ANALYSIS OF QUALITATIVE DATA(i.e., where they occur), or their relationship to a
major topic of interest?
For example, when studying working-class
life, LeMasters could have divided the general issue
of marriage into subparts (e.g., engagement, wed-
dings). He could mark all notes involving parts of
marriage and then relate marriage to themes of sex-
uality, division of labor in household tasks, views
on children, and so on. When the theme reappeared
in different places, he could have made comparisons
to see new themes (e.g., men and women have dif-
ferent attitudes toward marriage).
In the example of historical research on the
Knights of Labor, I looked for themes related to
temperance. I looked for discussions of saloons,
drinking or drunkenness, and relations between the
movement and political parties that supported or
opposed temperance. Themes that clustered around
temperance included drinking as a form of recrea-
tion, drinking as part of ethnic culture, different reli-
gious views on drinking, and gender differences
regarding drinking.
Graneheim and Lundman (2003) used a pro-
cess of axial coding in their study of interview data
on diabetes. They compared codes based on differ-
ences and similarities and sorted them into six sub-
categories and three categories. The two researchers
discussed tentative categories and revised them.
A process of reflection and discussion resulted in
agreement about how to sort the codes. Finally, the
researchers identified underlying meaning—that is,
the latent content—of the categories that they for-
mulated into a broader theme.
Axial coding not only stimulates thinking
about linkages between concepts or themes, but also
raises new questions. It can suggest dropping some
themes or examining others in more depth. It also
reinforces the connections between evidence and
concepts. As you consolidate codes, you may find
evidence in many places for core themes and build
a dense web of support in the qualitative data for
them. This is analogous to the idea of multiple indi-
cators described with regard to reliability and mea-
suring variables. The connection between a theme
and data is strengthened by multiple instances of
empirical evidence.^5
When I studied the Knights of Labor, I made
the movement’s failure to form alliances with other
political groups a major theme. I reviewed notes
looking for compromise and conflict between the
Knights and other political parties, including tem-
perance groups and the Prohibition Party. The array
of concepts and themes related to temperance in
axial coding helped me to see how the temperance
issue facilitated or inhibited alliances.Selective Coding.By the time you are ready for
this last pass through the data, you have identified
the major themes. Selective codinginvolves scan-
ning all the data and previous codes, looking selec-
tively for cases that illustrate themes, and making
comparisons after most or all data collection has
been completed. Selective coding should begin after
concepts have been well developed and several core
generalizations or ideas have been identified.
For example, as LeMasters studied working-
class life in a tavern, he decided to make gender rela-
tions a major theme. In selective coding, he could
have gone through his field notes, looking for dif-
ferences in how men and women talked about dat-
ing, engagements, weddings, divorce, extramarital
affairs, or husband/wife relations. He could then
compare male and female attitudes on each part of
the theme of marriage.
Graneheim and Lundman (2003) may have
used selective coding in their study of interview data
on diabetes. They provided readers of their study
examples of codes, subcategories, categories, and a
theme taken from text narratives about hyper-
glycemia, offering very explicit examples of each.
During selective coding, major themes or con-
cepts ultimately guide the search process. You reor-
ganize specific themes identified in earlier coding
and elaborate more than one major theme. For
example, in the working-class tavern study, LeMas-
ters could have examined opinions on marriage to
understand both the theme of gender relations
and the theme of different stages of the life cycle.Selective coding The last stage in coding qualitative
data that examines previous codes to identify and
select data that will support the conceptual coding cat-
egories that were developed.