Coders: Emily Cohn and Caryn Ganeles (yellow)
In our study of inter-coder reliability, we read through 20 posts on an online support group on marriage support. We then individually determined whether or not each of the 5 support types in addition to the unique support behavior, humor, were present in the message and compared our results, calculating inter-coder reliability. Following the model presented in Braithwaite’s article, “Communication of Social Support in Computer-Mediated Communication for People with Disabilities,” we analyzed messages and concluded that our inter-rater reliability was 73.3%.
When going about this experiment, we came across several messages that had to be discarded. For the most part, about half of the messages had to be discarded due to the fact that many posts on the support group were simply individuals proclaiming their problems and therefore failed to provide any sort of support.
Here were our results:
% inter-rater reliability 0.733333333
frequency % of msgs
Information 16 0.8
Tangible assistance 2 0.1
Esteem support 4 0.2
Network support 8 0.4
Emotional support 6 0.3
Humor 5 0.25
As you can see from our results chart, our content analysis differs in many ways from Braithwaite's results. In the article "Communication of Social Support in Computer-Mediated Groups for People With Disabilities", Braithwaite found that the order of prevalence of different support types was emotional support, information, esteem support, network support, and tangible assistance, respectively. However, our results show that our order of prevalence is information, network support, emotional support, esteem support, and tangible support. Humor was also added to our analysis, and this was found in 25% of the posts.
We found that information was prevalent in 80% of the posts. This is very different from Braithwaite’s 31.7% figure. This difference may be accounted for by the type of support network we were analyzing. In Braithwaite, the research team analyzed messages from a support network for people with disabilities. On page 24 of the article Braithwaite
writes, “…we studied only one computer support group, targeted toward persons with disabilities. One obvious direction for future research is to examine communication in other groups to determine whether the patterns we found are generalizable.” Due to our results, we conclude that the pattern is not generalizable. Because we analyzed a marriage support group,
members’ posts were more directed towards asking for advice than seeking social connections. Our support network members did not have barriers to FtF communication and may not have been looking for the same type of online relationship as those in the disability network.
We believe the difference in support networks is accountable for our differences in data from Braithwaite. Our second most prevalent support type, network esteem at 40%, does not match Braithwaite’s 7.1%. The posters in our support group liked to refer others to books or articles to help answer their questions. Emotional support was offered 30% of the time, which is slightly different from Braithwaite’s 40%. The subjects in the marriage network did not need as much emotional support as those with life-altering physical disability problems. Esteem support occurred 20% of the time in our results and 18.6% is Braithwaite’s. Lastly, both our group and Braithwaite ranked tangible assistance last, 2.7% in Braithwaite’s data and 10% in ours. It is difficult to offer physical assistance overthe CMC support network.
We can relate our findings to McKenna’s relationship facilitation factors. First, the factor of connecting to similar others was quite evident in the support group. Many people in the support group were there for the same reason; problematic marriages caused the message providers to all congregate in the group due to the principle of common group. Such a connection contributed to much of the emotional support and network support that was present in such messages.
Additionally, it is evident that identifiability played a role in the exchange of messages. This factor suggests that disinhibited behavior emerges online due to increased private self-awareness and decreased public self-awareness, both of which are caused by visual anonymity. It is evident that the members of this support group were disinhibited by their anonymity and therefore more inclined to reveal information that they would otherwise consider private information. If this support group were to exist in FtF interaction, such individuals would be discouraged from revealing such information, keeping in mind that what they say could reflect on themselves more personally.
Lastly, the removal of gating features also contributed to the messages relayed in the support group. Due to the fact that physical attractiveness, status cues, and social anxieties were not apparent on the site, contributors were not hesitant to express themselves in the online forum.
Emily's comments:
http://comm245yellow.blogspot.com/2007/10/assignment-8_9699.html
http://comm245yellow.blogspot.com/2007/10/8-sage-parenting-advice-from-usenet.html
2 comments:
Emily,
Nice job with this assignment. It was interesting to read your analysis of the information in which you got from the Excel spreadsheet. My group experienced very similar statistics: we had high information and emotional support from the messages we looked at. What I liked most about this post was the way that you connected all of the information to theory. Using McKenna’s Relationship Facilitation Factors (connecting to similar others) was clever and I did not think about this theory until reading your post. My group thought along the lines of Walther’s main theories, so it was good to gain another perspective. Overall, nice job on the post and I enjoyed reading it. Looking forward to reading more…
It is very interesting to read your post as your results were very different my groups results. The primary reason for the difference is, as you said, because different support networks have different needs. My group looked at a thread about quitting smoking where there was very little information and almost no network support. I would think that the biggest distinction between the two types of support groups is that information and networking is more easily available for quitting smoking. It can be seen as less emotional and personal than marriage problems, but that would be where the people would need the most support. Whereas with marriage problems, people are probably more hesitant to look for help offline as they have too much public self-awareness. Great job with your post.
Post a Comment