Professional Documents
Culture Documents
net/publication/274694952
CITATIONS READS
20 10,121
1 author:
Sarah Kaplan
University of Toronto
62 PUBLICATIONS 5,920 CITATIONS
SEE PROFILE
All content following this page was uploaded by Sarah Kaplan on 08 April 2015.
Sarah Kaplan
Rotman School, University of Toronto
105 St. George St., Room 7068,
Toronto, ON, M5S 3E6
sarah.kaplan@rotman.utoronto.ca
Forthcoming
Handbook of Innovative Qualitative Research Methods:
Pathways to Cool Ideas and Interesting Papers
K. Elsbach and R. Kramer, eds.
October 1, 2014
The author would like to acknowledge useful comments from Valentina Assenova, Jillian Chown, Kim
Elsbach, Rod Kramer, Michael Lounsbury, Chad McPherson, Matthew Regele, Michael Sauder, Amy
Wrzesniewski and the slump_management research group. All errors and omissions remain my own.
Abstract: This chapter offers several models of ways that scholars can usefully integrate
qualitative and quantitative research in both single works and broader research programs: using
qualitative methods to identify constructs or hypotheses that can subsequently be examined in quantitative
studies; using quantitative studies to identify patterns that can be explored in qualitative work; and,
quantification of qualitative data. Through this analysis, I call into question existing characterizations of
the relationships among the two types of research and offer a more nuanced portrayal of the links between
them. I suggest that the advantage of mixed methods studies is that they recognize the value of iterating
between that which can be counted and that which cannot in order to generate richer insights about the
phenomena of interest.
In the field of management, there is often a gulf between qualitative and quantitative research.
Despite a wealth of literature on mixing qualitative and quantitative methods of research (Bryman, 2006;
Creswell & Clark, 2007; Jick, 1979; Small, 2011), rarely do individual scholars conduct both forms of
research and even more rarely do they present them together in a single scholarly work. Quantitative
researchers may view qualitative research as suspect, or only as a means to generate hypotheses to test in
quantitative analyses. According to this view, qualitative research is but an intermediate output supporting
“empirical” regression analysis. On the other hand, qualitative researchers often view quantitative work as
too reductive. They may only include quantitative data for rhetorical purposes, recognizing that they are
outnumbered (no pun intended) in a world dominated by scholars untrained in qualitative methods.
There are many understandings of the words “qualitative” and “quantitative” within various
research traditions. For the purposes of this chapter, I take the broadest of definitions of each, where
qualitative research is any analysis based on words (e.g., texts, interviews, or field notes from
observations), and quantitative research as any analysis using numbers, either descriptively or in
regression analysis.
In this chapter, I offer several models of ways that scholars can usefully integrate qualitative and
quantitative research with the hopes of expanding the perspectives of scholars in both camps. In doing so,
I call existing characterizations of the relationships among the two types of research into question and
offer a more nuanced portrayal of the relationship between them. That is, qualitative methods can be used
to understand mechanisms that underlie relationships identified in quantitative research as much as they
can be sources of hypotheses or constructs for “large n” analyses (see Figure). Further, quantifications of
findings. Small (2011) and Creswell and Clark (2007) among others provide very detailed analyses of
mixed methods research and the various controversies within it. I have a much narrower goal in this
chapter: to identify several models for integrating qualitative and quantitative data and analyses into a
research paper (or in a research program) (see Table), highlighting the potential benefits and pitfalls of the
various approaches. Through this analysis, I demonstrate that the line in the sand between qualitative and
quantitative research may get in the way of interesting and compelling research projects and programs.
Use*qualita've*methods*to*understand*
mechanisms*underlying*rela'onships*
iden'fied*in*quan'ta've*regression*analysis*
Qualita've*research* Quan'ta've*research*
Use*qualita've*findings*to*generate*
hypotheses*to*be*tested*with*a*“large*n”*
sample*in*quan'ta've*regression*analysis*
*
Quan'fy*qualita've*data*to*validate*
observed*pa>erns*
Small (2011) makes important distinctions between complementary and confirmatory purposes
for integrating qualitative and quantitative methods, and also between sequential and concurrent research
designs. When planning research, it is certainly important to think about these choices. However, my own
experience and that of many of the scholars’ whose work I describe below (as told to me in interviews), is
that these purposes evolve over time as the research unfolds. Research intended to be confirmatory ends
up offering new, complementary insights. Research that was intended only to be quantitative becomes
multimethod as reviewers ask for evidence to elucidate the proposed mechanisms. The published articles
or books may not reflect these vagaries of the research process, but researchers engaging in mixed
Mixing
qualitative
and
quantitative
methods
in
a
Mixing
qualitative
and
quantitative
methods
in
a
single
research
paper
research
program
Qualitative
before
quantitative:
Qualitative
data
to
Lounsbury
(1997,
1998,
2001):
qualitative
fieldwork
justify
the
development
and
testing
of
a
construct…
used
to
develop
a
rich
understanding
of
the
context
• in
archival
regression
analysis
(Tripsas
1997;
Briscoe
which
is
presented
in
a
series
of
papers
taking
2007;
Canales
2014;
Doering
2014)
different
theoretical
angles.
The
qualitative
evidence
• in
a
survey
(Ely
1994;
Edmondson
1999;
Lounsbury
is
combined
with
a
subsequent
survey
for
quantitative
2001;
Bresman
2010)
analysis
in
on
paper.
• in
an
experiment
(Elsbach
1994;
Ranganathan
2014)
Kaplan
(2008
a,
b):
concurrent
qualitative
and
Quantitative
before
qualitative:
Survey
or
archival
quantitative
studies,
iterating
between
a
quantitative
data
based
regressions
to
establish
phenomenon
and
analysis
of
relationships
and
a
qualitative
analysis
of
fieldwork
to
understand
the
mechanisms
(Fernandez-‐ mechanisms.
Published
as
two
separate
papers,
but
Mateo
2009,
Bidwell
2009,
Edmondson
1999)
the
emerging
insights
in
each
shaped
the
research
strategies
and
analysis
of
the
other.
Quantifying
qualitative
data
to
enrich
analysis,
validate
qualitative
insights
(Barley
1986,
Kellogg
Mische
(2008):
rich
ethnographic
detail
combined
2009)
or
to
generate
constructs
to
be
tested
in
with
social
network
analysis
(based
on
network
data
regressions
(Wageman
2001)
collected
as
part
of
the
fieldwork).
Presented
in
a
book
format
in
order
to
accommodate
the
richness
of
the
two
forms
of
analysis.
Amongst the most traditional views of the relationship between qualitative and quantitative research is the
use of the former to generate hypotheses and constructs to be measured and tested in the latter. That is,
technological change or differences in team learning), existing explanations are not satisfactory, and
scholars use qualitative evidence to identify potential alternative explanations that are more complete or
resolve tensions that exist. Researchers then translate these insights into hypotheses that they then test in a
subsequent quantitative, regression analysis. The quantitative component can take multiple forms, from
Tripsas’ (1997) study of creative destruction in the typesetter industry starts with a puzzle: if a new
generation of technology is competence-destroying for incumbent firms, why do some survive and even
thrive? She uses a detailed historical analysis of the companies in the typesetter industry – using company
and trade association archives, interviews, personal records of retired employees, industry trade journals,
and other sources – to uncover the reason. She found that the differences are due to the degree to which a
firm’s specialized complementary assets are aligned with the new technological regime. She
demonstrated this in a series of tables identifying the technological competencies and complementary
assets required for each generation of technology. In the second half of the paper, she used these
qualitative insights to generate measures of competencies for each of the firms in a complete dataset of all
of the firms and products in the industry over time (note that she developed this quantitative dataset using
many of the same resources that provided the qualitative insights). In the regression analysis, she
demonstrated that incumbents whose complementary assets were not devalued had higher market shares
in new technological regimes than those whose complementary assets lost value. In this study, the
qualitative and quantitative analyses were both complementary and confirmatory. The analysis of the
qualitative data uncovered relationships and helped to generate constructs for a quantitative analysis that
Briscoe’s (2007) study of professional service workers follows a similar approach. He used a
field study of primary care physicians to develop a model of how organizations influence the temporal
flexibility of their work. The fieldwork – observations, archival data collection and interviews with 55
informants – allowed him to identify the primary source of inflexibility – inability to hand off patients at
the end of a shift – and locate specific practices associated with bureaucratic formalization that might
create more flexibility. This first study uncovered the mechanisms but did not allow Briscoe to rule out
other potential factors shaping temporal flexibility. Thus, in the second study, he used survey data on
physicians and their organizations in order to evaluate the effect of bureaucratic formalization on three
measures associated with temporal flexibility. He did not design or conduct this survey of over 6,000
data for other purposes. Matching his qualitative insights with these more comprehensive data enabled
him to confirm that bureaucratization was indeed associated with greater temporal flexibility, controlling
for other organizational characteristics such as size, type of ownership and age.
Doering’s (2014) study of a microfinance institution explored the ways that escalation of
commitment can actually be productive for organizational outcomes. She used ethnographic observations
and interviews to provide insights into loan officers’ behaviors as they interacted with clients, from which
she developed a series of hypotheses. As she noted, such data are “are poorly suited to address the
frequency of these processes or their long-term, broader consequences for the organization” (pp. 12-13).
So, she used a database from the organization on lending and repayment behaviors over time in order to
show that delinquent clients who work with original loan officers (who are presumed to escalate their
commitment if they keep these clients on the books rather than sending them to collections) are more
likely to repay loans on time and take out subsequent loans (evidence of higher levels of performance).
This regression analysis enabled her to connect escalation of commitment to performance as well as to
account for alternative explanations and selection bias not possible in the field study. In this study,
Doering iterated between complementary and confirmatory goals and between qualitative and
quantitative data collection and analysis over time. (See Canales (2014) for a similar style study of
microfinance using ethnography and archival evidence from loan databases to understand the tensions
In these examples, qualitative research enabled the researchers to identify a potential explanation
not previously explored in the literature and measures that could then be tested in a regression. However,
the ways that the scholars made the connections between their qualitative and quantitative data differ. In
Tripsas’ (1997) study, the sources of data for the qualitative and quantitative studies were highly
overlapping. Her comprehensive database of firms and product models was not available off-the-shelf, so
she necessarily had to dig into many forms of archival evidence to construct the quantitative variables for
testing the hypothesis she generated from a qualitative analysis of these same materials. For Briscoe
(2007), he was able to match insights from his own fieldwork with a previously conducted survey that he
Using Qualitative Data for the Development of a Construct and/or Model, which is then Tested in a
Survey Designed Based on those Constructs.
Edmondson’s (1999) paper on team psychological safety started with observations and interviews of 8
teams to verify that the constructs of team psychological safety and team learning behavior could be
operationalized in her research site. She then administered a survey measuring these constructs to
members of 53 teams and a separate survey to recipients of the teams’ work in order to measure team
performance. As a final step, she then turned back to more fieldwork to conduct a comparison of extreme
cases (four high- and three low-learning teams) to the relationships between the two focal constructs. The
goal, as she stated, “was to learn more about how they functioned as teams rather than to confirm or
disconfirm a model” (1999: 370). The most powerful part of this third phase of qualitative research was
her use of outliers to show how the lack of psychological safety hindered team learning even in contexts
Similarly, Bresman’s (2010) study of external learning by pharmaceutical teams used qualitative
data from 94 interviews of members of 6 teams to develop measures of different forms of external
additional teams. The survey data then allowed him to examine the performance implications of these
external learning activities, showing particularly that vicarious learning only leads to higher performance
Ely’s (1994) study worked slightly differently in that she interviewed and surveyed the exact
same sample of people (women in law firms) to understand the impact of the presence of women in upper
echelons of management on hierarchical and peer relationships among professional women. She first
conducted 4-5 hours of interviews for each of 30 women professionals from 4 matched pairs of law firms
(matched according to whether the upper management was gender inclusive or not). She then conducted
content analysis of the interviews (with inter-rater reliability tests) to identify the key dimensions of the
theoretical mechanisms of interest. She subsequently used the natural language of the respondents to
increased construct validity. Based on the survey results, she conducted a series of t-tests on the average
scores by type of firm in order to confirm that women in firms with few women leaders did not
experience gender as a positive basis for identification, did not find senior women to be positive role
models, and were more likely to compete with their peers. She complemented these statistical tests with
rich quotes from the interviews that showed the grounding of these identified relationships in the lived
Lounsbury’s (2001) study of the introduction of recycling programs at universities started with
interviews with program leaders across 60 different universities. In this exploratory study, he found that
the implementation of recycling came in two forms, either programs staffed by a full-time recycling
coordinator or ones where the responsibilities were added to existing work roles. The fieldwork (in
iteration with the literature) suggested several reasons why there might be variation in staffing models for
recycling programs. To test these hypotheses, he then conducted a survey of 154 schools in order to
confirm the role that status and social activism had in shaping adoption patterns.
Thus, qualitative fieldwork and survey studies can be useful complements (Sieber, 1973, provides
survey; or a survey can be seen as a useful supplement to fieldwork. Surveys are very tricky to get right.
Assuring the validity of the questions and the scales is essential for generating insight. For those scholars
whose initial intention is to do survey research, conducting in-depth fieldwork in order to develop a
survey may lead to a more valid set of measures than if they had just pre-tested a survey designed in the
absence of these insights. Fieldwork can also help in the interpretation and theorizing of the survey data.
For scholars starting with qualitative fieldwork, such as Lounsbury (2001), the idea to conduct a survey
might only emerge during the research process at a point when it becomes beneficial to collect data over a
larger number of sites. Surveys can contribute to the analysis of qualitative data from the field by
Elsbach (1994) studied the California beef industry in order to understand how organizations can manage
perceptions of negative events in order to maintain their legitimacy. Her paper presents three studies
sequentially. In the first two, she conducted interviews of those attempting to shape perceptions (15
executives in beef producer organizations and industry associations) and the intended audiences (15
members of the press, nutritionists, and politicians). The analysis of these interviews (as well as of
supplemental press articles and press releases) allowed her to produce a typology of different responses to
negative events based on the form and content of the communications and to construct three propositions
of the relationships between the types of negative events, the types of responses and their effectiveness.
In the third study, she translated these insights into a vignette-based lab experiment in which she
tested the effects on various measures of organizational legitimacy of each of the different types of
accounts. The experimental results largely confirm the findings from the two qualitative studies. This
combination of studies proves to be effective in a way that none would be on its own. With only 15
interviewees for each of Study 1 and 2, the reader might not be convinced of the findings; but
complementing these insights with a lab experiment validates the qualitative findings.
A recent study by Ranganathan (2014) on craftsmen and traders of lacquer-work in India follows
a similar template, using ethnographic evidence to show how different orientations towards work shape
prices charged and then conducting a field experiment to validate the findings and hone in on the
mechanism (with the counterintuitive insight that the craftsmen will charge less if they think the customer
Coupling fieldwork and experiments is not frequently found in the literature but offers great
potential. As Fine and Elsbach (2000) point out, qualitative field data can contribute a richer,
contextualized set of theories to test in experiments where lab work alone is a poor proxy for the real
world. On the other hand, they argue, qualitative data alone may produce theories that are highly complex
and too specific to the case or cases studied. They suggest, as the examples above illustrate, that coupling
fieldwork and experiments in an iterative process can lead to more nuanced analyses and more rigorous
An alternative model of mixing qualitative and quantitative methods is to start with a statistical regression
and then follow with qualitative evidence. This approach can have two effects. The first is to lend face
validity to regressions. If the relationships identified using necessarily abstracted quantitative measures
make sense in the context of the lived experience of people in those contexts, one can feel more
comfortable that the measures capture the constructs they are meant to capture and that the associations
(perhaps even causal relationships) identified are not spurious. Second, qualitative analysis may also
allow the researcher to explain the underlying mechanisms behind a relationship identified in a
This study is based on data from the placement records of a temporary staffing agency showing how the
gender wage gap evolves over time. She used the quantitative data in a series of analyses to show that
men accrue benefits from tenure at a greater rate than do women. She also ruled out a number of possible
mechanisms – such as firm-specific skills that develop over time – previously suggested by the literature.
In the second part of the paper, she identifies the two remaining possible mechanisms (the supply-side
explanation of general skills acquisition and the demand-side explanation of implicit biases). She then
used information from observations and 49 interviews at the staffing agency to adjudicate between them,
suggesting that implicit bias on the part of employers is the most plausible explanation for the cumulative
In an interesting twist on the use of surveys, in Bidwell’s (2009) effort to collect survey data
using in-person surveys in IT outsourcing, he ended up accumulating a good deal of rich qualitative
evidence that emerged in the discussions associated with collecting the quantitative survey data. He used
the qualitative data to give richness to the explanations about the differences between contractors and
employees and when each would be deployed on a project. And, as illustrated above, Edmondson’s
(1999) study came full circle, having used qualitative evidence to generate a survey and then returning to
fieldwork to understand the mechanisms underlying the effects from the survey-based quantitative
While it is certainly true that not all quantitative, regression analyses demand a qualitative
component in the published version of the paper, it is also true that nearly any quantitative scholar would
benefit from entering into the field. Even the most high-tech, sophisticated regression analysis can benefit
from some fieldwork to assure the regression results make sense and are grounded in reality. This is
important because variables used in quantitative studies are always simplifications or approximations of
the underlying constructs and the sanity check with the field helps assure the scholar of the validity of the
measures. Unfortunately, few quantitative scholars receive training in qualitative methods and therefore
A further model for the relationship of qualitative and quantitative research is in the quantification of
qualitative data (see McPherson & Sauder, this volume, for further detail on this subject). Scholars may
undertake this for the purposes of conducting subsequent statistical analyses to confirm findings.
Quantification may also help to represent qualitative evidence compactly, potentially revealing patterns
Barley’s (1986) study of the introduction of CT scanners into two hospitals is largely an
ethnographic study of the different organizational forms that can emerge from the adoption of the same
technology, but he introduced quantitative elements to reinforce or validate his analyses. He based the
bulk of the story on his observations of the negotiations about who makes the decisions in the scanning
process in which he documents a series of scripts (different by hospital) that were enacted over time. To
demonstrate the differences in the two hospitals, he counted the occurrences of these various scripts in his
field notes and showed that different patterns emerged in the different contexts.
Over the course of these observations, he recorded 400 radiological examinations of which 91
were CT scans sufficiently documented to understand whether the radiologist or the technologist was the
primary decision maker in the various decisions required by each scan. He calculated the percent of
decisions made by the radiologist for each scan (49 scans in one hospital and 42 in the other) and used the
plots of these ratios over time to demonstrate not only the shifts towards the technologist over time in
correlations established in two very simple regressions to support his qualitative conclusions about the
differences in rates between the two hospitals and also to validate his identification of the different phases
of structuration in each.
In a similar study of organizational change in two hospitals, Kellogg (2009) examined the degree
to which new practices associated with work hour reductions (such as handing off patients to doctors
coming on in the next shift) were implemented. She counted the different kinds of change mobilization
activities in 202 different meetings and the percent of handoffs that occurred in 101 sign-out encounters
she observed. She used a comparison of these counts to demonstrate that doctors in the successful hospital
engaged in a set of practices that those in the unsuccessful hospital did not.
In Wageman’s (2001) study of how leaders foster team effectiveness, she transformed interview
responses into quantitative variables by using multiple coders to assess the presence of various team
characteristics – several dimensions of team design, self-managing behaviors and leadership coaching –
as being either high, medium or low. She then matches these quantified data with survey responses and
archival information on team performance in a regression analysis in order to conclude that the
effectiveness of leaders’ coaching depending on how well teams were designed to begin with. In this
study, the qualitative data are not presented in a separate narrative but rather only as they are quantified.
Because of the rhetorical power of numbers, scholars are often tempted to generate counts from
qualitative, even ethnographic, data. As the Barley (1986) and Kellogg (2009) studies show, such
numbers can provide useful confirmation of the patterns developed through the qualitative analysis. In
each of these studies, counting is appropriate because the researchers were able to observe repeated
equivalent events collected in a systematic way. Other forms of quantification might be more problematic.
For example, generating a quantitative rating of degree (of importance, of intensity, etc.) using a Likert
scale should be seen as suspect especially if it is generated by the researcher herself. Better versions of
this approach might involve multiple coders not informed of the research question with tests for inter-
rater reliability (as in Wageman 2001). Alternatively, one could get the informants themselves to provide
a numerical rating through a survey instrument (perhaps used in the context of an interview, e.g.,
to offer quantifications, and indeed, such strategies should only be deployed when it is appropriate and
Recent trends in quantification of texts, from simple word counts to more sophisticated computer
science-based methods such as topic modeling or feature extraction, are another form of quantification of
qualitative data (e.g., Kaplan & Vakili, Forthcoming). With the increasing availability of electronic texts,
these approaches are only likely to grow over time (increasingly blurring the distinctions between
qualitative and quantitative evidence and analysis). Some of the early approaches have usefully mixed
these quantifications with qualitative evidence, such as in DiMaggio, Nag and Blei’s (2013) study of
controversies in arts funding in the US where they use topic modeling to identify topics emerging from
over 8,000 newspaper articles and then mobilize these topics in a rich historically informed explanation of
These same principles can also be applied within a research program rather than in a single paper. As
those who have attempted to shepherd a mixed method paper through the review process have discovered,
it is very difficult to introduce the required detail for each method in a journal-length article. The rewards
for succeeding in this endeavor are high, but there are also times when it is more appropriate to iterate
between quantitative and qualitative work, focusing single articles on specific data sets and
methodologies. The accumulation of these studies over the course of a research program can lead to rich
insights that might not always be possible to generate in a single mixed-method article.
study reported above included only some of the rich detail he gained through the fieldwork that preceded
it. In other papers, he took advantage of the more than 160 interviews he conducted to report on how
collective entrepreneurship contributed to the creation of a new occupational category (Lounsbury, 1998)
and to illustrate different forms of institutional analysis (Lounsbury, 1997). Kaplan conducted an
ethnography of strategy making in one organization in the communications industry (Kaplan, 2008b)
concurrently with a quantitative study of more than 70 firms in the industry over 20 years (Kaplan,
respond to radical shifts in the industry – the first looking at one firm’s response to the crash of 2001-
2002 and the other looking at the response to the emergence of fiber-optics in an industry previously
dominated by copper wiring. Insights from the field study that cognition and incentives were tightly
intertwined in how decision-making unfolded (in process she called “framing contests”) led her to explore
how variables measuring CEO cognition and organizational incentives might interact in the large-sample
study. As a result, she was able to demonstrate that CEO attention can compensate for poorly aligned
organizational incentives. The ethnographic study of the decision making processes in one organization
While in both of these examples the papers were published separately, the research process itself
was highly iterative. Rather than splitting up these results into multiple papers, another alternative (taken
less frequently in management scholarship) is to write a book that collects these insights together.
Mische’s (2008) sociological study of mobilization of Brazilian youths combining ethnography and social
Across these examples, the mixing of methods takes two forms: one is confirmation and the other is
complementarity (Small 2011). (This is the simplest typology. Greene, Caracelli and Graham (1989) and
Bryman (2006) offer more granular categorizations for those who are interested.) Sometimes the
quantitative analysis confirms a pattern observed in the qualitative data, either through quantification of
that data or through triangulation in a subsequent study. Or, the qualitative data can offer confirmation of
qualitative analyses can get at mechanisms underlying relationships identified in regression analyses, or
quantitative studies can establish connections to performance of the constructs identified in qualitative
studies.
Thus, the popular dichotomy – that quantitative research is deductive and qualitative is inductive
or exploratory – is less useful when we consider the research discussed above. Sometimes qualitative
evidence is used to confirm quantitative findings, sometimes it is used to explore a phenomenon in order
is there a relationship between these two variables? – and qualitative research can explain why. In other
words, once we mixed qualitative and quantitative methods, we see that each type of research plays a
variety of roles. These relationships between quantitative and qualitative data and analysis can be built
into the initial research design (e.g., Kaplan’s studies which were designed as multi-method from the
outset) or may emerge over time (e.g., Lounsbury’s survey work which only emerged as he was analyzing
his interview data or Fernandez-Mateo’s fieldwork which was instigated by the desire to sort through
alternative explanations of her quantitative results). Where one starts may be as much a matter of taste as
of any hard and fast rules about matching methods to inductive or deductive research problems.
Sometimes mixing qualitative and quantitative methods is based on the same data and involves
different forms of analysis, as in Tripsas’ (1997) historical study of the typesetter industry or Barley’s
(1986) study of CT scanners in hospitals. Other times, the mix requires two different data collection
strategies, as in Edmondson’s (1999) interviews and surveys or Elsbach’s (1994) interviews and lab
experiments. Small (2011) defines the former case as “mixed data analysis” studies and the latter as
“mixed data collection” studies. The challenge with either case is that the researcher must give adequate
treatment to each method. The risk, in the constraining form of a journal article, is in glossing over some
of the methodological details in order to save space. But, for a multi-method paper to be convincing, the
reader must be privy to the methodological details in the same way she would be in a single method
paper. Elsbach’s (1994) study is a model here as she presents each of her three studies sequentially,
including very rich detail on the data sources and analytical methods: she does not give short shrift to any
As Edmondson and McManus (2007) have pointed out, mixing qualitative and quantitative
evidence does not always lead to better research or a stronger paper. Mixing methods is often worthwhile
research, the scholar risks being jack-of-all-trades and master of none. In the presentation of research
results, if the researcher gives each dataset inadequate treatment, the evidence may remain superficial or
unconvincing. Adding qualitative data to a quantitative analysis must serve a purpose other than
rigorously and add enough insight so as not to appear to be a transparent rhetorical strategy to convince
than qualitative evidence (Denis, Langley, & Rouleau, 2006; Porter 2005). Mixing methods is often a
strategy that qualitative researchers are tempted to deploy for rhetorical reasons in a field (management)
where qualitative work is still in the minority. That is, because we are socialized to be convinced by
quantitative evidence, many scholars are likely to pursue quantitative research strategies. But, as Cameron
(1963) points out, “not everything that can be counted counts, and not everything that counts can be
counted.” The advantage of the mixed methods studies highlighted here is that they recognize the value of
iterating between that which can be counted and that which cannot in order to generate richer insights
References
31(1), 78-108.
Bidwell, M. (2009). Do Peripheral Workers Do Peripheral Work? Comparing the Use of Highly Skilled
Contractors and Regular Employees. Industrial & Labor Relations Review, 62(2), 200-225.
Bresman, H. (2010). External Learning Activities and Team Performance: A Multimethod Field Study.
Briscoe, F. (2007). From iron cage to iron shield? How bureaucracy enables temporal flexibility for
Bryman, A. (2006). Integrating quantitative and qualitative research: How is it done? Qualitative
Random House.
Canales, R. (2014). Weaving Straw into Gold: Managing Organizational Tensions Between
Creswell, J. W., & Clark, V. L. P. (2007). Designing and conducting mixed methods research. Thousand
Denis, J. L., Langley, A., & Rouleau, G. A. (2006). The power of numbers in strategizing. Strategic
DiMaggio, P., Nag, M., & Blei, D. (2013). Exploiting affinities between topic modeling and the
Paper.
Edmondson, A. (1999, Jun). Psychological safety and learning behavior in work teams. Administrative
Edmondson, A. C., & Mcmanus, S. E. (2007). Methodological fit in management field research. Academy
Eisenhardt, K. M., & Bourgeois, L. J., III. (1988). Politics of Strategic Decision Making in High-Velocity
Elsbach, K. D. (1994). Managing Organizational Legitimacy in the California Cattle Industry - the
Construction and Effectiveness of Verbal Accounts. Administrative Science Quarterly, 39(1), 57-
88.
Ely, R. J. (1994). The Effects of Organizational Demographics and Social Identity on Relationships
building: Tactics for integrating qualitative field data with quantitative lab data. Journal of
Greene, J. C., Caracelli, V. J., & Graham, W. (1989). Toward a conceptual framework for mixed-method
Jick, T. D. (1979). Mixing Qualitative and Quantitative Methods: Triangulation in Action. Administrative
Kaplan, S. (2008a). Cognition, capabilities, and incentives: Assessing firm response to the fiber-optic
Kaplan, S. (2008b). Framing contests: strategy making under uncertainty. Organization Science, 19(5),
729-752.
Kaplan, S., & Vakili, K. (Forthcoming). The double-edged sword of recombination in breakthrough
Kellogg, K. C. (2009). Operating Room: Relational Spaces and Micro-Institutional Change in Two
Lounsbury, M. (1997). Exploring the institutional tool kit - The rise of recycling in the US solid waste
Lounsbury, M. (1998). Collective entrepreneurship: the mobilization of college and university recycling
Lounsbury, M. (2001). Institutional sources of practice variation: Staffing college and university
McPherson, C.M., M. Sauder. 2015. The Quantification of Qualitative Data, Handbook of Innovative
Qualitative Research Methods: Pathways to Cool Ideas and Interesting Papers, K. Elsbach and
R. Kramer, eds.
Mische, A. (2008). Partisan publics: communication and contention across Brazilian youth activist
Ranganathan, A. (2014). Choosing Meaning over Money? Evidence from a Field Audit Study with
Sieber, S. D. (1973). Integration of Fieldwork and Survey Methods. American Journal of Sociology,
78(6), 1335-1359. .
Small, M. L. (2011). How to Conduct a Mixed Methods Study: Recent Trends in a Rapidly Growing
Tripsas, M. (1997). Unraveling the process of creative destruction: Complementary assets and incumbent
Wageman, R. (2001). How leaders foster self-managing team effectiveness: Design choices versus hands-
Author bio: Sarah Kaplan is Associate Professor of Strategic Management at the Rotman School of
participate in and respond to the emergence of new technologies and fields, with a particular focus on the
interpretive processes that shape choice and action. Her studies examine the biotechnology, fiber optics,
personal digital assistant, and nanotechnology fields. Most recently, she has turned her attention to the
new field emerging at the nexus of gender and finance. She received her PhD from MIT’s Sloan School
of Management and previously served on the faculty of the Wharton School, University of Pennsylvania
where she remains a Senior Fellow. She is Senior Editor at Organization Science, Guest Editor for the
Special Issue on New Research Methods for the Strategic Management Journal and formerly Associate