Content-oriented or persona-oriented? A text analytics of endorsement strategies on public willingness to participate in citizen science

This paper applies text analytics to study how the orientation of an endorsement strategy affects the public's willingness to participate in citizen science projects. Using 850 citizen science projects with 1,243 endorsements from an online citizen science platform Experiment.com as corpus, the orientation of the endorsement strategies is detected using the naïve Bayesian inference model with a Laplace estimator. Our results inform that 39% of the endorsements are persona-oriented while 61% are content-oriented. A persona-oriented endorsement strategy draws more participants but reduces the per capita invested. A content-oriented endorsement strategy has the opposite effect. Further, the project initiator's identity strengthens or weakens the effect of the endorsement orientation strategy. In the number of participants model, the projects initiated by the professional scientists and research students are positively moderated by persona-orientation endorsements, whereas the projects initiated by the amateur researchers are negatively moderated. The converse holds for the per capita invested model.