The number of research crowdfunding projects included in this analysis is only small – just the first round of eight Research My World projects. Within those eight projects, there are significant variations in their characteristics,i.e., amount of funding sought, duration of campaign, research topic area,and so on. Given these limitations it is possible to say that the observed mea-sures of correlation indicate a potential association between variables, but donot definitively represent a causal link between them, and they do not providea total basis for predicting project success.Acknowledging the limitations of the analysis above, from the first trans-formed variable, it seems clear that social media can play an important partin contributing to the success of a crowdfunding campaign. Anecdotal-ly, those describing their research crowdfunding experiences identify socialmedia communication as significantly improving their prospects of projectsuccess (Perlstein, 2013). Notably, while crowdfunding per se is not new,
PCA reduces the eight variables above to three that have the following factor structure:1. the diameter of the Twitter network, the average directed path length of the Twitter network, the average undirected path length of the Twitter network, and the average Twitter network Erdó ́s number for project principal – explaining approximately 48% of the variation in the original data (i.e., factors relating to the reach of the project’s Twitter network);2. the number of social media shares from the Pozible project website, the total page view count for the Pozible project website and the total unique page view count for the Pozible project website – explaining approximately 37% of the variation in the original data (i.e., factors relating to the ability of the project to attract eyes to its Pozible website and then get the website on-shared); and3. the average pledge amount – explaining approximately 14% of the variation in the original data. If these three transformed variables are computed for each project and the associations of these new variables with project success status are tested, then significant associations can be shown between transformed variables one and two, and project success status. On the other hand, variable three relating to average pledge amount did not have a significant association.<br>
正在翻译中..