Start Date
16-8-2018 12:00 AM
Description
Online communities have become an important knowledge sharing tool for users of the internet (Faraj et al., 2016). Leaders emerge in online communities in ways that are different from traditional person to person interactions. The research on online communities has been largely focused on relational online communities. Relational online communities can include things like forums, blogs (with comments), etc. In relational online communities leaders can be identified using factors like network position (Lee et al., 2018, Johnson et al., 2017); post features such as quality, sentiment, and uncertainty (Siering et al., 2018); post content such as linguistic indicators (syllables, dictionary words, etc.) (Malik et al., 2018). \ \ In relational online communities there are many social aspects that can be analyzed to measure the extent to which a contributor is a leader in the community. Research by Johnson et al. focused on the emergence of online leaders in an online forum, the leaders were identified using surveys sent to all users, a subjective measure of leadership. \ \ In this research we look to transition from relational communities to communities where there is not a social aspect to the community, such as online product review communities. Using toy review data from Amazon.com, we will use the number of helpful votes received to objectively identify discursive leaders in this non-relational online community. The dataset that we will use will be interesting because many of the reviews are written, not by the actual end-user of the product, but by an observer such as a parent, this will add an extra layer of complexity to the process of analyzing the review content. \ \ Once the discursive leaders in the community have been identified, we will use natural language processing, machine learning, and statistical methods to determine the differences that exist between discursive leaders and non-leaders in the online community. \
Recommended Citation
Diaz II, Chad; Conlon, Sumali; Novicevic, Milorad; and Ammeter, Tony, "The Emergence of Discursive Leaders in Online Communities" (2018). AMCIS 2018 Proceedings. 63.
https://aisel.aisnet.org/amcis2018/TREOsPDS/Presentations/63
The Emergence of Discursive Leaders in Online Communities
Online communities have become an important knowledge sharing tool for users of the internet (Faraj et al., 2016). Leaders emerge in online communities in ways that are different from traditional person to person interactions. The research on online communities has been largely focused on relational online communities. Relational online communities can include things like forums, blogs (with comments), etc. In relational online communities leaders can be identified using factors like network position (Lee et al., 2018, Johnson et al., 2017); post features such as quality, sentiment, and uncertainty (Siering et al., 2018); post content such as linguistic indicators (syllables, dictionary words, etc.) (Malik et al., 2018). \ \ In relational online communities there are many social aspects that can be analyzed to measure the extent to which a contributor is a leader in the community. Research by Johnson et al. focused on the emergence of online leaders in an online forum, the leaders were identified using surveys sent to all users, a subjective measure of leadership. \ \ In this research we look to transition from relational communities to communities where there is not a social aspect to the community, such as online product review communities. Using toy review data from Amazon.com, we will use the number of helpful votes received to objectively identify discursive leaders in this non-relational online community. The dataset that we will use will be interesting because many of the reviews are written, not by the actual end-user of the product, but by an observer such as a parent, this will add an extra layer of complexity to the process of analyzing the review content. \ \ Once the discursive leaders in the community have been identified, we will use natural language processing, machine learning, and statistical methods to determine the differences that exist between discursive leaders and non-leaders in the online community. \