Page 147 - 臺大管理論叢第32卷第2期

NTU Management Review Vol. 32 No. 2 Aug. 2022

are more willing to support it.
Table 3 presents the regression results. Model 1 examines the effect of the bystander
ratio. As shown in Model 1 of Table 3, the daily bystander ratio was negatively associated
with the daily number of pledges (B = -4.311, p < 0.01). The results suggest that bystander

effects do harm daily fundraising, thus supporting our Hypothesis 1.
Model 2 of Table 3 tests the moderating effect of project legitimacy on the daily
amount of fundraising. The results suggest projects initiated by incumbent firms mitigate

the negative impact of the bystander ratio on the daily amount of fundraising (B = 0.669, p
< 0.05). We thus receive qualified support for Hypothesis 2.
Figure 3 visualizes the moderating effect of project legitimacy on the relationship
between the bystander ratio and pledge amount based on Model 2 of Table 3. Specifically,
we plot the effect of the bystander ratio when project legitimacy takes the value of one

(“project legitimacy”) and zero (“no project legitimacy”). The downward sloping line
clearly shows that the negative effect of the bystander ratio is weakened when the project
is of a higher level of project legitimacy, i.e., projects initiated by incumbent firms.

Model 3 of Table 3 tests Hypothesis 3, which predicts that the negative effect of
the bystander ratio on daily fundraising will become weaker for projects with a longer
funding duration. The results show that project duration weakens the negative relationship
between the bystander ratio and the daily amount of fundraising (B = 0.972, p < 0.01), thus
supporting Hypothesis 3.

Based on Model 3 of Table 3, Figure 4 plots the moderating effect of project duration
when it takes the value of two standard deviations above (long project duration) and below
(short project duration) the mean value of project duration. It shows an apparent positive

moderating effect so that the negative impact of the bystander ratio becomes smaller as
the project duration is extended. However, the negative impact of the bystander ratio is
stronger for those projects with a shorter period. Therefore, we find support for Hypothesis
3.
Concerning the controls, we interpret our results based on Model 1 of Table 3. We

found that the coefficient of project legitimacy is negative (B = -0.556, p < 0.01). While
firm incumbency may provide crowd funders with information about the project creator,
it may also indicate that these incumbent firms have financial resources to support their

campaign projects; therefore, crowd funders might hold that there is less need for these

139