c


 


The activity of comparing countries centers on four main objectives:

1.p contextual description

2.p classification
3.p Hypothesis-testing
4.p Prediction

Uitleg begrippen:

1.p contextual description: the process of describing the political phenomena and events of a particular
country, or group of countries. The comparison to the researcher͛s own country is either implicit or
explicit, and the goal of contextual description is either more knowledge about the nation studied,
more knowledge about one͛s own political system, or both. ALL SYSTEMETIc RESEARcH BEGINS WITH
GOOD DEScRIPTION; purely descriptive studies serve as the raw data for those comparative studies
that aspire to higher levels of explanation and provide initial hunches about which topic of research
may be of interest and which factors may be important to explain observed phenomena that are
related to those topics.
2.p classification: In the search for cognitive simplification, comparativists often establish different
conceptual classifications in order to group vast number of countries, political systems, events, etc
into distinct categories with identifiable and shared characteristics. classification is a necessary
component of systematic comparison, but in many ways represents a higher level of comparison since
it seeks to group many separate descriptive entities into simpler categories. It reduces the complexity
of the world by seeking out those qualities that countries share and those that they do not share.
classification seeks to describe and simplify a more complex reality by identifying key features
common to each type. categories that make up a classification scheme can be derived 

from careful consideration of available evidence or through a process of  in which ͚ideal͛
types are generated.
3.p Hypothesis-testing: despite the differences between contextual description and classification, both
forms of activity contribute to the next objective of comparison: hypothesis-testing. comparison of
countries allows rival explanations to be ruled out and hypotheses derived from certain theoretical
perspectives to be tested through examining cross-national similarities and differences. comparison of
countries and testing hypotheses contributes to the progressive accumulation of knowledge about the
political world. The key explanatory and outcome variables are carefully defined and the relationships
between them are demonstrated through comparison of empirical evidence
4.p Prediction: the final and most difficult objective of comparative politics is a logical extension of
hypothesis-testing namely to make predictions about outcomes in other countries based on the
generalization from the initial comparison or to make claims about future political outcomes.
Predictions in comparative politics tends to be made in probabilistic terms.

Inference: making an inference is ͚using facts we know to learn something about facts we do not know͛.
comparative politics seeks to achieve the goal of inference about politics through comparing countries.

The strong case for a science of politics suggest that both (comparative) political science and natural science
share the same basic goals, namely (1) description; systematic collection of evidence, (2) classification; an
ordering of the evidence and the search for discernible patterns, (3) hypothesis-testing; formulation and testing
of contending explanations for the occurrence of the patterns and (4) prediction; the building of more general
theory.

 p Thus, a science of politics always contains this ͚evide nce-inference methodological core͛ or the
͚customary pair͛ of theory and observation.
  p Thus, both the natural and political sciences seek to make inferences based on the empirical world
they observe and both seek to maximize the certainty of these inferences

Two important differences between the natural sciences and the political science:

1.p The first difference is the role of experimentation. comparative politics, in particular, cannot use
experimentation for both practical and ethical reasons. This demonstrates the need for using
counterfactuals, or situations in which the researcher imagines a state of affairs where the antecedent
factors to a given event are absent and where an alternative course of events or outcomes is
considered. In this way, comparative research ͚stimulates͛ experimentation.
2.p The second difference between natural science and political science involves the law-like status that is
given to certain scientific theories. Law-like generalization of theories of political sciences are rare

Three famous ͚laws͛ of political science:

1.p Michels͛ ͚Iron Law of Oligarchy͛; suggest that the natural processes observable in the dynamics of
organization and small groups are such that over time, all groups and organizations develop a
hierarchical structure of authority with a small elite at their head
2.p ͚Duverger͛s Law͛; states that electoral systems based on single-member districts tend to produce two
parties while systems with proportional representation tend to produce multiple parties.
3.p ͚The democratic peace͛; states that democracies do not go to war with other democracies, while a
corollary law claims that democracies are less likely to be involved in militarized disputes than non -
democracies.

The main conclusions about comparative politics that can be drawn from comparison with natural science:

1.p For practical and ethical reasons, comparative politics relaxes some of the rigours of natural science,
but still employs the same logic of inference
2.p comparative politics is a non-experimental (or quasi-experimental) social science that seeks to make
generalizations based on the best available evidence
3.p As a substitute for experimentation, comparison allows for control, holding certain thing constant
while examining and accounting for observed differences.
4.p While not seeking ironclad laws, comparative politics seeks clarity, understanding and explanation of
political phenomena about which it can be reasonably certain.

Two basic types of theory in political science:

1.p Normative theory: specifies how things in society ought to be, given a desired set of outcomes and
philosophical position
2.p Empirical theory: seeks to establish relationships between two or more concepts in an effort to explain
the occurrence of observed political phenomena.

Theories in the political science can be:

1.p Deductive theories arrive ad their conclusion by applying reason to a given set of premises
2.p Inductive theories, on the other hand, arrive at their conclusion through observation of known facts

Method: is the means by which a theory is derived and tested, including the collection of evidence, formulation
and testing of hypotheses, and the arrival at substantive conclusions.

Ontology: is quite literally the study of being, or the metaphysical concern with the essence of things, including
the ͚nature, constitution and structure of the objects͛ of comparative inquiry. For cP, ontology concerns the
countries, events, actors, institution and processes among other things that are observable and in need of
explanation.

Epistemology: is the study of the nature of knowledge or how scholars come to know the world, both through a
priori means and through a posteriori means of observation, sense impression and experience. It concerns
what knowledge of the political world is possible and what rules of inquiry scholars fo llow in knowing the
political world

Methodology: concerns the ways in which knowledge of the political world is acquired. Methodology is the
study of different methods or systems of methods in a given field of inquiry.

 p Ontology establishes what is knowable, epistemology how it is knowable and methodology how it is
acquired systematically

cases: are those countries that feature in the comparative analysis.

Unit of analysis: are the object on which the scholar collects data

Variables: are those concepts whose value change over a given set of units

Observation: are the values of the variables for each unit, which can be numeric, verbal or even visual.

 p Vb: the cases are the countries. The units of analysis are the movements. The variable is ͚strategy͛ and
the observations is the value of the strategy variable for a given movement in a given country.

Dependent variables: alternatively referred to as outcome variables, endogenous variables or the explanandum
are those political outcomes that the research is trying to explain.

Independent variables: is that which explains the dependent variable.

It is possible to have more than one independent variables for a given dependent variable. The dependent
variable is often depicted by a y, and the independent variable is often depicted by an x.

Levels of analysis in political science are divided between the micro or individual level and the macro or system
level. Micro-analyst believe that the world of politics is shaped by the actions of ͚structure-less agents͛ while
macro-analyst believe that the world is shaped by the unstoppable processes of ͚agentless-structures͛

Quantitative methods seek to show differences in number between certain object of analysis. Quantitative
methods are based on the distribution these data exhibit and the relationship that can be established between
numeric variables using simple and advances statistical methods.

Qualitative methods seek to show differences in kind. Qualitative methods seek to identify and understand the
attributes, characteristic and traits of the object of inquiry, and the nature of the method necessarily requires a
focus on a small number of countries. In comparative politics there are three broad types of qualitative
methods:

1.p Macro-historical comparison (three subtypes: parallel demonstration of theory, contrast of context
and macrocausal explanation)
2.p In-depth interviews and participant observation
3.p Interpretivism, hermeneutics and ͚thick description͛ (are concerned with interpretation,
understanding and the deeper structure of meanings associated with the object of inquiry)
c
  


The distinction between different comparative methods should be seen as:

-p A function of the particular research question: different research questions require different methods
-p The time and resources of the researcher: the time and other resources of researchers are often
constrained, which limits the number of countries that van be feasibly researched in any one project
-p The method with which the researcher is comfortable: some are comfortable using quantitative
methods while others are not. Some enjoy large comparisons while others enjoy researching the fine
details of particular countries.
-p The epistemological position he or she adopts: researchers who adhere to deductive theory may use
different methods to those adhering to inductive theory. Those seeking more universal generalizations
may use different methods from those who seek more contextually specific levels of explanation.

The central distinction between comparative methods depends on the key trade-off between the level of
abstraction and the scope of countries being studied. The higher the level of conceptual abstraction, the more
potential there is for the inclusion of a large number of countries in a study, where political science concepts
travel across different contexts.

comparing many countries is commonly referred to as ͚large-n͛ comparison and comparing few countries
͚small-n͛ comparison, where n is the number of countries.

c 




comparing many countries most closely approximates the experimental method found in natural science. The
large number of countries makes this method of comparison particularly suited to quantitative analysis of
aggregate data collected on different measures that vary across many countries.

This method requires a higher level of abstraction in its specification of concepts

The main advantages The main disadvantages

- p Its ability to use statistical controls to rule
out rival explanations and control for
confounding factors
- p Its extensive coverage of countries over time
and space
- p Its ability to make strong inferences that
hold for more cases
- p Its ability to identify so -called deviant
countries or outliers that do not have the
outcomes expected from the theory being
tested

-p The limited availability of data for many
countries and many times periods
-p The validity of measures that are often crude
approximations of social scientific concepts
-p The mathematical and computing skills
needed to analyze increasingly complicated
data set whose structure and properties
violate many of the assumption of standard
statistical method of analysis
-p Many see this method of comparison as
inappropriate for analyzing many topics
involving complex causal mechanisms,
historical processes and deeper meanings
and understandings that are highly
dependent on the contextual specificities of
discrete country cases.

c 
 

comparing few countries involves the intentional selection of a few countries for comparison. This selection
may involve anywhere between 2 to more than 20 countries. The defining feature of this method of
comparison is the intentional selection of countries from the universe of possible cases. This method has been
variously called the comparative method, the ͚comparable cases strategy͛ or focused comparison, and it
achieves control through the careful selection of countries that are analyzed using a middle level of conceptual
abstraction. Studies using this method are more intensive and less extensive since they encompass more of the
nuances specific to each country. This type of comparison has thus also been referred to as ͚case-oriented͛,
since the country is often the unit of analysis, and the focus tends to be on the similarities and differences
among countries rather than the analytical relationship between variables.

There are two main types of research design that are included under this method of comparison:

1.p Most Similar Systems Design (MSSD): different outcomes across similar countries
2.p Most Different Systems Design (MDSD): similar outcomes across different countries

ë

 


A single-country study is considered comparative if it uses concepts that are applicable to other countries,
develops concepts that are applicable to other countries, and/or seeks to make larger inferences that stretch
beyond the original country uses in the study. Beyond this, however, single-country studies provide:

1.p contextual description

2.p Develop new classifications
3.p Generate hypotheses
4.p confirm and infirm theories
5.p Explain the presence of deviant countries identified through cross-national comparison

This section of the chapter has shown that all three methods should be grouped under the umbrella of
comparative politics if they seek to make generalizations through explicit comparison or if they use and
develop concepts applicable to other countries through implicit comparison. comparing many countries is the
best method for drawing inferences that have more global applicability. Through use of the method of
difference and method of agreement, comparing few countries can lead to inferences that are better informed
by the contextual specificities of the countries under scrutiny. Single-country studies can provide contextual
description, generate hypotheses, confirm and infirm theories and enrich our understanding of deviant
countries identified through other comparisons. It has also been made clear that different strategies of
comparison should be seen as the product of the trade-off between the level of conceptual abstraction and the
scope of countries, as well as the arbitrary and practical factors surrounding any comparative research project.

·

 

Too many variables and not enough countries/too many inferences and not enough observations: this problem
arises when more factors of explanation for the observed outcome have been identified than there are
countries (or observations) in the study, leading to what is called an indeterminate research design.

There are three solutions to the problem of too many variables and not enough countries, all of which are
based on the principle that the number of variables (or inferences) must be less than the number of countries
(or observations):

1.p Raise the number of observations to allow key factors of the study greater overall variation.
2.p Use the Most Similar System Design (MSSD) to achieve focused comparison of few countries.
3.p Reduce the number of variables by focusing on the key explanatory factors that are hypothesized as
important for explaining the outcome. This can be achieved either by using the most different system
design (MDSD) or by having stronger theoretical specifications.
Establishing equivalence:

The second problem is one of establishing equivalence both in the theoretical concepts that are used and the
operational indicators of those concepts as they are applied in multiple contexts.

There are three intellectual positions that offer insight into this problem:

1.p The universalist position: argues that if theoretical concepts and their indicators are to have any
explanatory power, they must be able to travel to all parts of the globe.
2.p The relativist position: argues that all meaning is locally determined and that a general ͚science͛ of
comparative politics is necessarily limited if not impossib le
3.p The middle position: argues that comparativists must not abandon all their concepts, but should
modify them to be more sensitive to the cultural specificities of the contexts they are studying.

Solutions for those seeking to make larger inferences through comparison (universal & middle position):

-p Raising the level of abstraction   allows the study to be more inclusive

-p Focusing on smaller number of countries for which the comparativists has thorough substantive
knowledge: suggests that the analyst be extremely cautious about engaging in cross-national
comparative research. Local knowledge can identify gaps between theoretical concepts and their
application.
-p Using specialist teams in compiling cross-national data sets: if truly informed comparison of many
countries is limited, then those seeking to compare many countries should venture out of the security
of the familiar if they are prepared to collaborate with other scholars͛ who possess specialist
knowledge of the countries under scrutiny.
-p Specifying the functional equivalence between concepts or indicators: identification of functional
equivalence of concepts and indicators. This solution does not envisage concepts as identical of even
similar, but functionally equivalent. The functional equivalence allows entities with seemingly
dissimilar characteristics to be grouped into useful and exclusive categories.

The key is careful specification of concepts, thoughtful construction of indicators that operationalize them,
careful application of them to multiple co ntexts and honest recognition of their limitations.

Selection bias:

2


 

 
 arises from the intentional choice of countries which often violates the
crucial scientific principle of using random samples. Selection bias in comparative politics occurs through the
non-random choice of countries for comparison or the deliberate selection by the comparativist. The most
blatant form of selection occurs when a study includes only those cases that support the theory. More subtle
forms of selection bias occur when the choice of countries relies on values of the dependent variable.

Studies that compare many countries usually have a sufficient number of observations to avoid the problem of
selection and quantitative studies of many countries can use a number of statistical techniques to eliminate the
problem.

Selection on the dependent variable led to an overestimation of the importance of certain explanatory factors.
There are three solutions to the problem of choosing on the dependent variable:

1.p Have a dependent variable that varies

2.p When comparing few countries, the choice of countries ought to reflect substantive knowledge of
parallel cases
3.p Stronger theory may specify more accurately a range of countries in which certain outcomes and their
explanations would obtain
 4.p Strong theory will also identify which countries represent ͚least likely͛ insta nces of the phenomenon
under investigation

½
  

 

 arises in qualitative studies that rely on historical sources, where the analyst
chooses historical accounts either intentionally or unintentionally whose description of events fits the
particular theory being tested.

Historiography varies in its description of how the past actually unfolded, which events receive emphasis, as
well as the different theoretical dispositions of the historians themselves.

Solutions to this form of selection bias include using multiple sources to arrive at a ͚mean͛ account of the
events and identifying the tendencies within each source to acknowledge possible sources of bias.

½   

 

 can occur from the time periods that are used in the comparison, especially for
those studies seeking to analyze social behavior that has a very long history, such as warfare, trade and the
emergence of states and regimes.

The solution is the attempt to provide generalizations about an important aspect of politics by comparing
whole systems over long periods of time.

Spuriousness:

The forth problem is spuriousness, or the omission of key variables that may account for both the outcome and
other explanatory factors already identified. A spurious explanation is one in which some unidentified factor is
responsible for the outcome, while the identified factor is mistakenly attributed to having an effect on the
outcome.

The solution to the problem of spuriousness are related to the number of countries in a comparative study:

1.p The easiest solution for spuriousness is to specify all the relevant variables that may account for the
observed outcome
2.p Select countries that fit the criteria of the theory that has been specified

Ecological and individualist fallacies:

The fifth problem arises when a study seeks to make inferences about one level of analysis using evidence from
another.

There are two types of data in the social sciences:

1.p Individual data: comprise information on individual people

2.p Ecological data: comprise information that has been aggregated for territorial units, such as voting
districts, municipalities, countries and states.

An ecological fallacy: occurs when results obtained through the analysis of aggregate-level data are used to
make inferences about individual-level behavior.

An individualist fallacy: occurs when results obtained through analysis of individual-level data are used to make
inferences about aggregate-level phenomena.

Both fallacies are a problem since analysis carried at one level may overestimate relationshops at another level
and both fallacies originate from the same sources, namely, the ontological predispositions of the researcher
and data availability.
Rationalist: may collect information on individuals to make larger claims about groups

Structuralist: may collect information on groups of people to make larger statements about individuals.

Data availability is the second source of ecological and individualist fall acies, since scholars may be forced to
substitute data form one level to examine a research question specified at another level.

The solutions for avoiding both fallacies is straightforward. The data used in any research ought to minimize
the chain of inferences between the theoretical concepts that are specified and the measures of those
concepts that are ultimately adopted in the analyses   principle of direct measurement

Value bias:

The final problem for all comparativist to consider is that of value bia s, where the particular cultural, political
and philosophical predisposition of the researcher necessarily biases the conduct and conclusions of the
enquiry.

The key to making valid comparisons is to be as public as possible in terms of the judgements that have been
made in the overall construction of the comparative study.

Summary (overgenomen van boek):

Method Strengths Weaknesses/challenges

comparing -p Statistical control -p Invalid measures
many countries -p Limited selection bias -p Data availability
-p Extensive scope -p Too abstract/high level of
-p Strong inferences and good for generality
theory-building -p Time-consuming
-p Identify deviant countries -p Mathematical and computer
training
comparing few -p control by selecting -p Less secure inferences
countries 1. Most similar systems design -p Selection bias:
(MSSD) 1. choice of countries
2. Most different systems design 2. choice of historical account
(MDSD) -p Language training
-p Good for theory-building -p Field research
-p Intensive, less variable oriented
-p Avoid ͚conceptual stretching͛
-p Thick description
-p Areas studies
-p configurative analysis
-p Macro-history
case Study -p Intensive ideographic, path- -p Insecure inferences
dependent, and configurative -p Selection bias:
analysis. 1. choice of countries
-p Six types: 2. choice of historical account
1. Atheoretical -p Language training
2. Interpretive -p Field research
3. Hypothesis-generating
4. Theory-confirming
5. Theory-infirming
6. Deviant countries
c
c 
 

Assumptions

The universe of countries has been intentionally selected an is significantly smaller than a global selection.

Various qualifications of comparing few countries:

1.p It is case-oriented rather than variable-oriented since the focus of the analysis is much more on the
specific unfolding events and variation in political developments within each country than variation in
macro-variables between countries
2.p It is intensive rather than extensive since those factors considered do not vary across a wide range of
countries, but vary over time and across sub-national units within a smaller sample of countries, which
allow the researcher to probe more deeply into each individual case comprising the sample
3.p While the comparison of few countries sacrifices the ability to make broad empirical generalization, it
means that they may well be located at a lower level of abstraction in which concepts and ideas are
operationalized in ways that fit more closely with the contextual specificities of the countries used in
the comparison

Most similar systems design (MSSD)/method of Seeks to compare political systems that share a host
difference of common features in an effort to neutralize some
differences while highlighting others.
Seeks to identify the key features that are different
among similar countries and which can account for
the observed political outcome
 p Area studies
Most different system design (MDSD)/method of compares countries that do not share any common
agreement features apart from the political outcome to be
explained and one or two of the explanatory factors
seen to be important for that outcome.
Seeks to identify those features that are the same
among different countries in an effort to account for
a particular outcome.

Where quantitative analysis requires mathematical and computer skills, area studies require language training
and extensive field research, which have been seen as distinct disadvantages to comparing countries from a
given region.

The intentional choice of countries based on the presence of the same outcome constitutes one form of
selection bias, which necessarily limits the types of inferences that can be drawn from comparison.

In certain instances, selecting on the dependent variable is acceptable, if and only if, the research is seeking to
test for necessary (as opposed to sufficient) conditions for a particular outcome of interest. In cases where
strong causal language is adopted in ways that imply the presence of necessary and sufficient independent
variables, then selecting on the dependent variable is not acceptable.

While it is typical for scholars to use either MSSD or MDSD some comparativists use both system designs.

What remains important to all these methods of comparing few countries is the proper specification of the
outcome that is to be explained, the reasons for adopting either system design, as well as the choice of the
particular countries under scrutiny.
Apart from the inherent problems of too many variables and not enough countries for both methods, as well as
the problem of selection bias in MDSD, there is a further complication that is related to the number of
countries that are compared:

-p Increasing the number of countries will affect the ability of a scholar to meet the conditions required
by doing either MSSD or MDSD   at some point the inclusion of more and more countries forces both
research designs to move into the realm of a many-country comparative design since MSSD loses its
ability to hold shared characteristics constant and MDSD runs out of countries from which to choose.

If a study want to include instances in which the outcome of interest does not occur (known as a negative case)
it is not entirely clear which negative cases ought to be chosen.

(X1,X2,Y) YES X1 NO
Y  YES NO YES NO
YES YES I: positive (1,1,1) II: disconfirming III: disconfirming IV: negative
(1,1,0) (0,1,1) (0,1,0)
X2
NO V: disconfirming VI: negative (1,0,0) VII: impossible, but VIII: irrelevant
(1,0,1) happens (0,0,1) (0,0,0)
NO
The possibility principle suggest that cases need to be selected in which there is a possibility of the outcome of
interest occurring (above: cells I-VI).

This consideration of the logical combinations of variables for a typical comparison of few countries is
instructive for two reasons:

1.p It identifies cases beyond the original comparison that ought to be examined to provide more
definitive support for a theory
2.p It identifies those cases that do not need to be examined to test the theory.

Eliminating the possible cases for inclusion is vital for scholars seeking to compare a small number of countries
in greater depth while nevertheless providing an evidence base from which more secure inferences can be
drawn and support for a general theory can be made.

Hypothetical truth table:

causal conditions Outcome

cases A B c D E F
1 A b c d e f
2 A b c d e f
3 A B c d E f
4 A B c D e F
5 A B c D E F
6 A B c D e F
7 A B c d E f
8 A B c d e f

9 A B c D E F

10 A B c D e F

½     
     
 



 
c




      


 
 






      


The purpose of the truth table and subsequent analysis is to determine whether there are any discernible
patterns across the causal conditions and the different outcomes

A Boolean truth table demonstrates that qualitative comparative analysis offer a powerful tool for the
comparison of few countries for four reasons:

1.p It allows for the inclusion of information that has not been measured precisely, but that is represented
through reasonable judgements and the application of criteria that are defensible
2.p It uses the combinatory logics of binary variables found in Boolean algebra to simplify the complexity
of the world in order to tease out the set of necessary and sufficient conditions that account for an
outcome of interest
3.p It allows for an assessment to demonstrate how certain causal conditions contribute to an outcome,
and how such a contribution needs to take place alongside the presence of other important factors in
order for the outcome to take place
4.p Beyond identifying this set of necessary and sufficient conditions, the technique also allows the
assessment to determine the reason for the outcome did not occur in certain cases, and thus draws on
the insights provided in the discussion of negative cases above

There are a number of limitations to the methods for comparing few countries that once again reinforce the
general position advanced in this book that comparative scholars will always face a trade-off between the
scope of countries included in any one study and the level of abstraction and strength of the inferences that
result from the number of countries that are compared.

-p The problem of selection bias looms large

-p The choice of most similar and most different cases can appear at times arbitrary
-p The inclusion of negative cases, while laudable, may nonetheless not have exhausted all cases that
ought to be considered when analyzing particular outcome of interest

It appears that most MDSD remains somewhat weaker than MSSD, because of the selection bias and since its
inferences relate only to the confirmation of the importance of the presence of certain explanatory factors and
in some cases the identification of necessary conditions for a particular outcome.

By definition MDSD does not provide a framework in which negative cases can be included.
c


 


By a single-country study, I refer primarily to any study in which a single country forms the basic unit of
analysis, but which may also be broken down into smaller units across time and space, by examining sub-
national variation across states in federal countries, other administrative units in unitary systems, as well as
other appropriate units of analysis, such as individuals.

Single-country studies are necessarily more intensive (have a lower level of abstraction)

Functions of a single-country studies:

-p contextual description
-p classification
-p Hypothesis generation and plausibility probes: The generation of hypotheses often comes from the
analysis of political events, outcomes and behavior in single countries that are well known to the
scholar that present new research puzzles for wider debates in the field, and that either explicitly or
implicitly suggest that the generated hypothesis be tested in a larger selection of countries
-p Theory informing and theory confirming: theory-confirming and theory infirming studies draw on
known findings from existing studies that have been conducted in a larger sample of countries
-p The analysis of deviant and outlier cases: deviant country studies can weaken existing theories as well
as further refine the concepts and measures used in the original comparative analysis
-p Process tracing and the elaboration of causal mechanisms

Single-country studies are thus not plucked from thin air, but are specifically chosen for the merit in
contributing to larger sets of questions in the field

The identification of outliers and the intentional selection of countries on that basis comprise what is often
called the least likely of most likely method of comparison:

-p Most likely studies are carried out on countries where a theory suggest a particular outcome is
definitely meant to occur. If the outcome is observed as expected, then the theory is confirmed. If the
outcome is not observed, then the theory is infirmed
-p Least likely studies are carried out on a country where a particular theory suggest the outcome is not
likely to occur. If the outcome is not observed, then the theory is confirmed. If the outcome is
observed, then the theory is infirmed

These crucial country studies do not definitively prove or disprove a theory but merely confirm or infirm its
applicability to other countries

Single-countries studies can be used to trace significant political processes and examine possible causal
mechanisms that lie between two or more variables of interest. In both instances, the intensiveness of the
single country study allows for a more detailed look at underlying processes and mechanisms that simply
cannot be investigated in studies that compare more countries

The explanatory leverage of a single country study can be increased through strategies that raise the number of
observations. These strategies are based on the general assumption that increasing the variation of any
observable phenomena enhances the ability to make inferences and provide more systematic explanation.

It is clear that any strategy to increase the number of observations rests on three different but related
parameters:

1.p Time: it is possible and desirable to raise the number of observations by comparing across different
historical periods in the country of across continuous units of time, such as years, months or days
 2.p Space: all countries have sub-national political units over which significant variables can be compared,
such as states in federal systems, counties, regions and cities
3.p Level of analysis: a single-country study can examine groups of people at a lower level of aggregation
than states or administrative units, as well as analyze mass publics themselves with individual data.

Generalization from single-country studies will always be limited since the country unit itself is bound by
particular characteristics, while the potential for comparing variation in political phenomena across units is
bound by time and space. However many sub-national units are compared over however many days, months
and years, the inferences one draws from them will have to made with care.

Single-country studies by definition trade their ability to provide in-depth knowledge and understanding of
particular context against the ability to draw generalization that have wider applicability.