Social Research Methods: Qualitative and Quantitative Approaches

ANALYSIS OF QUALITATIVE DATA

EXPANSION BOX 1

Themes and Coding Qualitative Data

condensed them. They abstracted the condensed
meaning units and labeled each with a code.
Although we can begin coding with a list of
concepts, we usually generate most coding themes
while reading data notes. Regardless of whether we
begin with a list of themes, we list themes afterfin-
ishing the open coding. Such a list serves three pur-
poses:

1. It helps to see the emerging themes at a glance.


2. It stimulates us to find themes in future open
   coding.


3. We can use the list to build a universe of all
   themes in the study, which we reorganize, sort,
   combine, discard, or extend in further analysis.

We vary in how completely and in how much detail

to code. Some researchers code every line or every

few words; others code paragraphs or pages. Some

of the data are not coded and are dross, or left over.

The degree of detail in coding depends on the

research question, the “richness” of the data, and

the research purposes (see Expansion Box 2, The

Process of Coding Qualitative Data).

Open-ended coding extends to analytic notes

or memos that you write to yourself while collect-

ing data. You should write memos on your codes

(see the later discussion of analytic memo writing).

Axial Coding.This is a “second pass” through the

data. During open coding, you focus on the actual

data and assigning code labels for themes. You are

little concerned about making connections among

themes or elaborating the concepts that the themes

represent. In contrast, you begin axial coding

with an organized set of initial codes or preliminary

Axial coding A second stage of coding of qualitative

data during which the researcher organizes the codes,

links them, and discovers key analytic categories.

“A good thematic code is one that captures the qual-

itative richness of the phenomenon. It is usable in the

analysis, the interpretation, and the presentation of

research” (Boyatzis, 1998:31). To code data into themes,

a researcher first needs to learn how “to see” or rec-

ognize themes in the data. Seeing themes rests on

four abilities: (1) recognizing patterns in the data,

(2) thinking in terms of systems and concepts, (3) hav-

ing tacit knowledge or in-depth background knowl-

edge (e.g., it helps to know Greek myths to

understand Shakespeare’s plays), and (4) possessing

relevant information (e.g., one needs to know a lot

about rock musicians and music to code themes

about a rock music concert) (see Boyatzis, 1998:7–8).

Three errors to avoid when coding (see Schwandt,

1997:17) are (1) staying at a descriptive level only (not

being analytic), (2) treating coding as a purely

mechanical process, and (3) keeping codes fixed and

inflexible. Codes have five parts: (1) a one- to three-

word label or name, (2) a definition with a main char-

acteristic, (3) a “flag” description of how to recognize

the code in the data, (4) any exclusions or qualifica-

tions, and (5) an example.

ILLUSTRATION OF FIVE PARTS

Label. Gender-role disputes.

Definition. Interpersonal verbal disagreements are

an example as are conflicts or disputes over what is

proper or acceptable behavior for males and females

in their interactions together or separately because

he or she is male or female.

Flag. An example is making sarcastic remarks or

jokes, or having disagreements (very mild to angry

arguments) over what a male or female should do

because he or she is male or female.

Qualifications. Only disputes among same gen-

dered persons are considered. Any type of behavior

(verbal or nonverbal) can be the target of a dispute.

Interactions among overtly homosexual and trans-

gendered persons are not included.

Example. Outside a classroom, Sara and Jessica,

16 years old, discuss their dates last night. Sara says,

“We went out for pizza—of course he paid.” Jessica

remarks, “Of course? You mean you expect the guy

to pay?” Sara answers, “Oh, forget it.”