On the relationship between number of votes and sentiment in crowdsourcing ideas and comments for innovation: A case study of Canada's digital compass

Recently, PWC facilitated an innovation crowdsourcing effort entitled "Canada's Digital Compass" aimed at generating creative directions for Canada to pursue in its future. As part of this effort, over 70 ideas were submitted and over five hundred votes on those ideas were generated. This paper analyzes data developed as part of that effort, investigating a number of relationships between different variables elicited from the analysis and the number of votes for those ideas. The analysis confirms some previous results, provides a structure for interpreting those results and generates some new results. In particular, this paper generates sentiment measures for both the innovation ideas and for the comments. Using a decision tree approach we find that the number of comments and the extent to which the sentiment in those comments is positive are statistically significantly related to the number of votes. In addition, using regression analysis, this paper finds that the number of votes is statistically significantly related to the interaction between sentiment measures for ideas and comments. Finally, a rationale for these results, based on game theory is proposed and investigated. Statistically significant relationship between number of comments & number of votesTransparency theory based on the game theoretic notion of "tit-for-tat"Information about the idea is related to the number of votes.Comment sentiment is related to number of votes.Interaction between idea and comment sentiment is related to number of votes.