An article posted on Research Design Review back in 2010 discussed the work of William James and, specifically, his concept that consciousness “flows” like a river or stream. The article goes on to say that James’ “stream of consciousness” is relevant to researchers of every stripe because we all share in the goal of designing research “to understand the subjective links within each individual.” Yet these subjective links come at a price, not the least of which is the “messiness” of the analysis as we work towards identifying these links and finding meaning that addresses our objectives.
Whether it is the verbatim comments from survey respondents to open-end questions or the transcripts from focus group discussions or ethnographic interviews, the researcher is faced with the daunting job of conducting a content analysis that reveals how people think while at the same time answers the research question and takes the sponsoring client to the next step. An analysis of our research content must, at the minimum, be based on the:
Development of codes/labels grounded in the data & the criteria that define these codes/labels are established
The dataset is coded
Categorical “buckets” are developed & contrasted/compared to categorical constructs based on what we know related to the research question
Patterns or themes across categories are identified
An analysis of the patterns in conjunction with earlier known research can be conducted
Final interpretations & implications from the data are derived
This simplified version of the process sounds, well, pretty simple, doesn’t it? Until, however, you consider the decisions that need to be made along the way. For instance,
- What data sources will you use in the content analysis? If qualitative, do you include just the text responses in the online group discussion or do you integrate the uploaded images and video into the analysis process?
- What unit(s) within the data sources will you include in the analysis?
- What role is the manifest (i.e., the visible and obvious) versus latent (i.e., the underlying meaning) content going to play in the coding and analysis?
- What level of inference is permissible and how do you control for that?
- Are the categorical labels mutually exclusive (as they should be) and, if not, where is the overlap and what adjustments need to be made to account for that?
- How will patterns be identified and how will this procedure go beyond the fact of mere magnitude?
- What role will earlier known research play in analyzing patterns and drawing conclusions?
Navigating the wide and divergent streams of our participants’ consciousness is a noble and necessary goal of all researchers. And the downed trees, boulders, and debris that lie in our way are there to remind us of the difficult task before us.