Notes on Qualitative Research
Qualitative vs. Quantitative Research
Qualitative Research
|
Quantitative Research
|
Quantitative research measures
differences in number for variables, and usually studies a large
number of cases (Large “N”). |
Qualitative research measures
differences in kind for variables, and usually studies a small
number of cases (Small “N”).
|
Because it covers a broad range of
cases, quantitative research yields conclusions that can be
generalized (it has the strongest external validity).
|
Because it looks closely at a few
cases and traces causal pathways, qualitative research
outperforms quantitative research in its measurement validity and
internal validity.
|
When selecting cases for your
quantitative research sample, it is imperative that you use random
selection.
|
In qualitative research, “selection
must be done in an intentional fashion, consistent with
research objectives and strategy.” (King, Keohane, and Verba, 1994,
p.139)
|
|
|
|
l
Qualitative and quantitative studies
are both types of observational studies.
A. Selecting Cases on the
Independent Variable
l
“Selecting on the independent
variable” means “selecting your cases according to the values of the
independent variable that they take on.”
l
In order to do this, you have to know
a little bit about all of your potential cases.
l
In order to do this right, you cannot
act as if you also know the values that the dependent variable takes
on.
|
The
Most Similar Systems |
The Most Different Systems
|
The Most Similar Systems method
selects cases that take on similar values of confounding variables,
but different values of a key independent variable.
l
This “holds constant” the confounds
because they take on the same values in all of the cases.
l
This is the design recommended by
King, Keohane, and Verba.
l
The cases that you might select in a
Most Similar Systems design will differ based on the research
question you ask.
l
USA, Canada, and Britain in a study of
life expectancy.
l
Bill Clinton, Jimmy Carter, and JFK in
a study of regional appeal in voting.
l
Genghis Kahn, Hannibal, and Hitler in
a study of how to defend invasions.
|
The Most Different Systems
method selects cases that take on very different values for multiple
independent variables.
l
If it turns out that these cases all
take on the same value of a dependent variable, then we can rule out
the independent variables as causes of the dependent variable.
l
Less useful since it can only disprove
a hypothesis.
Example: Income Inequality and Civil War
|
| |
|
B. Selecting Cases on the Dependent
Variable
l
Selecting cases according to the value
of the dependent variable that they take on is more controversial
than selecting on the independent variable.
l
It allows you to look at extreme
values or divergent cases.
l
“However, if this design is to lead to
meaningful … causal inferences, it is crucial to select observations
without regard to values of the explanatory variables. K.K.V.”
|
Method
of Agreement |
Method of
Difference |
When
you use Mills’ Method of Agreement, you select cases that
take on the same values of the dependent variable.
l
This helps you to rule out possible
causes, because independent variables that vary over these cases
can’t cause the dependent var.
l
This method can only disprove a
hypothesis, because it can’t find a correlation.
|
When you use Mills’ Method of
Difference, you select cases that take on different
values of the dependent variable.
l
After you have selected your cases,
you determine what values they take on for some independent
variables.
l
Perhaps one independent variable will
vary across your cases, and explain the D.V.
|
Example: Clem Miller
l
This Congressman writing home to his
supporters tells the story of two different lobbying efforts on
behalf of farmers.
l
Let’s assume that his process of case
selection came in the same order that his narrative is written.
l
If so, he selected on independent
variables and used the “Most Similar Systems” design.
Example:
Dreze and Sen
l
Both countries began a new political
regime at mid-century with large populations and little wealth.
l
They have diverged since then: “There
is little doubt that as far a morbidity, mortality, and longevity
are concerned, China has a large and decisive lead over India. (p.
205)”
l
“What has brought about that lead is a
matter of very considerable interest. (p. 206)”
Comparing Many Research Methods:
Measurement Validity
l
Measurement validity
judges the gap between your conceptual definition and your
operational definition.
l
Qualitative – Highest measurement
validity
l
The measurement validity of other
research methods (Lab experiments, quasi/natural experiments, and
quantitative research) really depends on what you are trying to
measure.
Comparing Many Research Methods:
| Internal Validity |
External Validity |
Internal validity judges
how well a research design has tested a causal relationship, in the
cases examined. Random assignment is the key.
Lab Experiments – highest
Quasi/Natural experiments – medium
Qualitative research – medium
Quantitative research – lowest
|
Comparing Many Research Methods:
External validity is how
confident we can be that a causal relationship identified in our
cases can be generalized to the outside world. Random sampling is
key
Quantitative research – highest
Quasi/Natural experiments – medium
Lab experiments – low
Qualitative research – low
|
l
The best
research design depends on your research question and the particular
problems that it poses.
l
The best
research uses a mixture of methods to test a hypothesis.
|
