Online task 8

8 What Did You Hear? Data Analysis

Excerpts and thoughts from: Rubin, H. J. and I. S. Rubin. 1995. Qualitative Interviewing. The Art Of Hearing Data. Thousand Oaks: Sage. Pp. 226-256.

Data analysis is exciting because you discover themes and concepts embedded throughout your interviews. As you continue with the data analysis, you weave these themes and concepts into a broader explanation of theoretical or practical import to guide your final report. Data analysis begins while the interviewing is still under way. After completing each interview, you examine the data you have heard, pull out the concepts and themes that describe the world of the interviewee and decide which areas should be examined in more detail.

• To begin the final data analysis, put into one category all the material from all your interviews that speaks to one theme or concept. Compare material within the categories to look for variations and nuances in meanings. Compare across the categories to discover connections between themes. The goal is to integrate the themes and concepts into a theory that offers an accurate, detailed, yet subtle interpretation of your research arena.

The steps of analysis: Let's take an example. An interviewer has collected over 2 million words of interview data, along with 12 boxes of documents. He reads the interviews, paragraph by paragraph and word by word, marking off each time a particular idea or concept is mentioned or explained, and indicating in a code the subject of each paragraph. Then he groups together responses describing the same idea or process and examines everything he has put in the same category. This is detailed work, but important because many of the examples are illuminating and periodically the interviewer may get an insight by seeing quotations and descriptions from separate interviews side by side.

• We start the coding process by rereading the interviews, so we have their general content clearly in mind. When we reread, we also think about the themes, concepts, and ideas we were trying to explore in each interview. Some of these efforts panned out; others we dropped for lack of support. We come up with a starting point for our coding categories by reconstructing which themes, concepts, and ideas we were able to successfully examine. We go back to the interviews, and this time mark off these concepts, themes, and ideas each time they occur in an interview. As we do this coding, we discover other themes, concepts, and ideas and designate new coding categories, we have to go back to the original interviews and look for and mark each place that is and example of the material that now belongs in the new categories. When the coding is complete, the data are grouped in categories that allow us compare what different people said, what themes were discussed, and how concepts wee understood. Through examining the overall descriptions of the cultural arena or explanations of the topic we are studying. We then seek out the broader significance by asking if our data support, modify, or contradict an existing theory or policy. Which themes we emphasize in the analysis depends in part on the audience for the report.

How do you hear and recognize what the data say? We code the material to group similar ideas together and figure out how the themes relate to each other.

- Recognizing Concepts: Conversational partners refer to their ideas or concepts through which they understand their worlds by using a specialized vocabulary or giving a particular tilt or a more common word or phrase. As you begin the data analysis, you look for these explanatory concepts by picking out the words the interviewees frequently use that sound different from your ordinary vocabulary. When looking for concepts and core ideas that interviewees have not labeled with a specific word, ask yourself first, what is the interviewee talking about? Then is the idea important? If it is important, can I summarize this idea with a word or phrase that suggests the meaning of the underlying idea? If the answer to that question is also yes, you have yourself a concept.

• Hearing Stories: When you are looking for underlying meanings and themes, it can be useful to pick out and analyze stories. Stories here are refined versions of events that may have been condensed or altered to make a point indirectly. Narratives are straightforward efforts to answer the question, "What happened?" But a story is often thought out in advance and designed to make a point, usually one that cannot be made in a direct way. Interviewees do not always provide stories, but when they do, you should pay attention, because stories often communicate significant themes that explain a topical or cultural arena. Stories have some or all the following characteristics:

1. Stories are told smoothly, with little fumbling or backtracking. The interviewees may have told them many times before, so they are familiar with their main lines.
2. Stories are often told as adventures, such as how my grandparents came from the old country without any money and started a business; or how Herb and Irene got, and got rid of, Teddy the Himalayan black bear.
3. Longer stores are carefully structured. They may begin with a time or place setting, introduce characters, describe some event or complexity, and then offer a resolution. Not every story ahs all these parts, but it would not be much of a story if it did not contain a dramatic event.
4. Stories are often marked by haunting symbols-condensed, summary images that convey a great deal of emotion and multiple meanings.
5. Stories might be marked by a change in speaking tone. Short incomplete sentences are replaced by fully elaborated ideas, or the interviewee sits back, takes center stage, and begins a long description,
6. Sometimes the researcher asks a question and gets and extended response and response can be a clue that the interviewee is telling a story.

Once you notice a story, try to figure out what lessons it is meant to communicate. In almost any kind of interview, people can use stories to answer difficult or threatening questions indirectly; in cultural studies, stories often present moral themes. Stories are often intentionally indirect, so you may have to figure out their meaning yourself. It may help to ask yourself, why did the interviewee choose to tell a story now, in answer to this question. Is the material emotionally intense? Is it embarrassing? Is it in some sense dangerous or forbidden? Is it a modified version of events that allows the interviewee to live with an uncomfortable or mixed background of some sort?

• Hearing Themes: Themes offer descriptions of how people do or should behave. Some themes are explicitly stated by conversational partners. E.g. " Gang membership is a form of self defense." Or faculty in research universities may tell you, "We don't get promoted if we don't publish." When themes are not explicitly stated, you can often deduce them from several illustrations. In coding the interviews, you can mark each passage with a brief summary of what the person was telling you. Some of these summaries that you make suggest themes, because they seem to fit a common pattern.
• Coding Interview Data: Coding is the process of grouping interviewees' responses into categories that bring together the similar ideas, concepts, or themes you have discovered, or steps or stages in a process. Coding proceeds in stages. First, you set up a few main coding categories, suggested by the original reading of the interviews and the intended purposes of the report. You can make yourself a list as you go through your interviews. Okay, you might say, here is an illustration of "frustration," which was an important idea I was looking at in some of the interviews. I will look for that one through my notes or transcripts. Oh, and yes, there is "uncontrollability of the work," I remember when I was looking for that one, that was important, too. You can start out with a relatively small list of these categories. As you go through the interviews, you can put in brackets, or underline, or otherwise mark off each word, phrase, sentence, paragraph, or extended story or example that you are using as a single coding unit that fits the coding categories you are listing. After you have designated the beginning and end of the coding unit in some way, you mark the unit with an indicator for the coding category into which you are putting the material. You can identify the category by putting a word in the margins. As you sort the data into the categories you choose, you might find that important information doesn't fit into these categories or that one of your categories blurs two or more separate concepts, themes or stages. Then you have to add new categories, you have to go back and recode the material already examined. When you first start coding, you are simultaneously placing data into preliminary coding categories and deciding whether these preliminary categories provide a good fit to the data. If the fit is inadequate, you change the coding categories to reflect what you discover in your data. When designing and applying your coding categories, you have to keep your mind engaged; avoid turning coding into an automated task. Coding encourages hearing the meaning in the data.
• Mechanics of Coding: After you have set up the major coding categories, make a copy of the text and then mark each word, phrase, sentence, paragraph, or example that belongs in each coding category. Sometimes the interviewee uses a word or phrase such as empowerment that indicates a concept. At other times, you have to recognize that a concept is described but not named. The phrases "I feel in charge" and "They had to hear me out" are expressions of empowerment and need to be labeled as such, even though the word empowerment is not used. Or if you are coding for people's names, you may find references to them that do not mention their names. You have to recognize that the phrase "the crooked politician" refers to Mr. R. N. and mark the phrase automatically. If you are going to write out themes and concepts next to the interview material, you should probably leave some extra space between the paragraphs, or split the page into columns, with the text on one side and your codings on the other.

The final stages of analysis: In the final stages of analysis, you organize the data in ways that help you formulate themes, refine concepts, and link them together to create a clear description or explanation of a culture or a topic. This material is then interpreted in terms of the literature and theories in the researcher's field.

• Building Toward Overarching Themes: The coding process fragments the interviews into separate categories of themes, concepts, events, or stages. Coding forces you have to look at each detail, each quote, to see what it adds to your understanding. Once you find the individual concepts and themes, you have to put them together to build an integrated explanation. You follow a two-stage process of thinking about the data. In the first, you examine and compare the material within categories. In the second, you compare material across categories. Within one category, you ask yourself, how uniform are the examples? Do the illustrations suggest some nuance of meaning in concept or theme? Is what was coded as one concept actually two related concepts? If you have coded for events, a single coding category should contain different perspectives on each particular event. You can use the information within one coding category to compare separate versions of an event or combine the separate parts of an event into one narrative. If you coded for time sequence, by looking at a file that contains all those sequence markers you can put together a time line that can organize a narrative. After you have analyzed the material inside each of the coding categories, you look for linkages across coding categories. Sometimes, you can see linkages easily, because the interviewee raises concepts or themes together and describes their relationship. Analysis across your coding categories lets you look for linkages that are much less obvious by putting related ideas in proximity to each other that were not raised at the same time or y the same interviewee. You can string together two, three, four, or even more themes and concepts, thinking through the implications each time you add one more to your existing synthesis. You can start with any of your themes or concepts and ask how they relate to each of the other themes and concepts that you have worked out. You should aim to come up with a number of themes that are linked together and collectively describe or analyze your research arena.

Preparing the themes for the final report: A goal of analysis is to find themes that both explain the research arena and fit together in a way that a reader can understand. Sometimes you can find an overarching theme that ties the individual pieces together. To create an integrative theme you step back and examine the smaller themes to see what, if anything, ties them together. Each of these separate themes would constitute a section of his report, replete with numerous examples. Once you have developed your overarching themes, you need to think about the implications. The data analysis ends when you have found overarching themes and put them in the context of broader theory and answered the question "So what?" Then you write the final report.