r Academy of Management Review

2021, Vol. 46, No. 3, 591–612.

https://doi.org/10.5465/amr.2018.0482

ACHIEVING RIGOR IN QUALITATIVE ANALYSIS:

THE ROLE OF ACTIVE CATEGORIZATION IN
THEORY BUILDING
STINE GRODAL
Northeastern University

MICHEL ANTEBY
Boston University

AUDREY L. HOLM
Boston University

Scholars have long debated how rigor can be achieved in qualitative analysis. To answer
this question, we need to better understand how theory is generated from data. Qualita-
tive analysis is, at its core, a categorization process. Nevertheless, despite a surge of inter-
est in categorization within the social sciences, insights from categorization theory have
not yet been applied to our understanding of qualitative analysis. Drawing from categori-
zation theory, we argue that the movement from data to theory is an active process in
which researchers choose between multiple moves that help them to make sense of their
data. In addition, we develop a framework of the main moves that people use when they
categorize data and demonstrate that evidence of these moves can also be found in past
qualitative scholarship. Our framework emphasizes that, if we are not sufficiently reflex-
ive and explicit about the active analytical processes that generate theoretical insights,
we cannot be transparent and, thus, rigorous about how we analyze data. We discuss the
implications of our framework for increasing rigor in qualitative analysis, for actively
constructing categories from data, and for spurring more methodological plurality with-
in qualitative theory building.

Qualitative analysis is a central tool for developing rigor, then, qualitative scholars need to detail more ef-
new theory (Edmondson & McManus, 2007; Eisen- fectively “the actual strategies used for collecting, cod-
hardt, 1989). In recent years, there has been a call for ing, analyzing, and presenting data when generating
increasing the rigor of qualitative research (Lamont & theory” (Glaser & Strauss, 1967: 244). When conduct-
White, 2008; Lubet, 2017; Pratt, Kaplan, & Whittington, ing qualitative analysis, we identify categories in our
2020; Small, 2013). This debate on rigor has led some data. These categories are generally labeled “codes” or
organizational scholars to ask, more specifically, “How groupings of codes, such as the first- and second-order
can inductive researchers apply systematic conceptual codes and overarching categories described in classical
and analytical discipline that leads to credible inter- grounded theory (Strauss & Corbin, 1990). These cate-
pretations of data and also helps to convince readers gories generate the concepts and mechanisms that
that the conclusions are plausible and defensible?” form the foundation for theory building (Eisenhardt,
(Gioia, Corley, & Hamilton, 2013: 15). Scholars often 1989; Yin, 2003).
assess rigor in qualitative research by examining quali- Given the centrality of categorization to qualitative
tative analysts’ descriptions of how they moved from analysis, it is surprising that we have not yet drawn
data to theory (Bansal & Corley, 2011). To demonstrate more purposely on categorization theory to penetrate
the challenges faced by qualitative scholars. A long-
standing and vibrant line of inquiry among psychol-
We would like to thank associate editor Heather Haveman ogists, sociologists, and management scholars has
and the three anonymous reviewers for their guidance, as well focused on understanding how humans construct
as Beth Bechky, Christine Beckman, Julia DiBenigno, Karen
Golden-Biddle, Gerardo Okhuysen, Siobhan O’Mahony, Melis-
and categorize the components of their world
sa Mazmanian, John Van Maanen, and Mark Zbaracki for their (Bingham & Kahl, 2013; Bloom, 2000; Khaire &
feedback on this paper. Wadhwani, 2010; Lamont & Molnar, 2002; Murphy &
591
Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder's express
written permission. Users may print, download, or email articles for individual use only.
592 Academy of Management Review July

Lassaline, 1997; Rosch, 1978; Vergne & Wry,

Relating Sequencing or dropping Contribution to theory

theoretical insights

Creating a theoretical
challenge existing
categories to past
2014). Categorization is the process through which

development

Linking tentative
individuals group elements together to generate an

categories to
Linking initial

elaborate or
understanding of their world (Bowker & Star, 2000).

theories

scaffold
Importantly, individuals have been shown to actively
construct categories based on the existing knowledge
and the intentions they bring to the categorization
process (Berger & Luckmann, 1967; Searle & Willis,

Developing

categories categories hypotheses

1995). Yet, often this active categorization process is
not evident in how qualitative researchers report their
research, which makes it difficult for readers to assess
the researchers’ analytical process. Instead, many
qualitative researchers draw on seemingly proven
templates when describing their work even though

Note: Within the table, the darker the shading, the more attention is paid to that particular move during a given analytical stage.
the structure of these templates might be a far cry
from the researchers’ actual analytical process.

An Active Categorization Framework for Theory Development

Bringing categorization theory into the debate on
how qualitative scholars achieve rigor allows us to
reflect on the active role that we as researchers play
in the construction of the categories critical to theory

Researchers’ active moves

categories
building. Such reflexivity allows us to articulate ex-

Splitting
actly what we did so that others can better under-
stand how we moved from data to theory.
Oftentimes, “we do not really know how the re-
categories
searcher got from 1,000 pages of field notes and tran- Merging
TABLE 1

scriptions to the final conclusions, as sprinkled with

vivid illustrations as they may be” (Miles, Huber-
man, & Salda~ na, 2014: 5). This is problematic for all
categories
Dropping

scholars trying to evaluate the results of such pur-

suits, but even more so for junior scholars, who look
to and try to learn from published pieces for guid-
ance on how to generate theoretical insights. The
on puzzles
Focusing

current lack of reflexivity also stands in the way of

theory development by limiting the repertoire into
which scholars tap when analyzing data and by arti-
ficially constraining the pathways that newcomers
Questions
Asking

drawing upon qualitative data believe they can

pursue.
In this article, we integrate insights from both cate-
gorization theory and existing scholarship on quali-
possible added data
Initial data collection

Further analyses and

tative analysis to develop a framework emphasizing

Analytical stages

the researchers’ active role in theory building. The

Re-analyses and
and analyses

framework includes the main analytical moves upon

integration
theoretical
collection

which scholars might rely to sift through their data

and serves as an invitation for them to both be more
reflexive and more transparent about their analytical
processes. Table 1 provides an overview of this
framework.
Refining tentative
Generating initial

We define “moves” as the micro-processes that re-

Categorization

categories

categories

categories

searchers undertake during qualitative theory build-

Stabilizing

ing. By integrating categorization theory with

process

insights from existing but at times neglected qualita-

tive scholarship, we identify specific moves. In other
2021 Grodal, Anteby, and Holm 593

words, while evidence of corresponding analytical including a definition of it and examples of how it
strategies exists within qualitative scholarship, these has been applied in qualitative data analysis.
moves have not, as such, been explicated. Specifi- Ultimately, we suggest that researchers can dem-
cally, we highlight eight main moves—(1) asking onstrate rigor by detailing more precisely how they
questions, (2) focusing on puzzles, (3) dropping cate- have purposefully drawn on a broad and diverse set
gories, (4) merging categories, (5) splitting categories, of moves to engage with their data. This approach ad-
(6) relating or contrasting categories, (7) sequencing vocates that we are more candid and explicit about
categories, and (8) developing or dropping working both the goals and the process driving theory devel-
hypotheses—that can provide analytical scaffolding opment. Moreover, we argue that different moves
to facilitate the often-daunting task of analyzing can be used to generate insights during different
qualitative data and crafting theoretical contribu- phases of the research process and that no one tem-
tions. Table 2 provides an overview of each move, plate should be reified as the true way to achieve

TABLE 2
An Overview of Eight Possible Analytical Moves
Move Definition Example

Asking questions Approaching the data with specific questions that How do actors “construct, navigate, and capitalize on
researchers want answers to timing norms in their attempts to change
institutions”? (Granqvist & Gustafsson, 2016: 1009)
Focusing on puzzles Focusing on the part of the data that is most It is puzzling that parties challenging established
surprising or salient to the researchers social systems collaborate with defenders of those
same systems. (O’Mahony & Bechky, 2008)
Dropping categories Dropping categories that were generated during the Upon attending his first role-playing game event, the
initial part of data analysis but that turned out not researcher noted that “there was no organization to
to have theoretical traction the group: there was no membership chairman, no
one that one had to meet to gain access; one
simply walked in” (Fine, 1983: 244). Yet, rapidly,
that category lost relevance as others gained more
theoretical traction.
Merging categories Uniting two or more existing categories to create a “In reviewing our first-level constructs and relating
superordinate category these to prior research, we concluded that all of
them represented different phases and forms of
identity work.” Thus “identity work” was adopted
as the label of a merged code. (Creed, Dejordy, &
Lok, 2010: 1342)
Splitting categories Separating a category into two or more subordinate Splitting the category of Total Quality Management
categories (TQM) “tools” into four subcategories, ranging
from the least technical (general TQM methods) to
the most technical (statistical tools). (Zbaracki,
1998: 610)
Relating or contrasting Comparing several categories with one another to Contrasting “grass-fed” and “conventional” to
categories identify relationships between them (or the lack of identify their similarities and differences. (Weber
such relationships) et al., 2008)
Sequencing categories Temporally organizing categories that researchers Researchers “sought evidence of boundary and
have identified in the data practice work patterns that co-occurred in time, by
actor type and by objective. We identified four
cycles of interconnected boundary work and
practice work. … We constructed raw data tables
for each cycle to provide another iteration between
the raw data and this higher level of abstraction.
… These cycles together formed a complete
lifecycle of institutional stability and change.”
(Zietsma & Lawrence, 2010: 200)
Developing or dropping Formulating an overarching theory and, by iterating “Throughout this cyclical process, we actively and
working hypotheses through the data, either finding increasing continually called into question our emerging
evidence for it, leading to its elaboration, or theoretical understanding by exposing it to further
finding contradictory or unsupportive evidence, data analysis.” (de Rond & Lok, 2016: 1971)
leading to its abandonment
594 Academy of Management Review July

rigor (Abbott, 2004: 52). That is true for Table 1 as cheated. Our advice is to let it happen. The rigor and
well—this is not a framework that should be applied vigor will follow. (Strauss & Corbin, 1990: 129)
as a template. Instead, researchers should apply the Put another way, qualitative scholars may adopt
moves that allow them to generate insightful catego- different approaches when moving from data to theo-
ries from their data and subsequently report verbally ry; we argue that these various approaches may
or visually the often messy process that resulted. In nonetheless be rigorous.
short, we not only echo the call for increased rigor Here, we draw on categorization theory to expli-
but also encourage scholars to deploy their full and cate how researchers decipher their data and ac-
diverse analytical imagination in pursuit of strong tively construct categories at different stages of the
theoretical insights. analytical process. In the process, we unpack how
a qualitative scholar might “let it happen” (in
ACHIEVING RIGOR IN Strauss and Corbin’s terminology) while still en-
QUALITATIVE ANALYSIS suring that readers can both evaluate the scholar’s
analytical process and be better guided in their
Qualitative scholars have long tried to achieve a
own future pursuits. Instead of promoting a unique
better understanding of how readers are convinced
template, we posit that rigorous qualitative analy-
by ethnographic texts and how limited cases can
sis can be achieved by being transparent and de-
lend themselves to rigorous interpretation (Golden-
tailed about individuals’ active categorization
Biddle & Locke, 1993; Small, 2009, 2013; Staw,
choices during the process of discovery (Glaser &
1995). A debate has more recently ignited about how
Strauss, 1967: 244). Such an approach to rigor
to improve methodological rigor (Gioia et al., 2013).
means that readers assessing the move from data to
To preempt skepticism about qualitative rigor, schol-
theory are able to grasp and evaluate more clearly
ars have articulated broad guidelines on how to con-
the many decisions made by researchers in their
duct qualitative research properly by calling on
analytical pursuits (Eisenhardt, Graebner, & So-
scholars to develop “a coding scheme and, insofar as
nenshein, 2016). This approach also encourages us
possible, provide a sample of likely coding catego-
to be more reflexive about our own research pro-
ries” (Lamont & White, 2008: 143)—yet they have
cesses and how we build insights from our data
been less explicit about how to do so.
(Bourdieu & Wacquant, 1992), which can help us
As observed by Bansal and Corley (2011) and
achieve a stronger theoretical contribution.
Langley and Abdallah (2015), in practice, many man-
agement scholars have translated the call for rigor
into a need to follow seemingly proven and tested Qualitative Analysis as a Categorization Process
templates in written accounts of their analytical pro- At its core, the process of qualitative analysis en-
cesses. While templates might prove useful in some tails sifting through data to generate new catego-
circumstances, they can also obfuscate the research- ries that can form the foundation for new
er’s generative role in the analytical process when theoretical insights (Becker, 2008; Charmaz, 1983;
applied indiscriminately. Moreover, using templates Fine, 1993; Golden-Biddle & Locke, 2007; Sprad-
can hinder scholarly output because it constrains the ley, 1979; Strauss & Corbin, 1990; Van Maanen,
multiple ways in which rich qualitative data can be 2011). As Charmaz (2006: 186) explained, “As the
mined for insights and have the unintended conse- researcher categorizes, he or she raises the concep-
quence of generated undue homogeneity in qualita- tual level of the analysis from description to a more
tive theory building. As Lamont and Swidler (2014: abstract theoretical level.” Van Maanen (1979:
157) argued, “Method debates are in fact theory de- 541) likewise emphasized “categorizing” as a step
bates.” Even Strauss and Corbin (1990: 129) cau- to move from data to more general findings. In ad-
tioned against the use of templates, first dition, Corbin and Strauss (1990: 7) stressed the
acknowledging: “We realize that beginners need importance of progressing from examining raw
structure and that placing the data into discrete box- data to constructing and labeling larger groupings
es makes them feel more in control of their analyses.” that capture instances of emerging categories to de-
They then added: velop theory. To generate “overarching catego-
Analysts who rigidify the process are like artists who ries,” they advocated identifying similarities
try too hard: although their creations might be techni- between subordinate categories and merging them
cally correct, they fail to capture the essence of the to create larger groupings. Importantly, within
objects represented, leaving viewers feeling slightly qualitative data, these categories are sometimes
2021 Grodal, Anteby, and Holm 595

concepts and at other times mechanisms, where This active construction of categories not only
mechanisms are categories of actions or behaviors shapes which categories are generated but also the
that relate the concepts to each other (Davis & process through which they are created. As the
Marquis, 2005; Hedstr€ om & Swedberg, 1996). knowledge and goals of the categorizer change, so do
The challenge of creating categories is exacerbated the categories created (Vygotsky, 1987). Further-
when data are rich and layered; this is true of much more, the process of categorization is not linear and
qualitative data. The richer and more layered the rational but characterized by iteration, detours, and
data, the more decisions the researcher needs to regression (Bloom, 2000). Categories are often later
make in order to generate categories that form the ba- forgotten. Some of these forgotten categories are lost
sis for novel insights. Qualitative data tend to be rich forever, whereas others are recreated at a later stage,
and layered because they are often longitudinal when the need for the same category arises (Siegler,
(such that categories might vary over time) (Langley, 1998).
1999), are collected on an ongoing basis (such that This active process also shapes the internal rela-
the categories created initially no longer prove rele- tionship between categories. Initially, scholars as-
vant later in the collection phase) (Lopez & Phillips, serted that categories were organized hierarchically:
2019), or derive from participants or data sources each overarching category encompassed all the other
with differing perspectives (such that categories categories underlying it (Murphy & Lassaline, 1997).
across participants or sources might differ) (Lofland, For example, “beef,” “lamb,” and “pork” can all be
1971). Nevertheless, while categorization is at the viewed as subsets of the overarching category
heart of qualitative analysis, we have not yet looked “meat,” which in turn belongs to the overarching cat-
to theories of categorization as a way to reflect on the egory “food.” This view, however, has been chal-
process of qualitative theory building. lenged. First, scholars have found that not all
categories are organized around hierarchies (Rosch,
1978). Second, more recent developments in catego-
Drawing on Categorization Research to Inform rization theory have emphasized that such sorting is
Theory Building often contextual and dependent on the goal. If the
goal is to consume only grass-fed meat (Weber,
Scholars have long recognized that humans cre-
Heinze, & DeSoucey, 2008), then one might reorder
ate categories to reduce the amount of information
these elements into grass-fed and traditionally fed
they need to process when navigating a complex,
meat, regardless of the type of animal from which the
information-satiated world (e.g., Bowker & Star,
meat is derived. Such a grouping can recast elements
2000; Lakoff, 2008; Rosch, 1978). Categories are ac-
into very different categories and produce new rela-
tively and socially constructed through a complex
tions between them.
and multifaceted process that is driven by the
Applied to qualitative analysis, such agency in
knowledge, goals, and contexts in which the catego-
generating categories suggests that there could be an
rization process unfolds (Barsalou, 1983; Berger &
infinite number of perspectives on a particular data
Luckmann, 1967; Durand & Paolella, 2013; Murphy
set, depending on the context and on the coder’s
& Medin, 1985; Searle & Willis, 1995). For example,
knowledge and goals. Furthermore, people’s per-
when antique dealers source pieces to acquire for
spectives and goals evolve as categories are created,
their collection, they might distinguish them by cen-
forgotten, and at times resurrected throughout the
tury rather than grouping them by function, thus
analysis process. Theory development encompasses
sorting the pieces, in effect, based on their goals and
many decisions that are anything but predefined and
knowledge structures. By contrast, lay people would
are therefore difficult to fit into templates. The schol-
most likely sort antique furniture by salient function-
ar is always at the center of the process and is active-
al features, for example, grouping all chairs together
ly involved in making these choices.
or all tables together. In other words, faced with the
same set of elements, and depending on people’s
Drawing on Past Qualitative Scholarship to
goals and existing knowledge, the sorting process
Inform Theory Building
will proceed differently, and people will ultimately
most likely generate distinct groupings (Barsalou, While an active perspective on theory building
1983). The creation of categories can therefore not be has not been at the center of qualitative analysis pub-
decoupled from the person(s) who created them or lished in management scholarship (Bansal & Corley,
the context in which they were created. 2011; Langley & Abdallah, 2015), a close reading
596 Academy of Management Review July

across decades of scholarship on qualitative analysis relation to three general stages in the analytical pro-
reveals that there is support for the active approach cess: (1) generating initial categories, (2) refining ten-
as an analytical strategy (Becker, 2008; Eisenhardt, tative categories, and (3) stabilizing categories. Since
1989; Fine, 1993; Golden-Biddle & Locke, 2007; analyzing data is a “live” and iterative process
Langley, 1999; Miles et al., 2014; Spradley, 1979; (Locke, Feldman, & Golden-Biddle, 2015), all moves
Van Maanen, 2011). We propose that many of the can be useful at all stages of the process. Some
moves used in qualitative analysis mirror processes moves, however, might be more relevant and more
identified in categorization theory. For instance, likely to be deployed at specific stages. In Table 1,
Becker (2008) outlined “tricks” that qualitative re- the shading of each move during a particular analyti-
searchers can use to build theory. He encouraged re- cal stage represents our expectations of its likelihood
searchers to actively start elaborating categories by to occur at that time. In the following section, we pre-
asking themselves what questions their data can an- sent the moves, following this possible order of
swer. As he articulated it, “The reformulated ques- appearance.
tions constitute the beginnings of conceptual
construction” (Becker, 2008: 122). Overall, Becker Generating Initial Categories: Initial Data
suggested that asking questions may be what paves Collection and Analysis
the way for theoretical development. Similarly, Lof-
land, Snow, Anderson, and Lofland (2006: 201) ad- Very few qualitative scholars approach field set-
vised that, “where the rubber hits the road, so to tings as fully blank slates (Lofland et al., 2006). Typi-
speak, [is the point at which] you begin to condense cally, qualitative scholars engage in early moves
and organize your data into categories that make aimed at generating insights to prime the theory-
sense in terms of your relevant interests, commit- building process (Spradley, 1979). When they are se-
ments, literatures, and/or perspectives.” Thus, many lecting field topics and starting to collect data, they
foundational recommendations in qualitative analy- draw on their existing knowledge and experiences to
sis point to a range of ways in which qualitative re- imagine possible pathways, setting aside others that
searchers can actively approach their data to derive do not trigger their curiosity. Thus, the choices and
insights. These recommendations inform the moves focus that researchers bring to their empirical inqui-
(detailed next) that constitute the building blocks of ry seed the categorization process. We identified as
our proposed active categorization framework. central to this early seeding process two moves asso-
ciated with generating initial categories: (1) asking
questions and (2) focusing on puzzles. Here, we de-
AN ACTIVE CATEGORIZATION FRAMEWORK
fine these moves, relate them to methodological
FOR THEORY BUILDING
strategies that we have identified in existing qualita-
In the following sections, we integrate insights from tive scholarship, and provide empirical examples.
categorization theory with insights from the at times Asking questions. Categorization theory sug-
neglected qualitative scholarship to spotlight eight gests that how humans make sense of the world
main moves in which qualitative researchers engage depends on their prior understanding and initial
when sifting through data and developing theory. We goals (Barsalou, 1983; Durand & Paolella, 2013;
detail how these eight moves can generate multiple Murphy & Medin, 1985). Deprived of existing cate-
pathways toward theory development. Like different gories, we would not be able to integrate or under-
swimming strokes, these moves enable researchers to stand the world around us. Our existing categories
swim differently through their data and reach diverse are the foundations for the new categories that we
destinations. As such, they should be viewed as ex- form when presented with novel stimuli (Arm-
amples of possible moves. Analytical moves can be strong, Gleitman, & Gleitman, 1983). In particular,
actively recombined in many ways to create multiple we categorize the same objects differently depend-
paths to insight, depending on the unfolding of the re- ing on the goals we are trying to achieve and the
search process or the phenomenon in question. Re- questions we ask of the world. For example, the
searchers will then themselves be able to identify categories people formulate to prepare to go camp-
more moves and to recombine them in potentially in- ing consist of heterogeneous objects—such as
finite ways to create a rigorous data-analysis descrip- boots, tents, and sleeping bags—that are united be-
tion with a fully transparent analytical process. cause they are useful for achieving the same goal.
One way to think about the active categorization By contrast, given a different goal (such as going to
framework is to consider these eight moves in work), boots and tents would not be viewed as part
2021 Grodal, Anteby, and Holm 597

of the same category. “Asking questions” is the the initial process of category construction and theo-
move in which researchers draw on their existing ry building.
categories to select and approach their field set- Our analysis of published qualitative papers re-
tings with specific questions to which they would vealed examples of how authors have used questions
like answers. By asking questions, and thereby re- to jump-start their theory development. Grodal, Nel-
stricting their goals up front, researchers begin to son, and Siino (2015: 140) described their process as
craft initial (possible) categories while they are follows:
collecting and analyzing their data. This is not to
Initially, our observations focused on [the] grand tour
say that early questions are not revised or that new question (Spradley, 1979): “How do helping behav-
ones do not surface during analysis, but the initial iors unfold?” Over time, however, as our inquiry be-
path through the data is determined by which came more focused on specific aspects of helping
questions are asked. behavior, the nature of these questions changed to re-
A close read of existing scholarship on qualitative flect our growing understanding.
methods reveals that asking questions has been sug-
gested as a move to initiate theory building (Glaser & Because the authors asked this question, helping
Strauss, 1967; Spradley, 1979; Strauss, 1987). Al- behavior became the focus of their study, even
though many qualitative scholars allow themselves though their data also spoke to other phenomena.
to be guided by their field data and interactions, they Deeper into their analysis, they began to wonder
still enter the field with some preconceptions and a how engagement is sustained during helping rou-
history of past research designs and questions (Gold- tines. Hence, their questions shifted at each stage
en-Biddle & Locke, 2007). As Cohen (2013: 435) lu- and in many ways shaped the categories on which
cidly reported, “Many of the choices made in my they honed in during their finer-grained analysis.
[qualitative] research design followed directly from This approach of iterative questioning is not unique:
past research.” This pattern explains why Strauss when describing the analytical process she followed
(1987: 306) labeled one common approach to qualita- in her study of employees’ use of mobile devices,
tive data analysis as a “theory elaboration exercise,” Mazmanian (2013: 1231) similarly reported asking
one in which past literature serves as a springboard questions of her data. Likewise, in her study of wom-
for asking questions to spur new lines of research “in en doing “unpaid” work in VIP nightclubs, Mears
service of discovering” a new and more encompass- (2015: 1099) asked the core question “Why do work-
ing theory. Similarly, Spradley (1979) suggested that ers participate in their own exploitation?” as a trigger
asking questions is a core part of the early discovery to her analytical process. With such questions in
process because it is through questions that research- mind, these scholars were able to spearhead unique
ers create initial categories for the problem they are trajectories through their data. Asking questions is
trying to understand. thus an important move in qualitative research be-
Likewise, many scholars recommend ongoing ef- cause it allows researchers to proactively direct their
forts to answer the key question in generalizing be- analysis toward a specific theoretical end, making it
yond the chosen empirical setting: “What is this a more likely that the initial categories that emerge
case of?” Asking this question allows for an explora- from the data-analysis process will be of theoretical
tion of the conceptual possibilities offered by the significance.
data (Glaser & Strauss, 1967: 24), thereby generating Focusing on puzzles. Initial categories are also
theoretical insights that “bump things up a level of generated by “focusing on puzzles,” the move in
generality” (Luker, 2008: 138). Another reason why which researchers concentrate on the parts of the
asking questions is an important part of the research data they find most surprising or salient. Not all in-
process is that it helps the researcher to manage the formation is equally important in the categorization
complexity and overload of data, which can stand in process (Bowker & Star, 2000). To categorize the
the way of theoretical insights. Without such ques- world around us, we bring with us the knowledge
tioning, “The result is death by data asphyxiation— we have already accumulated about a particular
the slow and inexorable sinking into the swimming space (Durkheim & Mauss, 2009); thus, we categorize
pool which started so cool, clear, and inviting and objects by making inferences based on the knowl-
now has become a clinging mass of maple syrup” edge base we already possess (Hirschfeld & Gelman,
(Pettigrew, 1990: 281). Even if questions are not fully 1994; Murphy, 2004). If we observe that a particular
articulated, they are nonetheless ubiquitous in the animal has wings, we assume that it can fly, lays
mind of many researchers and offer gateways into eggs, and will tend to its young in nests. When our
598 Academy of Management Review July

knowledge of the world is confirmed in actual be- Miles et al. (2014: 301) also called attention to
havior, we hardly pay attention: these categories “outliers” and “surprises” when analyzing data.
have become part of our taken-for-granted under- Many other scholars recommend honing in on puz-
standing of the world (Clark & Wilkes-Gibbs, 1986; zles as well. For example, Turco (2016: 206) de-
Colyvas & Powell, 2006; Vygotsky, 1987). When the scribed generating insight by “reading her field notes
bird with wings takes off in flight, we happily pro- and memos several times through” and never being
ceed undisturbed. able to “jump right into line-by-line coding.” Instead,
Nevertheless, our observations do not always cor- she suggested “looking for empirical puzzles” (p.
respond to our existing knowledge about the world. 206). In short, focusing on puzzling cases is an ana-
If we encounter a bird with wings that cannot fly lytical strategy that researchers can use to stimulate
(such as a penguin or an ostrich), we pause in puzzle- new lines of inquiry.
ment. The lack of flight is a salient feature in an ani- When Desmond (2012: 1302) conducted his eth-
mal with wings; it forces us to pay attention. nography of evicted tenants in high-poverty neigh-
“Salience refers to the intensity of a feature, the ex- borhoods, he noticed that these tenants relied more
tent to which it presents a high amplitude” (Sloman, on new acquaintances than on kin ties to meet their
Love, & Ahn, 1998: 193). Salience arises when an ob- more pressing needs, regardless of the strength or
servation is at odds with our existing knowledge weakness of those ties (Granovetter, 1973):
about a category, and it plays an important role in During everyday conversation, people in the trailer
category formation because salience helps us to fo- park and the inner city claimed to have no friends or
cus on the novelty and uniqueness of a specific situa- an abundance of them, to be surrounded by support-
tion (Ahn, 1999). For example, we might believe that ive kinsmen or estranged from them. I came to view
all swans are white. Encountering a black swan cre- these accounts skeptically, interpreting them as a
ates dissonance with our existing cognitive struc- kind of data in their own right but not as accurate
tures, and we pay attention to the phenomenon in an evaluations of people’s social relationships.
attempt to find an explanation for the discrepancy. This puzzling observation prompted Desmond to
Salience might take the form of a contradictory ex- develop the notion of “disposable ties” as a new cate-
emplar—such as a puzzling person, event, or utter- gory of social ties not fully captured by the notion of
ance (Smith & Zarate, 1992)—that challenges the weakness.
conceptual coherence through which we view the In a similar fashion, when describing her analyti-
world (Sloman et al., 1998). Focusing on puzzles cal process, Vaughan (1996: 461) highlighted how fo-
might help us abduct new insights about a novel and cusing on puzzles led her to generate theory about
untheorized category which may be present in the “normalizing deviance” in organizations. As she ex-
data (Behfar & Okhuysen, 2018; Locke, Golden-Bid- plained in relation to her study of the space shuttle
dle, & Feldman, 2008). In other words, a puzzling Challenger disaster:
piece of data can function as a lightning rod, attract-
ing insights around which new categories can form I began coding [the data], but soon realized … that
(Zhao, Ishihara, Jennings, & Lounsbury, 2018). this strategy was flawed: aggregating statements from
When developing theory, it is important to focus all interviews by topic (a practice I often used) would
extract part of each interview from its whole.
on what is surprising and unexpected in how the
data relate to existing theory. Scholars have called Instead, to elaborate her analytical categories, she
such puzzles “negative cases” or “unusual” inci- focused on the puzzle that informants’ statements re-
dents (Katz, 2001: 331) and have emphasized their garding a critical part of the space shuttle often con-
importance in the theorizing process, even if there is tradicted the archival record—a move that proved to
“only one case” (Emerson, Fretz, & Shaw, 2011: 193). be core to her theorizing. Focusing on puzzles thus
Focusing on such cases can elucidate the patterns directs researchers’ categorization efforts toward
and variations in the meanings that members attri- data elements that diverge from expectations and
bute to a given social setting. Stated otherwise, might therefore form the basis for generating new
strong qualitative research might result precisely theory.
from focusing on the unexpected. In Tricks of the During their initial data collection and analysis,
Trade, Becker (2008) described a move he called researchers generate initial categories by asking
over-focusing on strange elements: he explained questions and focusing on puzzles. This is important
how “weird” findings (p. 208) or a “finding that does in theory development because the initial questions
not fit” (p. 83) prove integral to theory development. and puzzles arise from discrepancies between what
2021 Grodal, Anteby, and Holm 599

researchers notice in the data in light of their under- theory a coherent detailed and worthwhile story to
standing of past theories, which can help them ab- tell then … they should drop the category.” Like-
duct new theoretical insights (Locke et al., 2008; wise, Miles et al. (2014) discussed the possible dis-
Timmermans & Tavory, 2012). In this stage, research- connect between the researcher’s emerging
ers therefore juxtapose initial categories with past understanding of a phenomenon and the categories
theoretical insights. identified in the early stages of a project. In particu-
lar, “some codes do not work; others decay. No field
Refining Tentative Categories: Further Analyses material fits them, or the way they slice up the phe-
and Possible Added Data Collection nomenon is not the way the phenomenon appears
empirically” (Miles et al., 2014: 82). The researcher
As research progresses, the process of data analy- realizes that “some codes do not work” and the asso-
sis shifts from generating initial categories toward re- ciated categories should be either dropped or trans-
fining tentative categories through moves such as formed to best reflect the data.
“dropping categories,” “merging categories,” When researchers write their methods sections,
“splitting categories,” “relating or contrasting cate- they tend to focus on the categories that remained in
gories,” and “sequencing categories.” These moves their analysis more than the ones that were dropped.
also help the analyst to begin distinguishing catego- This is particularly true of published work wherein
ries that might function as mechanisms from those methods sections are often shorter than in working
that function as concepts and to begin mapping out papers in which authors are more focused on con-
the links between them. The combination of these vincing readers of their methodological astuteness.
moves starts to elaborate and challenge existing However, examples from the coauthors’ own work
theories. suggest the importance of dropping categories for the
Dropping categories. When we categorize the categorization process.
world around us, we must often sort through an over- In her research on constructing social and symbol-
whelming amount of information (Murphy, 2004; ic boundaries in the nanotechnology field, Grodal
Rosch, 1978). A consequence of the overload of in- (2018: 792) described that she “developed some
formation is that people initially generate categories broad and some very specific codes … through sev-
that turn out not to be relevant to the categorization eral rounds of iteration and moving back and forth
process. They might, thus, initially create categories between broad and specific codes.” This process in-
that are faulty, biased, or irrelevant in explaining the volved in fact dropping several codes that turned out
phenomena that they are trying to categorize. When not to have theoretical traction. For example, the ini-
humans focus in on salient cues to guide in the cate- tial category “commercializing scientific knowl-
gorization process, these initial categories might be edge,” which was abundant in the data, turned out to
salient for accidental reasons instead of representing be only peripherally related to the overarching theo-
stable patterns (Weick, Sutcliffe, & Obstfeld, 2005). retical story of how the social and symbolic bound-
An important part of the categorization process aries of a field expand and contract over time; this
therefore entails dropping initial categories to focus category was therefore dropped from the analysis.
the categorization process on categories that are Similarly, when studying ghostwriters, Anteby
more important and meaningful in explaining the and Occhiuto (2020) initially created the category
phenomena at hand (Murphy, 2004). When people “previously published” to track whether the ghost-
“drop categories,” they stop paying attention to cate- writers they interviewed had published under their
gories that are no longer relevant in explaining the own name prior to agreeing to write for others and
phenomenon that they are paying attention to. how ghostwriters spoke about such past publica-
In the process of generating categories, researchers tions. The intuition was that ghostwriters who had
often create a multitude of categories, only a few of published under their own name might resent (more
which will be essential in explaining the puzzle that than others) being asked to remain invisible from
they have identified in the data. Several qualitative public view. Ultimately, however, the category
scholars suggest that an important part of qualitative “previously published” proved partly irrelevant to
analysis is to revise and reduce the number of catego- the authors’ main findings. The ways in which ghost-
ries that are used to explain the phenomena. Locke writers spoke about their own writing did not illumi-
(2001: 79) for example suggested that “selecting out nate much about their work experience of producing
categories” is an important part of the analytical pro- someone else’s self, or what the authors labeled
cess and that, if “analysts feel they have in their “stand-in labor.” While Anteby and Occhiuto (2020:
600 Academy of Management Review July

1295) noted that they “developed coding categories workable set of themes and concepts is in hand … we
inductively,” they could have also added that they investigate whether it is possible to distill the emer-
dropped several categories while developing their gent 2nd-order themes even further into 2nd-order
theoretical insights. “aggregate dimensions.”
Merging categories. When people “merge catego- Many qualitative papers have used the move
ries,” they unite two or more existing categories to “merging categories” in their data analysis, although
create a superordinate one. According to categoriza- they may not label it as such. Some have referred to
tion theory, people initially tend to create elaborate merging as “combining” and “bundling” (Huising,
and detailed categorization structures upon encoun- 2015: 270); others have described “consolidating”
tering unfamiliar objects or actions (Bloom, 2000). (e.g., Anteby & Molnar, 2012: 522; Pratt, Rockmann,
Subsequently, to refine the initial categorization & Kaufmann, 2006: 240), and still others have re-
structure and optimize the processing of informa- ported that they “assemble” codes into “aggregate
tion, people often merge these detailed categories to dimensions” (Nelson & Irwin, 2014: 900). Regardless
form superordinate ones. Over time, this merging of terminology, these researchers have proceeded
process can result in intricate hierarchies of catego- from having many categories to having fewer catego-
ries (Murphy, 2004). For example, after creating the ries, often referencing first-order categories, second-
categories “ants,” “cows,” and “jellyfish,” a person order categories, and overarching categories. For
might realize that all three can be assigned to a super- example, Ramus, Vaccaro, and Brusoni (2017:
ordinate category, “animal,” because they all move 1264–1265) described initially creating a large set of
and depend on oxygen for their survival. categories, which they gradually aggregated through
Historically, many qualitative scholars have char- subsequent data analysis:
acterized merging categories as a fundamental move
in qualitative analysis (Charmaz, 2006; Locke, 2001; First, we performed open coding (Strauss & Corbin,
1990: 61) of each document to identify initial, empir-
Strauss & Corbin, 1990). These scholars advocate be-
ical themes. … This stage of analysis drove us to
ginning qualitative analysis with “open coding”;
identify many empirical themes … Once a workable
that is, generation of a plethora of specific categories set of conceptual categories had been developed, we
that closely adhere to the phrasing articulated in the moved to more deliberate theorizing in an effort to
data. Ultimately, qualitative researchers should aggregate categories in an empirically grounded mod-
merge these categories into superordinate categories el that explained the process.
that capture the essence of their meaning. For exam-
ple, Strauss and Corbin (1990: 223, 229) recom- As an analytical process, merging categories can
mended that researchers progress from “open thus be found in both the categorization literature
coding” to ultimately generating “overarching cate- and the qualitative methodology literature as a strat-
gories.” As others have clarified, such a “clustering” egy for refining categories. Due to its importance for
process enables researchers to “clump” items into theory development, this move has been widely
“classes, categories, bins” and to progress from used in recent empirical work. It appears to be the
“lower” to “more complex” categories (Miles et al., most prevalent step that authors currently report
2014: 279). when conducting qualitative organizational re-
More recently, merging categories has become a search. Such an emphasis has generated an almost
common way to analyze qualitative data in organiza- self-fulfilling prophecy (Merton, 1948), in that the
tional scholarship. Gioia et al. (2013: 20) emphasized sheer assumption that theory development rests
both merging and the value of a “data structure” that mainly on merging categories minimizes other
depicts the conceptual movement from a multitude moves, such as splitting, which we turn to next.
to a reduced set of categories: Splitting categories. From studies of categoriza-
tion, we know that humans’ category formation does
In this 1st-order analysis, which tries to adhere faith- not progress uniformly from more categories to fewer
fully to informant terms, we make little attempt to dis- through merging; often, progression occurs in the op-
till categories, so the number of categories tends to
posite direction, from fewer to more (Murphy, 2004).
explode on the front end of a study. There could easily
Researchers studying children’s language develop-
be 50 to 100 1st-order categories. … As the research
progresses, we start seeking similarities and differ- ment have noted that children often
ences among the many categories … a process that “overgeneralize”; that is, they use a category more
eventually reduces the germane categories to a more broadly than is generally accepted. Over time, a child
manageable number (e.g., 25 or 30). … Once a splits the overgeneralized category into its
2021 Grodal, Anteby, and Holm 601

component parts to create categories that are aligned emotional labor than the others, with important im-
with conventional usage (MacWhinney, 1987). For plications for their professional standing. Likewise,
example, a young child might use the word “car” to Petriglieri (2015) described how, in her study of BP
refer to any object that moves, including a skate- executives during and after the 2010 Gulf of Mexico
board, a bicycle, and a ball. Over time, however, the oil rig explosion and spill, she came to split a catego-
child breaks down the category “car” and distin- ry she had developed during the first stage of her
guishes between “skateboard,” “bicycle,” and “ball.” analysis to gain a more nuanced understanding of
“Splitting categories” is a move defined as separating the relationship between an organization and its
a category into two or more subordinate categories. members. While she initially developed the category
Lofland and colleagues (2006: 119–143) asserted “questioning fit between self and organizational
that probing the nuances of categories to create finer identity” as a theme, she realized that dividing the
distinctions is an important component of theory category into emotional and cognitive components
generation. Spradley (1979: 144) also emphasized could lead to even more insights (Petriglieri, 2015:
that identifying “subsets within a domain and the re- 526). Others have also described how they have split
lationship between these subsets” is an important el- categories in order to specify their understanding of
ement in qualitative analysis. Miles et al. (2014: 284) a category (e.g., Crosina & Pratt, 2019). Broadly
wrote about “partitioning” and “unbundling” as crit- speaking, splitting can help researchers to recognize
ical ways to analyze data, adding that “there are nuances in their data and unpack a specific category
many times when differentiation [emphasis in origi- that has piqued their curiosity.
nal] is more important than integration.” As an illus- Relating or contrasting categories. Categoriza-
tration, researchers who have identified the category tion scholars have suggested that categories are not
“people above” in their data might, through further created independently of one another but often
analysis, split the category into such components as evolve in parallel. In particular, many scholars have
people “earning more” and people with “more asserted that categories are typically interrelated in
ambition” than their informants (Lamont, 2000: semantic networks (Collins & Loftus, 1975; Quillian,
102–145). Spradley (1979: 115–116) suggested that 1969). Categories that often cooccur in discourse be-
the process of splitting categories can begin during come more closely related in the semantic network,
data collection: if informants mention a category whereas categories that seldom occur together be-
during data collection, researchers might ask if they come more distantly related. For instance, if the cate-
can provide other examples of the category. By gory “dog” is often used in conjunction with the
means of splitting, researchers can reach a more nu- category “leash,” the two words will ultimately be
anced understanding of the different ways a category closely related in the semantic network. Neverthe-
is manifested (or not) in their data. less, it is not merely categories’ association with one
Several qualitative studies provide examples of another that creates categorical meaning; it can also
how splitting categories can advance theory. For ex- be their perceived opposition. Categories often exist
ample, a vibrant stream of research has developed in contrasting pairs (Douglas, 1966; Levi-Strauss,
around the notion of “emotional labor” (Hochschild, 1969). For example, people construct the meaning of
1983); until recently, however, analysis of this form the category “natural” in part from its contrast with
of labor remained quite monolithic. In her study of the category “artificial” (Weber et al., 2008). The
psychotherapy practices, Craciun (2018: 261) honed meaning of categories is therefore determined by the
in on a single category—the use of emotions at way people relate and oppose categories to one an-
work—that had emerged from her coding. Through other (Kahneman & Tversky, 1984).
further data analysis, she discovered that the use of “Relating or contrasting categories” is the move
emotions could be split into three categories: (1) that researchers use when they compare several cate-
“didactic,” as a tool of intervention (e.g., when pro- gories to specify the relationships (or lack thereof)
fessionals use emotion to foster particular disposi- among them. It is important to clarify how relating or
tions in their clients); (2) “supportive,” as a tool that contrasting categories differs from merging:
helps professionals to foster trust in their patients “merging” identifies underlying features that belong
(e.g., when professionals are driven by passion for to the same category; “relating or contrasting catego-
their work); or (3) “inductive” (e.g., when professio- ries” identifies relationships among categories that
nals rely on their emotions as epistemic tools to iden- might not belong to the same overarching category.
tify problems). Craciun also demonstrated that some Many qualitative scholars have emphasized the
professionals rely more heavily on one form of importance of comparing or contrasting categories
602 Academy of Management Review July

(Golden-Biddle & Locke, 2007: 45; Suddaby, 2006), sense in relation to each other: one entailed anchor-
thus implicitly acknowledging that categories can ing oneself in an existing identity; the other involved
only operate in relation to one another. Strauss and building ties across identity groups. Importantly,
Corbin (1990: 125) noted that categories are only anchoring contributed to taming intergroup
“related”; Becker (2008: 132–133) similarly called conflict and encouraging the medical providers to
all categories “relational,” pointing out that properly serve the soldiers. Relating or contrasting
“working class” and “middle class” labels have categories helps researchers to identify in their data
meaning only in relation to each other. Glaser and the boundaries of the categories and the relation-
Strauss (1967: 106) thus provided the following ships between categories.
recommendation: Sequencing categories. Categories are not static;
While coding an incident for a category, compare they are dynamically related to the world that people
it with the previous incidents in the same and dif- inhabit. An essential part of the categorization pro-
ferent groups coded in the same category. This cess is to create dynamic understandings of this un-
constant comparison of the incidents very soon folding reality (Durand & Paolella, 2013; Nakamura,
starts to generate theoretical properties of the cate- 1985). Categories are not merely cross-sectional
gory. The analyst starts thinking in terms of the groupings of objects; they encode causal relation-
full range of types or continua of the category … ships between objects and actions (Murphy & Medin,
[including] its relation to other categories, and its 1985). We categorize an event as a “birthday party”
other perspectives. not only due to the presence of a birthday cake and
Many other qualitative scholars would concur: the presents but also because the event follows a se-
recommendation of Booth, Colomb, and Williams quence of actions: (a) guests arrive, (b) the guests
(2003: 46) to scholars progressing from a focused top- make congratulatory statements, (c) the guests sing a
ic to a research question is to “identify the parts [of a birthday song, (d) cake is served, (e) candles are
topic]” and “how they relate to each other”; Miles blown out, (f) cake is eaten, and (g) guests leave.
et al. (2014: 287) suggested that, once one is Breaking this sequence (e.g., by serving and eating
“reasonably clear about what variables may be in cake before the guests arrive) violates elements of the
play in a situation, the natural next query is, ‘How do category “birthday party” and raises questions about
they relate to each other?’” whether the event falls into this category. Indeed, a
Powerful examples of how categories relate to one central part of the categorization process is to create
another appear in L evi-Strauss’s (1969) The Raw and sequential relationships among categories, be they
the Cooked and Douglas’s (1966) Purity and Danger. concepts or mechanisms (Ahn, 1999; Murphy &
Levi-Strauss’s study illustrates that myths cannot be Medin, 1985). In qualitative theory building, re-
understood in isolation: he describes a collection of searchers can use the sequencing move to temporally
myths from tropical South America and points out organize the categories they have identified in the
how duality and oppositions (such as raw vs. data. The categories created by researchers consist of
cooked) are fundamental to humans’ understanding sequences of actions (mechanisms), objects, persons,
of society. Douglas’s (1966) seminal work on purity and events (concepts) that together form complex
makes a similar relational point: specifically, that theories of reality.
the categories of “pure” and “impure” function as a A rich tradition in qualitative research has focused
dyad, creating oppositions that enable individuals to on sequential processes as a lens through which to
make sense of their worlds; in other words, we know achieve a better understanding of various phenome-
what is impure by comparing it to what is pure. na (Langley, 1999; Langley, Smallman, Tsoukas, &
A more recent example of such relational dynam- Van de Ven, 2013; Tsoukas & Chia, 2002). From this
ics between categories appears in a study of profes- viewpoint, the sequencing of categories matters deci-
sionals’ social standing in organizations. The study sively: an [A–B–C] sequence might differ from a
examined commander–medical provider interac- [C–B–A] sequence. As Locke (2001: 42) noted, pro-
tions in a military setting. As DiBenigno (2018: 18) cess-oriented research can be described in many
explained, she coded whether observed interactions ways, but all “reflect a common element, namely
entailed “developing personalized relationships time.” Langley (1999: 696) reminded us that events,
across groups” or “anchoring group members in their phases, incidents, and ordering are all fundamental
home group identity.” These two categories could be to the coding process. Process scholars position time
understood separately, but they only fully made at the forefront of their analysis, but many other
2021 Grodal, Anteby, and Holm 603

qualitative researchers also recognize the impor- offer a novel perspective on the interrelationship be-
tance of time in theory elaboration. In The Discovery tween concepts and mechanisms. Such efforts can
of Grounded Theory, Glaser and Strauss (1967: add to or even contradict what other theoretical
213–214) discussed sequencing in conjunction with lenses would predict, thereby fulfilling the goal of
studying research productivity. Longer hours spent achieving a more comprehensive understanding of
on research, they noted, can foster a higher level of our social worlds.
motivation and, thus, superior performance, but
those who perform at a higher level might be more Stabilizing Categories: Re-Analyses and
motivated than their peers and might therefore Theoretical Integration
spend more time conducting research. Thus, how
the categories “hours spent at work,” “motivation,” In the final analytical stage, the researcher aims to
and “productivity” are temporally aligned can yield create a theoretical scaffold to explain the studied
two very different theoretical stories, one led by the phenomena by reanalyzing existing categories and
hours worked and the other by high performance. integrating identified mechanisms and concepts.
Such distinct theoretical interpretations point to This stabilizing stage often allows researchers to pro-
why qualitative researchers need to pay close atten- vide answers to or explanations for their initial ques-
tion to the sequencing of categories when analyzing tions or puzzles. One of the critical moves in this
their data. stage entails “developing or dropping working hy-
Sequencing is important in qualitative analysis be- potheses.” This move allows the researcher to exam-
cause the interrelationship between categories is ine whether the data truly support the theoretical
essential to identifying processes and their temporal- conclusions reached in the prior stages and to be
ity. In Jarzabkowski’s (2008) investigation of top more actively reflexive of how these conclusions
managers’ strategizing behavior, an important part of were reached.
the analysis was to create a temporal sequence. The Developing or dropping working hypotheses.
author described how she “decomposed each chro- Categories are part of larger “theories” or “working
nology into analytical periods time 1, time 2, and, if hypotheses” that we have formed to explain how the
relevant, time 3 according to key strategic responses world works on the basis of prior experiences (Hirsch-
or a discernable shift in top manager behavior” feld & Gelman, 1994; Murphy, 2004). We know, for
(p. 626). She then generated categories to character- example, that, when apples are dropped, they fall to
ize the behavior within each period and sequenced the ground because of gravity. When encountering
them in time. Similarly, to understand how organiza- other instances of dropped objects, we assume that
tions legitimated the new market category “satellite gravity is also involved and look for evidence of gravi-
radio,” Navis and Glynn (2010) compared patterns of ty in these new contexts. Gravity is thus part of our
categories associated with data collected over a period theory of the category “apple,” and we would pause
of six years. By sequencing categories, the authors un- in astonishment if an apple did not fall but flew into
covered specific shifts in organizational identity and the sky when dropped. In this instance, we would
audience attention. Finally, Barley’s (1986: 86) exami- probably begin to question whether the object we had
nation of technology change as an occasion for social dropped actually belonged to the category “apple,”
structuration relied on data analysis that “traced the because its behavior did not conform to our “theory”
analytic logic suggested by the sequential model of of the apple category. Thus, how we construct catego-
the structuring process”; this indicated that the se- ries cannot be decoupled from the broader theoretical
quencing of scripts (or categories) mattered as much conjectures that form an integral part of a category
as the categories per se. In all three examples, se- (Durand & Paolella, 2013; Murphy & Medin, 1985). In
quencing enabled the researchers to develop novel in- our interactions with the world, we are constantly up-
sights into the temporal relationship between dating our working hypotheses and fitting them to our
categories. lived experiences (Hirschfeld & Gelman, 1994). Such
In the process of juggling these merging, splitting, updating might lead us to replace hypotheses about
relating or contrasting, and sequencing moves, the relationships that we once considered true with new
researcher begins to tease out some categories that ones (Ahn, 1999). In developing or dropping a work-
function more like mechanisms and others that func- ing hypothesis, researchers formulate an overarching
tion more like concepts. This is important because understanding of their data. As they iterate through
one of the fundamental elements of theory develop- their data, they find either confirmatory evidence,
ment is not only to identify categories but also to spurring them to elaborate the hypothesis, or
604 Academy of Management Review July

contradictory or unsupportive evidence, prompting returned to the data and found support for this
them to drop it. hypothesis.
Past advice on how best to approach qualitative Unsupported working hypotheses, however, often
data analysis emphasizes the need to develop, but disappear from the public view: typically, only suc-
also to drop, working hypotheses. As Luker (2008: cessful hypotheses graduate to presentation in print.
199) elegantly articulated it, “We make private bets Some authors’ retrospective comments on published
with ourselves about what features … will turn out pieces allow us nonetheless to see how working hy-
to be actual patterns. We know we will often be potheses wax and wane. For instance, Chan and An-
wrong … but we hang in there.” The frequent rec- teby (2016) began studying frontline officers at the
ommendation is for qualitative scholars to “abduct” U.S. Transportation Security Administration a de-
insights from their data (Timmermans & Tavory, cade after its creation. Initially, they focused on de-
2012) without fully committing to them, so as to be veloping categories capturing how employees had
able to redirect their analytical eye and remain open been socialized into a possible new profession. Find-
to surprises. Locke and her coauthors (2008: 907) ing little support for the hypothesis that their data
noted the coexistence of “doubt and belief” when an- spoke to this theme, however, they rapidly refocused
alyzing data, suggesting that both knowing and not the categorical analysis onto what interviewees
knowing are central to the research process. The deemed most salient with respect to gender discrimi-
sense of knowing might lead to what some authors nation and surveillance at work, and developed new
call “propositions,” but propositions ought to be working hypotheses (Anteby & Chan, 2018; Chan &
grasped loosely until each is either supported by fur- Anteby, 2016). In general, when researchers progress
ther data analysis or dropped (Lofland et al., 2006: from data analysis to theory development, develop-
176). This “back-and-forth … in which concepts, ing or dropping working hypotheses is an important
conjectures, and data are in continuous interplay” move: it prompts researchers via multiple iterations
serves as the backbone of many analyses (Van Maa- through the data either to marshal strong support
nen, Sørensen, & Mitchell, 2007: 1146). Such a shift- and validity for their assumptions or to realize that
ing picture explains why Miles et al. (2014: 99) their initial assumptions were faulty and to discover
recommended developing propositions to new ways of interpreting their data (Locke et al.,
“formalize and systematize the researcher’s thinking 2015). Once categories are stabilized and hypotheses
into a coherent set of explanations,” yet immediately are supported, researchers can propose a scaffold
warned researchers to adopt “safeguards against pre- that both fits their data and advances theory.
mature and unwarranted closure.” One way to build
such safeguards, they observed, is to purposely rate DISCUSSION
each conjecture as strong, qualified, neutral, or con-
tradictory, thus allowing the researcher to scrutinize The most important, yet mysterious, aspect of
those most likely to emerge from the data and discard qualitative data analysis is rigorous theory genera-
those that no longer fit. tion. In this paper, we have integrated insights from
In their study of women crying at work, Elsbach categorization theory with often-neglected methodo-
and Bechky (2018: 134–135) started their analysis by logical strategies to develop a framework for how
identifying and coding events when women cried at scholars can achieve and demonstrate rigor in quali-
work to form a richer understanding of this phenome- tative analysis. By being reflexive about their active
na. After the initial coding, they “developed a prelim- role in confronting and creating categories, scholars
inary model of defining the common situations in can be more transparent about their choice of moves,
which women cried at work” and hypothesized that and thus increase the rigor of their analytical process
the forms of situational emotional display could pre- by making it easier for readers to assess the work.
dict the observers’ reactions. They then returned to Our framework provides an overview of how re-
the literature to make sense of their data and began to searchers’ purposeful use of multiple moves can gen-
understand their data in light of the theory of erate theory from data. By spotlighting the various
“emotional display rules.” They found, however, that moves in which qualitative researchers engage, our
their data “did not fit with these simplistic rules.” In- hope is both to help fuel more rigorous scholarship
stead, through engagement with existing theory, they and to allow readers of qualitative research to evalu-
went on to create the new hypothesis that observers’ ate qualitative research efforts more effectively.
reactions could be explained by their own “cognitive Although any of the moves can occur at any point
scripts.” After forming this new hypothesis, they in time during the analytical process, the generation
2021 Grodal, Anteby, and Holm 605

of initial categories often occurs when researchers others to replicate the same theoretical contribution
ask questions that elicit insights from the data or when presented with the same data (Aguinis & Solar-
when they focus on puzzles in which the data con- ino, 2019). Others, however, have posited that this
flict with existing theories. After this initial stage, concept of “replicability,” which is borrowed from
and in order to refine tentative categories, qualitative quantitative research, should be interpreted differ-
scholars can drop, merge, or split categories, to make ently within qualitative scholarship (Pratt et al.,
them more or less encompassing, or organize catego- 2020). As Small (2009: 28) noted, qualitative re-
ries, either by sequencing them temporally or by re- searchers might need to embrace “alternative episte-
lating or contrasting them with each other. Finally, mological assumptions better suited for their unique
in order to stabilize categories, researchers develop questions rather than retreat toward models de-
or drop working hypotheses as they build on their signed for statistical descriptive research.”
analysis to progress toward theoretical insights. The In sum, unless we assume that all scholars have
model of theory development that we suggest is ac- the same cognitive or experiential predispositions,
tive and iterative: researchers constantly cycle the question of replicability becomes partly mute. In-
through multiple and distinct moves while purpose- stead of focusing on replicability as the path to rigor,
fully probing and revisiting initial categories. We ar- we suggest that scholars can achieve and demon-
gue that qualitative researchers who adopt such an strate rigor in qualitative research by being, first,
approach to data analysis will not only improve the more reflexive and, subsequently, more transparent
transparency and rigor of their work but also be bet- about their metacognitive knowledge—that is,
ter equipped to develop powerful theories that break knowledge about their own knowledge and goals
with existing understandings of the world. When re- (Pintrich, 2002). As researchers, we might obtain a
searchers become more reflexive, they open the greater degree of rigor in our work if we are explicit
black box of qualitative analysis to gain greater in- about and carefully detail our own knowledge and
sight into their own analytical process, which allows goals before and during our data collection and anal-
them to question and challenge their own assump- ysis. It is thus important that researchers find ways
tions. It is this prodding that might allow them to de- to report and represent their data that increases the
velop greater theoretical insights. transparency of their research process. Researchers
can do this either verbally, by detailing the moves
Rigor via Reflexivity on Scholars’ Active Role in that they engaged in and the consequences each
Categorization move had for their theoretical development, or they
can create visual representations of the analytical
Qualitative researchers have long been attuned to process that are true to how the research unfolded
the challenges associated with achieving rigor in and that show the complexity and messiness of the
their work and the need to detail their movement categorization process.
from data to theory carefully (Glaser & Strauss, 1967:
244). Indeed, the most significant challenge for qual-
Unfolding the Analytical Process: An
itative researchers “is not to find more rigorous
Opportunity for Theory Building
methods … [instead] the challenge … is to convince
its practitioners that they owe it to themselves … to We argue that, instead of mapping out all the pos-
explicate their procedures fully” (Comaroff, 2005: sible and replicable categories in any given data set,
38). Nevertheless, recent debates, inspired in part by qualitative researchers achieve rigor by tracing and
the rise of the behavioral sciences, have made it detailing their unique pathway through the data. Ad-
more urgent to address these concerns (Pratt et al., ditionally, we argue that active categorization is
2020). We argue that, by being more reflexive about more likely to yield theoretical insights. Following
the active process of categorization, scholars can bet- categorization theory, we suggest that entering the
ter demonstrate transparency in their movement field with some theoretical understanding is both
from data to theory, which will allow readers to bet- unavoidable and generative, and that not all data are
ter evaluate the rigor of their work. equally valuable for generating theory. For example,
Taking an active perspective on the process of ana- categorization theory suggests that the process of cat-
lyzing qualitative data suggests that researchers egorizing fundamentally forces us to not pay atten-
bring their own experiences and goals to bare on the tion to all the stimuli we encounter (Murphy, 2004).
analytical process. Some scholars have argued that a If we paid attention to all the information at once, we
sign of rigor in qualitative analysis is the ability of would suffer from cognitive overload and be unable
606 Academy of Management Review July

to create a coherent representation of the world, puzzles do not evolve. Nonetheless, driving the ana-
let alone navigate through it (Rosch, 1978). Likewise, lytical process with tentative questions or puzzles is
because most areas of organizational life are already critical to theory development.
explained by existing theories, it is important for
qualitative researchers to not focus on the elements Plurality within Qualitative Analysis
of their data that can be explained within existing
theories. Instead, we need to focus on the questions Adopting an active categorization frame means
and puzzles that challenge the current understand- that researchers can approach the same data with dif-
ing of the phenomenon at hand and clearly explain ferent goals and analyze them in different ways. The
the path that helped us to reach our conclusions— theoretical insights that are drawn from the data are
that is, we need to focus on salient elements of the thus not simply “given” in the data but actively con-
data that cannot be explained within our existing cat- structed by researchers to address puzzles that they
egorical apparatus (Ahn, 1999; Sloman et al., 1998). find interesting and important. Other researchers
Theory represents an abstracted and simplified might view the same data very differently: as we
lens on reality that contributes to a specific body of know, many valuable insights emerge from the inter-
literature (Merton, 1973). Strong theoretical pieces action of different individuals in distinct fields (An-
shine because of what they teach us: they help us to teby, 2013; de Rond & Tunçalp, 2017; Hudson &
see new connections, revisit our preconceived view- Okhuysen, 2014; Ketokivi & Mantere, 2010; Louis &
points, and develop new takes on old questions. Bartunek, 1992). Building on the discussion by La-
When reading seminal academic articles and books, mont and Swidler (2014) of diversity in qualitative
we focus on their theoretical insights and can easily methodologies, we call for more plurality both with-
gloss over the analytical choices that underpin these in and across qualitative studies to fully leverage the
findings. Nevertheless, these insights are enabled by richness and potential of qualitative data. By com-
how scholars have generated theory from data. This bining analytical moves in different ways, research-
is particularly true of qualitative data analysis, ers are ultimately likely to generate more diverse
which relies on researchers’ interpretations of their insights from the same data set, thus contributing to
data and the analytical moves they make to generate a broader range of literature.
theory from data. Such an approach counters the il- Beyond the fact that different paths can be taken
lusion of a fully detached and comprehensive induc- through the data, if researchers foreground certain
tion of theory from data. The comprehensive yet moves over others, it might shape their theory devel-
unintentional categorization of qualitative data is opment. A plurality of analytical approaches is
unlikely to yield novel theoretical insights because it therefore likely to foster more diversity in contribu-
will mostly reproduce an existing understanding. tions. For example, if researchers focus more on puz-
An active categorization framework recognizes the zles, they are more likely to generate theory that
unfolding of the research process and implies that re- challenges existing theoretical understanding be-
searchers select different moves at different times in cause the puzzling phenomena in the data derive
the analytical process to generate rigorous theory. precisely from the contrast they present when juxta-
For example, “asking questions” might be particu- posed with existing theories. This is the case in the
larly well suited to early phases of data analysis, pri- analysis by McPherson and Sauder (2013: 165) of
or even to developing categories, or to very late how occupational groups can advantageously hijack
phases in which researchers feel stuck. Similarly, one another’s occupational logics. By focusing on
the development of working hypotheses is most like- the puzzle of why some occupational groups did not
ly to occur after several categories have emerged and uniformly display their own occupational logics,
been placed in relation to one another. Thus, focus- McPherson and Sauder (2013) were able to demon-
ing exclusively on one move only tilts a scholar’s ef- strate that logics can be used for strategic purposes to
forts and attention toward a single phase of the “manage institutional complexity, reach consensus,
analytical process, which, although important, and get the work … done.” Thus, because the au-
might not stand alone. By spotlighting a researcher’s thors focused on a specific puzzle in their data, they
potential toolkit of moves, we emphasize the sequen- developed an insight that broke with the prevailing
tial and iterative nature of any analysis and under- understanding of how logics operate.
line that concentrating on a single phase or move Likewise, if researchers ask questions of their data,
might restrict our ability to theorize in all phases of they will develop theory that might be more likely to
the research. This is not to say that questions or expand existing understanding. As an illustration,
2021 Grodal, Anteby, and Holm 607

Tripsas and Gavetti (2000) asked how cognitions (e.g., Bingham & Kahl, 2013; Bloom, 2000; Bowker &
(and not capabilities) might shape firms’ abilities to Star, 2000; Khaire & Wadhwani, 2010; Lakoff, 2008;
manage technological discontinuities. From this Murphy, 2004; Rosch, 1978). Gehman, Glaser, Eisen-
point of departure, they were able to develop a per- hardt, Gioia, Langley, and Corley (2018: 14) coun-
spective that departed radically from prior views of seled that “it is important to customize the method
how firms manage technological transitions. If re- for your research context. Research situations are dif-
searchers start by merging or splitting categories, ferent and require the use of tools and techniques in
without asking questions or focusing on puzzles, it is different ways.” A call for increased rigor should
often difficult for them to gain theoretical traction, therefore not be confused with a call for uniformity.
because, without a guiding question or puzzle, they
can easily become overwhelmed by the multiplicity
of possibilities and asphyxiate in their data (Petti- CONCLUSION
grew, 1990). Overall, this article posits that qualitative analysis
The role of dropping, merging, splitting, relating, is fundamentally a categorization process. By draw-
and sequencing categories is to refine categories that ing on categorization theory, we spotlight the range
are created around an initial puzzle. These moves of diverse ways in which categories are actively
are likely to generate greater specificity in the result- formed. To obtain rigor in qualitative analysis, we
ing theory because they provide the researcher with need to be transparent about these active moves be-
more nuances in the created categories. However, cause otherwise readers cannot assess the precision
prioritizing any of these moves is likely to generate of our work. We also suggest that adopting an active
different theoretical outcomes. An emphasis on categorization framework has critical consequences
dropping will, for example, narrow researchers’ fo- for how qualitative analysis unfolds and for the theo-
cus on a smaller part of the data, allowing them to de- retical insights we generate. If we do not problemat-
velop a more detailed explanation of a smaller piece ize the creation of analytical categories, we will be
of the reality that is captured in the data. This is often limited in the kinds of theoretical insights we are
a cornerstone of qualitative analysis, as any given
able to develop. In other words, an active categoriza-
qualitative data set is multifaceted and complex and
tion framework can liberate scholars to discover
can be used to address a wide variety of theoretical
pathways through the data that might be less heavily
questions. If researchers emphasize merging and
traveled and, thus, more innovative. Moving in lock-
splitting, the resulting theory is likely to be better at
step, using identical tools, we may appear aligned
explaining cross-sectional variation, as has been
and convey an impression of rigor when, in reality,
done in research drawing on multiple cases (e.g.,
we are trampling on the seeds of our own
Graebner & Eisenhardt, 2004; Ozcan & Eisenhardt,
imaginations.
2009) or a matched pairing of cases (e.g., Barley,
1986; Kellogg, 2009, 2019). In contrast, if researchers
focus to a greater degree on sequencing, their result- REFERENCES
ing theories are more likely to excel at explaining Abbott, A. 2004. Methods of discovery: Heuristics for the
processes and mechanisms (e.g., Navis & Glynn, social sciences (contemporary societies). New York,
2010; Zietsma & Lawrence, 2010). NY: W.W. Norton & Company.
The timing of our call for increased plurality in Aguinis, H., & Solarino, A. M. 2019. Transparency and rep-
qualitative research is important. Although both cat- licability in qualitative research: The case of inter-
egorization scholars and qualitative scholars have views with elite informants. Strategic Management
shown that categorization processes are diverse, a re- Journal, 40: 1291–1315.
newed positivist orientation has, in practice, engen-
Ahn, W.-K. 1999. Effect of causal structure on category
dered uniformity in analytical approaches to
construction. Memory & Cognition, 27: 1008–1023.
qualitative organizational research (Bansal & Corley,
2011; Gioia et al., 2013; Langley & Abdallah, 2015). Anteby, M. 2013. Relaxing the taboo on telling our own
Nevertheless, as we have demonstrated, scholars stories: Upholding professional distance and personal
across the social sciences have found that categories involvement. Organization Science, 24: 1277–1290.
are often generated not solely through applying a sin- Anteby, M., & Chan, C. K. 2018. A self-fulfilling cycle of
gle move but by focusing on puzzles that arise in the coercive surveillance: Workers’ invisibility practices
data and by applying a variety of possible moves to and managerial justification. Organization Science,
generate categories that yield answers to that puzzle 29: 247–263.
608 Academy of Management Review July

Anteby, M., & Molnar, V. 2012. Collective memory meets Contemporary field research: 109–126. Boston, MA:
organizational identity: Remembering to forget in a Little Brown.
firm’s rhetorical history. Academy of Management Charmaz, K. 2006. Constructing grounded theory: A prac-
Journal, 55: 515–540. tical guide through qualitative analysis. Los Ange-
Anteby, M., & Occhiuto, N. 2020. Stand-in labor and the les, CA: SAGE.
rising economy of self. Social Forces, 98: 1287–1310. Clark, H. H., & Wilkes-Gibbs, D. 1986. Referring as a collab-
Armstrong, S. L., Gleitman, L. R., & Gleitman, H. 1983. orative process. Cognition, 22: 1–39.
What some concepts might not be. Cognition, 13: Cohen, L. E. 2013. Assembling jobs: A model of how tasks
263–308. are bundled into and across jobs. Organization Sci-
Bansal, P., & Corley, K. 2011. From the editors—The com- ence, 24: 432–454.
ing of age for qualitative research: Embracing the di- Collins, A. M., & Loftus, E. F. 1975. A spreading-activation
versity of qualitative methods. Academy of theory of semantic processing. Psychological Review,
Management Journal, 54: 233–237. 82: 407–428.
Barley, S. R. 1986. Technology as an occasion for structur- Colyvas, J. A., & Powell, W. W. 2006. Roads to institution-
ing: Evidence from observations of CT scanners and alization: The remaking of boundaries between public
the social order of radiology departments. Adminis- and private science. Research in Organizational Be-
trative Science Quarterly, 31: 78–108. havior, 27: 305–353.
Barsalou, L. W. 1983. Ad hoc categories. Memory & Cogni- Comaroff, J. 2005. Notes on anthropological method,
tion, 11: 211–227. mainly in the key of E. Workshop on interdisci-
Becker, H. S. 2008. Tricks of the trade: How to think plinary standards for systematic qualitative re-
about your research while you’re doing it. Chicago, search: 28–36. Washington, DC: National Science
Foundation.
IL: University of Chicago Press.
Corbin, J. M., & Strauss, A. L. 1990. Grounded theory re-
Behfar, K., & Okhuysen, G. A. 2018. Discovery within vali-
search: Procedures, canons, and evaluative criteria.
dation logic: Deliberately surfacing, complementing,
Qualitative Sociology, 13: 3–21.
and substituting abductive reasoning in hypothetico-
deductive inquiry. Organization Science, 29: Craciun, M. 2018. Emotions and knowledge in expert
323–340. work: A comparison of two psychotherapies. Ameri-
can Journal of Sociology, 123: 959–1003.
Berger, P. L., & Luckmann, T. 1967. The social construc-
tion of reality: A treatise in the sociology of knowl- Creed, W. E. D., Dejordy, R., & Lok, J. 2010. Being the
edge. New York, NY: Anchor. change: Resolving institutional contradiction through
identity work. Academy of Management Journal, 53:
Bingham, C. B., & Kahl, S. J. 2013. The process of schema 1336–1364.
emergence: Assimilation, deconstruction, unitization
and the plurality of analogies. Academy of Manage- Crosina, E., & Pratt, M. G. 2019. Toward a model of organi-
ment Journal, 56: 14–34. zational mourning: The case of former Lehman Broth-
ers bankers. Academy of Management Journal, 62:
Bloom, P. 2000. How children learn the meanings of 66–98.
words. Cambridge, MA: MIT Press.
Davis, G. F., & Marquis, C. 2005. Prospects for organization
Booth, W. C., Colomb, G. G., & Williams, J. M. 2003. The theory in the early twenty-first century: Institutional
craft of research. Chicago, IL: University of Chicago fields and mechanisms. Organization Science, 16:
press. 332–343.
Bourdieu, P., & Wacquant, L. 1992. An invitation to reflex- de Rond, M., & Lok, J. 2016. Some things can never be un-
ive sociology. Chicago, IL: University of Chicago seen: The role of context in psychological injury at
Press. war. Academy of Management Journal, 59:
Bowker, G. C., & Star, S. L. 2000. Sorting things out: Clas- 1965–1993.
sification and its consequences. Cambridge, MA: de Rond, M., & Tunçalp, D. 2017. Where the wild things
MIT press. are: How dreams can help identify countertransfer-
Chan, C. K., & Anteby, M. 2016. Task segregation as a ence in organizational research. Organizational Re-
mechanism for within-job inequality: Women and search Methods, 20: 413–437.
men of the Transportation Security Administration. Desmond, M. 2012. Disposable ties and the urban poor.
Administrative Science Quarterly, 61: 184–216. American Journal of Sociology, 117: 1295–1335.
Charmaz, K. 1983. The grounded theory method: An expli- DiBenigno, J. 2018. Anchored personalization in managing
cation and interpretation. In R. M. Emerson (Ed.), goal conflict between professional groups: The case of
2021 Grodal, Anteby, and Holm 609

U.S. Army mental health care. Administrative Sci- as syndicate in entrepreneurial firms. Administrative
ence Quarterly, 63: 526–569. Science Quarterly, 49: 366–403.
Douglas, M. 1966. Purity and danger: An analysis of the Granovetter, M. S. 1973. The strength of weak ties. Ameri-
concepts of pollution and taboo. New York, NY: can Journal of Sociology, 78: 1360–1380.
Routledge.
Granqvist, N., & Gustafsson, R. 2016. Temporal institution-
Durand, R., & Paolella, L. 2013. Category stretching: Reor- al work. Academy of Management Journal, 59:
ienting research on categories in strategy, entrepre- 1009–1035.
neurship, and organization theory. Journal of
Grodal, S. 2018. Field expansion and contraction: How
Management Studies, 50: 1100–1123.
communities shape social and symbolic boundaries.
Durkheim, E., & Mauss, M. 2009. Primitive classification. Administrative Science Quarterly, 63: 783–818.
London, U.K.: Routledge.
Grodal, S., Nelson, A. J., & Siino, R. M. 2015. Help-seeking
Edmondson, A. C., & McManus, S. E. 2007. Methodologi- and help-giving as an organizational routine: Continu-
cal fit in management field research. Academy of al engagement in innovative work. Academy of Man-
Management Review, 32: 1246–1264. agement Journal, 58: 136–168.
Eisenhardt, K. M. 1989. Building theories from case study Hedstr€
om, P., & Swedberg, R. 1996. Social mechanisms.
research. Academy of Management Review, 14: Acta Sociologica, 29: 281–308.
532–550.
Hirschfeld, L. A., & Gelman, S. A. 1994. Mapping the
Eisenhardt, K. M., Graebner, M. E., & Sonenshein, S. 2016. mind: Domain specificity in cognition and culture.
Grand challenges and inductive methods: Rigor with- Cambridge, U.K.: Cambridge University Press.
out rigor mortis. Academy of Management Journal,
59: 1113–1123. Hochschild, A. R. 1983. The managed heart: Commer-
cialization of human feeling. Berkeley, CA: Universi-
Elsbach, K. D., & Bechky, B. A. 2018. How observers assess
ty of California Press.
women who cry in professional work contexts. Acad-
emy of Management Discoveries, 4: 127–154. Hudson, B. A., & Okhuysen, G. A. 2014. Taboo topics:
Structural barriers to the study of organizational stig-
Emerson, R. M., Fretz, R. I., & Shaw, L. L. 2011. Writing
ma. Journal of Management Inquiry, 23: 242–253.
ethnographic fieldnotes. Chicago, IL: University of
Chicago Press. Huising, R. 2015. To hive or to hold? Producing relational
authority through scut work. Administrative Science
Fine, G. A. 1983. Shared fantasy: Role-playing games as
Quarterly, 60: 263–299.
social worlds. Chicago, IL: University of Chicago
Press. Jarzabkowski, P. 2008. Shaping strategy as a structuration
process. Academy of Management Journal, 51:
Fine, G. A. 1993. Ten lies of ethnography: Moral dilemmas
621–650.
of field research. Journal of Contemporary Ethnogra-
phy, 22: 267–294. Kahneman, D., & Tversky, A. 1984. Choices, values, and
Gehman, J., Glaser, V. L., Eisenhardt, K. M., Gioia, D., frames. American Psychologist, 39: 341–350.
Langley, A., & Corley, K. G. 2018. Finding theory–me- Katz, J. 2001. Analytic induction revisited. In R. M. Emer-
thod fit: A comparison of three qualitative approaches son (Ed.), Contemporary field research: Perspec-
to theory building. Journal of Management Inquiry, tives and formulations: 331–334. Prospect Heights,
27: 284–300. IL: Waveland Press.
Gioia, D. A., Corley, K. G., & Hamilton, A. L. 2013. Seeking Kellogg, K. C. 2009. Operating room: Relational spaces and
qualitative rigor in inductive research: Notes on the microinstitutional change in surgery. American Jour-
Gioia methodology. Organizational Research Meth- nal of Sociology, 115: 657–711.
ods, 16: 15–31.
Kellogg, K. C. 2019. Subordinate activation tactics: Semi-
Glaser, B. G., & Strauss, A. L. 1967. The discovery of professionals and micro-level institutional change in
grounded theory: Strategies for qualitative re- professional organizations. Administrative Science
search. Chicago, IL: Aldine. Quarterly, 64: 928–975.
Golden-Biddle, K., & Locke, K. 1993. Appealing work: An Ketokivi, M., & Mantere, S. 2010. Two strategies for induc-
investigation of how ethnographic texts convince. Or- tive reasoning in organizational research. Academy of
ganization Science, 4: 595–616. Management Review, 35: 315–333.
Golden-Biddle, K., & Locke, K. 2007. Composing qualita- Khaire, M., & Wadhwani, R. D. 2010. Changing landscapes:
tive research. Thousand Oaks, CA: SAGE. The construction of meaning and value in a new mar-
Graebner, M. E., & Eisenhardt, K. M. 2004. The seller’s side ket category—modern Indian art. Academy of Man-
of the story: Acquisition as courtship and governance agement Journal, 53: 1281–1304.
610 Academy of Management Review July

Lakoff, G. 2008. Women, fire, and dangerous things. Chi- Lubet, S. 2017. Interrogating ethnography: Why evi-
cago, IL: University of Chicago Press. dence matters. New York, NY: Oxford University
Lamont, M. 2000. The dignity of working men: Morality Press.
and the boundaries of race, class, and immigration. Luker, K. 2008. Salsa dancing into the social sciences.
Cambridge, MA: Harvard University Press. Cambridge, MA: Harvard University Press.
Lamont, M., & Molnar, V. 2002. The study of boundaries in MacWhinney, B. (Ed.), 1987. Mechanisms of language ac-
the social sciences. Annual Review of Sociology, 28: quisition. Hillsdale, NJ: Lawrence Erlbaum.
167–195. Mazmanian, M. 2013. Avoiding the trap of constant con-
Lamont, M., & Swidler, A. 2014. Methodological pluralism nectivity: When congruent frames allow for heteroge-
and the possibilities and limits of interviewing. Quali- neous practices. Academy of Management Journal,
tative Sociology, 37: 153–171. 56: 1225–1250.
Lamont, M., & White, P. 2008. The evaluation of system- McPherson, C. M., & Sauder, M. 2013. Logics in action:
atic qualitative research in the social sciences. Managing institutional complexity in a drug court.
Washington, DC: National Science Foundation. Administrative Science Quarterly, 58: 165–196.

Langley, A. 1999. Strategies for theorizing from process Mears, A. 2015. Working for free in the VIP: Relational
data. Academy of Management Review, 24: work and the production of consent. American Socio-
logical Review, 80: 1099–1122.
691–710.
Merton, R. K. 1948. The self-fulfilling prophecy. Antioch
Langley, A., & Abdallah, C. 2015. Templates and turns in
Review, 8: 193–210.
qualitative studies of strategy and management. In
G. B. Dagnino & M. C. Cinici (Eds.), Research meth- Merton, R. K. 1973. The sociology of science: Theoretical
ods for strategic management: 155–184. London, and empirical investigations. Chicago, IL: University
U.K.: Routledge. of Chicago press.
Langley, A., Smallman, C., Tsoukas, H., & Van de Ven, Miles, M. B., Huberman, A. M., & Salda~
na, J. 2014. Quali-
A. H. 2013. Process studies of change in organization tative data analysis: A methods sourcebook. Thou-
and management: Unveiling temporality, activity, and sand Oaks, CA: SAGE.
flow. Academy of Management Journal, 56: 1–13. Murphy, G. L. 2004. The big book of concepts. Cambridge,
L
evi-Strauss, C. 1969. The raw and the cooked. New MA: MIT Press.
York, NY: Harper & Row. Murphy, G. L., & Lassaline, M. E. 1997. Hierarchical struc-
Locke, K. 2001. Grounded theory in management re- ture in concepts and the basic level of categorization.
In K. Lamberts & D. R. Shanks (Eds.), Knowledge, con-
search. London, U.K.: SAGE.
cepts, and categories: 93–131. Cambridge, MA: MIT
Locke, K., Feldman, M. S., & Golden-Biddle, K. 2015. Dis- Press.
covery, validation, and live coding. In K. D. Elsbach &
Murphy, G. L., & Medin, D. L. 1985. The role of theories in
R. M. Kramer (Eds.), Handbook of qualitative organi-
conceptual coherence. Psychological Review, 92:
zational research: Innovative pathways and meth-
289–316.
ods: 371–380. New York, NY: Routledge.
Nakamura, G. V. 1985. Knowledge-based classification of
Locke, K., Golden-Biddle, K., & Feldman, M. S. 2008. Mak-
ill-defined categories. Memory & Cognition, 13:
ing doubt generative: Rethinking the role of doubt in 377–384.
the research process. Organization Science, 19:
907–918. Navis, C., & Glynn, M. A. 2010. How new market catego-
ries emerge: Temporal dynamics of legitimacy, identi-
Lofland, J. 1971. Analyzing social settings. Belmont, CA: ty, and entrepreneurship in satellite radio,
Wadsworth Publishing. 1990–2005. Administrative Science Quarterly, 55:
Lofland, J., Snow, D., Anderson, L., & Lofland, L. H. 2006. 439–471.
Analyzing social settings: A guide to qualitative ob- Nelson, A. J., & Irwin, J. 2014. “Defining what we do—all
servation and analysis (4th ed.). Belmont, CA: Wads- over again”: Occupational identity, technological
worth/Thomson Learning. change, and the librarian/Internet-search relationship.
Lopez, S. H., & Phillips, L. A. 2019. Unemployed: White- Academy of Management Journal, 57: 892–928.
collar job searching after the Great Recession. Work O'Mahony, S., & Bechky, B. A. 2008. Boundary organiza-
and Occupations, 46: 470–510. tions: Enabling collaboration among unexpected al-
Louis, M. R., & Bartunek, J. M. 1992. Insider/outsider re- lies. Administrative Science Quarterly, 53: 422–459.
search teams: Collaboration across diverse perspec- Ozcan, P., & Eisenhardt, K. M. 2009. Origin of alliance
tives. Journal of Management Inquiry, 1: 101–110. portfolios: Entrepreneurs, network strategies, and firm
2021 Grodal, Anteby, and Holm 611

performance. Academy of Management Journal, 52: in the organizational sciences: 85–97. London, U.K.:
246–279. SAGE.
Petriglieri, J. L. 2015. Co-creating relationship repair: Path- Strauss, A. L. 1987. Qualitative analysis for social scien-
ways to reconstructing destabilized organizational tists. Cambridge, U.K.: Cambridge University Press.
identification. Administrative Science Quarterly, 60: Strauss, A. L., & Corbin, J. M. 1990. Basics of qualitative
518–557. research: Grounded theory procedures and techni-
Pettigrew, A. M. 1990. Longitudinal field research on ques. Los Angeles, CA: SAGE.
change: Theory and practice. Organization Science, Suddaby, R. 2006. What grounded theory is not. Academy
1: 267–292. of Management Journal, 49: 633–642.
Pintrich, P. R. 2002. The role of metacognitive knowledge Timmermans, S., & Tavory, I. 2012. Theory construction in
in learning, teaching, and assessing. Theory into Prac- qualitative research: From grounded theory to abduc-
tice, 41: 219–225. tive analysis. Sociological Theory, 30: 167–186.
Pratt, M. G., Kaplan, S., & Whittington, R. 2020. The tumult Tripsas, M., & Gavetti, G. 2000. Capabilities, cognition,
over transparency: Decoupling transparency from rep- and inertia: Evidence from digital imaging. Strategic
lication in establishing trustworthy qualitative re- Management Journal, 21: 1147–1161.
search. Administrative Science Quarterly, 65: 1–19.
Tsoukas, H., & Chia, R. 2002. On organizational becoming:
Pratt, M. G., Rockmann, K. W., & Kaufmann, J. B. 2006. Rethinking organizational change. Organization Sci-
Constructing professional identity: The role of work ence, 13: 567–582.
and identity learning cycles in the customization of
Turco, C. J. 2016. The conversational firm: Rethinking
identity among medical residents. Academy of Man- bureaucracy in the age of social media. New York,
agement Journal, 49: 235–262. NY: Columbia University Press.
Quillian, M. R. 1969. The teachable language compre- Van Maanen, J. 1979. The fact of fiction in organizational
hender: A simulation program and theory of language. ethnography. Administrative Science Quarterly, 24:
Communications of the ACM, 12: 459–476. 539–550.
Ramus, T., Vaccaro, A., & Brusoni, S. 2017. Institutional Van Maanen, J. 2011. Tales of the field: On writing eth-
complexity in turbulent times: Formalization, collabo- nography. Chicago, IL: University of Chicago Press.
ration, and the emergence of blended logics. Academy
of Management Journal, 60: 1253–1284. Van Maanen, J., Sørensen, J. B., & Mitchell, T. R. 2007. The
interplay between theory and method. Academy of
Rosch, E. 1978. Principles of categorization. In E. Rosch & Management Review, 32: 1145–1154.
B. B. Lloyd (Eds.), Cognition and categorization:
Vaughan, D. 1996. The Challenger launch decision: Risky
27–48. Hillsdale, NJ: Lawrence Erlbaum Associates.
technology, culture, and deviance at NASA. Chica-
Searle, J. R., & Willis, S. 1995. The construction of social go, IL: University of Chicago Press.
reality. New York, NY: Simon and Schuster.
Vergne, J.-P., & Wry, T. 2014. Categorizing categorization
Siegler, R. S. 1998. Emerging minds: The process of research: Review, integration, and future directions.
change in children’s thinking. New York, NY: Oxford Journal of Management Studies, 51: 56–94.
University Press.
Vygotsky, L. S. 1987. Thought and language. New York,
Sloman, S. A., Love, B. C., & Ahn, W.-K. 1998. Feature cen- NY: Plenum Press.
trality and conceptual coherence. Cognitive Science,
Weber, K., Heinze, K. L., & DeSoucey, M. 2008. Forage for
22: 189–228.
thought: Mobilizing codes in the movement for grass-
Small, M. L. 2009. “How many cases do I need?” On sci- fed meat and dairy products. Administrative Science
ence and the logic of case selection in field-based re- Quarterly, 53: 529–567.
search. Ethnography, 10: 5–38.
Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. 2005. Organiz-
Small, M. L. 2013. Causal thinking and ethnographic re- ing and the process of sensemaking. Organization
search. American Journal of Sociology, 119: Science, 16: 409–421.
597–601.
Yin, R. K. 2003. Case study research: Design and meth-
Smith, E. R., & Zarate, M. A. 1992. Exemplar-based model ods. Thousand Oaks, CA: SAGE.
of social judgment. Psychological Review, 99: 3–21. Zbaracki, M. J. 1998. The rhetoric and reality of total quali-
Spradley, J. P. 1979. The ethnographic interview. New ty management. Administrative Science Quarterly,
York, NY: Holt, Rinehart and Winston. 43: 602–636.
Staw, B. M. 1995. Repairs on the road to relevance and rig- Zhao, E. Y., Ishihara, M., Jennings, P. D., & Lounsbury, M.
or. In L. L. Cummings & P. J. Frost (Eds.), Publishing 2018. Optimal distinctiveness in the console video
612 Academy of Management Review July

game industry: An exemplar-based model of proto- Michel Anteby ([email protected]) is a professor of

category evolution. Organization Science, 29: management and organizations at Boston University’s
558–611. Questrom School of Business and (by courtesy) sociology at
Zietsma, C., & Lawrence, T. B. 2010. Institutional work in Boston University’s College of Arts & Sciences. His research
the transformation of an organizational field: The in- looks at how people relate to their work, their occupations,
terplay of boundary work and practice work. Admin- and the organizations to which they belong. Recent
istrative Science Quarterly, 55: 189–221. empirical foci of his inquiries have included correctional
officers, ghostwriters, and subway drivers.
Audrey L. Holm ([email protected]) is a doctoral candidate
in Management and Organizations at Boston University’s
Stine Grodal ([email protected]) is a distinguished Questrom School of Business. Her research focuses on
professor of entrepreneurship and innovation at shifting occupational dynamics, evolving relationships at
Northeastern University's D'Amore-McKim School of work, and new labor market challenges. Her dissertation
Business. Her research examines the emergence and looks at how members of intermediary-organizations
evolution of markets, industries and fields with a specific (such as back-to-work programs) accompany formerly
focus on the strategies firms and other industry incarcerated people in their job searches.
stakeholders use to shape and exploit new market
categories.
Copyright of Academy of Management Review is the property of Academy of Management
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.