Research Design
CH 1
The main steps of the research process: problem definition, theory development,
conceptualization, operationalization, variable selection, and case selection.
What is research design? Ensures that the answers we provide are as valid as possible and are
discovered as efficiently as possible.
What is research? Research is about providing answers to questions, that can be best
supported by facts and arguments.
What is science? A minimal definition of science need only highlight two crucial aspects:
publicness and adherence to the scientific method.
- Publicness means that research needs to be public and open to scrutiny. It is not only that
it is almost always, directly or indirectly, funded by the public and that it contributes to
the common good rather than to private gain. Scientists must always disclose the methods
they work with and the data they used. They have an obligation not to withhold findings
based on whether they suit or not their favoured hypotheses or ideological
predispositions.
Publicness and transparency of methods, data, and results are indispensable, because
science is a community affair. A scientist’s work is scrutinized by other academics (his or
her peers) and made open to critique by any valid argument. It is the way science
proceeds: by collective scrutiny and criticism that allow for gradual improvements and
the correction of mistakes; by replicating, adjusting, and, occasionally, overturning what
others have done before.
- The scientific method are the rules scientists have to follow to answer questions.
However, there are major disagreements on what those rules are.
o Logical positivists argue that scientists start with a theoretically motivated hypothesis,
test the hypothesis with data, and proceed to conclusions rejecting the hypothesis that
fail the empirical tests (Popper. Chapter 3 for details). This view is clear and simple,
but it has been disqualified both as a descriptive (how science works) and as a
prescriptive (how science should work) model.
o In contemporary political science:
Empirical puzzles and substantive problems are as common starting points for
research work as theories
Theory testing is not the only and perhaps not the most important goal of science
Ideas are not simply tossed in the garbage bin of history at the first sign of empirical
inadequacy
However, within contemporary Political science, there are disagreements still on the
scientific method:
Subjectivism vs. positivism (philosophy of science)
o Subjectivism: the essential, unique characteristic of human behavior is its subjective
meaningfulness. Thus, social science is not a value-free pursuit of objective truths
about the social / political world. Human perception and experience make science
subjective. They argue that research should be concerned with interpreting the
, meaning of and reflecting on the reasons for human action. They dislike ‘social
mechanisms’ and ‘causal factors’- why?
Social science can only function as radical social critique. Blurring the distinction
between scientific research and advocacy/social action is detrimental to the practice
of both.
E.g.: Interpretivists, critical theorists, feminists
o Positivists (broadly), accept that social reality is not set in stone, objectively given,
and directly accessible to human perception, but constructed and reconstructed
through a variety of social processes. In other words, social reality is to a large
degree inter-subjective. However, and here is the crucial difference from radical
subjectivism, social science remains possible. It is seen as a quest for the discovery
and explanation of the causes and mechanisms of social phenomena, including
individual events as well as broader regularities and patterns. And it is subject to
transparent rules, standards, and procedures that ensure reproducibility, reliability,
and validity of the results.
Majority of political scientists follow this method
Their debate: Subjectivists would object that one cannot entirely separate the
observation of social facts from values and theoretical notions (see Chapter 3), this
is the philosophical stronger argument. However, even if so, we can still do much to
acknowledge and limit the influence of our particular values in doing research on
politics and governance. It remains important to be aware of the inherent limitations
of social science highlighted by subjectivism, but within the territory outlined by
these limitations, there is plenty of scope for scientific research subject to rigorous
procedures and explicit standards.
Empericism vs. scientific realism (based in positivist philosophy, disagree about
the ways to conduct the study). (At a deeper philosophical level, empiricism and
realism imply different ontological views (what is reality), but in practice these are
only manifested as differences in epistemology (how to know reality).
o Empiricists deny reality to unobservable entities such as theoretical concepts and
causal structures and usually adopt an instrumentalist view of theoretical
assumptions. The latter means that the assumptions of our theories and models do
not need to be realistic (as they never are) as long as they work; that is, as long as
they prove useful for prediction and manipulation. Prediction and intervention is
more important than understanding the underlying structure of the world that
generates the observable outcomes. (Causes and effects cannot be directly
observed)
Models of political processes (based in economics and game theory. Extremely
simplified- instrumental- assumptions about rationality and social interactions). And
descriptive research, describing and exploring political phenomena, such as public
opinion, electoral campaigns, or policy implementation, rather than building and
testing abstract theoretical models.
o Scientific realists strive for a ‘deep understanding’ that goes beyond the
instrumental uses of scientific results (see Chapter 6). the focus is on explanation
and the discovery of causal structures. Causal structures and concepts, although not
directly observable, are considered to be the primary targets of scientific inference -
Majority of research
, Quantitative vs. qualitative research (the mode of research).
o Quantitatively oriented (that is, they use numbers and statistics)
o Qualitative research (that is, they do dense case studies)
- The scientific method is best defined considering the various activities that make up the
scientific process separately and putting forward some requirements for each step of the
way. This means adherence to certain rules regarding theory development, measurement,
inference from data, and other aspects of the research process.
The confirmation bias is perhaps the most relevant for the context of research design. It
relates to the human tendency to seek only information that would confirm a guess or a
hypothesis but no information that would contradict it. It is a hallmark of the scientific
method (research design) in contrast to casual human thinking to search for disconfirmatory
evidence and properly test intuitions and ideas. There’s also the hindsight bias, that makes us
believe that events have been more foreseeable than they actually were; framing effects that
lead people to make different inferences from the same information with only some
innocuous-looking change of words; the availability bias that makes us take into account
only the most salient information that comes first to mind. Research design helps discipline
the process of data collection – what kind of evidence should be sought and how – to
overcome the limitations of informal human cognition that have been shown to affect laymen
and experts alike.
CH 2: focus on understanding what each types, how the different types differ, and how the
different types of connected
Types of research:
- Normative (ethical questions): what ought to be (should). It examens the moral order of
the world, seeks to find answers to what is good, appropriate, just, fair, or otherwise
desirable. It's about ethics, values, and value judgements. All recommendations for
adopting certain policies, making some decisions, or taking particular actions inevitably
have implicit or explicit normative components. PRESCRIPTIVE.
- Positive: What is, used to be or will be. Focus on describing, understanding, explaining,
or predicting reality as it is. Value judgements are suspended. The focus is squarely on
empirical phenomena, on the links between them, and on the links between empirical
phenomena and theoretical concepts. Relationships between abstract concepts can also be
studied positively, if no ethical judgements are implied. (Majority of research)
They are dependent on each other: Normative concerns inevitably drive the selection of the
research question, and the conclusions from positive research have normative implications as
well. In short, results from positive scientific research bear on normative issues. E.g.:
Normative research can clarify what ‘better’ means, single out characteristics necessary for
a ‘good’ political system (equality, efficiency, responsiveness, accountability, and so on),
illuminate trade-offs between some of these (for example, accountability versus efficiency),
or possible contradictions within any ideal of a ‘good’ polity. Positive research can show
how certain elements of a political system (say, the electoral system or the rules for party
financing) influence other elements of the system and bear on normative notions like equality
or accountability. A purely positive empirical study of the impact of majoritarian versus
, proportional electoral systems on the representation of national minorities has huge
normative implications about the design of ‘good’ political systems, provided we agree that
minority representation is a characteristic of any ‘good’ political system. From positive
research, you cannot answer normative questions; such as whether minority representation is
needed for a just democracy. In summary, positive research is usually motivated by
normative concerns and its results contribute to ethical discussions, but in itself it is (or at
least should be) free from subjective, value-ridden judgements and biases. This distinction
helps separate subjective value judgements (‘democracy is good’) from objective statements
about facts or concepts. Social researchers should make a conscious effort to prevent
normative elements from creeping into positive research questions. Otherwise, the line
between advocacy and analysis becomes blurred, with negative consequences for both.
Theoretical vs. empirical research!
Positive political theory deals with the rigorous analysis of the properties of abstract
concepts, such as democracy, governance, and power, and the relationships between them. It
can stand on itself, or be a part of the research process with the aim of deriving empirical
implications from initial, often vaguely expressed, theoretical ideas (theory development).
