Qualitative Data: Cooking Without a
Recipe
Forthcoming, Strategic Organization
Melissa E. Graebner
Jeffrey A. Martin
Philip T. Roundy
Potential
潜力
2
Frustrations
挫折
©Jeffrey A. Martin, PhD
3
Confusion
混淆
Many descriptions of qualitative
research:
• Lump all qualitative research together and offer
lists of “typical” characteristics
–
–
–
–
–
–
Nascent theory
Naturalistic setting
Inductive analytical approach
Constructivist, relativist, interpretive stance
Acceptance of researcher bias
Interest in ordinary or everyday behavior
• Or group into “categories” with their own “typical
characteristics”
No “recipe” or “cookbook,” but not
“anything goes”
“While qualitative methods need to be elaborated
or modified for each new application, this does
not mean that anything goes or that the best
method is no method” (Gephardt, 2004, p. 458).
This creates some dilemmas:
1. Many studies do not fit in a single school of
qualitative research (e.g., Maxwell, 2005)
2. Few studies have all of the “typical” attributes
Researchers may feel compelled to force their
work to fit into a mold that isn’t appropriate
A potential solution: Focus on why
you are using qualitative data
• There are multiple, distinct reasons for using
qualitative data.
• Understanding why you are using qualitative
data helps avoid forcing your study to fit with
an inappropriate paradigm
• This reduces confusion and frustration
How qualitative data are different
1. Open-ended
2. Concrete and vivid
3. Rich and nuanced
Given these advantages, what research
goals may call for qualitative data?
New
Theory
Reason #1: Build
new theory
when prior
theory is absent,
underdeveloped,
or flawed
Reason #1: Build new theory when prior theory
is absent, underdeveloped, or flawed
• Advantage of qualitative data: open-endedness
• Findings in a theory-building study can take diverse forms:
– Variance predictions (e.g., Ozcan & Eisenhart, 2009)
– Process models (e.g., Graebner, 2009)
– Typologies that unpack important and poorly understood constructs
(e.g., Hite, 2003)
• And may be aimed at developing “objective,” positivist theory
Lived
experiences
Reason #2: Capture
individuals’ lived
experiences and
interpretations
©Jeffrey A. Martin, PhD
13
Reason #2: Capture individuals’ lived
experiences and interpretations
• Again, the advantage of qualitative data is openendedness
• But there are important differences vs. theory-building:
• Interpretive studies:
– Aim to preserve the subjective nature of their data throughout
the analytical process
– May use qualitative data even when substantial prior theory
exists
• e.g., Creed et al.. use qualitative data to “complement and extend” previous
theoretical work (2010: 1337) .
Complex
processes
15
Reason #3: Understand complex process issues
• Phenomena involving complex temporal dynamics or
causal mechanisms, often embedded in nuanced
social interactions
• Advantage of qualitative data: Richness
• In practice, many process studies involve some
theory-building – but qualitative data can also be
used for process studies in areas of relatively mature
theory
– E.g. Martin (2011) – top management team processes; Lumineau et al. (2011) –
organizational learning
– Process researchers may even use qualitative data to test theory (e.g.,
Greenwood et al., 1994)
Illuminate
©Jeffrey A. Martin, PhD
17
Reason #4: Illustrate an abstract idea
• Advantage of qualitative data: Vividness, concreteness and
richness
• Example: Siggelkow, 2001. “The framework proposed in the
paper emerged more from a conceptual exercise than from my
exposure to Liz Claiborne’s experiences. However, the case
turned out to be a very helpful illustration and was used in that
manner after the conceptual framework was presented.”
• Open-endedness is less important – these researchers may have
well-developed models prior to gathering their data (e.g.,
Kauppila, 2010).
Examine
language
Examine
narratives,
discourse or other
linguistic
phenomena
©Jeffrey A. Martin, PhD
19
Reason #5: Examine narratives, discourse or
other linguistic phenomena
• Phenomena that fundamentally involve words and
language
• May or may not be interested in individuals’ subjective
experiences
– May examine media accounts, annual reports, websites and
press releases
– And they may code their data in ways that enable statistical
analysis
– E.g., Martens et al.’s (2007) analysis of narratives in IPO
prospectuses
• Quantified narratives to test hypotheses using statistical estimation
Why is this
important?
Identifying your reason(s)
for using qualitative data
can help with navigating
the review process:
1. Writing the “front end”
2. Describing analysis
3. Addressing biases
©Jeffrey A. Martin, PhD
21
Writing the “front end”
• Illustrating an abstract idea: expect a lengthy
front end
• Process questions, interpretive perspectives or
language-related topics may also have a
lengthy front end
• Theory building studies may have a short front
end, or a longer one that identifies specific
conflicts or other problems in prior theory
Describing analysis
• Vast majority of qualitative studies describe
their analysis as “inductive”
– But in reality, they often use a blend of inductive
and deductive processes – may choose certain
constructs or theoretical frames prior to data
collection
– If using qualitative data for a reason other than
building theory, no reason to expect a purely
inductive approach
Addressing biases
• Qualitative authors need to convince reviewers
they have minimized biases
– Informant bias
– Researcher bias
• Identifying the rationale for working with
qualitative data can help
– For interpretive research, the greater risk may be
researcher bias
– For (positivist) theory-building research, the greater
risk is likely to be informant bias
• Can be addressed through multiple informants, focus on
facts, triangulation with archival data, etc.
In summary…
• Qualitative data can serve a number of different
purposes
–
–
–
–
–
Theory-building
Interpretive perspective
Process issues
Illuminate abstract ideas
Linguistic phenomena
• Not all qualitative studies will look alike
• Identifying the reason for working with
qualitative data can help avoid minefields during
review process
Questions?