532435

research-article2014
QIXXXX10.1177/1077800414532435Qualitative Inquiry St. Pierre and Jackson

Introduction
Qualitative Inquiry

Qualitative Data Analysis After Coding

2014, Vol. 20(6) 715719
The Author(s) 2014
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1077800414532435
qix.sagepub.com

Elizabeth A. St. Pierre1 and Alecia Y. Jackson2

In 1991, Patti Lather (1991) called data analysis the black multiple coders (to achieve interrater reliability) (Miller,
hole of qualitative research (p. 149), and, as co-editors of 2009, p. 170)coders who may have no knowledge either
this special issue on qualitative data analysis after coding, of the studys participants or the theoretical context of the
we suspect it still is. In fact, we think analysisthinking studyhas become a fetish, a superficial marker of a posi-
with theory (Jackson & Mazzei, 2012)is so difficult to tivist scientism that appeals to certain social scientists who
describe and explain to the non-positivistand to teach to demand systematicity, however mindless, in the name of
our studentsthat we have resorted to equating qualitative science.
data analysis with coding data. In other words, we teach There are several problems in equating qualitative data
analysis as coding because it is teachable. analysis with coding, and we will briefly discuss some of
The critiques qualitative methodology endured during those below. Of course, it is impossible to separate a discus-
the debates about scientifically based research in the first sion of data analysis from a discussion of what counts as
decade of the 21st century surely intensified its already con- data and how data are collected. With regard to the latter
fused epistemological and ontological commitments. The in what St. Pierre (2011) called conventional humanist
incommensurability in this methodology is that a social sci- qualitative inquiry (p. 611), we collect data (words) by
ence approach that claims to be interpretive supports a posi- interviewing and observing people. Interviewing has
tivist, quasi-statistical analytic practicecoding datathat become the predominant method of data collection in quali-
has, unfortunately, been proliferated and formalized in too tative research, and the authentic voices of participants are
many introductory textbooks and university research hallowed, treated reverently by researchers, as if their
courses. A question we might ask at the outset is whether wordssupposedly uncontaminated by theoretical inter-
one would code data if one had not been taught to do so. pretationcan serve as a foundation of knowledge (for a
We should state here at the beginning that we are not critique, see Jackson & Mazzei, 2009). Too often, however,
referring to the kind of analysis MacLure (2008) described the quality of data in that approach is taken for granted, and
as follows: data collection is an end in itself without sufficient theo-
retical analysis to determine appropriate criteria (Young,
I enjoy that part of the research process that involves poring 1969, p. 490) for which data to collect. Too often, we use
over the data, annotating, describing, linking, bringing theory the vacuum cleaner approach to data collectionsweeping
to bear, recalling what others have written, and seeing things up any and all data into our studies thereby ignoring the
from different angles. I like to do it manually too, with paper quality of the data base (Cicourel, 1979, p. 172). In this
and pen, scribbling a dense texture of notes in margins and
approach, all data are equal and worthy of analysis.
spilling over onto separate pages. (p. 174)
However, we recommend instead using theory to deter-
mine, first, what counts as data and, second, what counts as
Rather, we are concerned about analysis that treats words
good or appropriate data. In that way, we can do more
(e.g., participants words in interview transcripts) as brute
with less data, as P. Lather (personal communication,
data waiting to be coded, labeled with other brute words
1993) recommended, focusing on the difficult work of anal-
(and even counted), perhaps entered into statistical pro-
ysis rather than on conducting more and more interviews.
grams to be manipulated by computers, and so on. In some
Of course, collecting data presumes weve already deter-
cases, words are reduced to numbers.
mined what counts as data. As we mentioned earlier, the
We argue that coding data in that way is thinkable and
doable only in a Cartesian ontological realism that assumes 1
University of Georgia, Athens, USA
data exist out there somewhere in the real world to be found, 2
Appalachian State University, Boone, NC, USA
collected, and coded using the Cartesian principle of
breaking down the difficulty into as many parts as may be Corresponding Author:
Elizabeth A. St. Pierre, Department of Educational Theory and Practice,
necessary for finding the solution (Derrida, 1967/1978, Womens Studies Institute, University of Georgia, 604E Aderhold Hall,
p. 287). The intricate manipulation of a thicket of impene- Athens, GA 30602, USA.
trable data, the epitome of which is blind coding by Email: [email protected]
716 Qualitative Inquiry 20(6)

empiricism that guides conventional humanist qualitative analysis with practices of formalization (Pascale, 2011,
inquiry is grounded in the Cartesian dualism, subject/object, p. 17), such as coding, that are supposed to guarantee rigor
that implies the existence of an object separate from and and validity but instead encourage the decontextualization
independent of the collecting subject, a brute datumpure and fragmentation of interview discourse into codable ele-
data that exist in an external reality waiting to be collected ments (Nespor & Barylske, 1991, p. 810).
and analyzed (see, for example, St. Pierre, 2013). In con- There is something nonsensical in this practice of coding
ventional humanist qualitative research, words in interview if one thinks about it, because in the logic of coding, words
transcripts and in field notes are considered primary data, can be both databrute, waiting to be interpretedand
collected as they are in face-to-face encounters in the codemeaningful. Even if one accommodates that incom-
presence of participants in their natural settings. Again, mensurabilitywhich coders dosuch a project is think-
words spoken by participants are privileged regardless of able only in logical positivism that presumes that language
their adequacy to respond to the studys substantive and can, indeed, be brute and value-free. To restate, it is logical
theoretical demands. And, given Derridas critique of pres- positivism that claims a word can be brute data. Other social
ence, we are suspicious of that privileging. If a researcher is science approaches such as interpretive, critical, and post
studying prisons, for example, why would she not include approaches assume that language is always already entirely
as data Foucaults (1975/1979) words in his book, Discipline contaminated by meaning, exploding with meaning
and Punish: The Birth of the Prison, which is one of the deferred. In those approaches, language cannot be and
most famous studies of prisons? Why are Foucaults words never has been brute, and so, as Popkewitz (2004) explained,
consigned to the literature review? Is it because they are there are no data without theory that orders and gives clas-
written, textualized? However, that criteria cannot hold sification to the things of the world (p. 72).
because words spoken in face-to-face interviews do not To code data, then, one must assume that words textual-
count as data until they are written, textualized in interview ized in interview transcripts and field notes are not only
transcriptsuntil they have lost their presence. In fact, data but also brute data that can be broken apart and decon-
words can never retain presence. textualized by codingeven using existing coding schemes
In the end, both Foucaults and the participants words from others research projects. Once coded, words can be
are written, not present, so the scientific rationale for why sorted into categories and then organized into themes
some words and not others count as data escapes us. Again, that somehow naturally and miraculously emerge as if
this problem speaks to the quality of the database because anyone could see them. However, as Pascale (2011)
we can imagine that Foucaults wordsproduced by a warned, this is the language of scientific discovery, the
complex theoretical analysismight be more significant language of Cartesian dualism (p. 105). Furthermore,
and useful than an off-the-cuff, uninformed comment by a social scientists have learned the fallacy of peoples ten-
participant. Of course, this is not to say that participants dency to look for patterns when none exist (Goldstein,
dont have many interesting things to tell us, else we 2011, p. 81). In other words, if you think you have to find
wouldnt talk with them. Still, we insist on counting words a theme, you probably will. In the end, the result [of cod-
as data, and further, if were concerned with rigor and qual- ing] is an elaborate presentation of data whose purposes are
ity, we argue that Foucaults words should count as much as unclear and whose utility is undermined by the absence of
participants words during analysis. Fortunately, we dont an adequate conceptual foundation (Young, 1969, p. 489).
yet know of anyone who codes Foucaults words in his or This may be why those who engage in an elementary clas-
her texts along with participants words in interview tran- sification using low-level empirical materials (Cicourel,
scripts, and thats a good thing, given that were not in favor 1979, p. 170) resort to counting codestreating words as
of coding. Nonetheless, we are left with this problem of numbersand also to producing extended, superficial
why some words count as data and others dont in a particu- descriptions or stories in which the data speak for them-
lar kind of empirical research. selves. Having opted out of a theoretical analysis, these
So before we begin to discuss what counts as analysis in researchers have nothing much to say and, often, they are
conventional humanist qualitative inquiry, weve identified too exhausted from months of coding to theorize at all. In
at least two interlocking problems with what counts as data fact, we have learned in our own teaching that coding data
and how those data are collected. The first is that using can be an excuse not to read theory.
presence as a criterion for quality, we assume that data col- The ontologies and epistemologies that enable one to
lected face-to-face from participants are of high quality and code data as described above are, of course, unthinkable in
worthy of collection and analysis, thereby counting some the posts (e.g., post-positivism, post-subjective, post-
words as data but not others. Second, weve assumed an humanism, post-modernism, post-structuralism, post-
empirical reality in which those words exist as brute data foundationalism, post-empiricism/materialism) that use
independent of the interpretive desires of the data collec- different ontologies that describe the world as unstable
tor. These problems are exacerbated if we equate data and becoming (Deleuze & Guattari, 1980/1987); different
St. Pierre and Jackson 717

epistemologies that describe meaning, too, as unstable and We conclude with a comment from Barad (2007) who
unable to be contained in language (Derrida, 1967/1974); wrote, The space of possibilities does not represent a fixed
and different onto-epistemologies that do not separate event horizon . . . nor does it represent a homogeneous,
knowing and being but describe the world as intra- fixed, uniform container of choices (p. 246). These articles
actionsor entanglementsof subject and object, human offer ways of responsibly imagining and intervening in the
and non-human (Barad, 2007). configurations of power (p. 246) of a positivist scienticism
Given that the posts have been available for more than that has, for too long, equated qualitative data analysis with
half a century now and given that their theoretical coding data.
approaches have been used in qualitative research for at Svend Brinkman leads the special issue by troubling
least three decades, one might ask how it is possible that what counts as data. Rather than saying, on the one hand,
analysis in qualitative methodology continues to be mired that data are givens that we collect and code (induction)
in positivismparadigms behind (Patton, 2008, p. or, on the other, that everything is data (deduction), he sug-
269)almost untouched by the posts. We respond to this gests we think of data as any material we use to think about
gap between theory and methodology with this special issue an astonishment or breakdown in ones understanding
on qualitative analysis after coding. We asked contributors (abduction) of life events, big or small. In this approach,
to describe what they do, other than code data, when theyre there can be no line between life, research, theory, and
doing something they think is analysis. Their responses methods because research is part of the life process.
include a variety of interpretive, critical, and post-social Jerry Rosiek and Julia Hefferman explain that atheo-
science approaches and illustrate what post-coding analysis retical approaches to data analysis such as coding are inad-
looks like in post-qualitative inquiry that does not rely on equate for analyzing socially produced silences. In their
positivist techniques or methods. study of the social erasure of a middle school girl who does
As readers will see, there is no recipe for this kind of anal- not conform to local gender norms, they use feminist and
ysisfor thinking with theorybecause one has to first read queer post-structural theories to explain the meaningfulness
and study theory carefully and then put it to work in a particu- of the social silences and the power dynamics at work in her
lar project. Discussing his power/knowledge analysis, school.
Foucault (2000), for example, explained, What Ive written Bronwyn Davies uses Karen Barads concept, diffrac-
is never prescriptive either for me or for othersat most its tion, to show how the new materialism can be put to work
instrumental and tentative (p. 240). Following Foucault, the in analysis. Looking at anger in early childhood intra-
authors who contributed to this special issue do not follow a actions, she examines the entanglement of research prob-
particular analytic method; rather, they borrow concepts, lem, concepts, emotions, transcripts, memories, and images
invent approaches, and create new assemblages that demon- in coming to know something differently. In this analysis,
strate a range of analytic practices. Post-coding analysis, there are no methodological steps to be followed, and the
then, can be thought as non-technique and non-method that is analytic process cannot be defined in advance. Rather,
always in a process of becoming as theories interlink, inten- entanglement affects not only what it is possible to see but
sify, and increase territoryspreading out and overturning also what it is possible to be and do, epistemologically,
the very codes that structure [them] . . . putting them to ontologically, and ethically.
strange new uses (Deleuze & Guattari, 1980/1987, pp. 11, Lisa A. Mazzei also uses Karen Barads concept, dif-
15). So the work of post-coding analysis cannot be neat, tidy, fraction, and Deleuze and Guattaris concept, desire, to
and contained. Furthermore, it cannot be easily explained think with theory in analyzing interview data. Noting that
either during or after analysis. It certainly cannot be repli- coding takes us back to what is known, she argues for dif-
cated because it is emergent and experimental. In addition, its fraction that takes a rhizomatic form and leads the analyst in
spacetime cannot be secured in the traditional linear pro- different directions, keeping analysis and knowledge pro-
cess trajectory of data collection>analysis>representation. duction on the move by plugging data into theory into data
In sum, these articles problematize not only what consti- as they constitute each other.
tutes qualitative data analysis but also where and when it Sharon Augustine describes how her nave faith in
happens. Following Deleuze and Guattari (1980/1987), we both the theories that framed her study and the writing pro-
might say that post-coding analysis functions like a rhizome cess itself enabled her to abandon the coding in which shed
(instead of a tree) in that, as a non-method, it is open and been trained and to use the Deleuzo-Guattarian concept,
connectable in all its dimensions; it is detachable, reversible, assemblage, in an interview study with expert readers to
susceptible to constant modification (Deleuze & Guattari, study how reading can become a force that enables complex
1980/1987, p. 12). Post-coding analysis occurs in the mid- relationships with texts. She tracks how she managed data,
dle of things (Deleuze & Guattari, 1980/1987, p. 293), how she analyzed data in memo-writing, how that analysis
without a beginning or end, without origin or destination. In flew off in all directions, and how she abandoned initial
this way, analysis occurs everywhere and all the time. theories and found others.
718 Qualitative Inquiry 20(6)

Teri Holbrook and Nichole M. Pourchier call them- that resists inquiry as invasion and offer three axioms that
selves a/r/tographers who see their artist/researcher/teacher ground their work, providing pedagogical strategies that
practices as simultaneous and non-hierarchical. In their use art and literature as resistance narratives.
work, they use theory, writing, and collaging in a rigorous Sara Childers discusses her use of Foucault and critical
analytic articulation process in which sense, rather than race theory to assemble what she names promiscuous fem-
meaning, is tentatively fabricated. Using Deleuze and inist research. She explains how she used seemingly
Guattaris conceptsfolding and assemblageand their incommensurate theories during her fieldwork at a high-
own conceptshoarding and musteringin analysis, they achieving, high-poverty urban school. Through her analy-
describe the process they used to create a collaged text, the sis, she demonstrates the breaking and blurring of both
text itself, and their exegesis of the text. theoretical and methodological boundaries that emerge to
Hillevi Lenz Taguchi and Anna Palmer use the Deleuzo- create a promiscuous method that is defiant and deliberate.
Guattarian methodology of cartography to study young Melissa Freeman explains that Clifford Geertzs con-
school girls school-related ill health and well-being. Various cept, thick description, has been used at the end of qualita-
kinds of data from medicine, psychology, popular science, tive studies in the service of representation. Here, she draws
and media, as well as narrative data from the girls and the two on Gadamers philosophical hermeneutics to argue that
researchers themselves, constitute a machinic assemblage of thick description is an analytic used to articulate the condi-
public health in Sweden. Their analysis shows what kind of tions in which understanding takes place. Examining how
Bodies without Organs are fabricated and what kinds of sub- philosophical hermeneutics was used in a study of street
jectivity are possible in this assemblage. kids, she explains that its intention is not to describe typical
Jessica Ringrose and Emma Renold focus on their lived experience in thick description but to foster new ques-
research with teen girls to illustrate the affective intensities tions that loosen the grip tradition holds on our imagina-
that propel their analytic work in a Baradian, Deleuzian tions and demand that we consider how the topic we
framework. They argue that analysis is always already study is thinkable.
entangled with complex affective ethical and political rela-
tionalities that circulate in, through, and outside empirical Declaration of Conflicting Interests
research. Drawing from methodological work by Maggie The author(s) declared no potential conflicts of interest with
MacLure, they locate data hot spots that allow a particu- respect to the research, authorship, and/or publication of this
lar analytical mapping of the multiplicities of the discur- article.
sivematerial category slut.
Rachel Holmes offers a fleshy analysis of decom- Funding
posing ethnography and data. She uses Deleuzes event to The author(s) received no financial support for the research,
illustrate the analytic work of destroying cohesion as she authorship, and/or publication of this article.
composes and decomposes a playground observation and
the making of an experimental film. Using the disciplines of
References
art, philosophy, education, and science, Holmes conducts a
repetitious analysis that creates new ways to think about Barad, K. (2007). Meeting the universe halfway: Quantum physics
and the entanglement of matter and meaning. Durham, NC:
relations among bodies, (in)organic matter, and time.
Duke University Press.
Kelly Clark/Keefe introduces somatographic analysis
Cicourel, A. V. (1979). Field research: The need for stronger the-
as a practice of deeply attuning creativity to materialdis- ory and more control over the data base. In W. E. Snizek, E.
cursive data of first-generation college students and aca- R. Fahrman, & M. K. Miller (Eds.), Contemporary issues in
demic women, as well as different art mediums, to move theory and research (pp. 161-176). Westport, CT: Greenwood
away from the analytic question, What does the data Press.
mean? to explore the question, What does the data do? Deleuze, G., & Guattari, F. (1987). A thousand plateaus:
Stephanie Daza and Francyne Huckaby disrupt the Capitalism and schizophrenia (B. Massumi, Trans.).
mind/body dualism that permeates much of analytic work Minneapolis: University of Minnesota Press. (Original work
to offer instead an embodied practice that links both physi- published 1980)
cal and metaphysical acts. Naming their analysis terra Derrida, J. (1974). Of grammatology (G. C. Spivak, Trans.).
Baltimore, MD: The Johns Hopkins University Press.
incognita, they illustrate how they simultaneously use their
(Original work published 1967)
bodies and minds as analytical terrain.
Derrida, J. (1978). Writing and difference (A. Bass, Trans.).
K. Wayne Yang and Eve Tuck use the role of refusal Chicago, IL: The University of Chicago Press. (Original work
as a generative stance to conduct analytic resistance to set- published 1967)
tler regimes of representation. They argue that settler colo- Foucault, M. (1979). Discipline and punish: The birth of the
nialism seeks authentic, real research that pursues stories prison (A. Sheridan, Trans.). New York, NY: Vintage Books.
of pain and humiliation. They detail an analysis of refusal (Original work published 1975)
St. Pierre and Jackson 719

Foucault, M. (2000). Power: Essential works of Foucault, 1954- Popkewitz, T. S. (2004). Is the National Research Council commit-
1984 Vol. III (R. Hurley, Trans., & J. D. Faubion, Ed.). New tees report on Scientific Research in Education scientific?:
York, NY: The New Press. On trusting the manifesto. Qualitative Inquiry, 10, 62-78.
Goldstein, E. R. (2011, November 18). The anatomy of influence. St. Pierre, E. A. (2011). Post qualitative research: The critique and
Chronicle of Higher Education, LVIII(13), B5-B10. the coming after. In N. K. Denzin & Y. S. Lincoln (Eds.),
Jackson, A. Y., & Mazzei, L. A. (2009). Voice in qualitative inquiry: SAGE handbook of qualitative inquiry (4th ed., pp. 611-635).
Challenging conventional, interpretive, and critical concep- Los Angeles, CA: SAGE.
tions in qualitative research. New York, NY: Routledge. St. Pierre, E. A. (2013). The appearance of data. Cultural Studies
Jackson, A. Y., & Mazzei, L. A. (2012). Thinking with theory in Critical Methodologies, 13, 223-227.
qualitative research: Viewing data across multiple perspec- Young, O. (1969). Professor Russet: Industrious tailor to a naked
tives. London, UK: Routledge. emperor. World Politics, 23, 486-511.
Lather, P. (1991). Getting smart: Feminist research and pedagogy
with/in the postmodern. New York, NY: Routledge. Author Biographies
MacLure, M. (2008). Classification or wonder? Coding as an
Elizabeth A. St. Pierre is professor in the educational theory and
analytic practice in qualitative research. In R. Coleman & J.
practice department and affiliated professor of both the
Ringrose (Eds.), Deleuze and research methodologies (pp.
Interdisciplinary Qualitative Research Program and the Womens
164-183). Edinburgh, UK: Edinburgh University Press.
Studies Institute at the University of Georgia, USA. Her work
Miller, J. (2009). The status of qualitative research in criminology.
focuses on theories of language and the subject in what she has
In M. Lamont & P. White (Eds.), Workshop on interdisciplin-
called post-qualitative inquiry or post-inquirywhat might come
ary standards for systematic qualitative research (pp. 69-75).
after conventional humanist qualitative research methodology
Washington, DC. National Science Foundation. Retrieved from
especially on the new empiricisms/new materialisms enabled by
http://www.nsf.gov/sbe/ses/soc/ISSQR_workshop_rpt.pdf
the ontological turn.
Nespor, J., & Barylske, J. (1991). Narrative discourse and teacher
knowledge. American Educational Research Journal, 28, Alecia Y. Jackson is an associate professor of educational
805-823. research and affiliated faculty of womens studies at Appalachian
Pascale, C.-E. (2011). Cartographies of knowledge: Exploring State University, North Carolina, USA. Her research interests
qualitative epistemologies. Los Angeles, CA: SAGE. bring feminist, post-structural, and post-humanist theories of
Patton, C. (2008). Finding fields in the field: Normalcy, risk, power, materiality, and subjectivity to bear on a range of overlap-
and ethnographic inquiry. International Review of Qualitative ping topics: deconstructions of narrative and voice, ontological
Research, 1, 255-274. studies of girlhood, and qualitative data analysis in the posts.