This chapter contains the research design and methodology of the study. The chapter presents the research hypothesis derived from the literature review and research questions, explains the research framework and how the variables are measured and tested in the study. The chapter also provides details of the measurement instrument as well as validation. Sample, data collection, and data analysis methods are also introduced in this chapter.
Research Framework
The research framework in this study was developed following the literature review.
In this study, there are three dependent variables: knowledge sharing, knowledge creation, and international students’ performance. The independent variable in this study is organizational culture. The following theoretical framework shows the relationship between the variables.
In order to examine the effects of the variables, this study developed the Universities’
Culture, Knowledge Sharing, Knowledge Creation and Commitment Model by Cheng-Ping Shih and Vasquez. The model was partially adopted from Organizational Culture by Parsons (Sashkin & Rosenbach, 2002), Knowledge sharing by Ulyses Agustine (2013), Knowledge Creation by Nonaka (1994) and Organizational Commitment by Mowday, Steers and Porter (1979). The research framework developed can be found below.
29 Figure 3. 1. Research framework.
Research Hypotheses
Following the literature review and research framework, three hypotheses will be tested in this study. Based on the literature review we assumed that there is a relationship between the variables chosen and that relationship and influence will be tested in this study. Therefore, we will use null hypotheses to test the relationship between the variables.
The sampled data will be used to test the following:
30
H1. Organizational culture has no effect on knowledge sharing in international students in Taiwanese universities.
H2. Knowledge sharing has no effect on the knowledge creation of international students in Taiwanese universities.
H3. Knowledge creation has no effect on the commitment of international student in Taiwanese universities.
Research Procedures
An orderly research procedure has been clearly outlined for the purpose of making clear the steps for this study: After the literature review and constructs of definitions, the next step is to identify and define the research questions and research hypotheses of the study. The development of the research framework follows after this step and the research framework is adapted from models used by different researchers (Augustine, 2013) (Nonaka, Byosiere, Borucki, & Kono, 1994), (Jaros, 2007), (Sashkin & Rosenbach, 2002). The research instrument is developed following the constructs of the research framework, and adopting a language that can be applied and easily understood in the educational context. The instrument consists of a 5 point Likert scale that ranges from 1- Totally Disagree to 5- Totally Agree.
For the conduction of the pilot study, the instrument is sent online to international students studying in Taiwan. For the main study, the questionnaire is distributed online and paper to international students. The questionnaire includes a cover letter that explains the purposes of the study to the student. Instructions for how to complete the questionnaire were also included.
31
Once the questionnaires are collected, the data is coded and keyed into the system equal modeling program (Smart PLS) and the Statistical package for Social Science (SPSS). The coding helps the researchers to analyze the information as well as to test the hypotheses in order to answer the research questions. The research is conducted following this research process:
32 Figure 3. 2. Research procedures
1
• Research motivation
2
• Review literature
3
• Define constructs
4
• Identify research questions and hypotheses
5
• Develop research framework of the study
6
• Determine research methodology
7
• Develop instrument
8
• Conduct Pilot study
9
• Expert judgement and review
10
• Instrument review
11
• Data collection and coding
12
• Data analysis
13
• Conclusions and suggestions
33
Measurement Instrument
A quantitative data approach is used for the conduction of this study. The measurement of the constructs in this study is accomplished by the development of a questionnaire, with multiple items and using a 5 point Likert Scale. The instrument uses a total of 17 validated construct scales that were found in the literature.
The research instrument is a self-examined questionnaire in which the respondents answer based on their beliefs and opinions. The questionnaire is distributed online and in paper questionnaire. The items of the questionnaire are divided into sections that mark the different variables that will be measured and provide the surveyed person with a clear understanding of the instructions.
All the items in the questionnaire are adapted from pre-validated measures in existing related studies. The questions are also adapted to fit the educational context and to facilitate the comprehension of the questions. The survey contains four sections:
Organizational Culture, Knowledge Sharing, Knowledge creation, Commitment, as well as general information of the respondent.
The survey is presented in a 1 to 5 point Likert Scale; 1 corresponding to “Totally Disagree” and 5 corresponding to “Totally Agree”. The items in the sections of the questionnaire are adapted from different sources as detailed in the following:
The measurement instrument consists of 4 variables with a total of 61 questions including a section for demographic questions. Part I of the questionnaire contains University’s Culture (C); Part II contains Knowledge Sharing (KS); Part III contains
34
Knowledge Creation (KC); Part IV contains Commitment (CC), with a total of 17 research variables. The questionnaire uses a 5- point Likert scale, ranging from 1- Totally disagree to 5- Totally agree.
Organizational Culture (C)
This section contains 14 questions that are adapted from the Organizational Culture Assessment Questionnaire (OCAQ), which is based on the work of Dr. Talcott Parsons, a sociologist at Harvard University (Sashkin & Rosenbach, 2002). This section is categorized in four variables that are: Managing Change (OCUL-MG), Achieving Goals (OCUL-AG), Coordinated Teamwork (OCUL-CT) and Cultural Strength (OCUL-CS).
Knowledge Sharing (KS)
The items measuring Knowledge Sharing are 17 and adopted from the questionnaire found in the work of Measuring the Effects of Knowledge Sharing and Knowledge Transfer on the Survival and Competitiveness of Taiwan ICDF questionnaire (Augustine, 2013). The variables of this section are: Trust (KS-T), Collaboration (KS-COL), Cooperation (KS-COO), Team (KS-T), Mutual Concern (KS-MC) and Asking Questions (KS-AQ).
Knowledge Creation (KC)
There are a total of 11 questions in this section. The questions are adapted from the questionnaire found in Managing the Effects of Knowledge Management Strategies, Enablers, and Assets on TSMC’s Creation Process and Performance (Liu, 2013). The SECI model of Nonaka and Takeuchi is also used in this section to define the variables
35
which are: Socialization (KC-S), Externalization KC-E), Combination (KC-C) and Internalization (KC-I).
Commitment (CC)
This section consists of 10 items adopted from the Organizational Commitment Questionnaire developed by Mowday and Steers and Porter (1979). The section consists of 3 variables that follow the model developed by Meyer and Allen but that are widely used among many researchers: Normative Commitment (OCOM-NC), Affective Commitment (OCOM-AC) and Continuance Commitment (OCOM-CC).
Population and Samples
The population of this study consists of foreign nationals currently enrolled in a study program in Taiwan, or studying higher non-degree seeking education. The Ministry of Education reports that the population of international students in Taiwan is 92,685 international student as of 2014 (Ministry of Education [MOE], 2015). This number is predicted to increase for 2015.
The respondents of the study are selected with a convenience and snowball sample method. The students volunteer to answer the questionnaire which is a self-reported questionnaire. Paper and online questionnaire are used to gather data from the students.
The sample size used for the pilot study comprises a total of 41 questionnaires. For the main study, a total of 210 questionnaires are collected from which 180 are valid questionnaires.
36
Validity and Reliability Validity and Reliability of Pilot study
Partial Least Squares (PLS) method is used to analyze the sample surveys of this study and to test the reliability of the instrument. PLS is used to analyze simultaneously the interrelationships among all the constructs. PLS is a popular model used to analyze small samples because it helps to predict responses. The software focuses on maximizing the variance of the dependent variables explained by the independent variables. The model reflects the relationships between the latent variables and a measurement component, which shows how the latent variables and the indicators are related (Haenlein
& Kaplan, 2004).
All the items from Table 3.1 show that the there is a high internal consistency based on the alpha reliability of items. The pilot results shows that all of the values for Cronbach Alpha are higher than 0.72. Items for the section of university culture (14 items) had a value of 0.915; for the knowledge sharing section (17 items) Cronbach’s Alpha was 0.876; for the section of knowledge creation (11 items) the value was 0.857 and for the section of commitment (10 items) the Cronbach’s Alpha was 0.765. All the values can be observed below in Table 3.1.
37 Table 3.1.
Pilot Cronbach's Alpha Results for all Dimensions
Constructs Number of Items Cronbach Alpha
University Culture 14 0.915
Knowledge Sharing 17 0.876
Knowledge Creation 11 0.857
Commitment 10 0.765
The individual reliability of the items was evaluated in the pilot study by examining the outer loadings of the measurements with their corresponding constructs. Table 3.2 shows that all the indicators had outer loadings greater than 0.5, meaning that the items in the questionnaire where related to the items they were supposed to measure. No items were deleted in this section.
Table 3.2.
Pilot Outer Loading
Loadings
CC_AC .862 CC_CC .856 CC_NC .755
C_AG .909 C_CS .915 C_CT .869
C_MC .877
KC_C .765 KC_E .868 KC_I .852
KC_S .848
(Continued)
38 Table 3.2. (Continued)
Loadings
KS_AQ .575 KS_C .886 KS_CO .890
KS_MC .822 KS_T .900 KS_TC .867
Note: CC_AC= Affective Commitment; CC_CC= Continuance Commitment; CC_NC=
Normative Commitment; C_AG= Achieving Goals; C_CS= Cultural Strength; C_CT=
Coordinated Teamwork; C_MC= Managing Change; KC_C= Combination; KC_E=
Externalization; KC_ S= Socialization; KC_I= Internalization; KS_AQ= Asking Questions; KS_C= Collaboration; KS_CO= Cooperation; KS_MC= Mutual Concern;
KS_T= Trust; KS_TC= Team.
To evaluate the convergent validity of the pilot study, composite reliability and average variance were examined. Since Cronbach’s Alpha tends to provide a conservative measure of internal consistency reliability in PLS-SEM, it is also recommended to use composite reliability to confirm the internal consistency of the items (Wong, 2013). To show high levels of internal consistency the values must be greater than the minimum of 0.6 (Wong, 2013). Table 3.3 shows that all the items have values grater the minimum requirement for internal consistency. In order to verify the convergent validity of the instrument, we check the Average Variance Extracted values. Here, a minimum of 0.5 is acceptable in order to confirm convergent validity of the items. Lastly, R2 shows the explanatory power of the total variance which are also important for this study.
39 Table 3.3.
Pilot Average Variance, Internal Consistency and R2
Constructs Number of Items AVE Internal consistency
R2 (%)
University Culture 14 0.797 0.940
Knowledge Sharing 17 0.679 0.912 0.518
Knowledge Creation 11 0.695 0.901 0.462
Commitment 10 0.682 0.865 0.492
Validity and Reliability of Main Study
Reliability Test Analysis.
The Cronbach’s Alpha reliability test in Table 3.5 shows the coefficient values for each of the constructs as follows: University Culture (.83); Knowledge Sharing (.88);
Knowledge Creation (.84); Commitment (.66). Except commitment, all the constructs have Cronbach’s Alpha values higher than .70. Even though commitment has a value lower than the acceptable level of .70, it can be still be accepted because it’s not so low that it must be rejected. Therefore, it can be concluded that the scales for all dimensions have a high level of internal consistency and reliability.
40 Table 3.4.
Cronbach’s Alpha Results for all Dimensions (N=180)
Constructs Number of Items Cronbach Alpha
University Culture 14 0.838
Knowledge Sharing 17 0.888
Knowledge Creation 11 0.842
Commitment 10 0.666
Validity and Reliability.
The assessment of the validity of the measurement instrument is measured with convergent validity and discriminant validity. Composite reliability and average variance extracted were analyzed to determine the convergent validity and reliability of the measurement instrument. The minimum value recommended for composite reliability is .70 (Wong, 2013). Table 3.5 shows that all the values of composite reliability are higher than minimum value, which indicates that the values are accepted.
Average Variance extracted was evaluated to check the convergent validity of each of the variables in this study. AVE values are recommended to be higher than 0.5 as it indicates that each of the dimensions converge well. Table 3.5 also shows that the values are higher than the minimum requirement. Therefore, all the constructs met convergent validity. Note also that most of the values for Cronbach Alpha are higher than 0.70, except for Commitment (0.66). However, this value can still be accepted since this value is still regarded as sufficient.
41 Table 3.5.
Main Study Constructs' Reliability Analysis for this Study
Code Constructs Composite
Reliability Cronbach Alpha AVE
C University Culture 0.892 0.838 0.673
KS Knowledge Sharing 0.915 0.888 0.642
KC Knowledge Creation 0.894 0.842 0.679
CC Commitment 0.812 0.666 0.597
Individual factor loading of each of the items were assessed in this study by using Smart PLS. Outer loadings of 0.7 are preferred, but loadings higher than 0.4 are also acceptable. All the items loadings in Table 3.6 are greater than 0.57, so all the items met the required criteria.
Table 3.6.
PLS Outer Loadings
Loadings
CC_AC .859 CC_CC .851 CC_NC .574
C_AG .801 C_CS .836 C_CT .827
C_MC .816
KC_C .77 KC_E .871 KC_I .842
(Continued)
42 Table 3.6. (Continued)
Loadings KC_S .809
KS_AQ .77 KS_C .759 KS_CO .835
KS_MC .825 KS_T .802 KS_TC .812
Note: CC_AC= Affective Commitment; CC_CC= Continuance Commitment; CC_NC=
Normative Commitment; C_AG= Achieving Goals; C_CS= Cultural Strength; C_CT=
Coordinated Teamwork; C_MC= Managing Change; KC_C= Combination; KC_E=
Externalization; KC_ S= Socialization; KC_I= Internalization; KS_AQ= Asking Questions; KS_C= Collaboration; KS_CO= Cooperation; KS_MC= Mutual Concern;
KS_T= Trust; KS_TC= Team.
Data Analysis Method
Partial Least Squares (PLS) method was used to analyze the data of this study. PLS is used to analyze simultaneously the interrelationships among all the constructs and is a fundamentally more sophisticated model. PLS is a popular model used to analyze small samples because it helps to predict responses. PLS focuses on maximizing the variance of the dependent variables explained by the independent variables. The model reflects the relationships between the latent variables and a measurement component, which shows how the latent variables and the indicators are related (Haenlein & Kaplan, 2004).
Data Collection
The population for this study consists of international students in Taiwan. Taiwan is a popular destination in Asia for students, because of its attractive environment, safety,
43
commodities, etc. Every year more than 3,000 international students arrive to the country seeking a degree, to study language or on exchange – and the trend is increasing. In this study, we aim to measure the level of commitment of the students with their universities and how the organizational culture influences the knowledge processes in students.
Knowledge management is an important issue that has attracted the attention of many researchers, but the topic is mostly explored for organizations and not enough attention is given to the topic in the education field. According to the Ministry of Education of Taiwan, there are more than 92,000 international students in the island by the year of 2014. Based on these statistics we select a sample of 200 international students in Taiwan who are currently pursuing a degree level education, enrolled in a language program or are exchange students.
A list of the universities in Taiwan is retrieved from the website of the Ministry of Education (MOE) and based on this the questionnaire is sent to the office of international affairs from different universities around the country. Convenience and snowball sampling are the preferred methods of data collection used. The respondents are also advised to forward the online questionnaire to their friends and classmates.
Also, online and paper questionnaires are used to facilitate the collection of data. A total of 100 surveys have been collected via online questionnaire (see Appendix B) and the remaining 110 have been collected with paper questionnaires.
This section contains general information about the respondents such as age, gender, nationality, and level of education. A total of 210 responses are used for this study, from
44
which 180 of the responses are valid. The largest group of respondent is between the ages of 20-26 corresponding to a total of 68% of the respondents.
Regarding the gender of the respondents, there is not a significant difference between the genders, being 44% female and the other 56% male. Regarding the level of education, the majority of the respondents are enrolled in a Bachelors program corresponding to 51% of the participants, followed by 38% of students who are enrolled in a Master’s degree program, 4% of students doing PhD studies, 4% from language programs, and the remaining 3% students from other programs, including Military education, and other technical courses. Regarding the origin of the participants, the largest group is from Central and South American corresponding to 52% of the respondents, followed by Asia with 22%, and Europe with 12%. Africa and North America are the smallest groups of respondents with 8% and 6% respectively. Regarding the student status, a classification provided by the MOE is used for the study. The biggest group of respondents is Degree seeking students with a total of 71% of the respondents, after this 16% correspond to International exchange students, 8% study Mandarin Chinese and a total of 2% are Overseas compatriot students and Mainland China Students studying for a degree,. The remaining are Overseas compatriot Youth technical training classes 1%
and Mainland China students studying for short term courses at 1%. Students are sponsored by ICDF at 23%, self-sponsored students 23%, 17% are sponsored by the Ministry of Foreign Affairs (MOFA), 18% corresponds to students sponsored by their universities and equally 18% of students are sponsored by other institutions (this includes other NGO’s, students sponsored by other universities and other governments).
45 Table 3.7.
Data of Variables by Entry and Values
Variables Entries Percentage
Age Under 20 11 6%
20-26 123 68%
27-33 35 19%
34-40 11 6%
41+ 0 0%
Gender Female 79 44%
Male 101 56%
Region Central and South America 94 52%
North America 11 6%
Europe 21 12%
Africa 15 8%
Asia 39 22%
Student Status Degree Seeking Student 128 71%
Overseas Compatriot Student (Including students from Hong
Kong and Macao) 4 2%
(Continued)
46 Table 3.7. (Continued)
Variables Entries Percentage
Student Status Main Land China Student (Studying
for a degree) 3 2%
International Exchange Student 29 16%
Short Term Courses Student 0 0%
Studying Mandarin Chinese 14 8%
Main Land China Student (Studying
short term courses) 1 1%
Overseas Compatriot Youth
Technical Training Classes 1 1%
Level of education Undergraduate 91 51%
Master 69 38%
PhD 8 4%
Language Program 7 4%
Other 5 3%
Sponsoring ICDF 42 23%
MOFA 31 17%
University Scholarship 33 18%
Self-Sponsored 41 23%
Other 33 18%
47
Data Coding System
Before the analysis, the data was coded to facilitate the processing of information.
The 61 items of the questionnaire were coded using a 5-point Likert scale. The coding system for all the variables and demographic questions are included in Table 3.8.
Table 3.8.
Coding System Used in PLS Data Analysis
Variables
48
Region 1 = Central and South America
2 = North America 3 = Europe
4 = Africa 5 = Asia
Student Status 1 = Degree Seeking Student
2 = Overseas Compatriot Student (Including Student from Hong Kong and Macao)
3 = Main Land China Student (Studying for a Degree) 4 = International Exchange Student
5 = Short Term Courses Student 6 = Studying Mandarin Chinese
7 = Main Land China Student (Studying Short Term Courses) 8 = Overseas Compatriot Youth Technical Training Classes
(Continued)
49 Table 3.8. (Continued)
Variables
Education 1 = Undergraduate
2 = Master 3 = PhD
4 = Language Program 5 = Other
Sponsoring 1 = ICDF
2 = MOFA
3 = University Scholarship 4 = Self Sponsored
5 = Other
Descriptive Statistics
Descriptive statistics were used to summarized and describe the data gathered in this study. This process allowed the researcher to organize the data in more comprehensible and meaningful way by calculating numerical indexes such as means and standard deviation; which allow us an easier interpretation of the data collected.
Inferential Statistics
The researcher used inferential statistics in order to examine the relationships and differences between variables. This process is important because it aids to testing of statistical hypotheses and significance testing. Since, the number of international students in Taiwan is very large, inferential statistics help us to analyze the information gathered and make inferences about the population. This study investigate the relationship between
50
university culture, knowledge sharing, knowledge creation and commitment of international students in Taiwan.
Partial Least Squares Structural Equation Modeling
The relationship between the constructs was analyzed using structural equation modeling SEM and the partial least square PLS approach. Structural equation modeling is a general statistical modeling technique which allow us to infer about cause-effect relationships between different theoretical constructs and variables. The purpose of this model is to study complex relationships between variables, where some variables can be hypothetical and unobserved. SEM is also considered a framework for quantitative analysis that uses a statistical technique instead of a statistical method (Hox & Bechger, 2001) which also permits an evaluation of networks of directs and indirect effects.
SEM has different types of approaches, the first one is covariance-based SEM (CB-SEM, which is regularly used with software packages like AMOS, ISREL and MPlus;
SEM has different types of approaches, the first one is covariance-based SEM (CB-SEM, which is regularly used with software packages like AMOS, ISREL and MPlus;