Identifying Temporal Changes and Topics that Promote Growth Within Online Communities: A Prospective Study of Six Online Cancer Forums

Identifying Temporal Changes and Topics that Promote Growth Within Online Communities: A Prospective Study of Six Online Cancer Forums

Kathleen T. Durant (Silverlink Communications, USA), Alexa T. McCray (Harvard Medical School and Beth Israel Deaconess Medical Center, USA) and Charles Safran (Harvard Medical School and Beth Israel Deaconess Medical Center, USA)
DOI: 10.4018/jcmam.2011040101
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In this paper the authors have extended the methodology for temporal analysis of online forums and applied the methodology to six online cancer forums (melanoma, prostate cancer, testicular cancer, ovarian cancer and breast cancer). The goal was to develop, apply and improve methods that quantify the responsiveness of the interactions in online forums in order to identify the users and topics that promote use and usefulness of these online medical communities. The evolutional stages that gauge when a forum is expanding, contracting, or in a state of equilibrium were considered. The response function was thought to be an approximation of a discussion group’s utility to its members. By applying the evolutionary phase algorithm, it was determined that two out of six of the forums are in contracting phases, while four are in their largest growth phase. By analyzing the topics of the influential threads, the authors conclude that cancer treatment discussions as well as stage IV cancer discussions promote growth in the forums. It is observed that the discussion of treatment rather than diagnosis is important to help a cancer forum thrive.
Article Preview

We have previously defined a methodology (Durant et. al, 2010b) using a phase detection algorithm and response function and applied it to a melanoma forum. We extended the analysis of thread topics in order to compare the topics that are influential in each calendar year as well as for the duration of each cancer forum.

Temporal data models have been analyzed by Leskovec (2007, 2008a, 2008b, 2008c). However, their analysis is not completely applicable to communicative network models. The communicative networks we present in this paper do not follow Leskovec’s proposed growth power law. Our communicative networks are similar to Leskovec’s models since they do display heavy left tails (but not right tails) for in and out degree distributions. However, our models do not get denser over time as Lescovec’s do. Over certain time periods the example networks get sparser. The occurrence of this phenomenon is explained in subsequent sections and is related to the temporary nature of the data elements.

Complete Article List

Search this Journal:
Reset
Volume 4: 2 Issues (2014)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing