Page 426 -

14.2 Online research 417

researchers might need to resort to parsing HTML content. Once message content
is available, it can be analyzed via textual analysis approaches, including building
distributions of words or phrases used in different contexts. These data might then
be stored in a database and analyzed for frequency of occurrence, sequences, or
other patterns of interest. For a slightly more nuanced approach, natural-language
processing techniques might be applied to posts to distinguish between, for example,
cases where someone is discussing their own current concerns as opposed to those
faced in the past, or experienced by a family member (Harkema et al., 2009). As
with other triangulation approaches, these techniques might work best hand in hand,
with qualitative insights suggesting patterns that might be quantified and quantitative
identification of frequent behaviors driving new theories for qualitative exploration.
Diane Maloney-Krichmar and Jenny Preece's in-depth study of an online forum
for people with knee injuries provides a rich example of the use of both qualita-
tive and quantitative methods to understand the dynamics of a complex online com-
munity. Using a four-phase research plan, Maloney-Krichmar and Preece combined
preliminary observation with usability analysis, detailed quantitative analysis of
1 week's worth of messages, and interviews with members of the site. Results in-
cluded characterization of site features supporting sociability; membership patterns
encouraging the health of the community, including identifiable subgroups; task and
individual roles assumed by members; distribution of discussion length, including
the number of messages in each thread; and characterization of the role of group
participation in members' lives (Maloney-Krichmar and Preece, 2005). This detailed
picture provides an example of the possibilities of applying ethnographic techniques
to online communities.
HCI researchers have studied a wide range of online communities. Analyses of
content and interviews with participants were used to develop detailed descriptions
of a 2015 “protest” in the Reddit online community, during which volunteer modera-
tors protested changes in company policy and staff (Centivany and Glushko, 2016;
Matias, 2016). A study of contributions to a repository of projects developed using
the online programming tool Scratch used review of published user profiles and com-
ments on projects to explore the diversity of participants in the community (Richard
and Kafai, 2016).
Online communities can be useful research resources even if you are not willing
(or able) to undertake a detailed ethnographic analysis of a specific group's dynamics.
As shared resources for individuals with common interests, these communities can of-
ten be valuable tools for recruiting participants for studies of factors surrounding com-
munity goals. One study of the dynamics of conflict in free and open-source software
development conducted a survey involving participants in the software development
site GitHub, home to many open-source projects (Filippova and Cho, 2016). Other
studies might involve communities spanning multiple sites. A study of the credibility
of medical “crowdfunding” requests (using online sites to solicit contributions to off-
set medical expenses) examined Reddit discussions regarding campaigns posted on
other sites. Like other studies discussed earlier, this investigation combined content
analysis of postings with participant interviews (Kim et al., 2016a).