The thread our group examined was a google group that dealt with anorexia and marijuana. The people on the thread discussed how smoking marijuana increases appetite and how that could possibly cause anorexics to eat more and gain weight. The three people in our group are all girls, are all between the ages of 19 and 20, are all studying either psychology or communication, and are all in sororities. As you can see, all three of us lead very similar lives and as a result had very similar interpretations of the 20 messages we read in this thread.
The cutoff for an acceptable reliability score is .7. Our inter-rater reliability score was roughly .81, making our score more than acceptable. But, Braithwaite’s coding scheme percentages are quite a bit different than some of our coding scheme percentages. The schemes with the biggest differences were information, network support, and emotional support. Braithwaite found that information had a 31.3% presence, while we found it to have a 70% presence in the messages. Braithwaite found a 7.1% presence for network support, while we didn’t find any presence of network support in the messages. Lastly, Braithwaite found a 40.0% presence of emotional support and we only found a 2% emotional support presence. The schemes with the most similarities were tangible assistance and esteem support. Braithwaite found a 2.7% presence for tangible assistance and we found a 3% presence. Braithwaite also found an 18.6% presence of esteem support, and we found a 15% presence. Although two of the percentages were extremely similar, three of the percentages were extremely different.
We believe that our percentages vary from Braithwaite’s because the forum we observed and studied diverges from originally supporting the person who started the discussion to moving more towards being concerned with the accuracy of the content of anorexia on the thread. This shift in topic made our percentages shift in relation to Braithwaite.
Walther & Boyd’s theory of interaction management could also have affected the shift in discussion on this thread. People can spend much time editing what they write before posting and because of that, what they originally intend to say gets reconstructed over the course of their editing. Also, as there are more and more posts on a thread, people don’t keep referring to all of the posts before theirs, they really just look one or two people above to make a new point. Lastly, because the conversation on a thread is a asynchronous, people don’t follow the conversation as coherently as it would be in face to face. Completely new subtopics, related to the overall theme of a thread can be mentioned, and that is acceptable. Due to these characteristics of forums, our results vary slightly in some categories from Braithwaite’s. But, ultimately, the similarity of our group members allowed us to all interpret the messages in the same way.
No comments:
Post a Comment