Predicting leadership in online health communities
The purpose of this dissertation was to examine elements of leadership behaviors in online health and wellness communities, and be able to predict when leadership could occur in an online community. Communication behaviors of those who influence other members in terms of initiating comments, triggering replies, and conversations were analyzed to look for leadership characteristics. The research utilized a secondary dataset containing postings from participants across eight global online health and wellness communities on public social media websites: a) TEDMED, b) MyNetDiary, c) Nike+Fuelband, d) Forks Over Knives, e) Runner’s World, f) MyOptumHealth, g) America’s Health Rankings, and h) UHC TV. The observations took place over a six-month period, used content analysis, social network analysis (SNA), and inferential statistics to identify and predict the elements of leadership in online health and wellness communities. Content analysis was used to explore the leadership themes. Principal components analysis was performed to reduce the number of leadership attributes into the top themes, and logistic regression using leadership characteristic count significantly predicted leadership, where the odds ratio for leadership (eB = 1.324) indicated that as the variable leadership characteristic count increased by 1, subjects were 1.324 times more likely to be classified as leaders. The SNA was able to identify five of the eight communities as having a very strong degree of leadership. The methodology used in this research study could be used in future research as a common leadership data model, for sharing and aggregating data.