Friend-Invitation Promotion Scheme Used in Electric Carsharing: Empirical Analysis and Policy Implications

Table 1

Null hypotheses were used in the data analysis and their respective tests.

S/N	Contents	Tests
1	Ages in the samples of the new and old members originate from the same distribution	One-way ANOVA on ranks
2	The median age of the new members is not greater than that of the old members	Mann–Whitney Wilcoxon test
3	The median number of successful friend invitations of the old scheme participants is not greater than that of the new scheme participants	Mann–Whitney Wilcoxon test
4	The median number of successful friend invitations per month of the new scheme participants is not greater than that of the old scheme participants	Mann–Whitney Wilcoxon test
5	The number of friend invitations in the samples of the new scheme participants who were invited by friends to register as a member and the other new scheme participants originates from the same distribution	One-way ANOVA on ranks
6	The median number of friend invitations of the new scheme participants who were invited by friends to register as a member is not equal to that of the other new scheme participants	Mann–Whitney Wilcoxon test
7	The median age of the not-invited new members is not greater than that of the invited new members	Mann–Whitney Wilcoxon test
8	The median age of the scheme participants is not less than that of the other members	Mann–Whitney Wilcoxon test
9	During the first eight months in 2016, activeness in the samples of the old members and the invited new members originate from the same distribution	One-way ANOVA on ranks
10	During the first eight months in 2016, the median activeness of the old members is not equal to that of the invited new members	Mann–Whitney Wilcoxon test
11	During the first eight months in 2016, the median activeness of the not-invited new members is not greater than that of the invited new members	Mann–Whitney Wilcoxon test
12	During the first eight months in 2016, the median of the old active members is not greater than that of the invited new active members	Mann–Whitney Wilcoxon test
13	During the first eight months in 2016, the median of the invited new active members is not greater than that of the not-invited new active members	Mann–Whitney Wilcoxon test
14	The median within the period from June 2016 to August 2016 of the invited new active members is not greater than that of the old active members	Mann–Whitney Wilcoxon test
15	The median within the period from June 2016 to August 2016 of the invited new active members is not greater than that of the not-invited new active members	Mann–Whitney Wilcoxon test
16	During the first eight months in 2016, the median of the invited new active members is not greater than that of the old active members	Mann–Whitney Wilcoxon test
17	During the first eight months in 2016, the median of the invited new active members is not greater than that of the not-invited new active members	Mann–Whitney Wilcoxon test
18	During the first eight months in 2016, for the invited new active members, in the samples of who made successful friend invitations and who did not make successful friend invitations originate from the same distribution	One-way ANOVA on ranks
19	During the first eight months in 2016, for the invited new active members, the median of those who made successful friend invitations is not equal to that of those who did not make successful friend invitations	Mann–Whitney Wilcoxon test
20	During the first eight months in 2016, the median of the old active members is not greater than that of the invited new active members	Mann–Whitney Wilcoxon test
21	During the first eight months in 2016, the median of the not-invited new active members is not greater than that of the invited new active members	Mann–Whitney Wilcoxon test
22	During the first eight months in 2016, in the samples of the old active members and not-invited new active members originate from the same distribution	One-way ANOVA on ranks
23	During the first eight months in 2016, the median of the old active members is not equal to that of the not-invited new active members	Mann–Whitney Wilcoxon test
24	During the first eight months in 2016, for the invited new active members, in the samples of those who made successful friend invitations and those who did not make successful friend invitations originate from the same distribution	One-way ANOVA on ranks
25	During the first eight months in 2016, for the invited new active members, the median of those who made successful friend invitations is not equal to that of those who did not make successful friend invitations	Mann–Whitney Wilcoxon test

and denote that an estimate is statistically significant at the level of 1% and 5%, respectively.