Chapter 4: Findings
4.2 Descriptive Analysis of Independent Variables
4.2(i) Self-Identification
Question 7 (Table 9) looks at independent variable 1: ethnic self-identification as either Chinese or Taiwanese. As with the previous two questions above, the wording of this question is also consistent with the norm employed by the ESC. This dimension is treated as dichotomous: (1) Taiwanese only (2) Taiwanese and Chinese. Moreover, this item can be helpful in establishing external validity by comparing with other surveys conducted by the ESC that also employ this same item.
Table 9: Ethnic Self-Identification
Frequency Percent Valid Percent Cumulative Percent
Valid
Taiwanese 296 59.1 59.2 59.2
both 194 38.7 38.8 98.0
Chinese 10 2.0 2.0 100.0
Total 500 99.8 100.0
Missing don’t know/
refuse to answer 1 .2
Total 501 100.0
In order to provide a comparison between respondents’ perceived ethnic identity and their parents’ origins, a, a scale to measure the degree of parental origin from mainland or from Taiwan, shown in Table 10, was recoded: the newly created measure ranges from -2 (both parents non mainlanders) to +2 (both parents mainlanders), and is centered on 0 (for one parent of each origin).
10 This could be explained by the high number of respondents answering “none” on the question of religion: the place of religion in Asian culture is not as defined or as definable as it is in the Western conceptualization of religiosity, and hence results such as this are not surprising.
48 Table 10: Parents' Ethnic Background
Frequency Percent Valid Percent Cumulative Percent
Valid
Neither parent from mainland 394 78.6 78.6 78.6
One parent from mainland 66 13.2 13.2 91.8
Both parents from mainland 41 8.2 8.2 100.0
Total 501 100.0 100.0
Then this “parental ethnic background” value was regressed on respondent’s perceived identity. As can be seen on Graph B, the more likely it is that the respondent’s parents are mainlanders (2 vs. 1, vs. 0), the more likely the respondent is to self-identify as Chinese, which was expected.
Graph B: Parental Ethnic Background vs. Perceived Identity
The correlation is significant and positive r = .35: (see Table 11). The regression line’s slope (.125) is r2 or .354 x .354 =.125. While this remains a weak correlation, it attests to the face validity of the respondents’ perception of identity and confirms it is influenced by one’s parents’ ethnic origins.
49
4.2(ii) Identification (Taiwanese vs. Chinese) and Demographics
The next analysis (Table 11), reviews the correlations obtained between demographic variables.
Table 11: Correlations Between Demographic Variables
Ethnic self-identifi
cation a
Parents' ethnic background
b
Age Educational level c
Gender
Ethnic self-identificati on a
Pearson Correlation 1 .354** -.360** -.069 -.038
Sig. (2-tailed) .000 .000 .123 .397
N 500 500 500 499 500
Parents' ethnic background b
Pearson Correlation .354** 1 -.182** -.010 .056
Sig. (2-tailed) .000 .000 .826 .210
N 500 501 501 499 501
Age
Pearson Correlation -.360** -.182** 1 .330** .142**
Sig. (2-tailed) .000 .000 .000 .001
N 500 501 501 499 501
Educational level c
Pearson Correlation -.069 -.010 .330** 1 .005
Sig. (2-tailed) .123 .826 .000 .906
N 499 499 499 499 499
Gender
Pearson Correlation -.038 .056 .142** .005 1
Sig. (2-tailed) .397 .210 .001 .906
N 500 501 501 499 501
50
Table 12: Correlations Between Self-Identification and Religion
Ethnic self-identifi
cation a
Non- religious
Buddhist Taoist Christian
Ethnic self- identification a
Pearson Correlation 1 -.078 -.123** .284** -.051
Sig. (2-tailed) .083 .006 .000 .257
N 500 500 500 500 500
Non-religious
Pearson Correlation -.078 1 -.669** -.136** -.025
Sig. (2-tailed) .083 .000 .002 .575
N 500 501 501 501 501
Buddhist
Pearson Correlation -.123** -.669** 1 -.581** -.107*
Sig. (2-tailed) .006 .000 .000 .017
N 500 501 501 501 501
Taoist
Pearson Correlation .284** -.136** -.581** 1 -.022
Sig. (2-tailed) .000 .002 .000 .627
N 500 501 501 501 501
Christian
Pearson Correlation -.051 -.025 -.107* -.022 1
Sig. (2-tailed) .257 .575 .017 .627
N 500 501 501 501 501
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
a. Taiwanese = 1, both = 2, Chinese = 3
NB: Religious identification coded Yes = 1, No = 0
The year of birth is correlated to identity, with the results indicating that the older (small year of birth) a person is, the more they self-identify as Chinese. It should be pointed out, however that age and ethnic background are correlated. The results would seem to support the notion that the self-identification as Chinese ethnicity—common among older people—tends to slowly melt away as respondents become younger, and tend to self-identify as Taiwanese.
Education and gender are not related to identity.
Religion also appears related to self-identification, as illustrated in Table 12, with Buddhists reporting feeling more Taiwanese (r = .12, p < .01) and Taoists more Chinese (r = .28, p
< .01).
51
Another way to look at the correlation between religious appurtenance and identification is to review the religious denominations of those identifying to Taiwanese or Chinese/Taiwanese:
the population ratio (about 60% feel Taiwanese, 40% feel both) is replicated for non-religious and Taoists, but not for Buddhists and Christians, as shown in Table 13. It might be concluded, although not a main part of this study, that religious appurtenance is related to Table 13: Correlations Between Self-Identification and Non-Religious And Taoists
Religious belief Frequency Percent Valid
52
self-identity, with Taoism more represented among those feeling Taiwan and Buddhism more represented among those identifying to both Taiwanese and Chinese.
4.2(iii) Position on Unification/Independence
Table 14 shows the results for Question 8 (Concerning the relationship between Taiwan and mainland China, which of the following six positions do you agree with: (1) immediate unification, (2) immediate independence, (3) maintain the status quo, decide either unification or independence in the future, (4) maintain the status quo, move toward unification in the future, (5) maintain the status quo, move toward independence in the future, (6) maintain the status quo forever), measuring views on the unification vs. independence issue. The scale is modified to represent a Likert scale, from unification to independence, thus allowing regression analysis:
Immediate unification: 1
Status quo toward unification: 2
Status quo and decide in the future: 3
Status quo toward independence: 4
Immediate independence: 5
Status quo forever: treated as missing data
Table 14: Position On Unification/Independence
Frequency Percent Valid Percent
Cumulative Percent
Valid
immediate unification 9 1.8 2.0 2.0
Status quo toward unification 54 10.8 12.3 14.3
status quo, decide in the future 178 35.5 40.5 54.8
status quo toward independence 168 33.5 38.2 93.0
immediate independence 31 6.2 7.0 100.0
Total 440 87.8 100.0
Missing
maintain the status quo forever 59 11.8
don’t know / no opinion 2 .4
Total 61 12.2
Total 501 100.0
A regression is performed on control variables as predictors of feelings toward unification vs.
independence. Finally, the other variables are included in the regression equation.
Interestingly, the regression on control variables illustrates that age and ethnic background significantly predict feelings toward unification or independence (Table 15) The younger the respondents, the more they want independence (β = .26, p < .001) and those with more parents who are mainlanders are for unification (β = -.28, p < .001).
53
Table 15: Predictors of Opinion on Unification/Independence a
Model Unstandardized
Coefficients
Standardized Coefficients
t Sig.
B Std. Error Beta
1
(Constant) -32.197 6.513 -4.944 .000
Gender .088 .077 .050 1.135 .257
Age .018 .003 .256 5.299 .000
Parents' ethnic background b -.199 .032 -.278 -6.215 .000 Educational level c
-.003 .025 -.005 -.113 .910
a. Dependent Variable: Unification =1 to independence = 5
b. Neither parent from mainland = -2, one parent from mainland = 0, both parents from mainland = 2 c. illiterate up to post grad
When self-identification as Taiwanese/Chinese is tested as predictor, (Table 16), identity appears a significant predictor (β = -.63, p < .001), with those identifying as Chinese leaning toward unification, and vice versa for those who identify as Taiwanese. Identity explains 39%
of the variance, Figure 1.
Table 16: Unification/Independence: Self-Identification as Predictor a
Model Unstandardized
Coefficients
Standardized Coefficients
t Sig.
B Std. Error Beta
1
(Constant) 4.793 .091 52.629 .000
Ethnic self-identification b
-1.010 .060 -.626 -16.822 .000
a. Dependent Variable: Unification =1 to independence = 5 b. Taiwanese = 1, both = 2, Chinese = 3
Figure 1: Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .626a .392 .391 .67168
a. Predictors: (Constant), Ethnic self-identification
Finally, when all predictors are combined in order to control for gender, age, and especially ethnic background, and it is checked again to see if the identity effect is maintained, which it is: Table 17 illustrates that Chinese identity (β = -.56, p < .001) as well as ethnic background
54
(β = -.12, p = .004), both predict a preference for unification even after controlling for gender, age and level of education. In this case, the model explains 41% of variance, which demonstrates control variables have very little effect, Figure 2.
Figure 2: Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .642a .412 .405 .66454
a. Predictors: (Constant), Educational level, Parents' ethnic background, Gender, Ethnic self-identification, Age
The inclusion or not of control variables does not change the result: Identity is the strongest significant predictor of unification / independence positioning.
4.2(iv) Political Tendencies
Table 18 shows the results obtained for Question 9 (There are many political parties in Taiwan. Which one do you lean toward?), measuring Independent Variable 3, which may be treated like three distinct answers based on the realities of the Taiwan political spectrum: (1) KMT + NP + PFP + MKT + Pan blue, and (2) DPP + TSU + Green + NPP + SDP + pan green.
The rest are treated as missing (The Trees Party is not affiliated to either position on the blue-green spectrum).
Table 17: Unification/Independence: All Predictors a
Model Unstandardized
Coefficients
Standardized Coefficients
t Sig.
B Std. Error Beta
1
(Constant) -4.058 5.882 -.690 .491
Ethnic self-identification b -.904 .068 -.560 -13.331 .000
Gender .090 .065 .052 1.381 .168
Age .004 .003 .060 1.394 .164
Parents' ethnic background c -.082 .028 -.115 -2.895 .004
Educational level d
.014 .021 .027 .675 .500
a. Dependent Variable: Unification =1 to independence = 5 b. Taiwanese = 1, both = 2, Chinese = 3
c. Neither parent from mainland = -2, one parent from mainland = 0, both parents from mainland = 2 d. illiterate up to post grad
55 Table 18: Political Party Preference
Frequency Percent Valid Percent Cumulative Percent
Valid
KMT 127 25.3 25.4 25.4
DPP 129 25.7 25.8 51.2
NP 9 1.8 1.8 53.0
PFP 14 2.8 2.8 55.8
TSU 3 .6 .6 56.4
Green Party 14 2.8 2.8 59.2
New Power Party 26 5.2 5.2 64.4
Minkuotang 3 .6 .6 65.0
Social Democratic Party 4 .8 .8 65.8
pan-blue 13 2.6 2.6 68.4
pan-green 19 3.8 3.8 72.2
Is not partial 79 15.8 15.8 88.0
has preference for one political
party, but refuse to answer 7 1.4 1.4 89.4
Support all 8 1.6 1.6 91.0
Do not support any 45 9.0 9.0 100.0
Total 500 99.8 100.0
Missing Other political parties 1 .2
Total 501 100.0
For the analysis, only two responses are kept: two groups, the blue and green are created (Table 19).
Table 19: Blue and Green Groups a
Frequency Percent Valid Percent Cumulative Percent
Valid
Green 195 38.9 54.0 54.0
Blue 166 33.1 46.0 100.0
Total 361 72.1 100.0
Missing System 140 27.9
Total 501 100.0
a. Green = 0, Blue = 1
Political orientation is hence dichotomous and can be analyzed as response variable, on a Logit regression. First, control variables only are used in the logit regression, as in the
56
previous question (Table 20; Q1 = gender, Q2 year of birth, Q3_Q4 is parents’ ethnicity and Q5 is education level). As seen in Table 20, Q2 (year of birth) and Q3-Q4 (ethnicity of parents) are significant predictors of the green/blue variable. For these predictors, a EXP(B) >
1, an increase in Q2 (being younger) and Q3-Q4 (more parents being Chinese), both increases the odds of voting Blue. Education and gender are not significant predictors of the green/blue orientation.
Table 20: Political Party Preference: Control Variables
B S.E. Wald df Sig. Exp(B)
Step 1a
Gender -.050 .246 .042 1 .838 .951
Age -.050 .011 21.253 1 .000 .951
Parents’ Ethnicity .798 .146 30.073 1 .000 2.221
Education .066 .076 .759 1 .384 1.068
Constant 3.404 .978 12.121 1 .000 30.088
a. Variable(s) entered on step 1: Gender, Age, Parents’ Ethnicity, Education.
Then identity (Q7) is added in the logit regression, together with the other control variables.
It appears, based on the results tabulated in Table 21, that respondents with Chinese parents and those with more of a Chinese identity (vs. A Taiwanese identity) tend to lean toward the pan-Blue. Age, education, and gender do not matter. The variable Age loses its significance because it has been “controlled” by identity. In other words, identity was hiding inside the age effect. By employing the Logit regression we can determine that that Q3_Q4 (parents background) and Q7 (Identity) are enough to predict correctly 84% of the choices between the green or blue political preference.
Table 21: Political Party Preference: Identity Logit Regression
B S.E. Wald df Sig. Exp(B)
Step 1a
Gender -.134 .302 .195 1 .659 .875
Age -.020 .013 2.243 1 .134 .980
Parents’ Ethnicity .623 .163 14.641 1 .000 1.864
Education -.011 .094 .013 1 .908 .989
Self-identification 2.911 .309 89.029 1 .000 18.381
Constant -1.756 1.289 1.856 1 .173 .173
a. Variable(s) entered on step 1: Gender, Age, Parents’ Ethnicity, Education, Self-identification.
57 4.3 Correlations Analysis of Dependent Variables
4.3(i) The China Threat
An analysis reveals that 80% of respondents to Question 10 (Countries face many threats, but there is usually one main threat. What is the main threat facing Taiwan?) indicates that
“China,” is seen as the major threat (Table 22). Because China is threat for 79% of responses, the variables is treated like Dichotomous and recoded: (Table 23; 1: “the threat Is China; 2.
the threat is another country than China.”) Then a logit regression can be run, with the considered predictors.
Table 22: Perceived Threat
Frequency Percent Valid Percent Cumulative Percent
Valid
China 394 78.6 81.4 81.4
Japan 11 2.2 2.3 83.7
United States 30 6.0 6.2 89.9
other 18 3.6 3.7 93.6
Conflict over islands in the South
China Sea/East China Sea 31 6.2 6.4 100.0
Total 484 96.6 100.0
Missing Do not know / no opinion 17 3.4
Total 501 100.0
Table 23: Main Threat: Dichotomous Coding
Frequency Percent Valid Percent Cumulative Percent
Valid
others 90 18.0 18.6 18.6
China 394 78.6 81.4 100.0
Total 484 96.6 100.0
Missing 17 3.4
Total 501 100.0
Then a Logit regression is run with all predictors, as per the results in Table 24: EXP(B) > 1 increases the odds respondent will consider China as threat. EXP(B) < 1 increases the odds respondents consider “Others” as a threat. Results indicate that those with Chinese parents (Q3_Q4), and especially those identifying as Chinese (Q7), tend to think that China is not the threat (EXP(B) < 1). Moreover, those with a high education (Q5) tend to see China as more of a threat (EXP(B) > 1). Age and gender do not matter. Therefore, Hypothesis one (H1) is confirmed, and threat perception is impacted significantly by self-identification. In other
58
words, respondents with a stronger self-identification as Chinese tend to see China as not the main threat facing Taiwan (Table 24).
Table 24: Main Threat: Logit Regression With All Predictors
B S.E. Wald df Sig. Exp(B)
Step 1a
Gender -.058 .274 .045 1 .832 .944
Age .010 .012 .677 1 .410 1.010
Parents’ Ethnicity -.222 .098 5.151 1 .023 .801
Education .233 .080 8.562 1 .003 1.263
Self-identification -1.912 .289 43.898 1 .000 .148
Constant 1.168 1.167 1.002 1 .317 3.216
a. Variable(s) entered on step 1: Gender, Age, Parents’ Ethnicity, Education, Self-identification.
4.3(ii) Cross-strait worries
Analyzing Question 11 (Over the past year or two, as the situation across the Taiwan Straits has changed, the PRC has continually increased its military capabilities against Taiwan.
Concerning these military threats from the PRC, some people are worried, and other people are not worried. Are you worried or not worried?), the results of which are tabulated in Table 25, the Likert scale can be can be treated as continuous so that a linear regression can be run with several predictors.
Table 25: Cross-Strait Worry
Frequency Percent Valid Percent
Cumulative Percent
Valid
very worried 37 7.4 7.5 7.5
a little worried 214 42.7 43.1 50.6
Do not worry 210 41.9 42.3 92.9
Not worried at all 35 7.0 7.1 100.0
Total 496 99.0 100.0
Missing don’t know / no opinion 5 1.0
Total 501 100.0
A regression analysis is run directly with all control variables as well as identity as predictors.
The significant predictors are seen in Table 26, showing that the more educated the respondent, the more worried they are about the military threats from the PRC (β = -.10, p
< .05). Other predictors are not significant.
59
Table 26: Cross-Strait Worry: Control Variables, Identity as Predictors a
Model Unstandardized
Coefficients
Standardized Coefficients
t Sig.
B Std. Error Beta
1
(Constant) 7.473 5.890 1.269 .205
Gender -.067 .067 -.045 -1.009 .313
Age -.002 .003 -.038 -.753 .452
Parents' ethnic background b .049 .029 .081 1.710 .088
Educational level c -.044 .021 -.098 -2.087 .037
Ethnic self-identification d .110 .069 .080 1.600 .110
a. Dependent Variable: Are you very worried = 1, or not worried = 4
b. Neither parent from mainland = -2, one parent from mainland = 0, both parents from mainland = 2 c. Illiterate up to post grad
d. Taiwanese = 1, both = 2, Chinese = 3
The regression is controlling for each variable in a single equation. In order to also review the one-by-one correlations of all those variables, correlations factors are tabulated (Table 27).
Line by line, the results can be analyzed as follows:
Identity as Chinese (first line) strongly correlates to less tendency toward independence (r = -.63, p < .001), perception that threat is not from China (r = -.43, p < .001), and less worries (r =.13, p = .003).
For those with “more” mainlander parents (second line), they lean toward unification (r = -.32, p < .001), perceive that threat is not from China (r = -.26, p < .001), and they are less worried (r =.12, p = .008).
China threat (fourth line) is strongly correlated to independence (r =.46, p
< .001) Finally, respondents who perceive the threat is from China are more worried (r =-.27, p < .01) and those leaning toward independence are more worried (r = -.26, p < .001)
Therefore, a significant relationship exists between worries about the military imbalance in the Taiwan Strait (Q11) and self-identification: Taiwanese who have an ethnic self-identification as Chinese are significantly less likely to worry about the China threat than their counterparts who identify as Taiwanese and those who lean toward independence are more worried (Table 27).
60
Table 27: Cross-Strait Worry: One-By-One Correlations
Ethnic
self-identification a Parents' ethnic
background b Position on
Unification/
Independence c Threat perception d Cross-strait worry e
Ethnic self- identification a
Pearson Correlation 1 .354** -.626** -.430** .134**
Sig. (2-tailed) .000 .000 .000 .003
N 500 500 440 484 496
Parents' ethnic background b
Pearson Correlation .354** 1 -.319** -.261** .119**
Sig. (2-tailed) .000 .000 .000 .008
N 500 501 440 484 496
Position on Unification/
Independence c
Pearson Correlation -.626** -.319** 1 .459** -.258**
Sig. (2-tailed) .000 .000 .000 .000
N 440 440 440 426 436
Threat perception d
Pearson Correlation -.430** -.261** .459** 1 -.273**
Sig. (2-tailed) .000 .000 .000 .000
N 484 484 426 484 481
Cross-strait worry e
Pearson Correlation .134** .119** -.258** -.273** 1
Sig. (2-tailed) .003 .008 .000 .000
N 496 496 436 481 496
**. Correlation is significant at the 0.01 level (2-tailed).
a. Taiwanese = 1, both = 2, Chinese = 3
b. Neither parent from mainland = -2, one parent from mainland = 0, both parents from mainland = 2 c. Unification = 1 to Independence = 5
d. Threat is China = 1, Threat is a country other than Taiwan = 0 e. Are you very worried = 1, to are you not worried = 4
4.3(iii) Problems Faced by the Military
Looking at Question 12 (Some people said that your military faces many problems today.
What is the main problem facing your military?) presents some difficulties, because items are neither dichotomous nor continuous. The results are shown in Table 28. However, we can look at each answer separately and see the percentage of people from Q7, Q8, Q9 who choose which answers. This is more or less an exploratory exercise, as well as being somewhat descriptive.
61 Table 28: Problems Facing Military
Frequency Percent Valid Percent Cumulative Percent
Valid
insufficient budget 6 1.2 1.2 1.2
difficulty purchasing weapons 32 6.4 6.6 7.8
inability to train with other
militaries 9 1.8 1.8 9.6
espionage 8 1.6 1.6 11.3
outdated attitudes within the
military 149 29.7 30.5 41.8
low morale 48 9.6 9.8 51.6
confusion in national identity 186 37.1 38.1 89.8
lack of modernized training 50 10.0 10.2 100.0
Total 488 97.4 100.0
Missing Do not know / no opinion 13 2.6
Total 501 100.0
The analysis strategy here is to attempt to rank these problems after separating respondents in groups and compare the ranking afterwards. However, there proved to be very little difference in the answers. Table 29 shows that the only exception was by gender, in that men tend to report confusion over national identity as the main problem facing the ROC military.
62 Table 29: Problems Facing Military: By Gender
Gender Frequency Percent Valid
Percent
Cumulative Percent
male
Valid
confusion in national identity 119 41.6 42.0 42.0
outdated attitudes within the military 87 30.4 30.7 72.8
low morale 29 10.1 10.2 83.0
lack of modernized training 23 8.0 8.1 91.2
difficulty purchasing weapons 16 5.6 5.7 96.8
espionage 4 1.4 1.4 98.2
inability to train with other militaries 3 1.0 1.1 99.3
insufficient budget 2 .7 .7 100.0
Total 283 99.0 100.0
Missing Do not know / no opinion 3 1.0
Total 286 100.0
female Valid
confusion in national identity 67 31.2 32.7 32.7
outdated attitudes within the military 62 28.8 30.2 62.9
lack of modernized training 27 12.6 13.2 76.1
low morale 19 8.8 9.3 85.4
difficulty purchasing weapons 16 7.4 7.8 93.2
inability to train with other militaries 6 2.8 2.9 96.1
insufficient budget 4 1.9 2.0 98.0
espionage 4 1.9 2.0 100.0
Total 205 95.3 100.0
Missing Do not know / no opinion 10 4.7
Total 215 100.0
63
Likewise, those with mainland ancestry point more to national identity confusion, as seen in Table 30.
Table 30: Problems Facing Military: By Parents’ Ethnic Background
Parents' ethnic background Frequency Percent Valid
Percent
confusion in national identity 138 35.0 36.1 36.1 outdated attitudes within the
military 126 32.0 33.0 69.1
low morale 35 8.9 9.2 78.3
lack of modernized training 34 8.6 8.9 87.2
difficulty purchasing weapons 29 7.4 7.6 94.8
inability to train with other
64
outdated attitudes within the military 105 35.5 36.5 36.5
confusion in national identity 100 33.8 34.7 71.2
lack of modernized training 27 9.1 9.4 80.6
low morale 23 7.8 8.0 88.5
difficulty purchasing weapons 16 5.4 5.6 94.1
espionage 7 2.4 2.4 96.5
outdated attitudes within the military 43 22.2 22.6 65.8
low morale 23 11.9 12.1 77.9
lack of modernized training 22 11.3 11.6 89.5
difficulty purchasing weapons 15 7.7 7.9 97.4
inability to train with other militaries 3 1.5 1.6 98.9
difficulty purchasing weapons 1 10.0 10.0 70.0
espionage 1 10.0 10.0 80.0
outdated attitudes within the military 1 10.0 10.0 90.0
lack of modernized training 1 10.0 10.0 100.0
There appears to be an across-the-board consensus that the two main problems facing today’s ROC military is confusion about national identity and outdated attitudes in the military (Table 31).
65
Table 32: Problems Facing Military: By Political Preference
Political preference (Green = 0, Blue = 1) Frequency Percent Valid Percent
Cumulative Percent
Green
outdated attitudes within the military 78 40.0 40.8 40.8
confusion in national identity 67 34.4 35.1 75.9
lack of modernized training 14 7.2 7.3 83.2
low morale 11 5.6 5.8 89.0
difficulty purchasing weapons 10 5.1 5.2 94.2
inability to train with other militaries 4 2.1 2.1 96.3
espionage 4 2.1 2.1 98.4
insufficient budget 3 1.5 1.6 100.0
Total 191 97.9 100.0
Missing Do not know / no opinion 4 2.1
Total 195 100.0
Blue
confusion in national identity 76 45.8 46.3 46.3
outdated attitudes within the military 28 16.9 17.1 63.4
low morale 24 14.5 14.6 78.0
lack of modernized training 17 10.2 10.4 88.4
difficulty purchasing weapons 16 9.6 9.8 98.2
inability to train with other militaries 2 1.2 1.2 99.4
espionage 1 .6 .6 100.0
Total 164 98.8 100.0
Missing Do not know / no opinion 2 1.2
Total 166 100.0
Table 32 indicates that persons who identify more on the Blue side of the political spectrum appear slightly more concerned about “confusion in identity.”
Attitudes towards the main problem facing the military, measured by Question 12, does not show a significant relationship with self-identification. The conclusion that can be drawn from the responses to Question 12 is that men, Blue voters, and respondents with both identities (Chinese and Taiwanese) are more worried about confusion in national identity than are women, green voters, and those who identify as Taiwanese. Especially, Blue and Green voters identify a different problem facing militaries: Blue voters see confusion in national identity as the main issue whereas green voters see outdated attitudes as the main issue.
66 4.3(iv) Major Mission Definition
Table 33 shows the results obtained for Question 13 (Militaries have many jobs to do, but usually there is one main job. What should be the main job of your military?). For an analysis, it can be treated in the same way as Question 12; by forming several groups and comparing their opinions.
Table 33: Main Mission Definition
Frequency Percent Valid Percent
Cumulative Percent
Valid
defend against attack from China 253 50.5 52.8 52.8
defend an attack from another country 156 31.1 32.6 85.4
Search and Rescue 54 10.8 11.3 96.7
humanitarian relief 11 2.2 2.3 99.0
assist allies 5 1.0 1.0 100.0
Total 479 95.6 100.0
Missing Do not know / no opinion 22 4.4
Total 501 100.0
67 Table 34: Main Mission Definition: By Gender
Gender Frequency Percent Valid
Percent
Cumulative Percent
male
Valid
defend against attack from China 161 56.3 57.7 57.7
defend an attack from another country 88 30.8 31.5 89.2
Search and Rescue 25 8.7 9.0 98.2
humanitarian relief 3 1.0 1.1 99.3
assist allies 2 .7 .7 100.0
Total 279 97.6 100.0
Missing Do not know / no opinion 7 2.4
Missing Do not know / no opinion 7 2.4