10
K. Chinen et al.
1.4.1.3 Ethnocentrism and Control Variables
The high level of ethnocentrism (ET > 3) is 16.9% of the respondents (N = 178).
The proportion of females is 41.7% (N = 175). Traditional college-age respondents represent 75.1% of 177 respondents and White respondents are 38.8% of 165
respondents. Respondents who supported Republicans are 43.5% (N = 170).
1.4.1.4 Hypotheses H1a, b, c
The ordered logit model requires a test of parallel regression assumption that all slope
coefficients are statistically identical. We verify them by Wolfe-Gould test under the
null hypothesis is γ 1 = γ 2 . If the test does not reject the null hypothesis, we accept
the ordered logit model under γ 1 = γ 2 (the constrained model). The significance
probabilities of all countries are more than 0.15, therefore we accept the constrained
model.
The results in Table 1.2 indicate all countries’ PQ c s have significant influences on
PI c s. The factor change in the odds for a unit increase in PQ c is derived by exp
ˆ
γ
. For
example, ˆ
γ China is 1.776 and exp
ˆ
γ China
= 5.905. The odds of the outcome ‘agree
(neutral)’ toward CBEVs made in China are 5.905 times larger than the outcome
‘neutral (disagree)’. These results support the hypothesis H1a.
Most of the estimates of PB and PR are insignificant. These results indicate that
perceived benefits of buying EVs and perceived risks toward them do not have significant influences on his/her purchase intention toward each country’s CBEVs except
the effect of perceived benefit on Chinese CBEVs. The hypotheses H1b and H1c
cannot be verified.
Table 1.2 Results by ordered logit model
CH
US
SW
MX
KR
JP
PB
0.45**
0.22
0.04
– 0.07
0.20
0.40
PR
−0.26
−0.09
−0.07
0.19
−0.18
−0.03
PQ c
1.78***
1.79***
1.97***
2.60***
2.02***
2.09***
TSt 1
5.01***
4.49***
5.29***
7.61***
4.88**
6.06***
TSt 2
7.03***
7.05***
8.43***
10.39***
7.86***
8.89***
N
152
151
150
150
150
150
R 2
0.20
0.21
0.22
0.30
0.25
0.30
Note The response variable is the purchase intention toward country c’s CBEVs (PI c ). “c” means
China, the US, Sweden, Mexico, South Korea, and Japan. Asterisks (**, ***) indicate that the
estimate is significant at 5% and 1% levels, respectively. TSt 1 , t 2 correspond to Threshold t 1 , t 2 . R 2
is Pseudo R 2 . Control variables are included in each model
K. Chinen et al.
1.4.1.3 Ethnocentrism and Control Variables
The high level of ethnocentrism (ET > 3) is 16.9% of the respondents (N = 178).
The proportion of females is 41.7% (N = 175). Traditional college-age respondents represent 75.1% of 177 respondents and White respondents are 38.8% of 165
respondents. Respondents who supported Republicans are 43.5% (N = 170).
1.4.1.4 Hypotheses H1a, b, c
The ordered logit model requires a test of parallel regression assumption that all slope
coefficients are statistically identical. We verify them by Wolfe-Gould test under the
null hypothesis is γ 1 = γ 2 . If the test does not reject the null hypothesis, we accept
the ordered logit model under γ 1 = γ 2 (the constrained model). The significance
probabilities of all countries are more than 0.15, therefore we accept the constrained
model.
The results in Table 1.2 indicate all countries’ PQ c s have significant influences on
PI c s. The factor change in the odds for a unit increase in PQ c is derived by exp
ˆ
γ
. For
example, ˆ
γ China is 1.776 and exp
ˆ
γ China
= 5.905. The odds of the outcome ‘agree
(neutral)’ toward CBEVs made in China are 5.905 times larger than the outcome
‘neutral (disagree)’. These results support the hypothesis H1a.
Most of the estimates of PB and PR are insignificant. These results indicate that
perceived benefits of buying EVs and perceived risks toward them do not have significant influences on his/her purchase intention toward each country’s CBEVs except
the effect of perceived benefit on Chinese CBEVs. The hypotheses H1b and H1c
cannot be verified.
Table 1.2 Results by ordered logit model
CH
US
SW
MX
KR
JP
PB
0.45**
0.22
0.04
– 0.07
0.20
0.40
PR
−0.26
−0.09
−0.07
0.19
−0.18
−0.03
PQ c
1.78***
1.79***
1.97***
2.60***
2.02***
2.09***
TSt 1
5.01***
4.49***
5.29***
7.61***
4.88**
6.06***
TSt 2
7.03***
7.05***
8.43***
10.39***
7.86***
8.89***
N
152
151
150
150
150
150
R 2
0.20
0.21
0.22
0.30
0.25
0.30
Note The response variable is the purchase intention toward country c’s CBEVs (PI c ). “c” means
China, the US, Sweden, Mexico, South Korea, and Japan. Asterisks (**, ***) indicate that the
estimate is significant at 5% and 1% levels, respectively. TSt 1 , t 2 correspond to Threshold t 1 , t 2 . R 2
is Pseudo R 2 . Control variables are included in each model