October 5, 2004

Content and Discource Analysis Session

Here are the small group discussion notes for the

Table A:
 Is everything content in the new media...or how do we define content? Is it the product of conscious creation? Of unconscious creation? Is a network content or context?
 The rules for turn taking and other aspects of discourse analysis could be useful for content analysis but there are differences as discourse analysis is more structural than content analysis.
 Need to have more consideration of the receiver in context when considering the meaning and importance of content in new media.
 Was the war coverage study a content analysis or a cultural analysis? Perhaps a comparison across time, Iraq 1 - Iraq 2, would have been more possible... perhaps the variations in media make the research more important but we need to find a way to deal with these differences and measure them.
 Time is an important part of communication as pauses are meaningful this could be important to conversational analysis.

Table B:
Qualitative vs quantitative models: have used quantitative as precursor to a qualitative study
- Coding online interactions of children
- Study of content should be connected with study of newsrooms / production, how do you compare new and old media - maybe qualitative studies comparing old media users with new media users would be useful. New media users can alter the content - they can change ¨prominence¨ - understanding the meaning of ¨make your own news¨ is important. New media used according to perceived expectations of how it is to be used
- Social interactions shaped by this, even speed of connection, changes the media product / nature of interaction and must account for technical issues altering content - content is dynamic.
- Content Analysis is useful but in general must account for reception factors and must look at production - can´t take content for granted - how do you contextualize the specific content you sample?

Table C:
BW: In the problem of not being able to capture every image (in Susan Keith’s paper): you don’t really need to capture all images
AT: Sampling, but exchange is difficult to know when it happens, you can’t just take samples and assume they convey meaning. A fundamental problem is that sampling does not allow for ephemeral data and development of meaning.
BW: Still able to say something about how news is framed without having all of the images.
EG: Multiple editions have been with us for a long time, newspapers and tv. Then we had to make a selection for determining framing criteria. One old procedure was to take the final edition of a newspaper, this is not a new problem.
AT: Agrees re value of sampling of news products, but still holds that sampling of conversations is problematic. For example, if you are following a thread, you can begin with the point of entry and go to the end of thread, it is all interlocking conversation....recognizing this a serious problem, but not providing a solution.
KG: When they can´t follow the discourse on a list, the use of discourse analysis was not suitable as a method of study. Adaptation of qualitative content analysis. New point - reliability of age and gender, coming from third presentation. In a study of Singapore online and offline chat service, have much to learn from users...do they meet offline? People who use the medium are more sensible to some cues than others. Including users into research design is important in order to approach this issue of reliability. Have users as design participants...able to illuminate hidden communication. One method may help inform data from another method. Another problem is the overload of data...how to deal with this... how to select... selection technique has huge impact on outcome... We haven´t talked about the multimedia environment and how to integrate understanding of these sources. One possible approach would be to approach users and ask them how they íntegrate´ such multiple sources of data...

Table D:
Issue of the context of analysis, how in-depth should analysis go? Comparability between media (what is an image?). How to improve the method?

Table E:
Questions raised:
• How much data is enough (sampling)?
• Is content analysis suited to web content? Web material may or may not be appropriately conceptualized as a document.
• While the raising the bar of methodological rigor may be difficult, is, for the time being, methodological transparency sufficient?
• New methods?

Table F:
- Content analysis of conversational transcripts raises important ethical issues, such as those related to informed consent, age, and potential reaction (or lack thereof) to incidents discovered online
- Some questions raised: Can content analysis assess how people actually engage with web content? What are appropriate research questions for content analysis? Sheer volume of content on the web is revitalizing use of the method, but is it attached to theory? (See McMillan for discussion)
- Sampling: Need to focus on intention and purpose; uneven considerations thus far regarding how and why

Posted by npaul at October 5, 2004 4:16 PM