Health communication researchers synthesize complex qualitative data to create personalized text-message based interventions for behavior change, but the process is time-consuming, cognitively taxing and resources-intensive. Crowdsourcing shows the promise of supporting this process, through “on-demand harnessing of flexible and powerful human congnition”. In this work, we addresses the following challenges of crowd-sourcing research: getting behavior change message content that is context-appropriate and non-redundant. We are conducting an empirical study of a three-stage approach for enabling crowds to create context-appropriate, tone-sensitive and non-redundant behavior messages.
This work is under progress.