Page 357 - Contributed Paper Session (CPS) - Volume 4
P. 357
CPS2277 Iris M H Yeung et al.
likelihood-ratio chi-square test statistic = 194.31; p value < 0.0001
respectively). However, multinomial logit model has slightly bigger
generalized R-square value (0.281 versus 0.256) and the score test of the
proportional odds assumption shows that the ordinal logit model does not
meet the assumption (p value of 0.018). Despite this, the signs of the
parameter estimates for almost all the variables in the ordinal logit model
(column 8) are the same as those for multinomial logit model (columns 2 and
5), and the magnitude of the parameter estimates for most variables in the
ordinal logit model lie between those of the two contrasts in multinomial logit
model except for three knowledge variables. It suggests that the effects of
most variables on WTP follow the ordered sequence of WTP except the three
knowledge variables. Due to better model performance, the results of
multinomial logit models are discussed. The following observations are drawn
for the contrast between “WTP above HK$30” and “WTP below HK$30” being
given in columns 2-4 of Table 1:
1. The degree of support for waste charge policy (b = 1.045; odds ratio =
2.843) has the greatest positive impact on the contrast. However, the
degree of support for building a new incinerator (b = 0.226; odds ratio =
1.253) has the smallest positive impact on the contrast. Obviously the
waste charge policy is more relevant to WTP. The odds of WTP above
HK$30 increased by a greater factor (2.843) per unit increase in the degree
of support for the waste charge policy.
2. Income (b = 0.591; odds ratio = 1.806) and daily waste disposal (b = 0.504;
odds ratio = 1.655) have medium positive impact on the contrast.
Residents who are richer and have more waste disposal are 81% and 66%
more likely to be WTP above HK$30 per unit increase in income and waste
disposal. These residents can afford to pay more and they have greater
needs for waste disposal.
3. Age has a negative significant effect on the contrast (b = -0.031; odds
ratio = 0.969). The older the person is, the less likely they are WTP above
HK$30.
The parameter estimation results for comparing “WTP exactly HK$30”
versus “WTP below HK$30” cases are given in columns 5-7 of Table 1.
The following observations were drawn:
1. Like “WTP above HK$30” versus “WTP below HK$30” contrast, the degree
of support for waste charge policy (b = 0.611; odds ratio = 1.842) has the
greatest positive impact. However, the magnitude is smaller perhaps
because the WTP amount being compared (“exactly HK$30” vs. “below
HK$30”) is less extreme. This may explain why the degree of support for
building a new incinerator, age, income and daily waste disposal amount
have no significant positive impact on this contrast.
346 | I S I W S C 2 0 1 9