Result

In the manipulation check, the author make hedonic score minus utility score (eg.

“Make me feel happy”) minus (eg. “Must be useful and practical for me”). Define

"relative hedonic score" for the target product. The results of the T test showed that notebook (M=-0.62, t=-9.09) and luxury handbag (M=0.76, t=6.89) were utility product and hedonic product for the subjects. In addition, the consumer's perception of guilty (M=2.49) on the purchase (redeem) of the notebook is less than luxury handbag (M=4.01, t=10.11, p <0.01), which is in line with the product type.

18

(一)First part (situation－cash vs. situation－points) Main effect

In the first part, the payment formats (cash vs. points) are the independent variables and the gender is covariance. The general linear model (GLM) results are as follows：

(1) Non-medium and medium have significant differences in luxury handbag's willingness to pay (M^cash =8877.18 vs. M^points=13404.69, F(1；103)=6.04, p＜0.05)，

The subject's willingness to pay for luxury handbag in the points situation is significantly higher than the cash situation.；(2) Non-medium and medium have no significant differences in notebook's willingness to pay (M^cash =25397.93 vs. M^points

=24648.30, F(1；103)=0.17, p＞0.1).；(3) Non-medium and medium are significant marginal in luxury handbag's proportion of budget allocation (M ^cash =22.6% vs.

M^points=32.0%, F(1；103)=3.58, p=0.06)，the ratio of the budget allocation of luxury

handbag in the points situation is significant marginal higher than cash.；(4) Non-medium and Non-medium have no significant differences in notebook's proportion of budget allocation (M^cash =72.5% vs. M^points =73.0%, F(1；103)=0.01, p＞0.1). Finally, logistic regression analysis shows that the choice of non-medium and medium are significant marginal (the proportion of choice hedonic product: M^cash=5.5% vs. M ^points=15.7%, χ 2 (1) = 2.77, p=0.10). The results show that the impact of the medium is mainly on hedonic consumption (not on utility consumption).

Mediator analysis

The authors discuss their effects on possible mediator with the same independent variables and covariance. The GLM results show that: (1) Non-medium and medium are significant differences in windfall (M^cash=2.82 vs. M^points=3.63, F(1；103)=7.13, p

＜0.01)，windfall in points situation is significantly higher than cash situation; (2) Non-medium and Non-medium are nonsignificant in efforts (M^cash =5.74 vs. M^points =5.36, F(1；

103)=2.21, p＞0.1).；(3) Non-medium and medium have significant differences in pain

19

of paying in luxury handbag (M^cash = 4.32 vs. M^points=3.59, F(1；103)=5.1, p＜0.05)，

subjects’ pain of paying of luxury handbag in the cash situation is significantly higher than the points；(4) Non-medium and medium are nonsignificant in notebook's pain of paying (M^cash =2.81 vs. M^points=2.57, F(1；103)=0.7, p＞0.1).；(5) Non-medium and medium are nonsignificant in luxury handbag of guilt (M^cash =4.06 vs. M^points =4.03, F(1；

103)=0.01, p＞0.1)；(6) Non-medium and medium are nonsignificant in notebook (M^cash

=2.41 vs. M^points=2.49, F(1；103)=0.12, p＞0.1).

The authors used the same independent variables and covariance, and incorporated the two variables of windfall and hedonic product’s pain of paying into the model. The results showed: (1) Non-medium and medium have significant difference in luxury handbag's willingness to pay (p <0.05), but when pain of paying and windfall added to the model become significant marginal (p<0.1). Windfall is not a mediator, there is no significant difference in luxury handbag's willingness to pay (p>0.1). Therefore, hedonic product’s pain of paying is a mediator, because its willingness to pay for the hedonic product reaches significant marginal (F(1；101)=12.28, p＜0.1).；(2) Non-medium and Non-medium have significant difference in luxury handbag's proportion of budget allocation (p < 0.1), but when pain of paying and windfall added to the model become nosignificant (F(1;101)= 0.97, p>0.1). Windfall is not a mediator, there is nosignificant in budget allocation of the hedonic product (luxury handbag) (p

>0.1).However, pain of paying is a mediator (F(1；101)=17.25, p＜0.01)，because it affects the impact of different payment formats on the proportion of hedonic product (F(1；101)=12.28, p＜0.1).；(3) Non-medium and medium are significant marginal in choice (p<0.1), but when pain of paying and windfall added to the model become nosignificant (p＞0.1). Windfall is not a mediator, because it does not affect hedonic

20

consumption (p >0.1), but pain of paying is a mediator which affects the choice of products (χ2(1)=5.42, p <0.05 ).

(二)Second part (situation－cash vs. situation－informing points purchased by cash)

Main effect

In the second part, the payment formats (cash vs. informing points purchased by cash) is the independent variable and the gender is covariance. GLM results are as follows：(1) Non-medium and medium have significant differences in luxury handbag's willingness to pay (M^cash =8317.82 vs. M^points=15550.68, F(1；100)=21.9, p＜0.05)，

The subject's willingness to pay for luxury handbag in the points situation is significantly higher than the cash situation.；(2)Non-medium and medium have no significant differences in notebook's willingness to pay (M^cash =25562.86 vs. M^points

=23740.47, F(1；100)=1.914, p＞0.1).；(3)Non-medium and medium are significant marginal in luxury handbag's proportion of budget allocation (M^cash =21.4% vs.

M^points=38.6%, F(1；100)=15.45, p＜0.01)，the ratio of the budget allocation of luxury

handbag in the points situation is significan marginal t higher than cash.；(4) Non-medium and Non-medium have no significant differences in notebook's proportion of budget allocation (M^cash =73.3% vs. M^points =70.1%, F(1；100)=0.59, p＞0.1).；(5) Non-medium and medium have significant differences in product's choice (the proportion of choice hedonic product: M^cash=5.5% vs. M ^points=16.7%, χ 2 (1)=3.90, p＜0.05). This result shows that even if the subject knows that points are purchased from cash, it will still be affected by the medium effect, which is different from the cash payment in the hedonic consumption tendency.

Mediator analysis

In the second part, the payment formats (cash vs. informing points purchased by cash) is the independent variable and the gender is covariance. GLM results are as

21

follows：(1)Non-medium and medium are nosignificant differences in windfall (M^cash

=2.89 vs. M^points=3.06, F(1；100)=0.45, p＞0.1).；(2) Non-medium and medium are nosignificant differences in effort (M^cash=5.67 vs. M^points=5.55, F(1；100)=0.26, p＞0.1).；

(3) Non-medium and medium are significant differences in handbag's pain of paying (M^cash=4.23 vs. M^points=3.35, F(1；100)=9.42, p＜0.01) , handbag's pain of paying in the points situation is significant higher than cash.；(4) Non-medium and medium are significant differences in notebook's pain of paying (M^cash=2.90 vs. M^points=2.25, F(1；

100)=6.77, p＜0.05) , notebook's pain of paying in the points situation is significant higher than cash.；(5) Non-medium and medium are nosignificant in handbag's guilt (M^cash=4.01 vs. M^points=3.85, F(1；100)=0.27, p＞0.1).；(6) Non-medium and medium are nosignificant in notebook's guilt (M^cash=2.38 vs. M^points=2.61, F(1；100)=0.81, p＞

0.1).

The authors incorporated the two variables of handbag's pain of paying and notebook's pain of paying into the model. The results showed: (1) Non-medium and medium have significant difference in luxury handbag's willingness to pay (p<0.01), but when two pain of paying added to the model still have significant difference (Mcash=8193.02 vs. Mpoints =15665.25, F(1；98)=23.97, p＜0.01)；(2) Non-medium and medium have significant difference in luxury handbag's proportion of budget allocation (F(1 ； 100)=15.45, p ＜ 0.01), when two pain of paying added to the model have nosignificant difference (F(1；98)=1.98, p＞0.1). , then luxury handbag's pain of paying (F(1；98)=3.89, p=0.05) and notebook's pain of paying (F(1；98)=6.78, p=0.01) to hedonic product's budget allocation is significant. Therefore, hedonic product's pain of paying and utility product's pain of paying are mediators for payment formats to hedonic product's budget allocation.；(3) Non-medium and medium are significant marginal in choice (p<0.1), but when hedonic product's pain of paying and utility product's pain of

22

paying added to the model become significant (χ²(1)=6.70, p＜0.01). Therefore, hedonic product's pain of paying and utility product's pain of paying are not mediators.

Overall, the main effect shows that even if the subject knows that the points are purchased from cash, it still shows a tendency to differ from cash payments in hedonic consumption. However, mediator analysis shows that the mediator process seems to be different from when the subject is not told that the points are purchased from cash.