| Crowdfunding is popular among many start-ups due to its accessibility and high efficiency.However,its low success rate also plagues fundraisers.To figure out factors contributing to crowdfunding success,many scholars explored from various perspectives.But they failed to excavate the potential of interaction texts generated by investors and fundraisers.Only when fundraisers respond to reviewers will a complete communication happen.Considering the openness of the Internet,fundraisers’ responses not only affect receivers’ attitudes,but also attract the attention of potential investors.Therefore,the mining of fundraisers’ responses is of great significance to researches on crowdfunding performance.We attempt to help fundraisers understand potential investors and improve their performance by settling the following questions: What is the demand of crowdfunding reviewers? Will reviewers replied to change their decisions? What are the response strategies adopted by fundraisers? How do these response strategies affect reviewers’ investment decisions?This study aims to illuminate the relation between fundraisers’ response strategies and reviewers’ investment decisions.With a program written in Python,projects on Modian are crawled.Grounded in 6405 interaction texts,we induce and refine crowdfunding reviewers’ demands and fundraisers’ response strategies referring to relevant literature and researchers’ experience in the process of threelevel coding.And Naive Bayes Method is employed to classify them.On the basis,we examine the function of the existence of responses,the respondent identity and fundraisers’ response strategies progressively.In addition,the information specificity is chosen as the moderator.Finally,we give an explanation to the results combining Language Expectancy theory with Cognition-AffectionConation model.Besides,the robustness of the research conclusions is tested with the reclassification and subsample regression.The results of text mining show that crowdfunding users review either to collect information or to express themselves,and fundraisers often apply projectoriented strategies which devote to transmitting signals of projects’ quality and social-oriented response strategies which are meant to shorten potential investors’ psychological distance.With the supervised learning algorithm,they are classified at the average precision of 80 percent.Moreover,empirical results have shown that(1)Reviewers who receive responses are more likely to invest.(2)In line with ordinary consumption scenarios,crowdfunding users also have a great preference for the respondent identity,which means the effectiveness of other reviewers’ responses is inferior to that of fundraisers.(3)Compared with socialoriented strategies,project-oriented strategies are more popular among potential investors.And information specificity can enhance the positive effect.(4)We have obtained a model that can predict the investment willingness of crowdfunding reviewers with an accuracy of 75.6757%.Moreover,the robustness test has implied that the impact of response strategy on individual investment remains reliable.This study enriches and refines the research on crowdfunding performance and provides new perspectives for the text mining in crowdfunding texts.From the individual perspective,we figure out crowdfunding reviewers’ demands and their attitudes toward fundraisers’ response strategies.Meanwhile,this study also expands the application scope of response strategy research to crowdfunding.And Language Expectancy theory and Cognition-Affection-Conation model are creatively applied to explain the mechanism.In practice,it not only helps crowdfunding financiers understand the personalized information needs of potential investors,and reminds them to emphasize and manage online interactions,but also directs platforms to optimize their service for both ends. |
PDF Full Text Request