In this chapter, the research framework, hypothesis, research design, measurement, questionnaire design, and data analysis method were described. This research adopted quantitative research process to investigate the relationships among cultural intelligence (CQ), psychological capital (PsyCap) with four feathers, cross-cultural adjustment with three dimensions, and perceived supervisor support.
Research Framework
The independent variable was CQ. The dependent variable was cross-cultural adjustment which was consisted of three dimensions: general adjustment, interaction adjustment, and work adjustment. Psychological capital was the mediator to examine the relationship between CQ and cross-cultural adjustment. Perceived supervisor support served as the moderator to examine the relationship between PsyCap and cross-cultural adjustment with three dimensions. The research framework was shown below, see Figure 3.1.
28
Research Hypotheses
Based on previous literatures, the research purpose, and research questions, the hypotheses were described as followed.
Hypothesis 1. Cultural intelligence has a positive effect on cross-cultural adjustment.
H 1-1. Cultural intelligence has a positive effect on general adjustment.
H 1-2. Cultural intelligence has a positive effect on interaction adjustment.
H 1-3. Cultural intelligence has a positive effect on work adjustment are.
Hypothesis 2. Cultural intelligence has a positive effect on psychological capital.
Hypothesis 3. Psychological capital has a positive effect on cross-cultural adjustment.
H 3-1. Psychological capital has a positive effect on general adjustment.
H 3-2. Psychological capital has a positive effect on interaction adjustment.
H 3-3. Psychological capital has a positive effect on work adjustment.
Hypothesis 4. Psychological capital serves as a mediator between cultural intelligence and cross-cultural adjustment.
H 4-1. Psychological capital serves as a mediator between cultural intelligence and general adjustment.
H 4-2. Psychological capital serves as a mediator between cultural intelligence and interaction adjustment.
H 4-3. Psychological capital serves as a mediator between cultural intelligence and work adjustment.
Hypothesis 5. Perceived supervisor support positively moderates the relationship between psychological capital and cross-cultural adjustment.
29
H 5-1. Perceived supervisor support positively moderates the relationship between psychological capital and general adjustment.
H 5-2. Perceived supervisor support positively moderates the relationship between psychological capital and interaction adjustment.
H 5-3. Perceived supervisor support positively moderates the relationship between psychological capital and work adjustment.
Research Design Samples
Foreign labors face lots of adjustment challenges when moving from the home country to work in other countries since they were hired by a company located in a host country. In this study, foreign labors were defined as “legal blue collar labors being employed in Taiwan without the host nationality according to the Employment Services Act” (http://law.moj.gov.tw). The blue collar labors referred to those who engaged in physical works such as mining, construction, and manufacturing.
The total number of foreign labors in Taiwan was 454,171 according to the statistic data provided by Bureau of Employment and Vocational Training in May, 2013. Foreign labors who work in industry and social welfare sectors in Taiwan came from Indonesia, Philippian, Thailand, Vietnam, Malaysia, and Mongolia. As dividing labors according to industry, there were 9,460 (2.08%) labors in agricultural, forestry, fishery, and husbandry industry; 233,530 (51.41%) labors in manufacturing industry;
2,630 (0.01%) labors in construction industry; 208,551 (45.91%) labors in nursing and housing industry. Above all, Indonesia and Vietnam labors were mainly engage in nursing and housing industry while Thailand and Philippian labors were mainly engage in manufacturing and construction industry. In addition, all the Philippine labors engaged in the high technology industry were female.
30
Table 3.1.
Population Description of Foreign Labors in Taiwan
Nationality
Sectors Total Indonesia Philippian Thailand Vietnam Malaysia Mongolia
Total 454,171 199,971 87,808 62,591 103,797 4 0
Industry 245,620 36,347 63,664 64,865 82,621 4 0
Primary industrial 9,460 7,968 1,346 9 137 0 0
Manufacturing 233,530 28,246 63,484 59,700 82,096 0 0
Construction 2,630 133 35 2.074 388 0 0
Welfare 208,551 163,624 22,942 808 21,176 0 0
Note. Adapted from Bureau of Employment and Vocational Training. Retrieved May 28, 2013.
This study conducted purposive sampling concerning the language boundary and sample accessibility toward foreign labors. The target participants were Philippine labors in technology manufacturing industries.
The sample criteria included: fluent in English, high educational attainment, and working in group with supervisors. First of all, high technology manufacturing industry in Taiwan prefer to hire female Philippine labors because they are fluent in English and they got at least college degree. Secondly, in order to answer questions related to perception of supervisor, the target participants must have high connection with the supervisors in a workplace. Hence, this study chose the Philippine labors as our target samples.
Measurement
This study adopted matured scales to measure all variables and the precision of measurements was confirmed through the reliability analyses. The degree of precision indicated by Cronbach alpha coefficient (Cronbachα) is to show the “stability” and
31
“consistency” of instrument. The higher value of Cronbachα means a higher consistency of each dimension. The acceptable value is higher than 0.7 (Nunally &
Berstein, 1994).
Cultural Intelligence (CQ)
Cultural intelligence refers to an individual’s capability to adapt effectively to situations of cultural diversity (Earley, 2002). This research adopted the cultural intelligence scale (CQS) which developed by Ang et al. (2007). The scale contains 20 items in 4 dimensions: meta-cognitive CQ, cognitive CQ, motivational CQ, and behavioral CQ. Example items of each dimension were “I am aware of the cultural knowledge I use when interacting with people with different cultural backgrounds”
(meta-cognitive CQ); “I know the legal and economics of other cultures” (cognitive CQ); “I enjoy interacting with people from different cultures” (motivational CQ); “I change my verbal behavior (e.g. accent, tone) when a cross-cultural interaction requires it” (behavioral CQ). Responses were designed on a 7-point Likert scale, ranging from 1(strongly disagree) to 7 (strongly agree). The high score indicated that the person can deal with cultural diversity well in a new environment. The Cronbach’s alpha reliability of the CQ in this study was 0.92 and the four dimensions of CQ were ranged from 0.84 to 0.90 (See Table 4.2).
Cross-cultural Adjustment
Cross-cultural adjustment refers to a person’s feeling of comfort in the new role and the degree of mastering the role requirement (Black, 1988). The scale was developed by Black and Stephens (1989). It contains 14 items divided into three dimensions:
general adjustment with 7 items, interaction adjustment with 4 items, and work adjustment with 3 items. Example items of each dimension were “I am adjusted to the living condition in general” (general adjustment); “I am adjusted to socializing with Taiwanese” (interaction adjustment); “I am adjusted to my specific job responsibilities”
32
(work adjustment). Responses were designed on a 7-point Likert scale, ranging from 1(strongly disagree) to 7 (strongly agree). The high score indicated that the person can adjust to a new cultural environment. The Cronbach’s alpha reliability of the cross-cultural adjustment in this study was 0.94 and the three dimensions of cross-cultural adjustment were ranged from 0.89 to 0.91 (See Table 4.2).
Psychological Capital (PsyCap)
Psychological capital (Psycap) was a term derived from positive psychology that go beyond human and social capital. Individuals can gain competitive advantages through investing and developing of ‘who you are’. (Luthans, et al., 2004; Luthans &
Youssef, 2004). Psychological capital was adopted from Psychological capital (PsyCap) questionnaire (PCQ) self-rater version by Harms and Luthans (2013). There are 24 items (includes three reverse items) divided into four dimensions which were Hope, Resilience, Optimism, and Efficacy. Example item of hope was “If I should find myself in a jam at work, I could think of many ways to get out of it.”; example item of resilience was “I feel I can handle many things at a time at this job”, and the reverse item was “When I have a setback at work, I have trouble recovering from it, moving on.”; example item of optimism was “If something can go wrong for me work-wise, it will”, and it included two reverse items; example item of efficacy was “I feel confident analyzing a long-term problem to find a solution.” Each dimension was in the form of statement and responses were designed to answer by a 5-point Likert scale, ranging from 1(strongly disagree) to 5 (strongly agree). The high score of scale indicated that people have higher level of psychological capital (PsyCap). The Cronbach’s alpha reliability of the PsyCap in this study was 0.87 (See Table 4.2).
Perceived Supervisor Support (PSS)
Perceived supervisor support (PSS) refers to the degree a person’s feeling of how the supervisor values his/her contribution and well-being. This study adopted PSS scale
33
developed by Cheng, Jiang, and Riley (2003), which consists of four items with 7-point Likert scale ranged from 1(strongly disagree) to 7 (strongly agree). Example item included: “My supervisor offers help when I am in a personal crisis.” The high score indicated that the person’s perception of supervisor support is high. The Cronbach’s alpha reliability of the PsyCap in this study was 0.84 (See Table 4.2).
Control Variables
Prior research on cross-cultural adjustment had established that individual-level factors, such as age, educational attainment, work and overseas experience, duration in Taiwan, marriage, and language ability had influence on individual’s cross-cultural adjustment. Therefore, these control variables were included in the final section of the questionnaire to clarify the relationships among dependent, independent, moderator and mediator variables.
Age.
Researchers suggested that age is one demographic factor that would affect a person’s adaptation. The older workers are generally had better adaptation to the other culture and were more satisfied with their job than younger workers (Parker & McEvoy, 1993;
Sam, 1998). Therefore, the questionnaire included one blank for participants to fill in their age.
Educational attainment.
The highest education degree the participants had received was included in the personal information part in the questionnaire and it was measured by dummy variable. The coding was 1 (high school), 2 (vocational course), 3 (college).
Work and overseas experience.
Black and Gregersen (1991) and Black et al. (1991) pointed out that previous work-related experiences help individual to form accurate work expectations in mind;
on the other hand, previous non-work experiences facilitated the formation of accurate
34
non-work expectations. Besides, Bhaskar et al. (2005) stated that people who had longer experience in overseas assignment or related industry tended to leverage the experience better. The previous experience and the relocation skills acquired from the previous assignment helped people to go through trial-and-error processes and became more concentrated on the present work (Parker & McEvoy, 1993; Shaffer, Harrison & Gilley, 1999). Participants needed to answer whether they have ever worked or lived in other countries, and if yes, they filled out for how many years and months they’ve been there. The answer of “no” was coded as 0; the answer of “yes”
was coded as 1. The duration was analyzed by months.
Duration in Taiwan.
The more time individual stay in the host country and spend with the host-country nationals, the better adaptation level performed (Parker & McEvoy, 1993; Ramalu et al., 2010). The participants indicated how many times they have visited and for how long they have worked in Taiwan. The number of visits to Taiwan was coded as “1”
for once, “2” as “twice”, etc. The year they’ve worked in Taiwan was analyzed by months.
Language ability.
Previous literatures indicated that language ability will have great impact on a person’s adjustment (Bhaskar et al., 2005; Ramalu et al., 2010; Shaffer et al., 1999).
The participants rated their both their Chinese and English language ability according to the proficiency on listening, speaking, reading, and writing. The answers were coded as 1 “not at all”, 2 “poor”, 3 “average”, 4 “good”, 5 “excellent”.
Questionnaire Design
The questionnaire adopted maturing scales to ensure that all the desired traits were measured and enhance the content validity. Since the official language in Philippian is English, the questionnaire was designed in English and was reviewed by Philippian
35
translator to confirm whether all the Philippian labors can understand the statement of each item or not in order to reduce the language boundary and confirmed the face validity.
Aside from confirming the questionnaire validity, the questionnaire was also designed to avoid common method variance (CMV) (Podsakfoff, MacKenzie, Lee, &
Podsakoff, 2003) that may occur during the data collection process. The research purpose and variables were not revealed in the questionnaire; also, the questionnaire adopted both 5-point and 7-point Likert scale to measure different variables in order to reduce the CMV problem. In addition, reverse items were contained in the questionnaire to avoid random answers.
In addition to the prevention in advance, Harman’s one-factor analysis was conducted in this study as post-preventive to examine the CMV (Podsakoff & Organ, 1986). The general hypothesis of Harman’s one-factor analysis was to see if the majority of the variance can be explained by a single un-rotated factor.
Focus Group
Before conducting the pilot test, we interviewed one Philippine translator and five Philippine labors to know the possibilities to conduct this research. First of all, we interviewed the Philippine translator to have brief understanding of the characteristic of Philippine labors and to know if the circumstances of each question meet the Philippine labors’ current situation. Secondly, five Philippine labors were required to complete the whole questionnaire and gave us some feedback about the wording and questionnaire design. Also, we estimated the time to complete the questionnaire in order to facilitate the following questionnaire distributing process.
Pilot Test
The purpose of conducting the pilot test was to get understanding about the samples, to confirm the reliability of the measurement, and to examine the relationship between
36
all the variables. The data collected for the pilot test was from the central Taiwan through the help by the manpower agency and the sample size was 100. The period for data collection lasted for one week, and each of the participants rewarded a gift after handing in a complete questionnaire.
First of all, the descriptive statistics analysis was applied to understand the demographic information about the samples. Table 3.2 indicates that all of the participants were female. Most of them were single (66%) and aged form 26 to 35 years old (74%). Regarding to the international experiences, 92% of them did not work or live beyond their host nation, and most of them (47%) have been stayed in Taiwan for four to eight years. As for the educational background, 63% of the participants got the bachelor degree and 55% of the participants rated their English ability as good. The demographic information presented the characteristics of the target samples and indicated the criteria to fill in the questionnaire were met. For instance, their rating of the English listening-speaking and reading-writing ability showed that they had the ability to understand the questions and they have been stayed in Taiwan long enough to identify the difficulty they meet during the adaptation process.
Secondly, the reliability and Pearson correlation analysis were also conducted for the pilot test to confirm the reliability of the instruments and get brief understanding of the relationships among each variable. Table 3.3 presents the mean, standard deviation, correlations, and reliability for pilot test. The cronbach alpha above 0.7 is acceptable (reference). In this study the Cronbach alpha for each instrument were: 0.91 for CQ, 0.92 for cross-cultural adjustment, 0.86 for PsyCap, and 0.87 for PSS. The result of the correlation analysis indicated that all the main variables were significantly positive correlated (r=.56, p<.001 for the relationship between CQ and cross-cultural adjustment; r=.45, p<.001 for the relationship between
37
CQ and PsyCap; r=.42, p<.001 for the relationship between CQ and PSS; r=.30, p<.001 for the relationship between PsyCap and PSS).
Table 3.2.
Descriptive Statistics for Pilot Test (n=100)
Item Frequency Percentage Item Frequency Percentage
1. Age 6. Marriage
5. Work experience 9. Chinese ability
No 20 20% Not at all 20 20%
38
Table 3.3.
Mean, Standard Deviations, Correlations, and Reliability for Pilot Test (n=100)
Notes. Numbers in parentheses represent Cronbach’s alpha value. ***p < .001 **p < .01 *p < .05
Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12
1. Age (year) 32.20 4.32
2. Overseas experience (month) 2.50 13.77 -.10
3. Duration in Taiwan (month) 51.20 33.34 .62*** -.10
4. English ability 3.86 0.62 -.08 .02 -.08
5. Chinese ability 2.07 0.39 .05 .03 -.05 .01
6. CQ 4.87 0.71 .06 .01 .00 .12 -.01 (.91)
7. Cross-cultural adjustment 5.08 0.79 .14 .10 .13 .10 .05 .56*** (.92)
8. General adjustment 5.12 0.90 .10 .11 .10 .13 .04 .49*** .94*** (.86)
9. Interaction adjustment 4.84 0.93 .10 .05 .10 .05 .14 .50*** .85*** .67*** (.89)
10. Work adjustment 5.30 0.87 .19* .10 .14 .05 -.07 .44*** .74*** .57*** .53*** (.77)
11. Psychological capital 3.64 0.43 .09 .10 .02 .05 -.15 .45*** .40*** .34*** .33*** .40*** (.86)
12. Perceived supervisor support 4.26 1.34 .07 .15 -.08 .12 .08 .42*** .58*** .46*** .57*** .54*** .30*** (.87)
39
Data Collection
The channel to contact with foreign labors was through manpower agencies. In order to avoid the lawsuit in Taiwan and minimize the management cost, companies in the technology industry recruit Philippine labors through the manpower agency or outsourcing the related affairs to manpower agency. In Taiwan, most high technology manufacturing industries were located in north and central Taiwan. By contacting with managers and translators in the manpower agency, the number of labors can be estimated. Two manpower agencies in north and central Taiwan were willing to help to distribute 800 questionnaires to their foreign labors one by one and collect back around two months. Each respondent gained one small gift while completing the questionnaire. Through the data key in process, if over half of the questions were blank or if the responses were all the same, the questionnaires were regarded as invalid questionnaires. The number of valid questionnaires was 538, and return rate was 81.7%.
Data Analysis
This was a quantitative study being conducted in questionnaires. The statistic software of SPSS v18.0 and AMOS 18.0 were used in this study to analyze the data. The analysis method included:
Pearson correlation analysis
The Pearson correlation analysis was performed to understand the degree of linear relationship between two variables. Positive correlation indicated that both variables increase or decrease together, while negative correlation indicated that as one variable increases, the other one decreases, and vice versa. Pearson's r ranged from +1 to -1. A correlation of +1 means that there is a perfect positive linear relationship between variables; on the contrary, a correlation of -1 means that there is a perfect negative linear relationship between variables. A correlation of 0 means there is no linear
40
relationship between variable. A correlation of 0.4 to 0.8 means the variables are medium correlated.
Descriptive Statistics
This study adopted descriptive statistics to describe the characteristics of participants.
The frequency distribution and percentage number were used to identify the participants’ distribution pattern in age, educational attainment, work experience, overseas experience, duration in Taiwan, and language ability.
Confirmatory Factor Analysis (CFA)
The validity of the questionnaire was tested by confirmatory factor analysis (CFA) to confirm whether the data fit a hypothesized measurement model that based on theory or previous researches. This study adopted indicators had been proposed in previous researches to examine the construct validity of cultural intelligence, cross-cultural adjustment, psychological capital, and perceived supervisor support. The indexes were presented as follows:
Chi-square test (χ2) is to determine the overall model fit. According to Joreskog and Sorbom (1993), the value lower than 3 indicates a good model fit. In addition to the chi-square test, there are many other statistics for assessing the goodness of fit of the experimental value and the theoretical value.
Root Mean Square error of approximation (RMSEA) serves as a supplementary statistic to determine fit in large sample sizes. The value of RMSEA equals to 0 represents a perfect model fit. The value range from 0.05 to 0.1 represents a medium model fit; thus, the value higher than 0.1 represents a bad model fit (Brown & Cudeck, 1993; MacCallum, Browne & Sugawara, 1996). Goodness of fit index (GFI) higher than 0.8 describes the observed data well fit in the experimental distribution (Carmines & McIver, 1981). Normed fit index (NFI) represents a ratio of the difference in the value for the model and the independent model. Comparative fit
41
index (CFI) and non-normed fit index (NNFI) takes sample size into account based on NFI. CFI, NNFI, and NFI were higher than 0.9, which represent desirable indexes (Bentler & Bonett, 1980; Tucker & Lewis, 1973). To sum up, the above indexed range from 0 to 1; specifically, the value close to 1 represents a better model fit. Table 3.4 summarized the fit indexes.
Table 3.4.
Summary of Fit Indexes
Fit indexes Threshold References
χ²/df <3 Joreskog & Sorborm, 1993
RMSEA <.08 Brown & Cudeck, 1993; MacCallum, Brown & Sugawara, 1996 GFI >0.8 Carmines & McIver, 1981
NFI >0.9 Bentler & Bonett, 1980 Bentler & Bonett, 1980 Tucke & Lewis, 1973 CFI >0.9
NNFI >0.9
Structure Equation Model (SEM)
A complete structure equation model (SEM) includes two parts. The first one is the measurement model, and the second one is the structure model. First of all, the measurement model is to test the relationships among observed variables and latent variables, and confirm the validity of each model through confirmative factor analysis.
Secondly, structure model was adopted to examine the causal effects among latent variables through the path analyses. In this study, the latent independent variable was PsyCap; the latent dependent variable was cross-cultural adjustment; the mediating
Secondly, structure model was adopted to examine the causal effects among latent variables through the path analyses. In this study, the latent independent variable was PsyCap; the latent dependent variable was cross-cultural adjustment; the mediating