participants (186 employees and 15 managers) from 8 divisions were returned. There are 14 employees are missing because they have resigned from the organization, and some of them have reached retirement age. Notwithstanding, the surveys were collected during the period of January-February, 2019. With the assistance of CET’s director, the director was described the purpose of this study and handed out the paper-based questionnaires to each of employee and their immediate manager. Each of the questionnaire was attached a cover letter to explain the objective of the study and convinced them of the confidentiality of their responses. Additionally, the present study was reducing the impact of
33
common method bias (CMB) and common method variance (CMV) (Podsakoff, 2003) on the validity score of this study by collecting the data from two sources: managers and employees. The questionnaires for managers are differ from employees. And employees were asked to rate their intrinsic motivation and the quality of LMX relationship with their immediate managers.
Meanwhile, managers were asked to rate their subordinates' creativity. After the questionnaires were completed, it had returned to the director office. Furthermore, the questionnaires were asking about general information of the participants to ensure that they are working forCET organization, Thailand.
Measurement Design
The questionnaire survey was used for this study. And it divided into two kinds of the survey: managers and their subordinates. The questionnaire for managers contains two parts. Part one consists of 3-item reflecting the employees’ creativity. Part two asks about general information of the participants including age, gender, job tenure, and education level. Similarly, the questionnaire for subordinates also contains two parts. Part one consists of 14-item reflecting the two variables (7-item for the quality of LMX and 7-item for intrinsic motivation). And part two asks about general information of the participants including age, gender, job tenure, and education level. Moreover, the survey instruments were originally designed in English and finalize the English version with reliability and validity. Since the survey instruments were used with participants who are Thailand's nationality, it is necessary to translate from English to Thai language by two bilingual scholars in Thailand to guarantee the quality and coordinate of meaning based on a back-translation (Brislin, 1970). The first bilingual scholar was translating the scale into Thai and the second bilingual scholar was translating it back into English.And if the meaning of the English and Thai versions are homologous, then the translation was applied in this study.
Furthermore, the subordinates-reported variables are measured using a continuous rating scale of sum of five-point items and a five-point Likert Scale. Managers-rated variable is measured by a five-point Likert Scale.
According to the three variables, it is important to describe the measurement of these three variables: LMX, employee creativity, and intrinsic motivation. Each of the measurement is described below:
34
LMX
The LMX 7-item scale used by Graen and Scandura (1987) was adopted for the instrument of LMX. According to Erdogan and Bauer (2014), they claimed that the LMX 7-item scale is the most often used for measuring LMX quality. This instrument includes seven items measuring the quality of LMX relationship. Sample items of LMX were “How well does your leader understand your job problems and needs?” The Cronbach’s α for this scale was .89 (Lam, Xu, & Loi, 2018).
The scales use continuous rating scale of sum of five-point items (1 left to 5 right) to score based on the subordinate versions. The subordinates were asked to indicate to which extent they characterize the quality of their exchange relationships with their managers. The entire items of LMX are shown on the table 3.1.
Intrinsic Motivation
The measurement of intrinsic motivation is adopted from Jacobsen et al. (2014) adaptation.
The adaptation items used items from Jacobsen (2010). It is using seven items to indicate the different levels of interest and satisfaction in an employees’ everyday work. This instrument uses five-point Likert-type (1 = strongly disagree, 5 = strongly agree) and rate by subordinates. The sample items of intrinsic motivation were “I very much enjoy my daily work”, the Cronbach’s alpha for these items was .82 (Jacobsen et al., 2014). The entire items of intrinsic motivation are shown on the table 3.1.
Employee Creativity
The measurement of employee creativity is adopted from Wang et al. (2014) adaptation.
The adaptation items were developed from Oldham and Cummings (1996) three-item which is originally established. And the study also adopted two-item more from Tierney et al. (1999).
Totally, this instrument includes item scales of employee creativity in workplace. The five-item mention as original, adaptive, and creative of those employees. So, the managers are rating their subordinates’ creativity based on how frequent their subordinates perform the creativity behaviors on their task in an organization. This instrument uses five-point Likert scale (1 = strongly disagree, 5 = strongly agree). Sample items of employee creativity were “This employee often comes up with original and practical ideas to improve performance”, the Cronbach’s alpha for three-items (Wang et al., 2014) was .82 and two-item from Tierney et al. (1999) was .95. The entire items of intrinsic motivation are shown on the table 3.1.
35 Table 3.1.
Scale of Variables
Code Items
Leader-Member Exchange (LMX)
LMX1 Do you know where you stand with your leader….do you usually know how satisfied your leader is with what you do?
LMX2 How well does your leader understand your job problems and needs?
LMX3 How well does your leader recognize your potential?
LMX4 Regardless of how much formal authority he/she has built into his/ her position, what are the chances that your leader would use his/ her power to help you solve problems in your work?
LMX5 Again, regardless of the amount of formal authority your leader has, what are the chances that he/ she would “bail you out,” at his/ her expense?
LMX6 I have enough confidence in my leader that I would defend and justify his or her decision if he or she were not present to do so.
LMX7 How would you characterize your working relationship with your leader?
Intrinsic Motivation (IM)
IM1 I very much enjoy my daily work.
IM2 A rather large part of my tasks at work are boring.
IM3 I always look forward to going to work in the mornings.
IM4 My work is very exciting.
IM5 I like performing most of my work processes.
IM6 Doing my job, I feel a great personal satisfaction.
IM7 If I won 5 million in the lottery I would still keep my present job.
(continued)
36 Table 3.1. (continued)
Code Items
Employee Creativity (EC)
EC1 This employee often comes up with original and practical ideas to improve performance
EC2 This employee always suggests adaptive and practical ways to achieve goals or objectives
EC3 This employee often comes up with creative ideas, methods, or products that are both useful to the organization.
EC4 This employee identified opportunities for new products/processes.
EC5 This employee served as a good role model for creativity.
Control Variables
There are many studies (Farmer, Tierney, & Kung-Mcintyre, 2003; Gong et al., 2009; Shin
& Zhou, 2003; Tierney & Farmer, 2011) recommended that education level, age, job tenure, and gender may impact on the creativity of those employees behavior. According to Mumford and Gustafson (1988), they claimed that the employees who are older and working with an organization for a long time might influence the relationship with their leaders and creativity. And younger employees seem to concern the idea generation as a core concept, but middle age or older employees seem to concern career identification as a critical concept for producing creativity.
Additionally, Simonton (1984) also mentioned that education level can predict job performance and innovative competence. Moreover, Wang et al. (2014) suggested that job tenure can affect the way that managers perceive employee creativity. For example, the long tenure can allow the manager to have more time to notice their subordinates’ creative behaviors and also employees have more time to show their creative capability either. Thus, in this study, it is necessary to control these four variables as a control variable in the analysis because it is possible that socio-demographic differences might impact the result of the study.
37
Data Analysis
This study used IBM SPSS 23.0 and AMOS 22.0 to calculate the data analysis. Also, descriptive statistics analysis, confirmatory factor analysis (CFA), Structural equation modeling (SEM), and Harman's (1976) single-factor test.
First of all, Harman (1976) single-factor test was used to test the validity of research findings and CMV. According to Anderson and Gerbing (1988), the researcher adopted the two-step strategy to test the hypothesized model. Firstly, beforetesting the hypotheses, AMOS 22.0 was used to conduct CFA (t-test, composite reliability (CR), the average amount of variance extracted (AVE)) on the three variables for verifying the construct validity and reliability of the concepts in this research, also the scale. Then, it is performing structural equation modeling (SEM) to measure the fit and standardized path coefficients (t-value) of the hypothesized model.
Reliability and Validity Pilot Study
A pilot study in this study was conducted before the data collection process. It aimed to investigate the accuracy and quality of questionnaire for managers and employees in the Center Educational Technology (CET), Ministry of Education in Thailand. According to Yurdugül, (2008), he suggested that the minimum sample size for reliability test is 30. Then, the 30 respondents from other organizations undertaken by MOE in Thailand are used for the pilot test.
Additionally, the pilot test in this study determined the reliability of the instrument. The content reliability in the questionnaires was evaluated and tested, Cronbach’s alpha of each item obtain than .70 are acceptable, but Cronbach’s alpha lower than .70 must be revised (Nunnally, 1967).
So, the result has illustrated the Cronbach alpha value of each variable, which obtained a value above .70, indicating good reliability, each of those variables is described below.
The independent variable of leader-member exchange has 7-item with a Cronbach alpha value of .752. The items were rated on a continuous rating scale of the sum of 5-point items (1 left to 5 right) to score based on the subordinate versions.
The mediating variable of intrinsic motivation has 7-item with a Cronbach alpha value of .728. The items were rated on a five-point Likert scale (1 = strongly disagree, 5 = strongly agree) by subordinates.
38
The dependent variable of employee creativity has five-item with a Cronbach alpha value of .944. The items were rated on a five-point Likert scale (1 = strongly disagree, 5 = strongly agree) to score based on the manager versions. See table 3.1. for more details.
Table 3.2.
Reliability Results of the Pilot Study
Variable Name Cronbach Alpha Value
Leader–Member Exchange .752
Intrinsic Motivation .728
Employee Creativity .944
Confirmatory Factor Analysis (CFA)
According to the data collection methods in this study, it is required to examine the common method variance and the validity of the measurement model because all of the data are collected with the same questionnaire also in the same period of time with cross-sectional research design. So, the confirmatory factor analysis (CFA) was adopted, it is a part of SEM. In other words, it defines as measurement model which used to examine validity, common method variance, and common method biases. CFA was conducted by using AMOS 22.0. In this part, it can be divided into two steps. At first, Harman's (1976) single-factor test in factor analysis was conducted by SPSS 23.0 and it used to test common method variance and common method biases. Then, CFA was used to test the verification of common method variance, common method biases and validity for ensuring the quality of the measurement method in this study.
According to Eichhorn (2014), if the results show that the first item of latent factor accounts for more than 50% of the variance, then common method variance and bias may occur. After all of the 19-item were accessed into factor analysis, table 3.3. shows the result of Harman’s single factor test. And it is found that the first item did not explain for a majority of the variance (43.4%).
It is implied that there is no general item is apparent for the variance. So, the common method variance and biases are not found in this processing.
39 Table 3.3.
The Result of Harman’s Single Factor Test
Initial Eigenvalues Extraction Sums of Squared Loading Items Total % of Variance Cumulative % Total % of Variance Cumulative %
LMX 1 8.25 43.455 43.455 8.257 43.455 43.455
LMX 2 2.40 12.634 56.089
LMX 3 1.41 7.441 63.530
LMX 4 .999 5.259 68.790
LMX 5 .739 3.891 72.681
LMX 6 .654 3.443 76.123
LMX 7 .612 3.220 79.343
IM 8 .543 2.860 82.203
IM 9 .475 2.501 84.705
IM 10 .431 2.268 86.973
IM 11 .386 2.033 89.007
IM 12 .372 1.958 90.965
IM 13 .342 1.800 92.764
IM 14 .287 1.513 94.277
EC 15 .280 1.476 95.753
EC 16 .252 1.326 97.079
EC 17 .206 1.086 98.165
EC 18 .186 .981 99.146
EC 19 .162 .854 100.000
Note. 1. Extraction Method: Principal Component Analysis, 2. LMX = Leader-Member Exchange, 3. IM = Intrinsic Motivation, 4. EC = Employee Creativity.
40
For the common method variance and biases in CFA, the common method variance is occurring among those variables when the one-factor CFA model fits the data correctly (McFarlin
& Sweeney, 1992; Podsakoff & Organ, 1986). Table 3.4 shows the summary of good-fit criteria model for SEM analysis, but it also can use in CFA. The results of CFA models were robust and fitted with the structure based on the following criteria and conditions in this study. After 19 items loaded into one model to investigate the fit of the confirmatory factor analysis model, table 3.5 illustrates that the single-factor model did not fit the data correctly (χ2= 743.0 (df = 152, p < .01);
χ2/df = 4.88; GFI = .602; AGFI = .502; RMR = .083; RMSEA = .145; CFI = .713). So, the results indicate that the CFA model for Harman’s single factor test (see figure 4.1.) did not found the common method variance and biases in this study.
Table 3.4.
Summary of Good-Fit Criteria
Fit Index Acceptable Levels References
Chi-square χ2 p > .05 Kelloway (1998)
Chi-square χ2/df < 3.0 Diamantopoulos and Siguaw
(2013); Kline (2005)
GFI > .90 Kelloway (1998)
AGFI > 0.80-0.89 Reasonable fit
> 0.90 Good model fit
Kelloway (1998); Lattin, Carroll, and Green (2003)
RMR < 0.05 Diamantopoulos and Siguaw
(2013); Kelloway (1998)
RMSEA < 0.05 Kelloway (1998)
CFI > 0.90 Hu and Bentler (1999);
Kelloway (1998)
41 Table 3.5.
Confirmatory Factor Analysis Result for Harman’s Single Factor Test
χ2 df χ2/df GFI AGFI RMR RMSEA CFI
743.0** 152 4.88 .602 .502 .083 .145 .713
Note. χ2= Chi square. df= Degree of freedom. GFI= Goodness-of-fit index. AGFI= Adjusted goodness-of-fit index. RMR= Root mean square residual.RMSEA= Root mean-square error of approximation. CFI= Comparative fit index. **p < .01
Additionally, before testing the hypotheses of this research, CFA was conducted on the three main variables in order to verify the construct validity of the study framework. Table 3.6.
shows that the factor loadings of three variables and their observed variables.
The factor loading of LMX1 to LMX7 with leader-member exchange are ranged from .51 to .82, it is acceptable because the criteria for factor loading suggests that it should be greater than .40 (Hinkin, 1998; Kim & Mueller, 1978). And the results of t-test all over 1.96 which are reached the level of significance.It implies that the observed variables are reasonable to reflect the constructed latent variables (Hu & Bentler, 1999; Joreskog & Sorbom, 2006). Moreover, the average amount of variance extracted (AVE) and composite reliability (CR) were used to examine the reliability and validity of the scale. According to Hair, Anderson, Tatham, and Black (1998), the criteria of CR value suggests that it must be greater than .70 and the AVE value must be greater than .50 (Fornell & Larcker, 1981). The Leader-member exchange’s CR value in this study reached at .88 and AVE values reached at .44, it below than .50, but it can acceptable because of it close to .50.
The factor loading of IM1 to IM7 with intrinsic motivation are ranged from .33 to .88. IM5 is the only observed variable that has factor loading below .40, but it acceptable because the model fit-indices are met the satisfactory levels and IM5 factor loading is significant to the latent variables. In addition, the intrinsic motivation’s CR value in this study reaches at .89 and AVE values reach at .55. Moreover, the factor loading of E1 to E5 with employee creativity is ranged from .81 to .89. The employee creativity’s CR value in this study reaches at .94 and AVE values reach at .73. Additionally, the results indicate that the model has a good fit to the data (see figure 4.2) (χ2= 150.326 (df = 135, p = .174), χ2/df = 1.114, GFI = .924, AGFI = .894, RMR = .047, RMSEA = .025, CFI = .993). The overview of the results in table 3.6. illustrates the latent variables
42
in this research have great consistency, reliability, and convergent validity to the research
in this research have great consistency, reliability, and convergent validity to the research