This chapter presents the analyzed result of the data collected. First of all, the chapter examined the demographics of the study in order to have a comprehensive view of this study.
Second, the result showed the confirmatory factor analysis, reliability, common method variance to understand the study’s validity and reliability. Continually are the highlights of the study, which elaborated the result of this study’s hypothesis, including correlation, linear regression, hierarchical regression, and t-test. Finally, the chapter introduced the result of this paper and the discussions.
Demographic Statistic
For this study, there was a total of 256 questionnaires collected. The data was collected via an online survey. Out of those 256 responds, 220 had oversea studied experience. Which the data was analyzed for hypothesis 1 to 3. However, hypothesis 4 was examining the comparison between A group and B group (individuals with under 8 months of study abroad experience, including no experience), therefore it was analyzed from the total 256 responds. In the demographic statistics, the sample characteristics of age, continent, whether they had studied abroad, and whether they had studied abroad over 8 months were contained in order to have basic background information of this study. Table 4.1 presents a summary of descriptive statistics on sample demographics.
Among the 256 responders, the majority of the participants are 44 (65.2%) which are in the age of 21-25, the following are in the age under 20 (44/17.2%) and in the age of 26-30 (40/15.6%). However, there are 4 participants in the age of 31-35 (1.6%) and also 1 participant in the age of 46-50 (0.4%). As for the nationality of the participants, there are 124 (48.4%) from Asia, 54 (21.1%) from Europe, 42 (16.4%) from North America, 29 (11.3%) from South America, 4 (1.6%) from Africa, and 3 (1.2%) from Australia/ Oceania.
According to the hypothesis of this study, hypothesis 1 to 3 are testing the sample of ones who studied abroad. In the demographics summary showed that there are 220 (85.9%) participants studied abroad and 36 (14.1%) without studied abroad experience. As for hypothesis 4, the study is comparing the sample of the group with over 8 months of study abroad experience and with less than 8 months of studied abroad experience (including no experience). The demographics reported there are 210 (82%) had over 8 months of study abroad experience and 46 (18%) without the experience.
26
Table 4.1.
Descriptive Statistics on Sample Demographics (N = 256)
Item Categories Frequency Percentage (%)
Age Under 20 44 17.2
21-25 167 65.2
26-30 40 15.6
31-35 4 1.6
46-50 1 0.4
Continent Asia 124 48.4
Africa 4 1.6
Australia/ Oceania 3 1.2
Europe 54 21.1
North America 42 16.4
South America 29 11.3
Study abroad Yes 220 85.9
No 36 14.1
Study abroad over 8 months Yes 210 82.0
No 46 18.0
Common Method Variance (CMV)
When conducting quantitative research, data are often collected by self-report instruments.
Therefore, Common Method Variance (CMV) was often mentioned by researchers. CMV was defined as “the overlap in variance between two variables attributable to the type of measurement instrument used rather than due to a relationship between the underlying constructs”(Avolio, Yammarino, & Bass, 1991, p. 572). In order to check CMV, Harman one-factor analysis was frequently utilized. The criteria of the Harman one-one-factor analysis are under
27
50%. The Harman one-factor result of this study is 34.89%, which showed that there are no single factor accounts for the majority of the variance in the variables from this study.
Reliability Analysis
Mallery and George (2003) had mentioned that the number of Cronbach alpha is in charge of explaining the internal consistency of variables. The result ranging from .6 to .7 is considered to be acceptable while from .7 to .8 is good.
Table 4.2 presents the report of reliability analysis. According to the result, all the Cronbach’ alpha of the variables is higher than .7. The result of psychological flexibility’s Cronbach alpha is .871, self-efficacy’s Cronbach alpha is .865, and psychological well-being’s Cronbach alpha is .755. As a result, the measurement of this study had good reliability.
Table 4.2.
Cronbach’s Alpha Analysis (N = 220)
Variables Number of items Cronbach's Alpha
Psychological Flexibility 7 .871
Self-Efficacy 8 .865
Psychological Well-Being 6 .755
Confirmatory Factor Analysis
The study conducted the Confirmatory Factor Analysis (CFA) for checking the validity of the measurement model for each variable in this research. There are three criteria for checking after we finished the CFA. First, is about factor loading. When the factor loading is lower than .5, the item requires to be re-examined (Tinsley & Tinsley, 1987). Based on this situation, the researcher will think over other results and decide whether to delete the item or not.
Secondly, the model fitness indices are reviewed to ensure whether the variables had achieved the required levels. The criteria for the good model-fit indices are present in Table 4.2. Third, for construct validity, Fornell and Larcker (1981) suggested that Composite Reliability (CR) needs to larger than .6 and Average Variance Extracted (AVE) larger than .5.
28
Table 4.3.
Summary of Good-Fit Criteria
Fit index Acceptable levels
Chi-Square X2 No significant; >. 05
CFI > .90 Reasonable fit
> .95 Good model fit
NFI and GFI > .90 Reasonable fit
> .95 Good model fit
RMSEA < .05 Good model fit
> .05 but <. 08 Reasonable fit
Note. Adapted from “Significance tests and goodness of fit in the analysis of covariance structures.” By P. M. Bentler & D. G. Bonett. 1980, Psychological Bulletin, 88(3), 588-606.
Psychological Flexibility (PF)
Figure 4.1 presents the model of psychological flexibility and Table 4.3 shows the result of the CFA report. Without deleting any item, the factor loading of all items are larger than the criteria .5. When checking the model fit indices, the result shows as below: X2 = 53.85, df = 14, RMSEA = .11, CFI = .94, NFI = .92, GFI = .93. Although RMSEA didn’t achieve the criteria, however, other indices had achieved it. As for the result of CR and AVE, CR = .87 which is above the criteria .6. AVE = .49, the criteria of AVE is .5, therefore the result of .49 is still acceptable. To sum up the CFA result of psychological flexibility, except RMSEA and AVE’s result is under the criteria, other indices and result had achieved the standard. Based on the result, the researcher doesn’t delete any item in psychological flexibility.
29
Figure 4.1. Psychological flexibility model
Table 4.4.
CFA Results for Psychological Flexibility
Items Factor Loading CR AVE
PF1 .68***
.87 .49
PF2 .66***
PF3 .71***
PF4 .75***
PF5 .71***
PF6 .70***
PF7 .67***
X2 = 53.85, df = 14, RMSEA = .11, CFI = .94, NFI = .92, GFI = .93 Note. p*** < .001.
30
Self-Efficacy (SE)
Figure 4.2 shows the model of self-efficacy and Table 4.4 examines the result of the CFA report. The result of the factor loading of each item is all above .5, which are all achieved the criteria. Continually, the model fit indices report is presented as below: X2 = 60.37, df = 20, RMSEA = .10, CFI = .94, NFI = .91, GFI = .94. Same as psychological flexibility, the result of RMSEA is larger than .08. However, all the other indices has reached the criteria. The last report is CR and AVE, CR = .87 and AVE = .45. In self-efficacy CFA report, CR is larger than the criteria .6, but AVE is smaller than the criteria .5. Consider all the factor loading is all above the criteria and the result of RMSEA and AVE is actually closer to the goal. Therefore, the researcher doesn’t delete any items of self-efficacy.
Figure 4.2. Self-efficacy model
31
Figure 4.3 reports the model of psychological well-being, and Table 4.5 presents the CFA result of psychological well-being. According to Figure 4.3, there are two item’s factor loading lower than .5 which is PWB4 = .45 and PWB6 = .47. As for the other items result of psychological well-being are all above .5. The report of model fit indices are examined under:
X2 = 6.924, df = 9, RMSEA = 0, CFI = 1, NFI = .98, GFI = .99. And the report of CR = .76 and AVE = .36. As mentioned by the report, CFI = 1, the number shows that this model’s chi-square is bigger than the degree of freedom (Cheung & Rensvold, 2002). Moreover, according to the result of RMSEA = 0. Fan, Thompson, & Wang (1999) reported that when the sample size was too small, the result of RMSEA might very close to 0.
According to the report above, the factor loading of PWB4 and PWB6 is lower than .5.
And the AVE for the variable doesn’t achieve the criteria either. Therefore, the researcher tries to delete PWB4 and PWB6, to test the variable’s AVE and Cronbach alpha. However the result of AVE and Cronbach alpha doesn’t have a big difference than before deleting the items.
Moreover, the researcher does the exploratory factor analysis (EFA) in order to check whether all the items consist of the same factor. And the result of EFA shows that all six items consist of one component. Consider all the deleting result and the existing result of psychological well-being’s Cronbach alpha, the study decides not to delete any item.
32
Figure 4.3. Psychological well-being model
Table 4.6.
CFA Results for Psychological Well-Being
Items Factor Loading CR AVE
PWB1 .54***
.76 .36
PWB2 .78***
PWB3 .67***
PWB4 .45***
PWB5 .61***
PWB6 .47***
X2 = 6.924, df = 9, RMSEA = 0, CFI = 1, NFI = .98, GFI = .99 Note. p*** < .001.
33
In order to have a comprehensive result of CFA, Table 4.7 showed the detail of fit indices from different factors which refers to the measurement model. The researcher ran three-factor to one-factor. The result of the three-factor model presented an expected range of a strong measurement model (X2 = 364.196, df = 186, RMSEA = .066, CFI = .903, NFI = .822, GFI
= .865). And the result of the two-factor model (X2 = 729.655, df = 188, RMSEA = .115, CFI
= .706, NFI = .644, GFI = .668) was poorer than the three factor model. Moreover, the result of the one-factor model (X2 = 766.717, df = 189, RMSEA = .118, CFI = .686, NFI = .626, GFI
= .663) was also poorer than the two-factor model. According to Schermelleh-Engel, Moosbrugger, and Muller (2003), when ∆X2 is significant, the result presents the original model is better than the reformed model. To sum up, the proposed model present a strong model for this study to measure.
Table 4.7.
Comparison of Fit Indices in This Study
Models Fit Indices
X2 ∆X2 df RMSEA CFI NFI GFI
Three-Factor 364.196 -- 186 .066 .903 .822 .865
Two-Factor 729.655 365.459 188 .115 .706 .644 .668
One-Factor 766.717 37.062 189 .118 .686 .626 .663
Note. Three-factor model: Proposed measurement model; Two-factor model: Combines psychological flexibility and self-efficacy; One-factor model: Combines psychological flexibility, self-efficacy, and psychological well-being.
34
Correlation Analysis
The Pearson correlation analysis is presenting to examine the correlation relationship between the independent variable, dependent variable, and the two moderators. Table 4.7 shows the result of the correlation analysis, moreover, the table also presents variables mean and standard deviation. According to the result of correlation, psychological flexibility is highly correlate to self-efficacy (r = .412, p < .01) and psychological well-being (r = .473, p < .01).
However, there is no correlation effect between psychological flexibility and study abroad length (r = .056). As a result of psychological well-being, including there is a high correlation effect between psychological well-being and psychological flexibility. There are also correlation effect between psychological well-being and self-efficacy (r = .637, p < .01). To sum up the correlation result, except study abroad length doesn’t have any correlation effect with other variables, all other variables have highly correlated with each other.
Table 4.8.
Pearson Correlation Analysis (N = 220)
Mean S.D. 1 2 3 4
1 Psychological Flexibility 5.32 1.02
1
2 Self-Efficacy 4.03 .52 .412** 1
3 Study Abroad Length 16.7 13.93 .056 .071 1
4 Psychological Well-Being 5.94 .73 .473** .637** .088 1 Note. p** < .01.
35
Regression Analysis Simple Linear Regression Analysis
In order to examine the hypothesis, the researcher does a simple linear regression analysis.
Table 4.9 reports the result of a simple linear regression analysis. The result shows out that psychological flexibility is a statistically positive significant predictor of psychological well-being (β = .471, p < .001, F = 31.373, p < .001). Furthermore, psychological flexibility explains 22% of the variance in psychological well-being. Therefore, hypothesis 1 was accepted.
H1: Psychological flexibility is positively related to psychological well-being.
Table 4.9.
The Linear Regression Result for the Relationship between Psychological Flexibility and Psychological Well-Being
Variables Model 1
β
Model 2 β
Working Experience (Control) .078 .020
Psychological Flexibility (IV) .471***
Note. Dependent variable: Psychological Well-Being, ***p < .001.
Hierarchical Regression Analysis
The Hierarchical regression analysis was conducted to test the moderation effect of the variables. Hypothesis 2 and 3 are both moderation effect, which the result presents in Table 4.10 and Table 4.11.
Table 4.10 reports the result of the hierarchical regression analysis in the relationship between psychological flexibility, self-efficacy, and psychological well-being. According to Table 4.10, model 2 reports the result of the moderation effect (β = -2.18, p < .001, ∆F = 19.312, p < .001). As mentioned above, the result presents a significant moderating effect on the independent variable and the dependent variable. Moreover, together with the independent variable and the moderator explain 50% of the variance in psychological well-being. Therefore, there is a significant effect on self-efficacy to the relationship of psychological flexibility and
36
psychological well-being, However, the number of β = -2.18, which presents that the moderation effect is on a negative way. Figure 4.4 presents the moderation effect of self-efficacy. According to Figure 4.4, when someone’s self-efficacy is higher, the effect of his or her psychological flexibility to psychological well-being is smaller. To sum up the result, hypothesis 2 was rejected.
H2: Self-efficacy has a positive moderating effect on the relationship between psychological flexibility and psychological well-being.
Table 4.10.
Summary of Hierarchical Regression for Moderating Effect of Self-Efficacy on the Relationship between Psychological Flexibility and Psychological Well-Being
Variables Model 1
β
Model 2 β
Model 3 β
Working Experience (Control) .078 .017 .041
Psychological Flexibility (IV) .252*** .226***
Self-Efficacy (Moderation) .533*** .497***
Psychological Flexibility × Self-Efficacy -.218***
R² .006 .460 .504
Adjust R² .001 .452 .495
∆ R² .454 .045
F 1.328 61.263*** 54.671***
∆F 90.684*** 19.312***
Note. Dependent variable: Psychological Well-Being, ***p < .001.
37
Figure 4.4. The moderation effect of self-efficacy
As a result of the moderating effect of study abroad length to the relationship between psychological flexibility and psychological well-being, Table 4.11 presents as below.
According to Table 4.11, β = -.025 (no significant), ∆F = .175 (no significant). Therefore, the result examined there is no significant moderating effect of study abroad length to the relationship between psychological flexibility and psychological well-being. Altogether, hypothesis 3 was rejected.
H3: Study abroad length has a positive moderating effect on the relationship between psychological flexibility and psychological well-being.
4 4.5 5 5.5 6 6.5 7
Low Psychological Flexibility High Psychological Flexibility
Psychological Well-Being
Low Self-Efficacy
High Self-Efficacy
38
Table 4.11.
Summary of Hierarchical Regression for Moderating Effect of Study Abroad Experience on the Relationship between Psychological Flexibility and Psychological Well-Being
Variables Model 1
β
Model 2 β
Model 3 β
Working Experience (Control) .078 .013 .016
Psychological Flexibility (IV) .468*** .470***
Study Abroad Experience (Moderation) .060 .061
Psychological Flexibility × Study Abroad
Experience -.025
R² .006 .228 .228
Adjust R² .001 .217 .214
∆ R² .222 .001
F 1.328 21.241*** 15.913***
∆F 31.014*** .175
Note. Dependent variable: Psychological Well-Being, ***p < .001.
39
Independent T-test Analysis
Independent T-test was conducted to check hypothesis 4. Hypothesis 4 is to compare A group and B group’s means refers to the relationship with psychological well-being. Table 4.11 shows the result of independent T-test.
According to Table 4.12, there are 210 responders in A group, and 46 responders in B group. First, the t-value = -2.95 (p < .01) presents a significant difference between the two groups. Furthermore, when comparing the mean of these two groups, the group with over 8 months of study abroad experience (35.82) is larger than the other group (33.72). As a result, hypothesis 4 is accepted.
H4: The group with study abroad experience over 8 months has a higher level of psychological well-being than the group with study abroad experience under 8 months.
Table 4.12.
Independent T-test
Study abroad over 8 months (N = 210)
Study abroad under 8 months (N = 46)
Variable Mean Std. Dev. Mean Std. Dev. t-value
Psychological
Well-Being 35.82 4.28 33.72 4.84 -2.95**
Note. **p < .01.
40
Summary and Discussions of the Analysis Result
Based on the proposed hypothesis, the researcher runs a simple linear regression analysis, hierarchical regression analysis, and T-test. The overall result is presenting as below Table 4.13. Hypothesis 1 and 4 are accepted, however, hypothesis 2 and 3 are rejected.
Table 4.13.
Result of the Study
Hypothesis Accepted Rejected
H1 Psychological flexibility is positively related to psychological well-being. V Supported H2 Self-efficacy has a positive moderating effect on the relationship between
psychological flexibility and psychological well-being.
V Not Supported
H3 Study abroad experience has a positive moderating effect on the relationship between psychological flexibility and psychological well-being.
V Not Supported
H4 The group with study abroad experience over 8 months has a higher level of psychological well-being than the group with study abroad experience under 8 months.
V
Supported
41
Psychological Flexibility and Psychological Well-Being
This study is conducted to examine the relationship among psychological flexibility, self-efficacy, study abroad length, and psychological well-being. After doing the correlation analysis and simple linear regression analysis. The result shows that psychological flexibility has a significant positive effect on psychological well-being. In other words, when ones can better control their thoughts and feelings through different situation and values, they are possible to be happier and be more satisfy to their lives. Also, this result was supported by Biglan et al. (2012), they claimed that individuals are more willing to enhance their well-being when their psychological flexibility is better. Moreover, the study also consists of several researchers which claimed that both psychological flexibility and psychological well-being plays a big role in human health (Kashdan & Rottenberg, 2010; Ryff, 1989).
Moderation Effect of Self-Efficacy
Secondly, hypothesis 2 examines the positive moderation effect of self-efficacy to the relationship between psychological flexibility and psychological well-being. In addition, the result of this study has rejected the hypothesis by doing the hierarchical regression analysis. As a result of this study, although there was a moderation effect of self-efficacy, however, the effect is in a negative way. That is to say, when ones have stronger self-efficacy, it can be a negative mechanism to the effect of psychological flexibility to psychological well-being.
In literature, there are also several studies research about the relationship among self-efficacy, psychological flexibility, and psychological well-being, however, the result of this study showed the opposite way to the literature (Soysa & Wilcomb, 2015; Wei et al., 2015).
The possible reason for this result may because the effect power of self-efficacy is stronger than psychological flexibility. Therefore, there is the moderation effect of self-efficacy but in a negative way.
Moderation Effect of Study Abroad Experience
The third hypothesis is the moderation effect of study abroad length to the relationship between psychological well-being. Although there are some researchers claiming that study abroad experience had a positive effect on one’s well-being (Kauffmann & Kuh, 1984).
However, the result of this result is no moderation effect from study abroad length. There are some possible reasons: 1. The sample size may be too small. 2. This study only uses study abroad length to predict the experience of study abroad. The predictor might be too weak to measure the effect. Therefore, it may affect the result of the moderation effect of study abroad length to the relationship between psychological flexibility and psychological well-being.
42
Longer Study Abroad Length Effects More
The last hypothesis was comparing the group with over 8 months of study abroad experience and the group with less than 8 months of study abroad experience (including no experience). This study compares the two groups’ psychological well-being. T-test supported the hypothesis. In other words, if ones have longer study abroad experience and the experience has over 8 months, the person may have higher psychological well-being. This result is also explained by some researchers, for example, Dwyer and Peters (2004) reported that the longer students went to study abroad, the bigger effect they will gain. Kitsantas (2004) also examined that study abroad experience is one of well-being’s predictor. Furthermore, as Black and Mendenhall (1991) mentioned, they present the U-curve which examined the pattern of through the study abroad length longer, the bigger effect the person had. The U-curve showed that there was a significant difference after an individual went study abroad over 8 months which also explains hypothesis 4.
43