Class Notes

Qualitative Methods

11.19.2013

In terms of interviewing and information gathering, there is solicited and unsolicited information. There are questions of validity when you bring in equipment (cameras/recorders). However, this can be seen as a non-issue; the interview constructs a non-normal state anyway, and the camera doesn’t add anything that is necessarily more biased. Jackson uses the term ‘ethnographic sincerity’ to refer to the trust we establish between ethnographer and audience that reality is represented faithfully.

We are required to report to authorities any suggestions of child or elder abuse given by participants. It is also possible that the same is true for self-harm and suicide.

In deciding whether something is an artifact or document, we should look at how it is activated in a social context. We also need to be careful about assuming that any room is single purpose or that its artifacts are necessarily tied to the room. Often things are just around, but not used.

For the last class, we are going to each take a stand on issues about qualitative research. The proposition is that qualitative research is somewhere in between academic narcissism and unprofitable fiction. Our portfolio is due on 12/16.

Hammersly and Atkinson are fairly positivistic. Contemporary ethnography embraces a methodological holism, which is more concerned with processes of social interaction than the many individuals. While methodological individualism allows for static explanations of things like document (their existence is their own validity), methodological holism needs dynamic explanations that explain the document in use, made valid by its place in human interaction.

Our analysis of texts needs to look at facts, language/symbolism, framing, narrative structure/argument form, and discursive performance. Language always appears within framing, and narrative or argument is always some type of discursive performance. Texts are bodies of work, and sources are what we have access to. Sources are representative of the potential text. We should observe form, frame, language, narrative, and discourse when arguing that our source is indicative of the text.

Cases are sources that are evaluated for their qualities as a member of the congregation. This is the first step in coding. Units of analysis are pre-selected conceptual/formal properties of the text to be coded. We should report our units of analysis because it shows that the study is thorough. Codes identify properties, agents, actions, consequences, values, and theoretical propositions. A code is a piece of semiotic material and it establishes the interpretive demand of understanding.

Coding must reflect both breadth and depth. Breadth refers to the coverage of the material and depth refers to the layering of significance. Rhizomatic coding leads to overlapping codes, which is typically represented by the depth of coding. Through coding we can accomplish, describe, confirm, compare and contrast, change, entail, critique, and move an agenda.

Before coding, we need to acknowledge the burden of coding.  We must also understand that there are several types of work associated with different types of text, including things like digitization and work with QDA software. After this, you must assemble the needed resources, including hardware, software, services, and personnel.

When doing coding, we need to read sources several time, look for things like rhetorical force, intertextuality, and interpellation (hailing). Close reading is never the same as coding. There should be a classification of each case, and the unit of analysis needs to be established. We should avoid using a ‘shiny-thing’ approach to choosing units of analysis. Unitizing by paragraph or turn is systematic, but artificial. It is also important to collect statistics on your coding activity like the amount of time spent, the codes per unit, and the density of coding. This lets you estimate how long it will take on future projects.

In the second pass of coding—meta-coding—you codes the codes for things like convergence or divergence. The codes themselves become a text, with presences and absences. It should be emergent and theoretical.

If we are working with a large data-set, we shouldn’t be thinking of it all working toward a single coding exercise. Each level of coding can present one or two publishable papers before the coding is even finished.

If we believe in pure grounded theory, then we should be able to find the theory in the text. Coding is a process of exposing the theory within the artifact. This is different than emergent coding. Axial coding adds theory in the 2^nd level of coding, and interactive coding uses grounded codes in the case and then apply them to existing theory, aiming for a synthesis.  Two-phase coding involves first using emergent coding and then following it with theoretical coding. This is the kind most likely to get published.

When trying to get published, you need to be thorough in your analysis of your procedures and the statistics associated with your coding. The argument starts with the importance of the text, follows through with theory and literature, and uses multiple case warrants and should account for disconfirmations. We need to be careful when invoking the audience or author [of the artifact], and our conclusions need to be anchored within the text.