IRC Resource Paper No.

IRC Resource Paper No. 1

2013

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 1

 IRC Resource Paper No. 1

CODING
an overview and guide to qualitative data analysis for integral researchers
Version 1.0

by Nicholas Hedlund-de Witt1

What is Coding?

• Coding is a key part of qualitative data analysis, some researchers even

considering coding to be synonymous with analysis. It is part of the
interpretive process of moving “from the data to the idea, and from the idea to
all the data pertaining to that idea” (Richards & Morse, 2007, p. 137).

• Coding is primarily a hermeneutic (or interpretive) act that represents the

transitional process between data collection and more extensive data analysis
(Saldana, 2009, p. 4).

o Examples of data: interview transcripts, participant-observation field

notes, journals, documents, literature, artifacts, photographs, video,
websites, email correspondence, et cetera.

• The word coding is of Greek origin, meaning “to discover.” Yet coding
simultaneously highlights the constructivist dimension of research and is thus
as much about enacting—through the epistemic profile and methodological
orientation of the researcher—as it is about discovering. As Saldana puts it,
“the act of coding requires that you wear your researcher’s analytic lens. But
how you perceive and interpret what is happening in the data depends on what
type of filter covers that lens” (Saldana, 2009, p. 6). Thus, a key practice for
integral scholars engaging the coding process is to be conscious and self-
reflexive with respect to one’s own epistemological lens and methodological
approach. See the below section on integral coding for more detail.

o As Adler and Adler (1987) highlight, your level of involvement or

participation with your subjects will also impact or filter how you
interpret and code your data. Likewise, your data will be impacted by
the types of questions that you ask and the types of responses you
receive during interviews (Kvale, 1996; Rubin & Rubin, 1995), the
detail and structuring of your field notes (Emerson, Fretz, & Shaw,
1995), the gender and race/ethnicity of your participants—and yourself

1
Partly abstracted from The Coding Manual For Qualitative Researchers by Johnny Saldana (2009).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 2

 IRC Resource Paper No. 1

(Behar & Gordon, 1995; Stanfield & Dennis, 1993), and whether you
collect data from adults or children (Greene & Hogan, 2005); Zwiers &
Morrissette, 1999).

• Coding is often looking for repetitive patterns or consistency (including the

pattern of idiosyncrasy). Coding may proceed by grouping things together
based on similarity, but also can proceed based on commonality, which
paradoxically, can consist of difference.

o According to Hatch (2002), we can think of patterns not just as stable

regularities, but also as varying forms. A pattern can be characterized
by:

• Similarity (things happen the same way)

• Difference (they happen in predictably different ways)
• Frequency (they happen often or seldom)
• Sequence (they happen in a certain order)
• Correspondence (they happen in relation to other activities or
events)
• Causation (one appears to cause another) (p. 155).

• Creswell (2007) notes that codes can emerge in response to not only expected
patterning, but also what you find to be striking, surprising, unusual or
conceptually captivating (p. 153).

• As you identify patterns and construct categories in the coding process, keep in
mind that, according to Tesch (1990), “a confounding property of category
construction in qualitative inquiry is that data within them cannot always be
precisely and discretely bounded; they are within ‘fuzzy’ boundaries at best
(pp. 135-8).

What is a Code?

• A code, in qualitative inquiry, is most often a word or short phrase that

symbolically assigns a summative, salient, essence-capturing, and/or evocative
attribute for a portion of language-based or visual data (Saldana, 2009, p. 3).

• A code is a lower level of data analysis on the way to labels, categories, themes,
and theory. As Charmaz (2006) puts it metaphorically, coding “generates the
bones of your analysis…. [I]ntegration will assemble those bones into a
working skeleton” (Saldana, 2009, p. 45).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 3

 IRC Resource Paper No. 1

• A code functions to capture a datum’s primary content and essence, somewhat

like a (good) title of a book or film.

• Saldana, 2009 states, “…be aware that a code can sometimes summarize or
condense data, not simply reduce it…depending on the researcher’s academic
discipline, ontological and epistemological orientations, theoretical and
conceptual frameworks, and even the choice of coding method itself, some
codes can attribute more evocative meanings to the data” (p. 4).

• Decoding: when we reflect on a passage of data to decipher its core meaning

• Encoding: when we determine a passage’s appropriate code and label it

(Saldana, 2009, p.4)

Coding is a cyclical or iterative act:

• With each subsequent coding cycle or iteration we are further managing,

filtering, ordering, highlighting, and focusing the “salient features of the
qualitative data record for generating categories, themes, and concepts,
grasping meaning, and/or building theory” (Saldana, 2009, p. 8).

Preliminary Jottings and Pre-coding:

• You don’t have to wait until all your data has been collected and assembled to
begin your preliminary coding.
o As you are transcribing interviews, writing up field notes, or filing
relevant documents, you can jot down preliminary phrases or words as
“analytic memos” in a research journal for future reference.
o Memory doesn’t always serve us as well as we might like! Make sure
that such analytic memos are distinctively marked so as to avoid mixing
them with the raw data.
o Researchers may also choose to “precode” (Layder, 1998) by
underlining, highlighting, circling, bolding, or coloring salient passages
that are deemed as striking.

First Cycle Coding Processes:

• The magnitude of the data coded can range from a single word to a full
sentence to an entire page of text, to a stream of moving images. First cycle can
be repeated numerous times before proceeding to Second Cycle coding. If
higher-level categories, concepts, or themes pop out at you, that is fine, just
make note of them in a separate analytic memo (as it can sometimes inform

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 4

 IRC Resource Paper No. 1

later analysis) and don’t muddle higher-level analysis with the first cycle (a
kind of ‘pre/trans fallacy’ in Wilber’s terms).

Second Cycle Coding Processes:

• In terms of portions, you can code the exact same units coded in First Cycle,
longer passages of text, or even a reconfiguration of the first cycle codes.
Second Cycle coding can be repeated numerous times, since coding demands
that researchers pay meticulous attention to the nuances of language and reflect
deeply on the emergent patterning embedded within the data.

• Some researchers argue that every detail of the raw field data should be coded,
while others feel that only aspects of the data that are deemed to be salient
should be coded and even up to half of the record can be summarized or
deleted all together. In my opinion, researchers should use caution when
deleting data, as what appears in one coding cycle to be irrelevant may turn out
to hold keys to unlocking the larger emergent pattern that may come forth in
subsequent coding cycles. Generally, beginner researchers may want to
consider erring on the side of conserving (and not deleting) data while there
“data sense” is developing.

• In subsequent coding cycles, initial codes can be refined, relabeled, subsumed

by other codes, or dropped all together.

• Abbott (2004) likens the cycles of coding to the process of decorating a room:
“you try it, step back, move few things, step back again, try a serious
reorganization, and so on” (p. 215).

Some Brief Coding Examples: (see the section “28 Coding Methods” below for a
more comprehensive and detailed list of coding methods)

• Descriptive Coding: summarizes the primary topic of the excerpt/datum.

• In Vivo Code: when a code is taken verbatim directly from the data and placed in
quotation marks. (e.g., I really feel inspired around him. Code “INSPIRED”).

• Initial Coding: an open-ended approach to coding in which the researcher codes

for their first ‘hit’ or impression words or phrases in response to engaging the
datum.

• Process Coding: a word or phrase that captures action/behavior.

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 5

 IRC Resource Paper No. 1

• Simultaneous Coding: the application of two or more codes within a single datum
(e.g., when one code refers to an embedded or interconnected part of the single
datum).

• Values Coding: captures and labels subjective-value perspectives.

• Holistic or “Lumper” Coding: a broad-brush stroke representation of a (relatively

large) datum, intended to capture the gestalt or essence (e.g., using one code to
represent an entire 200-word excerpt from an interview transcript).

• Splitter Coding: “splitter” coding can be contrasted with so-called Holistic or

“lumper” coding—splitting the data such that more codes can be applied (e.g.,
applying 10 codes to a 200-word excerpt from an interview transcript).
Lumping or splitting, can both be appropriate relative to the research context.

Questions to code for:

• One general approach to coding is to code for your central research

question/research concern.

o In one exemplary approach, Emerson, Fretz, and Shaw (1995) propose

a general list of questions to consider in coding field notes:

o What are people doing? What are they trying to accomplish?

o How exactly, do they do this? What specific means and/or
strategies do they use?
o How do members talk about, characterize, and understand
what is going on?
o What assumptions are they making?
o What do I see going on here? What did I learn from these
notes?
o Why did I include them? (p. 146).

Codifying and Categorizing:

• Codify: “to arrange things in a systematic order, to make something part of a

system or classification, to categorize” (Saldana, 2009, p. 8).

• Codification helps to group and link data so as to enact and consolidate

meaning and explanatory interpretation.

o Such analysis or codification “is the search for patterns in the data and
for ideas that help explain why those patterns are there in the first
place” (Bernard, 2006, p. 452).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 6

 IRC Resource Paper No. 1

• Categories are generated by grouping similarly coded data based on shared

characteristics.

• Categories can be descriptive or speak to more conceptual processes,

depending on your method.

Themes:

• A theme is the result of coding, analytic reflection and contemplation,

interpretive categorization and so forth, not some thing that is “coded” as
such. Generating themes involves a higher-order of data analysis that
should not be confused with the lower-order coding process (in the proper
sense).

Theory Development:

• Themes, or concepts, are progressively abstracted from the categories,

which in turn are based on codes. “Our ability to show how these themes
and concepts systematically interrelate lead toward the development of
theory” (Corbin & Strauss, 2008, p. 55).

• However, pre-established theoretical suppositions can inform, or even

drive, the initial coding process itself. As Mason (2002) contends,
researchers should admit and acknowledge that preexisting theories drive
the entire research enterprise, whether one is aware of it or not.

Quantities of Qualities:

• A typical qualitative research study might generate 80-100 codes, which will be
organized into 15-20 categories, which are ultimately synthesized into 5-7 major
themes/concepts.

Coding Software:

• Saldana (2009) recommends learning to code manually (pen and paper) before
delving into the world of coding software, such as CAQDAS. CAQDAS is
highly recommended for larger scale research projects (multiple participant
interviews, et cetera).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 7

 IRC Resource Paper No. 1

• 3 major CAQDAS programs:

o ATLAS.ti: www.atlasti.com
o MAX QDA: www.maxqda.com
o NVivo: www.qsrinternational.com

* Refer to Lewins & Silver (2007) and Baseley (2007) for supporting literature on
these programs.

• Other coding software:

o Express Scribe: http://www.nch.com.au/scribe/index.html

• Transcription Foot pedals: http://www.nch.com.au/scribe/pedals.html

Keeping a Codebook or Code List:

• As you code, Saldana (2009) recommends that you keep a record of your
emergent codes, their content descriptions, and a brief data example in a
codebook, separate file, or via a qualitative analysis software program such as
CAQDAS. It can be quite useful in the analysis process to be able to see all
your codes together in one place, without having to sort through your raw data
documentation.

Collaborative Coding and Member Checks:

• While most coding efforts will likely be engaged alone, researchers can bring
additional researchers into the process so as to widen the sphere of possibilities
in terms of interpreting and enacting the data. Multiple researchers may then
code the same raw data and then attempt to bridge or synthesize the spheres of
divergence and/or move towards interpretive convergence.

• Researcher can also bring their research subjects into the coding/analytical
process to varying degrees, actually inviting them to collaboratively code the
data, or to simply crosscheck your interpretations with theirs (so-called
“member checks”). Such member checks, or triangulation of interpretations in
the coding process, will tend to increase the validity of your knowledge claims.

A Priori Orientations in Coding:

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 8

 IRC Resource Paper No. 1

• In choosing the most appropriate coding methods for your particular study, you
should clarify your a priori orientations or goals.

o If your goal is to construct an analysis that can directly address

particular research questions, you may want to look into the Structural,
Provisional, Hypothesis, Protocol, and Elaborative Coding methods.

o If, however, your goal is to engage in a more open-ended, exploratory

inquiry or develop a new theory related to your research topic, then the
coding methods associated with classic or re-envisioned grounded
theory (In Vivo, Process, Initial, Focused, Axial, and Theoretical
Coding) may be better suited to your goals. (Note: the methods listed
with respect to each of these two broad orientations are merely
suggestive, by no means definitive or exhaustive.)

28 Established Coding Methods:

First Cycle Coding Methods: are preliminary coding methods that serve to organize
the raw data. Keep in mind that the following methods are can overlap in various
ways and can also be “mixed and matched,” so long as you are aware of what you are
doing and specify that in the method section of your final research study.

• Grammatical Methods:

o Attribute Coding: is the notation, typically used at the beginning of the

data set, rather than embedded with in it, of the basic descriptive
information such as: work setting (e.g., name, location), participant
characteristics or demographics, data format (e.g., interview, transcript,
field note, document), time frame, or other relevant variables. Attribute
coding is, in part, a way of addressing the location of a data set within
corpus. It is a way of documenting “cover” information about
participants, the site, or other components of the study. Analysis based
on attribute coding can reveal organizational, hierarchical, or
chronological patterning within the data.

o Magnitude Coding: “consists of and adds a supplemental alphanumeric or

symbolic code or sub-code to an existing coded datum or category to
indicate its intensity, frequency, direction, presence, or evaluative
content. Magnitude codes can be qualitative, quantitative, and/or
nominal indicators to enhance description. For example, this can show
up as applying various scales (likert, binary, etc.) to an existing code or
category.

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 9

 IRC Resource Paper No. 1

o Simultaneous Coding: The application of two or more distinct codes to a

single datum-passage (or the overlapping of two or more codes applied
to sequential data passages. Simultaneous coding is useful when the
data’s content is interpreted to have multiple significant meanings that
cannot be captured in a single code. For example, Simultaneous Coding
might be useful if you interpret a passage to have both (lower-level)
descriptive and (higher-level) inferential meaning.

• Elemental Methods: are primary coding methods for qualitative data analysis,
providing basic, focused filters which help lay the foundation for subsequent
coding cycles.

o Structural Coding: is a question-based coding method that applies

conceptual or content-based phrases representing a topic of inquiry or
specific research question used to frame the interview (or other data
set). Structural coding is particularly useful for studies employing
multiple participants, standardized or semi-structured data gathering
protocols, hypothesis testing, or exploratory inquiries used to gather
used to gather major categories or themes. Structural coding does not
rely on margined entries, but rather inserts a particular research
question into the page and then articulates a structural code above its
corresponding data-passage. Structural coding can also be employed as
a partially quantitative technique, by determining the frequency of
various codes across participants, for example.

o Descriptive Coding: is a straightforward coding method used to assign

basic, descriptive labels to data to provide an inventory of their topics.
It summarizes in a word or short phrase (most often as a noun) the basic
topic of a data-passage. “Topic” here is distinguished from content:
topic being that which is being talked about, whereas content is the
actual substance of the passage. Descriptive coding is often employed
as a first step in the data analysis and is considered a good technique for
beginners.

o In Vivo Coding: when a code is taken verbatim directly from the data and
placed in quotation marks. (e.g., I really feel comfortable around him.
Code “COMFORTABLE”). P. 3 In Vivo coding is particularly well
suited for extracting and highlighting “folk” or “indigenous” terms
(participant-generated words indicative of a group, culture, or sub-
cultures categories of meaning). In Vivo coding is associated with
grounded theory.

o Process Coding: codes for a word or phrase that captures action/behavior

in the form of a “gerund” (or “-ing” word). Simple observable behavior

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 10

 IRC Resource Paper No. 1

(e.g., running, drinking water, watching a movie) as well as more

broadly conceived “action” (e.g., struggling, negotiating, adapting) can
be coded via Process Coding. This method can be particularly useful
for identifying responses to certain situations or problems, for example.
Process coding is associated with grounded theory.

o Initial Coding (or “Open Coding” also known in earlier grounded theory
publications) as an open-ended approach to coding in which the
researcher codes for their “first impression” words or phrases in
response to engaging the datum (p. 4). (As such, all codes are held as
tentative and provisional formulations used to inform further iterative
coding cycles.) Initial Coding is a fundamental approach to coding that
is not formulaic, yet employs some broad guidelines. Initial Coding is an
opportunity for the researcher to begin to deeply reflect on the contents
of the data, break it down into discrete parts, and examine them for
similarities and differences. Initial Coding is an important method for
grounded theory studies, as its goal is to “remain open to all possible
theoretical directions indicated by your readings of the data” (Charmaz,
2006, p. 46). As such, the typically phenomenological practice of
“bracketing” or “epoché” can be appropriately coupled with this
method. Note that initial coding can also make use of other (typically
basic, descriptive) coding methods, such as In Vivo Coding or Process
Coding.

• Affective Methods: explore subjective qualities of human experience (e.g.,

emotions, values, conflicts, judgments) by directly labeling and naming those
experiences.

o Emotion Coding: simply labels the emotions or feelings recalled and/or

experienced by the participant, or inferred/interpreted by the researcher
about the participant. Deep insight into the participant’s perspectives,
worldview, or life conditions can be gained by looking at the emotional
dimension. Engaging such a coding method underscores the
importance of cultivating empathic resonance with your participants
and “reading” their non-verbal cues. These can be initially recorded as
analytic notes during or after an interview.

o Values Coding: codes for data that appears to reflect a participant’s

underlying values, attitudes, and beliefs that represent or reflect his or her
perspectives or worldview. Values coding is particularly useful for
studies that seek to explore the interior dimensions of their subjects.

o Versus Coding: identifies dynamics of conflict by labeling for various polar

tensions and binaries at play in a given context (e.g., “Teachers vs.
Standards). Often Versus Coding is engaged with an eye for power

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 11

 IRC Resource Paper No. 1

imbalances and dynamics relative to mutually exclusive binary

oppositions. Often used in critical ethnography and some forms of
action research oriented around issues of injustice.

o Evaluation Coding: focuses on how data can be analyzed such that the
relative merit of various programs or polices can be assessed or judged.
Often is used in the context of program, organization, policy, or
personnel evaluation. Evaluation coding often focuses on the patterning
of participant responses relative to attributes or factors that can be used
to assess quality or performance. Particularly apt for policy, critical,
action, organizational, and evaluation studies.

• Literary and Language Methods: draw on methods from established approaches

to literary and discourse analysis to explore various sociological, cultural, and
psychological constructs and realities.

o Dramaturgical Coding: approaches observations and interview narratives

as “social drama” in its broadest sense. Dramaturgical codes apply the
terms and conventions of character, play script, and production analysis
onto qualitative data. For example, a participant might be looked at as a
character in a social drama and analyzed in terms of the participant-
“actor” objectives, conflicts, attitudes, emotions, and subtexts.
Dramaturgical Coding is particularly appropriate for case studies that
seek to understand the intra- and interpersonal experiences, as well as
the qualities, perspectives, and drives of a participant. They can also be
well suited for narrative or arts-based presentational forms.

o Motif Coding: Applies folk literature’s symbolic elements (types [e.g.,

hero, fool, ogres, widows) and other common mythic motifs) as codes,
leading to an evocative approach to analysis. Can be particularly well
suited to narrative or arts-based presentational forms.

o Narrative Coding: incorporates literary terms as codes to discover the

structural properties of participant’s stories. Such literary terms
include: Type (of narrative), Genre, Purpose, Setting, Plot, Character,
Form, Point of view (first-person, second-person, third-person,
omniscient, witness, etc.), Elements, Spoken Features, etc. Particularly
useful in postmodern narrative approaches, such as critical/feminist
studies, as well as inquiries into identity development and
documentation of life course.

o Verbal Exchange Coding: H.L. (Bud) Goodall, Jr.’s unique ethnographic

data analysis (in the form of conversation), through reflecting on social
practices and interpretive meanings. This approach to coding involves
the verbatim transcript analysis and interpretation of the types of
conversation and personal meanings associated with key moments. The

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 12

 IRC Resource Paper No. 1

approach is used in “writing a story of culture” (Goodall, 2000, p. 121).

The researcher uses Goodall’s particular typology/continuum of five
forms of verbal exchange to identify the unit or units: 1) Phatic
Communion or Ritual Interaction; 2) Ordinary Conversation; 3) Skilled
Conversation; 4) Personal Narratives; 5) Dialogue.

• Exploratory Methods: are preliminary, exploratory assignments of codes to the

data before more refined coding cycles are developed and applied.

o Holistic Coding: or “Lumper” Coding, is a broad-brush stroke

representation of a (relatively large) datum, intended to capture the
“whole,” gestalt, or essence (e.g., using one code to represent an entire
145-word excerpt from an interview transcript). Holistic Coding can
be contrasted with many of the more analytical, approaches that could
be called “Splitter Coding”—splitting the data into smaller “parts” such
that more codes can be applied (e.g., applying seven codes to a 145-
word excerpt from an interview transcript). Holistic coding is generally
used as a preparatory approach (of “chunking” the data) prior to a more
detailed coding or categorization process through First or Second Cycle
methods. Holistic coding can be applied to a data-unit as small as half a
page in length or as large as an entire transcript or completed study.

o Provisional Coding: begins by compiling a predetermined “start list of set

codes prior to field work” (Miles & Huberman, 1994, p. 58). These
codes can be anticipated categories generated from a variety of potential
preparatory investigative matters, including literature reviews related to
the study, the studies conceptual framework and research questions,
previous research findings, pilot study fieldwork, the researchers own
phenomenological data, and researcher-formulated hypotheses or
hunches. As the actual data is collected, coded, and analyzed,
Provisional Codes can be reworked, modified, deleted altogether, or
expanded to include new, a posteriori codes. Provisional Coding is
well-suited for studies that build on or attempt to corroborate previous
research

o Hypothesis Coding: is a deductive approach (as opposed to the inductive

approach of grounded theory methods) based on the application of a
researcher-generated, predetermined list of codes onto qualitative data
specifically to assess or evaluate a researcher-generated hypothesis.
This is a particularly useful approach for assessing hypotheses relative
to the rules, causes, or explanations in the data. Hypothesis Coding can
also be used subsequent to more inductive coding methods from which a
“grounded” hypothesis was generated (e.g., as a Second-Cycle method).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 13

 IRC Resource Paper No. 1

• Procedural Methods: are prescriptive methods consisting of pre-established

coding systems or very specific ways of analyzing qualitative data.

o OCM (Outline of Cultural Materials) Coding: developed by social scientists

at Yale University in the mid-twentieth century as a topical index
(consisting of thousands of entries) for anthropologists and
archeologists based on categories of social life traditionally found in
ethnographic description. This can be a powerful approach for
ethnographic studies that seek to relate their findings to the broader
conceptual streams of ethnographic description.

o Protocol Coding: is a general approach to coding in which the researcher

codes the data according to some pre-established, standardized, or
prescribed system. This approach generally involves a comprehensive
list of codes and categories that the researcher applies after data
collection. Some approaches have specific protocols for the application
of codes to the data. This approach is useful for operating within the
sphere of the pre-understandings and assumptions of a given theoretical
(or meta-theoretical) construct. Keep in mind that the pre-established
coding scheme can be the result of a previous inductive research (in the
style of grounded theory). Researchers should be attentive to the
potential that the pre-existing scheme does not accommodate emergent
aspects of the generated data.

o Domain and Taxonomic Coding: is an ethnographic method for discovering

the cultural knowledge people use to organize their behaviors and
interpret their experiences (Spradley, 1980, pp. 30-31). This approach
seeks to clarify the tacit, informal “coding” schemes or categories of
meaning associated with particular cultures or sub-cultures. Such meta-
categories (domains and taxonomies) are generated from a first cycle
that identifies participant-generated “folk-terms” while recording them
verbatim. If no folk-terms are given, they are developed by the
researcher and are known as “analytic terms.” Such folk-terms are then
analyzed according to Spradley’s (1979) scheme of semantic
relationships and forms to generate a “folk taxonomy.”

Second Cycle Coding Methods: are more advanced ways of reorganizing and
reanalyzing the data coded through First Cycle methods. Second Cycle methods
involve the exploration of the interrelationships across multiple codes and categories so
as to develop a coherent synthesis of the data. Second Cycle coding serves to develop a
sense of thematic, conceptual, and/or theoretical organization and coherence from your
First Cycle codes.

• Pattern Coding: are “explanatory or inferential codes, ones that identify an

emergent theme, configuration, or explanation. They pull together a lot of

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 14

 IRC Resource Paper No. 1

material into a more meaningful and parsimonious unit of analysis. They are a
sort of meta-code. Pattern Coding is a way of grouping those summaries into a
smaller number of sets, themes, or constructs” (Miles & Huberman, 1994, p.
69). According to some theorists, Pattern Coding works well following a First
Cycle coding method, such as Initial Coding.

• Focused Coding: done subsequent to Initial Coding, searches for the most
frequent or significant Initial Codes to develop the categories that are deemed
to be the most salient, largely based on an evaluation of which Initial Codes
appear to generate the most analytic traction. Focused coding is associated with
grounded theory.

• Axial Coding: attempts to strategically reassemble data that were split via the
analysis of Initial Coding. Axial Coding relates categories to sub-categories and
defines the properties (i.e., characteristics or attributes) and dimensions (the
location of a property along a continuum or range) of each category. This
coding process is intended to develop constructs that inform the researcher
around “if, when, how, and why” something happens (Charmaz, 2006, p. 60).
This is a well-suited approach for grounded theory studies in particular.

• Theoretical Coding: seeks to uncover or generate the primary umbrella theme—

the central or core category—which “consists of all the products of analysis
condensed into a few words that seem to explain what ‘this research is all
about’” (Strauss & Corbin, 1998, p. 146). Such a central theme attempts to
systematically link all categories and subcategories to it. This central theme
functions as a core thread that runs through and connects the entire data
corpus. Theoretical coding is, in many respects, the culminating step in the
development of a grounded theory.

• Elaborative Coding: also known as “top-down” coding, attempts to develop pre-

existing theory further. In this method, “one begins with the theoretical
constructs from [a] previous study in mind. This contrasts with the coding one
does in an initial study (bottom-up), where relevant text is selected without
preconceived ideas in mind [to develop grounded theory]. In elaborative
coding where the goal is to refine theoretical constructs from a previous study,
relevant text is selected with those constructs in mind” (Auerbach &
Silverstein, 2003, p. 104). So at least two distinct yet related studies (one
completed and one in progress) are required for Elaborative Coding. This
coding method is most appropriate in the context of a second study that seeks
to elaborate on the major theoretical conclusions/findings of the first.

• Longitudinal Coding: is the attribution or interpretation of selected change

processes to qualitative data collected and compared across time. Beyond the
quantitative analysis of change over time (including the statistical increase,
decrease, constancy, etc.) of certain variables or interest, Longitudinal Coding

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 15

 IRC Resource Paper No. 1

looks primarily at qualitative increase, decrease, constancy and so on within data

gathered over time. This method reviews the data corpus categorically and
comparatively over time to assess whether change may have occurred in
participants/subjects. According to Saldana’s (2003) approach to Longitudinal
Coding, there are 7 descriptive categories that are first used to organize the
data into a matrix cell: 1) Increase and Emerge; 2) Cumulative; 3) Surges,
Epiphanies, and Turning Points; 4) Decrease and Cease; 5) Constant and
Consistent; 6) Idiosyncratic; 7) Missing. This method is particularly apt for
studies that explore change and development in individuals, groups, and
organizations through extended periods of time.

A Brief Example of a Grounded Theory approach to the phases of the coding

process:

1. Pre-coding
2. First Cycle: Initial coding
a. including In Vivo and Process coding;
3. Second Cycle: axial coding, theoretical coding

New and Hybrid Coding Methods:

Keep in mind that you can adapt, synthesize, or develop totally new coding methods to
suit the unique needs and goals of your research study. Don’t be afraid to be creative!
You only need to be methodologically transparent by clearly delineating each aspect of
your approach in the method section of your research report, drawing linkages to
existing approaches when appropriate.

Integral Coding (A Provisional Sketch): In relation to data analysis or coding, integral

researchers to date have done little to explore the ways in which integral principles
could be more explicitly and skillfully applied. The following represents a ‘first pass’
with respect to the formulation of what a very general integral approach to coding
might look like relative to the principles of meta-patterning, non-exclusion, and enactment.
With respect to the principle that I call meta-patterning, Integral Theory posits that
there are at least five key recurring meta-patterns, or “elements,” which constitute
essential aspects of any occasion: quadrants, levels, lines, states, and types (Wilber,
2006). In the context of coding, any of these five key recurring meta-patterns can be
used as an a priori or “protocol” heuristic for coding data (Saldana, 2009). These
elements of the AQAL model can be held, for purposes of coding, as major units of
organization and can be coded for in the data. For example, an integral researcher
could code raw data (e.g., a transcript) from a zone 3 interview, analyzing the raw data
in terms of quadrants/quadrivia (e.g., labeling passages that deal with Upper-Left
quadrant psychological content; Upper-Right quadrant behavioral content; Lower-
Left quadrant cultural content; or Lower-Right quadrant systems content). This kind

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 16

 IRC Resource Paper No. 1

of quadratic or quadrivial analysis (Fuhs, 2008) could be seen as a variation of “IMP

coding,” the novel coding method pioneered by Edward Kelly (2008) and Jeff Cohen
(2008) that makes use of all eight zones. One approach to applying the Integral
principle of non-exclusion within the context of the aforementioned 28 coding methods
might suggest the value of making use of multiple coding methods relative to a single
data pool. For example, one could choose a more inductive, grounded theory–oriented
coding method to generate a number of initial codes in the first cycle of coding and
then could apply a more deductive, theoretically structured approach to coding those
labels into categories, then categories into themes in the second cycle. At the very least,
being aware of the multiplicity of coding approaches available could allow Integral
Researchers to justify their choices and situate the partiality of their data analysis
therein. The principle of enactment can be applied to coding in terms of an emphasis on
reflexivity in service of situating the researcher’s interpretations and illuminating their
idiosyncratic elements. The application of the Integral principle of enactment vis-à-vis
coding highlights the epistemological and methodological orientation of the researcher
in participatory interaction with the data. Such an understanding of coding is
grounded in an Integral Post-Metaphysical perspective that reality (ontology) is not
simply a singular pre-given perception but can support multiple enactments
(methodology) by inquiring subjects (epistemology). In other words, ontological
objects are disclosed through the structures of the epistemological subject and the
methodological injunctions with which they are engaged (Hedlund, 2008, p. 46). As
such, self-reflexively exploring the structure of the epistemic “filter(s)” and “lens’”
through which data perspectives are enacted can be seen as a foundational element of
an Integral approach to coding. More specifically, reflexively referencing salient ways
in which the researcher’s epistemic lens is impacting their coding in the
discussion/interpretation of the data could be a powerful post-metaphysical move.
Such (self-)reflexive research into the epistemological structures of the researcher can
be used to engage a systematic method (e.g., the iterative-reflective method) for
mitigating the otherwise ubiquitous challenge of inter-individual variability. Integral
Research thereby holds the potential to bring more reflexive transparency, and thereby
integrity and rigor, to the overall research process. This might also serve to increase
the legitimacy of qualitative methods in particular, which remain somewhat
marginalized in the eyes of some traditional quantitative researchers. However, the
extent to which this would be useful depends on a number of contextual variables,
including the aims of the study, the audience in question, rhetorical strategy, and so on.

• Self-Reflexive Coding: As stated above, the word coding is of Greek origin,

meaning “to discover.” However, from an Integral perspective, coding is as
much about enacting, disclosing, and bringing data-perspectives into being as it
is about “discovering.” In this way, integral coding would highlight the
epistemological and methodological orientation of the researcher (their epistemic
profile) in participatory interaction with the data. Such an understanding of
coding is grounded in an Integral-Postmetaphysical perspective that reality
(ontology) and is not simply a pre-given perception but is, in part, disclosed
(methodology) by the inquiring subject (epistemology).

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 17

 IRC Resource Paper No. 1

o Within the context of coding, the qualitative researcher Johnny Saldana

(2009) proposes that, “the act of coding requires that you wear your
researcher’s analytic lens. But how you perceive and interpret what is
happening in the data depends on what type of filter covers that lens”
(p. 6). As such, self-reflexively exploring the structure of the
epistemological “filter(s)” and “lens’” through which data perspectives
are being enacted can be seen as a foundational element of an Integral
approach to coding.

o The following are a number of factors (organized according to Integral

Theory’s quadrants) to consider in engaging an Integral approach to
coding:

o Psychological-Enactive Factors in Coding:

 What is your felt-sense of some of the most significant ways

that your coding been influenced by the uniqueness of your
subjective being and history?

 What is your felt-sense of any subtle or hidden agendas in

your research (e.g., how it relates to your history and
background, your object relations, your shadow dynamics, et
cetera) that may be influencing how you are coding your
data?

 In what ways are your passions and/or spiritual intuitions

influencing your approach to coding?

 How is it driven by career ambitions? (Need to differentiate

UL from UR here).

 What are the horizontal structures of knowing that you, the

researcher, tend to inhabit (e.g., the Enneagram, Myers
Briggs, astrology, et cetera)? And How specifically do they
influence your data-interpretation in the coding process?

 What are the vertical structures of knowing (e.g., levels in

the various lines of development: cognitive, self/ego-identity,
morals, values, needs, interpersonal, emotional, spiritual, et
cetera) that you tend to inhabit help to shape and fashion the
data-perspectives that are generated through your coding
process?

o Cultural-Enactive Factors in Coding:

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 18

 IRC Resource Paper No. 1

 Paradigmatic inquiry: what is the shared “normal” scientific

paradigm I am operating from? In what ways might I tend to
disregard or pre-filter anomalous data-perspectives?

 Cultural inquiry: how does my cultural (European-American

rational “scientismic” worldview, Bay Area “415 paradigm”
postmodern) background influence my enactment of my
coding process? What are the norms and values within the
various sub-cultures you are embedded in and in what ways
are they influencing your approach to coding?

• Ex: Culture of Scientism influences research as a

tendency to deny or hide cultural and psychological
dimensions/factors influencing research—the
tendency to overlook this very conversation! Or
more subtly, the tendency to lend relatively more
credence to 3-p methods in mixed methods
research…to use 3-p as the privileged perspective in
mixed methods triangulation (SCTi is more valid than
hermeneutic assessment or even systematic
introspection).
• Ex 2: What are the norms, values, and assumptions
associated with the culture of the particular academic
setting you are embedded, such as JFKU? At JFKU,
there is a sense of embeddedness in a more qualitative
research paradigm, an open and avant-guard
sensibility, and even a spiritually driven inquiry.

o Behavioral-Enactive Factors in Coding:

 What are some of your habitual behavior patterns that

influence your approach to coding the data? (work habits,
organizational skills, attention to tracking details in data
collection and analysis)
 What are some salient cognitive biases that impact your
perception and how do they influence your coding of the
data? (See
http://en.wikipedia.org/wiki/List_of_cognitive_biases for a
list of cognitive biases.)
 What is your level of reading speed and comprehension and
how does that influence your relationship to the coding
process?
 What is your level of technological fluency and skill (e.g.,
with coding software, recording devises, et cetera) and how

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 19

 IRC Resource Paper No. 1

does that impact the data perspectives brought forth through

your coding?

o Sociological-Enactive Factors in Coding:

 How do the sociological dynamics of the educational system

you’re enmeshed in influence your research and approach to
coding? (See Bruno Latour’s “Science in Action” )

 How does your funding source(s) influence your approach to

coding the data?

 How do techno-economic factors (e.g., being enmeshed in a

industrial-capitalist economy) influence your approach to
research and coding?

 How do political dynamics (at the local, state/provincial,

national, or international level) influence your approach to
coding?

• 5 Element Coding: Any of the 5 key patterns, or “elements,” in Wilber’s AQAL

model can be used as evaluative heuristics for coding the data. The 5 elements,
of the AQAL model can be held, for purposes of coding, as major units of
organization and can be coded for in the data, both in terms of their presence
and/or absence in the data. Below I take each element and briefly
explore/sketch some possible ways of applying them in the coding process.

1. Quadrant Coding (Quadrants):

 Perspectival Coding (or Quadrivial Coding):

• Coding the content for the quadratic perspectives
showing up in the data-content (e.g., First-, Second-,
Third-person perspective, 4 quadrants, 8 zones). That is,
coding for the perspective that research subject(s) are
looking as.

 Dimensional Coding (Quadratic Coding):

• Coding the content for the quadratic dimensions showing
up in the data-content (e.g., 4 quadrants, 8 zones). That is
coding for the dimensions the research subject(s) are
looking at.

2. Complexity Coding (Levels):

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 20

 IRC Resource Paper No. 1

 Coding the content according to any number of level schemes,

such as Wilber’s altitudes, Cook-Greuter’s levels of ego-
development, Kegan’s Orders of Consciousness, or various
esoteric schemes [physical aspects, etheric aspects, emotional
aspects, mental aspects, spiritual aspects].

3. Intelligence Coding (Lines):

 Coding the content for the multiple intelligences at play in the

data. That is, coding the content for cognitive, emotional, moral,
or spiritual intelligence, to name some.

4. Consciousness Coding (States):

 Coding the content for the various states of consciousness

apparent in the data. That is, coding the content according to
various taxonomies of states (e.g., Gross [Waking], Subtle
[Dreaming], Causal [Deep Dreamless Sleep], Witnessing
[Turyia], and Nondual [Turyiatita]; Endogenous/Exogenous;
Temporary States/Stable State-Stage acquisitions).

5. Style Coding (Types):

 Coding the content for various horizontal styles or apparent

typological structures and dynamics (e.g., the Nine Enneagram
types and subtypes, Myers-Briggs, Masculine/Feminine
dynamics, et cetera.)

Nick Hedlund-de Witt, M.A., is a scholar-practitioner specializing in the intersection

of metatheory and the cultural and psychological dimensions of sustainability. As a
Ph.D. researcher studying under Roy Bhaskar at the University of London Institute of
Education, he is synthesizing Critical Realism and Integral Theory into a new
metatheoretical framework for integrative knowledge production and emancipatory
social research, while applying it to address global climate change. Nick has served as
adjunct professor at John F. Kennedy University, associate director of the Integral
Ecology Center, associate organizer of the biennial Integral Theory Conference, and
co-organizer of the International Critical Realism & Integral Theory Symposium. His
articles have appeared in journals such as the Journal of Integral Theory and
Practice, where he was guest editor of a special issue on Integral Research. He is
currently editing a book with Roy Bhaskar and Sean Esbjörn-Hargens
entitled Metatheory for the 21st-Century: Critical Realism and Integral Theory in
Dialogue (Routledge). Nick holds a Bachelor’s degree (Summa Cum Laude) in Culture,
Ecology, and Consciousness from the University of Colorado at Boulder, a Master’s in
Integral Psychology from John F. Kennedy University, and a (second) Master’s in

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 21

 IRC Resource Paper No. 1

Philosophy, Cosmology, and Consciousness from the California Institute of Integral

Studies. Through the course of his studies, he has received several academic honors
including the Jacob Van Ek Scholar Award and the Honors Graduating Senior
Scholarship Award for his undergraduate thesis. He currently lives in Amsterdam, the
Netherlands, with his wife and research collaborator Annick Hedlund-de Witt. Nick
can be reached at: [email protected].

Copyright © Nicholas Hedlund-de Witt, 2013, All Rights Reserved. 22