Doing Research in Political Science

An Introduction to Comparative Methods and Statistics

Paul Pennings, Hans Keman, Jan Kleinnijenhuis

• Publisher: SAGE Publications Ltd
• Year: 2005
• Print ISBN: 9781412903776, 9781412903769
• eBook ISBN: 9781446226902
• Edition: 2
• Topic: Politics & International Relations
• Subtopic: Politics
• Index: Politics & International Relations

This is an immensely helpful book for students starting their own research… an excellent introduction to the comparative method giving an authoritative overview over the research process

- Klaus Armingeon, University of Bern

Doing Research in Political Science is the book for mastering the comparative method in all the social sciences

- Jan-Erik Lane, University of Geneva

This book has established itself as a concise and well-readable text on comparative methods and statistics in political science I…strongly recommend it.

- Dirk Berg-Schlosser, Philipps-University Marburg

This thoroughly revised edition of the popular textbook offers an accessible but comprehensive introduction to comparative research methods and statistics for students of political science. Clearly organized around three parts, the text introduces the main theories and methodologies used in the discipline.

Part 1 frames the comparative approach within the methodological framework of the political and social sciences.

Part 2 introduces basic descriptive and inferential statistical methods as well as more advanced multivariate methods used in quantitative political analysis.

Part 3 applies the methods and techniques of Parts 1 & 2 to research questions drawn from contemporary themes and issues in political science.

Incorporating practice exercises, ideas for further reading and summary questions throughout, Doing Research in Political Science provides an invaluable step-by-step guide for students and researchers in political science, comparative politics and empirical political analysis.

Part 1

Comparative Methodology

1

Comparative methodology and statistics in political science

CONTENTS

1.1 Introduction

1.2 The Comparative Approach to Political and Social Science: Theory and Method

1.3 Comparing Data: Selecting Cases and Variables

1.4 Developing Empirical-Analytical Comparative Analysis

1.5 How to Use This Book

1.6 Endmatter

Topics highlighted

Questions

Exercises

Further reading

1.1 Introduction

Almost everyone watches daily TV, regularly reads a daily newspaper and often discusses what goes on in the world. These activities shape our views on society and, in particular, influence our views on and perspective of the role and impact of politics on societal developments. In this era of easy access to electronic communication (e.g. Internet), worldwide TV coverage of events (e.g. CNN) and rapid changes in the political mapping of the world (globalization), one is confronted not only with a multitude of bits and pieces of information, but also with various and often conflicting opinionated views what events may mean and what consequences they may have for our lives and the society we are part of and live in.

Although we do not realize it all the time (or at all) we use this information in its multifarious forms in a comparative way. Both the ‘messengers’ (e.g. journalists, political spokesmen and so-called opinion leaders) and the ‘receivers’ (readers, TV watchers, person-to-person communicators) are, more or less consciously, using the ‘art’ of comparing in order to come to a more or less well-founded interpretation of what goes on in public life.

The first point of departure of this book is therefore not only that students of social and political sciences are in fact comparing information to form an opinion, but that everyone is doing this in assessing the facts of life around him or her. For instance, how often do you use the words ‘more’ and ‘less’ or ‘bigger’ and ‘smaller’, and this is ‘different’ from or ‘similar’ to that, and so on? All these expressions, used by everyone in their daily conversation, basically imply that you (seem to) have a comparative idea about what occurs in reality. And not only that – most of the time, if not always, you do deliver a statement about, for instance, politics and society that is, more or less, implicitly of an evaluating nature. To give an example: in New Zealand in 1996 the first elections were held under a new system (it used to be ‘First Past the Post’ and it is now a variation of a proportional representation electoral system). The electoral outcome necessitated the formation of a coalition government instead of a one-party government. Apart from the fact that this type of government and the related procedure of government formation were new to both the public and the politicians, everyone could now compare the actual result of changing the electoral system and what it implies in reality. Hence, one could now evaluate what goes on by means of comparing the old with the new situation.

The ‘art of comparing’ is thus one of the most important cornerstones to develop knowledge about society and politics and insights into what is going on, how things develop and, more often than not, the formulation of statements about why this is the case and what it may mean to all of us. To take another example: in a number of Western European democracies one can witness recently a rise of so-called ‘populist’ parties (e.g. in Austria, Belgium, France, Italy, and the Netherlands; see Mair, 2002). The problem that emerged was how to define ‘populism’ as such in order to indicate which party was more (or less) populist, or – for instance – extreme right-wing or not, and therefore a threat to the existing party system (Mény and Surel, 2002). Hence, the problem was less to observe the phenomenon, and more how to measure it properly from a comparative point of view.

Yet, and this is our second point of departure, the use and application of the comparative method is often not systematic, nor is it applied rigorously in most cases. This may result not only in unfounded opinions or flawed conclusions, but also in biased views of reality as well as in inappropriate generalizations about what goes on in society. In this book we wish to introduce you to the comparative method and related statistical tools in order to help you to reduce these hazards and to develop standards for you and others to gain a more sustainable view on the world. In addition, we shall provide you with a clear schedule to develop an adequate research design that helps to avoid the mistakes and biases. This is the assignment of Part I.

In this chapter we shall therefore discuss how to do research in ‘comparative politics’. This means that the focus is on the development of a proper research design that enables one to translate questions about real-world events into observations, which allow for drawing systematically conclusions that can be generalized. For instance: is there a relationship between the (electoral) rise of populist parties and a growing dissatisfaction of the public with the working of parliamentary democracy? This type of Research Question can and should be elaborated in a proper Research Design. This crucial step in doing research in political science is the subject of the next chapter. It requires the elaboration of the phenomenon under review (e.g. what is populism, and which parties can be viewed as ‘populist’ or ‘right wing’?), the mode of analysis that makes a comparison useful and meaningful (e.g. relating the emergence of populist parties to subsequent events such as elections and stable government), and – in addition – the empirical investigation of all relevant cases (in comparing political systems that allow for corroborating hypotheses). Hence, instead of focusing on ‘events’ or isolated developments, the point of departure of our approach is:

• developing systematic knowledge that transcends mere description and allows for generalizations (i.e. external validity);
• deriving answers to questions on the basis of existing theory or, if possible, plausible hypotheses (i.e. theory guidance);
• striving for exact information and comparable indicators that are reliable and open to replication (i.e. internal validity).

In summary: without a proper research question and research design, the ‘art of comparing’ becomes meaningless and – which is worse – may lead to dubious evidence and conclusions that affect many in society. Max Weber – the famous German sociologist – warned against these practices in 1918 in his major work Economy and Society (Weber, 1972), by discussing value-free science vis-à-vis ideologically driven analysis, which would not only harm scientific progress, but also jeopardize the correct use and application of social scientific results in practice (see Bendix, 1977; Giddens, 1971).

From this follows, as the third point of our presentation, that it is crucial to know from the beginning what, when and how to compare. Seemingly this triad goes almost without saying. Yet, it is vital for any comparative analysis to ask him or she whether or not there is indeed a proper answer to these methodological questions. If not, the chances to come up with valid and reliable answers will be reduced and the quality of knowledge advanced will be less. Hence, you must know beforehand what the phenomenon is that you wish to research, when – or at what point of time or period under review – the phenomenon can be best studied, and how to do this.

This highlights perhaps the most important message we wish to emphasize. We view the ‘art of comparing’, or what is generally called the ‘comparative approach’ to political and social science, not as an ‘art’ in itself (or a method per se), but as one of the most adequate ways to connect ideas (theory) about society and politics with what is actually going on in the world we live in (i.e. empirically founded facts). In short, we wish to introduce you to the comparative approach in such a way that one can explain convincingly and plausibly what is going on in the real world of politics and society.

Box 1.1 Comparing as a basic tool of the social sciences

The British poet Rudyard Kipling (1865–1936) wrote: ‘And what should they know of England who only England know?’ He meant to say that without comparing there is little to gain from a description only. Therefore the ‘art of comparing’ is a basic tool for linking ideas and, eventually, theory to evidence. Conversely, without theory a comparison remains meaningless. Our view is thus that ‘doing research’ in the social sciences always implies – be it implicitly or explicitly – the application of the comparative approach to gain knowledge of politics and society and to assess its plausibility.

1.2 The Comparative Approach to Political and Social Science: Theory and Method

We contend that the comparative approach and its methodological application must be conducted by means of theory-driven research questions. This is to say: a research question must be formulated as a point of departure of comparative investigation, which enables the student to reflect on what, when and how to compare and for what purpose. If not, the comparison becomes a recording instrument only. This, however, is not our goal, nor is it in our view scientific. Scientific activities always imply the quest for explanations, which are not only empirically based and yield systematic results, but also lead to results which are plausible. It is vital to realize that throughout this book we shall contend that empirical-analytical analysis is an instrument to develop social and political knowledge that is both scientifically valid and plausible for a wider audience.

Valid means here not only whether or not it is devoid of mistakes of the ‘Third Order’ (Blalock, 1979), i.e. avoiding wrong operationalizations, incorrect indicators and inadequate levels of measurement and inferring false causal conclusions – these matters will be dealt with in Part II of this book – but primarily whether or not the research design is indeed adequately derived from the research question which underlies the comparative research. Validity in comparative (and other types of) research is a very central concept. However, more often than not, it is used in different ways and its use may well confuse the student. Throughout this book we shall employ the concept as follows:

• Internal validity concerns the question whether or not the measurements used in a given research are properly, i.e. correctly, operationalized in view of the theoretical concept as intended. For instance: in a research project on political parties, can all the parties under review be considered to be identical in terms of their properties (e.g. participating in elections by putting forward candidates for office), and can they be seen as unique entities and not be confused with other types of social and political movements (like interest groups or new social movements)? Hence research results are internally valid if and when they are truly comparative, i.e. yield the same results for all cases under review (if not, then a case is ‘deviant’).
• External validity presupposes that the concepts used in a given piece of research, and the related outcomes, apply not only to the cases under review but to all similar cases that satisfy the conditions set out in the research question and related research design. Similarity implies here comparability through space or time. For example, the factors found to explain the variations in government formation in terms of the resulting types of government (e.g. majority or minority and one-party versus multi-party governments) should also apply to those cases that were not included or in periods that were not covered in the original analysis. Another example would be the study of populism, right-wing parties and party system development (see, for instance, Kitschelt, 2002; Pennings and Keman, 2003). Obviously this requires careful and qualified arguments and spills over into the quality of operationalization and measurement (i.e. internal validity!). Hence research results are viewed as externally valid if they yield truly comparable results for similar cases that have not yet been under review. This implies that one would expect that a replication of such a research should produce by and large the same results (King et al., 1994: 100).

It should be realized that the concepts of internal and external validity are of an ideal-typical nature: in a perfect world with complete information the standards of validity may well be met, but in practice this is not a realistic goal. Yet, and this is what we put forward, one should try to get as close as feasible to these standards (see Mayer, 1989: 55; King et al., 1994). Only by keeping these standards is it possible to strive for positive theory development, that is, systematically relating extant theory to evidence and so improving the theory.

To enhance this process of theory development we argue throughout this book that one needs to formulate a Research Question (RQ) first, in order to be able to decide what, how and when to compare. This leads in turn to the development of a Research Design (RD) in which these matters are addressed and elaborated in such a way that the research results will be valid, reliable and plausible. It is also important to note that the comparative approach allows for two types of analysis: one is the explorative type that aims at identifying relationships which may be conducive to theory formation; the other is driven by theory and aims at testing causal relationships, which is necessary to corroborate extant theory and to develop these further. Only then it is possible to decide which data must be collected to carry out the empirical and statistical analysis for a meaningful comparison that may produce substantial explanations of why societal and political events and developments have taken place. In short: substan...

Table of contents

• Cover Page
• Title
• Copyright
• Contents
• Preface
• Part 1 Comparative Methodology
• Part 2 Statistics in Political Science
• Part 3 Doing Political Research
• Appendix
• Bibliography
• Index