An Analysis on the Importance of Persuasion Strategies in Environmental-Oriented Online Crowdfunding Projects

Abstract

This research aims to investigate linguistic features of online crowdfunding projects from the perspective of persuasion strategies. To understand how a project is described to persuade sponsors to make investments, especially with the highly concerned environmental issues, this research targets projects grouped in the Environmental category on Kickstarter. A total of 566 projects (522,280 tokens) were collected from January 2020 to August 2021. Features classified in accord with four persuasion strategies are extracted, such as psychological distancing (e.g., first, second and third singular and plural personal pronouns), concreteness (e.g., time and percentages), message framing (e.g., sentiment tendency of project descriptions and frequencies of positive and negative words), and impact and value (whether a project is assigned with an environmental commitment label by the fundraiser). Five machine learning models (e.g., Naïve Bayes, Decision Trees, Random Forest, Support Vector Machine, and Logistic Regression) were used and trained with features to perform classification tasks. The results display that Decision Tree (F1-score = 0.76), Random Forest (F1-score = 0.67), Logistic Regression (F1-score = 0.61), and Support Vector Machine (F1-score = 0.71) have outperformed the baseline models trained with Term Frequency-Inverse Document Frequency (TF-IDF) as the baseline with F1-score of 0.59, 0.48, 0.55, and N/A respectively. The results indicate that the usage of persuasion strategies play important roles in online environment-related crowdfunding projects. Based on the findings of this study, crowdfunding project raisers may refer to the usage of these persuasion strategies while writing project descriptions to increase the success rate.