SHORTCOMINGS IN METHODOLOGICAL APPROACHES

Quantitative research on the basis of mass data nowadays is very fashionable. However, the quantitative analyses of regional or worldwide surveys (Eurobarometer, World Values Survey/European Values Survey) and statistical material has problems. The data are often provided by specific institutions (such as the United Nations, World Bank, OECD, European Union), but they do not provide perhaps a consistently accurate picture of every country. 'The lower the level of development, the lower is also the validity of quantitative data' (Dogan, 1994: 41). In addition, the data may not illustrate the problem in the right way for testing hypotheses because they are not collected within the framework of the particular research project. If the latter is done, the process takes so much time that social and economic realities may have changed by the time the data are ready. This is true above all for worldwide designs. Another problem is that the links of the data to their context are absent, with the result that the significance of statistical information can vary from country to country.

Frequently, surveys aim at measuring political and social attitudes, values and behaviour. People in different countries may interpret researchers' questions in

'The lower the level of development, the lower is also the validity of quantitative data'.
'mass data may have "lured some researchers into a false sense of security"'.

different ways, for example people may have different ideas about democratic political systems. This problem will occur more frequently in data that stem from countries with political cultures different from each other. The same phenomenon, for example political participation, is seen in Islamic countries in a different light than in the western world, as the role of men and women concerning public matters differs strongly (Norris and Inglehart, 2004). Considering this, mass data may have 'lured some researchers into a false sense of security' (Dogan, 1994: 36) and may have prevented them from seeing the real problems. Dogan (1994: 37) stresses the problem of an 'imbalance of quantitative data and the technological capacity of statistical treatment'. It follows that it would be appropriate to make comparisons with data only of the same background, for example affluent liberal democracies of the first world, countries in transformation from communist rule, countries of the Islamic world and pre-modern states. There is a wide consensus among political scientists on the question as to which countries belong to the core of each category. The debate concentrates mainly on those countries that are in transformation. Nevertheless, worldwide comparisons have become fashionable in the past two decades.

Finally, political researchers using mass data have to be aware that data can only reproduce facts that are easy to quantify, for example expenditure figures. However these data never provide information as to how they came about. Furthermore, for many research questions, suitable and comparable data are not available for different political systems. As a consequence, the collection of data is usually limited to sub-systems for a small number of countries (e.g. political parties, such as the party research of Katz and Mair, 1992, 1994) or even party programmes (as done by Klingemann et al, 1994) in a worldwide approach (Klingemann and Hofferbert, 2000).

An alternative to the quantitative approach is the case-study approach. For some scholars the case-study approach means that only one entity is analysed: 'Case studies are based on the sample of one' (Hague et al, 1993: 37). However, results from analysing a single case are seen to be of limited value. The scientific community holds the method in low regard. This view persists even though many single case studies have become classic works (see Gerring, 2004: 341; Walton, 1992: 125). Case studies are only estimated as explorative, preparing the ground for quantitative analyses (Lijphart, 1975: 160). They are not seen as suitable for generalisation of the results, but only 'as useful to disconfirm a regularity to a limited degree' (Sartori, 1994: 23).

Yet, the aim of empirical comparative research must be to formulate rules that apply to all similar cases or those of a particular type. As a consequence, for decades, a tendency has taken hold in the discipline to analyse and compare more than one case in research. This is often seen as a stepping stone for generalisation to a larger set of units. The effort of Gerring to reach this goal on the basis of a single unit, defining case studies as 'an intensive study of a single unit for the purpose of understanding a larger class of (similar) units', seems to be a setback (Gerring, 2004: 342). Furthermore, it only provides a partial solution to the problem of how to create an adequate research setting, in order to meet the above-formulated demands. Yet, many who do not think case studies are particularly valuable even regard comparisons on the basis of small N's as being deterministic because, it is argued, it is impossible to examine interaction effects (Lieberson, 1992: 109, 111) and there are doubts that a representative sample can actually be achieved. A genuine debate among scholars of political science about these problems has not yet been set in train. Moreover, in methodological matters political scientists still make use of the common methodological instruments of social science, which have been provided by sociologists and are not adapted to special research problems that need to be mastered for the needs of political science.

In short, both strategies (using survey data for quantitative analysis as well as case studies, using a mix of qualitative and quantitative methods) bear imponderables. We shall now explore a way out of this dilemma.

STEPS TO OVERCOME THE DILEMMA

In the first place, at least two nations or sub-systems have to be analysed. In case studies, it is usual to collect data using different methods according to the hypotheses being tested, which is always time-consuming. In cross-national designs, it is common to use the expertise of scholars close to the cases, to review data from different sources before analysing data sets. Therefore, most research will end up with a small number of cases under study. In most investigations, more than that would not be practicable, as financial resources are usually limited. In addition, 'the problems of reliability and validity are smaller', if the researcher analyses a 'smaller number of cases more thoroughly, and he is less dependent on data that he cannot properly evaluate' (Lijphart, 1975: 170).

Nevertheless, many researchers hope that findings from a large number of case studies, carried out by numerous scholars across different projects, might one day contribute to theory building. This is especially true for those who have been active in the different policy fields for years,¹ where an atomisation of research is often noticeable (e.g. Hofferbert, 1990: 135). The selection of these cases is often done 'in a haphazard manner' (Peters, 1998: 56), for example the research is conducted in the neighbourhood of the university of the researchers, in their own country, the political party to which scholars are affiliated, etc. An explanation as to why these cases are selected is often not given. The results of these case studies are compared with findings from other political systems. Most research in political science is undertaken in an environment that is easily accessible to the researcher. At present, the problem with comparisons of a small number of case studies is that a well-prepared design is often missing, and this might be the major reason why empirical research using case studies is held in low esteem.

SELECTING THE CASES

The major aim must be to select the right cases for the design. 'A case study method is correctly understood as a particular way of defining cases, ...' (Gerring, 2004: 341). As there are only a limited number of potential cases that are comparable – for example states of the world (this limitation also applies to cities and towns, as will be pointed out later) – selection at random is not appropriate. Only if there are enough potential cases to choose from will all likely factors (even those one might not have thought of) be in the sample.

'the problems of reliability and validity are smaller', if the researcher analyses a 'smaller number of cases more thoroughly, and he is less dependent on data that he cannot properly evaluate'.
'A case study method is correctly understood as a particular way of defining cases...'.

The larger the sample, the more likely it is that all characteristics of the 'whole' will be contained in the sample and therefore the sample will be an accurate representation of the 'whole'.

As empirical investigations aim at generalisation of the results, scholars need to select their cases with care. They should initially be aware of any consensus on the major features of the given political systems (i.e. features that are typical and essential), for these provide a guide to the selection of cases. In addition, one has to bear in mind and make use of what is already known about the functioning of political systems. Commencing on the basis of the current 'state of the art' is indispensable: 'existing theory is a good guide' (Peters, 1998: 34, 109). This helps to place particular emphasis on specific variables that are classified as dependent or independent variables according to the research question. The assumption, therefore, that a case study method is an approach that is not variable-oriented is wrong.

There is a consensus in the scientific community of political scientists that the problems of large states are not comparable to those of very small ones. Yet, while every political system has to cope with, say, tough environmental challenges, and the smaller even more so than the larger ones (Katzenstein, 1984, 1985), the value of comparing Liechtenstein or Luxembourg, on the one hand, with the US or Canada, on the other, is questionable. The size of any communities for comparisons is important: mega-cities, big cities, middle-sized cities, small towns, as measured by the number of inhabitants. The reason is that communities have to face major problems that change according to size. Very large cities are global players, with specific infrastructures and economic structures (Sassen, 2000, 2001). They (as well as big cities) have to deal with problems caused by foreign immigrants and the affluent families leaving these cities to live in the suburbs. Middle-sized cities are fighting to defend their central functions against towns in their neighbourhoods, competing for administrative authorities, providing modern infrastructures for housing, shopping, etc. A mono-structured economy, or a town dominated by one large employer, may produce a special lobby structure. The struggle among middle-sized cities and small towns concerns competition for inhabitants by offering building sites for houses. Nevertheless, the size of cities, measured by the number of inhabitants, is not the only criterion for classification. The location of the communities has to be taken into account as well. In a rural area, a middle-sized town has more central functions than in an urban region, where other middle-sized and larger cities in the neighborhood are competitors for these functions.

The entities to compare have to be similar to each other along most variables (the contextual variables). This introduces the opportunity of looking for more specific variables according to the hypotheses that have to be verified or falsified. Because of the complex structure of political systems, the inclusion of all variables would be impossible: the smaller the sample, the more the variables can be taken into account. But how does one identify the important or key variables that structure political decision making? There is an absence of research into this. Lijphart (1971, 1975: 166) came close to it, discussing the problem of comparing small N's, but he never named the key variables to which the comparison should be restricted, that is, omitting those of marginal importance (Lijphart, 1975: 159). This is also true for Peters (1998: 37ff.) and King et al (1994: 124–49). The latter named some variables from a behavioural perspective (1994: 178). Ragin (2000: 316), with his 'fuzzy set' approach, has tried to bridge the gap between analysing mass data on the one hand and conducting case studies on the other. He argues that the 'great strength of the fuzzy-set approach is that the key decisions that structure the analyses must be made explicit by the investigator and are open to evaluation by the audiences for the research', but he only gives examples of membership criteria for the cases to select, or for 'defining and selecting relevant aspects of the case'. Major variables are also missing in other recently published considerations of this issue (Collier et al, 2004: 85–102; Ragin, 2004: 125, 133). We will now try to address these gaps in the literature.

HYPOTHESES ABOUT MAJOR VARIABLES

Hypotheses regarding the interaction of variables are provided by the results of different approaches in political science. Scholars have studied the complexity of political systems from different perspectives, formulating and testing similar hypotheses. This has resulted in other aspects being neglected, despite their relevance from other perspectives. Specific views, while analysing the reality, end up in specific results that contribute to our overall knowledge of political systems. Utilising different approaches, our knowledge about the complexity of the phenomena has become more profound (see also Peters, 1998: 109). Good research projects, while adopting different perspectives to earlier research, nonetheless take into account the findings of that research. To avoid starting again from scratch, earlier findings have to be borne in mind (Wieviorka, 1992: 163).

A variable-oriented design for comparison should not only be used in the quantitative analysis of mass data but should also be routine in the design of comparative case studies (Ragin, 1987: 55). As Ragin (2000: 70) notes, the case-study research strategy should be 'extraordinarily theory-dependent. Theory is used to make sense of the case as a configuration of theoretically relevant aspects, which cohere as a package'. This requires a close analysis of the political system. However, the problem has often been that interdependences have been overlooked. For example, researchers starting from the standpoint of individual behaviour subsequently accepted that institutions influence political behaviour. This awareness should dissuade scholars from assuming that their approaches or paradigms are definitive and self-contained. Nevertheless, it remains true that scholars still, after many decades, highlight the same sort of factors that structure activities in political systems. This should be the basis for shared standards. Without these shared standards, comparative political science will lose its way.

Not all comparisons produce results from which generalisations can be drawn. Whether findings in comparative research are produced that are valid for political systems beyond those analysed depends on the selection. The design has to be rooted in knowledge about political systems. Adopting correctly the methods typical of comparative research (the method of difference and the method of agreement) means that the variables representing differences and similarities have to be able to be identified.

At present, the method is seen as appropriate only if we are sure that we have identified all plausible causal factors.² However, as innovative research continues, discussion about the major variables should not be viewed as exhausted. One never knows whether new problems will create new approaches in which new variables are examined. Therefore, the variables should be identified by the researcher while constructing the design for comparison, so that evaluation of the selection is possible. Only if we bear this in mind from the start can the comparison as method be carried out correctly. This is even true for investigations that try to include all political systems in the comparison, as sub-divisions usually have to be identified and created.

Although in the political science literature there are always slight differences concerning classification and combination of research paradigms, there is nonetheless a broad consensus on approaches used in empirical research: institutionalism, pluralism, corporatism, behaviourism, cultural perspectives and policy analyses. Following a renewed interest in political institutions (dubbed 'new institutionalism'), these approaches have been more or less combined. The 'new institutionalism' emphasises the need to integrate findings of the various efforts in order to achieve a better understanding of how institutions shape, order and modify individual choices. The various approaches to political systems have contributed to building a more complex picture about political systems and the major factors that structure their complexity. Moreover, these analyses have provided a foundation for typologies and classifications, providing guidance for the selection of political systems for comparative research. A consensus about the strategy of selecting cases would be a suitable basis for comparing the results of different empirical research projects and for working on desiderata. Existing treatises on comparative research do not elaborate on this issue.

GENERATING VARIABLES FROM COMMON APPROACHES

At present, scholars have to work on the basis of tentative typologies, as in many fields there is a deficit of empirical research and findings. In any case, different approaches have brought about major similarities concerning the importance of certain factors and variables.³ Institutionalism has highlighted institutions as independent variables. Polities are shaped by constitutions, laws, standing orders, etc. In earlier studies, scholars had concentrated their attention on formal rules: 'Constitutions and formal organisations of government were examined, while informal relationships between political actors remained more or less unstudied' (Hague et al, 1993: 31). This was also true for collective actors. However, political scientists then focused more on the impact of the rules. Institutions were then regarded as structural incentives for political action. In using the various opportunities presented to them, actors interpret their environmental macro and micro situation. These influence the behaviour of actors, while bringing about 'shared concepts used by humans in repetitive situations' (Ostrom, 1999: 37). This means that, for political scientists, the impact of institutions is more important than written rules.

Owing to the fact that in most established democracies the basic institutional arrangements do not change fundamentally over time, there is considerable knowledge of democratic institutions and their performance. Political systems are frequently analysed by the way in which the division of power is institutionalised. For Western democracies, two ideal types have become standard: the parliamentary (with a division of power between the parliamentary majority and the opposition) and the presidential (with a separation of power between president and parliament). Presidential systems have two counteracting centres of political legitimacy stemming from the people's vote. The president is the effective head of the government as well as the head of the nation. He is formally independent of congress, while the prime minister in the parliamentary system is dependent on the support of the majority in parliament. As all established liberal Western democracies have adopted the parliamentary form of government, except the United States, this country is often seen as unique. Much research has been carried out comparing the different states within the United States.

In parliamentary systems, there are factors of minor importance (see Siaroff, 2003; Strøm, 2000: 262–66). Not all of those highlighted in the articles of Strøm (2000) and Siaroff (2003) are important for the selection of cases, since they are relevant only to focused investigations, for example the procedure of appointment of the government (nomination by the head of state or election by parliament). A more important factor would appear to be the relationship of cabinet government or individual ministerial responsibility with the majority in parliament. The impact of the method of appointment of the head of state (hereditary, or election by the people or by a special committee) is of less importance than the relationship between the head of the state and the government. Cases such as France, where the head of state dominates the political process, are often seen as hybrids of presidential and parliamentary systems. Nevertheless, the French President remains dependent on the majority in parliament.

This typology (parliamentary versus presidential system) is less useful at the local level, as (especially under federalism) differences in the institutional arrangements of cities and towns are common and combinations of elements of parliamentarism and presidentialism occur. The elements vary from country to country as well as between some states (the Länder of Germany) and towns (in the USA). The smaller the town, the more important the informal networks among the actors. Therefore, the impact of the written institutional arrangements is of less importance in small towns for the selection of cases. In small towns, political parties play a minor role in decision making, even in European countries.

At all levels of the political system, the party system and the structure of opposition are of major importance to the system's functioning, especially in relation to stable governments and governmental turnover. Electoral systems (different sorts of majority voting systems versus proportional representation systems) are widely regarded as the most important explanation for centripetal or centrifugal parliaments (for presidential systems worldwide see Riggs, 1994: 72ff.). This thesis, that electoral systems matter, has been recently verified in Italy and New Zealand. There seems to be only one other factor to counteract this common picture: regional strongholds of parties based on specific socio-economic structures, for example ethnic minorities, historical entities (and often with different languages). As a consequence, we have to distinguish between political systems based on chance majorities and those with safe working majorities. The index of Laakso and Taagepera (1979), as applied to parliamentary seats helps to separate the cases from each other.

Pluralists have emphasised the unequal aggregation and articulation of interests as a result of the freedom to organise special interests. As noted above, socio-economic structures bring about differences in the strongholds of parties and the strength of interest groups. With regard to the major interest groups, their embedding in special decision-making structures is of importance. This may range from informal participation in decision-making to corporatist or consociational structures. Although different usages of the term corporatism abound (Siaroff, 1999: 176ff.), there is a consensus that it constitutes a systematic management of significant national policies (e.g. economy, health) by cooperation and coordination of public authorities and major interest groups in specific bodies to the benefit of all actors. The contrasting way to influence decision-making is through pressure.⁴ These attempts to take part in decision-making seem to be more common in some countries than in others. Scholars have therefore tried to list Western democracies on a continuum, from those characterised by fragmented pluralism (Australia, Canada, Great Britain, US, France and Italy), to integrated pluralism (Belgium, Germany and Finland), to corporatism (Netherlands, Norway, Austria, Sweden) and consociationalism (Switzerland). The latter countries are those that prefer consensual decision-making, with characteristics such as an oversized coalition government, formal and informal decision-making in special bodies and proportionality in political representation, while most of the fragmented and those with integrated pluralist structures are representative of majoritarian democracies. As Siaroff (1999: 183, 185) argues, there are many different rankings available, and the most controversial issue is how to classify Switzerland, Japan and France. However, there is a consensus in the scientific community that Switzerland has more consensual behaviour than Japan. In any case, classifications have to take into account changes concerning the intensity of corporatism over time. Corporatism as a concept has been under attack in traditional fields and 'it reappeared under a different guise in new policy fields' (Waarden, 2002: 59). Rankings often depend on informed estimates. For most countries, the findings of Lijphart and Crepaz (1991: 235–46) resemble the above picture.⁵

For decades, the concept of political culture has been on the agenda of political science. Political culture reflects the prevailing political beliefs and values of the people. The focus is behaviour towards the political system, arising from beliefs and values. The important factors are:

• the historical ownership of one's estate and the law of succession;
• natural resources and social and economic development;
• the current social and economic structure of a region;
• the value system embedded in religious institutions, with either strong ties of the people to these or a high degree of secularisation; and
• strong social democratic parties, with preferences for a strong state to promote social equality.

Political culture shapes trust in, and satisfaction with, the political system. It also influences decision-making, factionalism or cohesion of parties in government and opposition, and whether a system is more competitive/majority-oriented or characterised by a more cooperative/consensual style. In order to understand the role of political actors, empirical results generated by a multitude of different qualitative findings and surveys on different aspects of political culture have to be taken into account.

'Electoral systems...are widely regarded as the most important explanation for centripetal or centrifugal parliaments'.

Research inevitably has to extend far beyond institutional arrangements.

This is also true with regard to regional differences that complicate decision-making in political systems, especially if they are not traditionally centralised via institutional arrangements. The theoretical model of the political system (Easton, 1953) has highlighted interrelations with the environment as well as the interactions of subsystems within it. Scholars have realised that variations in institutional arrangements mean that subsystems are more or less powerful according to their functions and resources (e.g. tax income). This has entailed analysis of internal and external communications with different sub-systems and (collective) actors on the outside, the formulation of demands and the provision of support. The more functions subsystems are entitled to perform, the more important they become. This makes decision-making and governance in the system as a whole more complicated to analyse. Certainly, for comparing political systems, we must consider the factor of federalism versus centralisation.⁶

Finally, research in policy fields has adopted Easton's model in order to look more closely at decision-making procedures, introducing the policy cycle as a new model of interpretation of political processes.⁷ Starting with Lowi's (1972) categories, which differentiated between distributive, re-distributive and regulative policies and the model emphasising different steps in decision-making, a considerable amount of research material has been produced. Analyses of decision-making include individual networks and arenas and the involvement of collective and single actors in particular decisions. Furthermore, policy analysis brought in the administration as an important actor. This produced findings concerning typical structures in decision-making, including the role of the administration in different policy fields. Policy strategies and policy outputs in numerous countries have been compared. National differences as well as the effects following implementation of policies have been highlighted. The findings have formed the basis for different typologies.

In order to explore the performance of political systems in specific policy fields, the research design has to be linked to the variables highlighted by institutionalists and pluralists/corporatists as well as those identified by research into political culture. In the meantime, behaviourists, who started with considerations based on rational decision-making of actors and conducting research in specific policy fields, have recognised that behaviour follows an individual rationality that is termed 'bounded rationality'. Individual resources (like origin, gender, education and milieu) shape people's rationality according to values and information stemming from the environment. Furthermore, behaviourists have identified institutional arrangements structuring participation in formal or informal groups and organisations as well as in elections and decision-making (e.g. Immergut, 1992). In this way, they have become the vanguard of the 'new institutionalism'. The different actors in the decision-making process are either supporters or veto players (Tsebelis, 1999). Which role they play depends on the institutional arrangements, as well as the actors' intellectual or persuasive power and their political conviction or affiliation. The number of veto players is important in determining a change in policies. In short, analysis of the activities of the major individual, collective and corporate actors in any institutional arrangements has promised to provide new insights into the political process. The policy style of the major actors has become a new focus.

CONCLUSION

This article has highlighted the importance of different approaches to the creation of a comparative design for case studies, and one that will produce results that do not apply just to the analysed cases but to others that are similar according to the identified variables. The high degree of overlap in the results from different approaches tends to confirm that the variables identified are of major importance to the functioning of the political system and its performance in specific policy fields. At present, the major variables, to which particular attention should be paid, are:

• institutional arrangements (presidential, parliamentary) including electoral systems;
• interest aggregation, articulation and participation (pressure or corporatism), including the collective actors involved and other powerful (especially economic) actors;
• centralisation and decentralisation of political systems;
• political culture; and
• outstanding individual actors in the political arena, for example populists, long-standing and successful office-holders.

Decision-making structures in political systems are shaped by various factors: historical development; socio-economic bases; the values of actors and the people; and the functioning of institutional arrangements. People acting within these arrangements, either as individuals or collective actors, are shaped by these arrangements and they can shape them. Formal structures are more important in large systems than in small (e.g. small towns). In spite of a wide consensus on the relevance of these variables, researchers need to reveal on what grounds the classification of cases is made. There are frequent attempts to establish new approaches in political science, and these have to be taken into consideration concerning results and selection of cases. A close connection is needed between theoretical efforts, former empirical findings and focused investigations. Furthermore, the 'key decisions that structure the analysis must be made explicit by the investigator and are open to evaluation by the audiences for research' (Ragin, 2000: 316).

Usually, researchers prefer most similar cases or a most similar system design. This means that some variables that are similar may be ignored, while the number of variables that are different can be controlled more easily. Consequently, Sartori's recommendation is: 'choose entities that are similar, if possible, in all variables, with the exception of the phenomenon to be investigated' (Sartori, 1994: 22). It is true that comparisons of contrasting systems are possible too. Yet, comparisons of political systems from different worlds do not seem to be particularly useful, as everything in the Third World is different from that in the First World, starting, of course, with political culture (Huntington, 1993) and socio-economic development. However, proceeding with the research may produce knowledge about differences and similarities, thus enabling the researcher to find categories that provide a more precise understanding of the cases and an ability to refine them.

After selecting case studies on the basis of this theoretical background, researchers then need to adopt appropriate methods from the arsenal of quantitative and qualitative methods available, in order to work on research questions and try to verify or falsify hypotheses: 'There are not enough countries in the world to allow theories to be tested precisely' (Hague et al, 1993: 30). Comparisons are far from being perfect from the perspective of researchers using mass data for quantitative analyses. Yet, at least a degree of consensus on the impact of an increasing number of variables can be achieved, so that our overall knowledge of political systems can progress. The different research strategies, case studies and analysis of mass data should reinforce rather than oppose each other, and the method or approach chosen should depend on the research questions to be answered or the hypotheses that have to be verified or falsified.

Notes

¹ See the critical statement on this by Blomquist (1999: 223).

² Lieberson (1992: 112, 114, 116), from the perspective of a sociologist, is doubtful about the possibility of complying with this.

³ The strategy does not start with definitions, as proposed by Collier and Levitsky (1997), but with empirical findings.

⁴ The different means of interest intermediation (corporatism versus pressure) was an innovation in the work of Schmitter (1981).

⁵ See also Armingeon (2002).

⁶ An index was developed by Lijphart (1999: 185ff.).

⁷ Because this model was quickly criticised, the more complex model of the 'garbage can', introduced by Cohen et al (1972), was adopted. 'It views the policy process as composed of three streams of actors and processes: a problem stream ..., a policy stream involving the proponents of solutions to policy problems, and the politics stream...'. These streams normally 'operate independently of each other, except when "a window of opportunity" permits policy entrepreneurs to couple the various streams' (Sabatier, 1999: 9). However, in the eyes of many scholars this model seems to be too complex for empirical research.

References

1. Armingeon, K. (2002) 'Interest Intermediation: The Cases of Consociational Democracy and Corporatism', in H. Keman (ed.) Comparative Democratic Politics. A Guide to Contemporary Theory and Research, London: Sage, pp. 143–165.
2. Blomquist, W. (1999) 'The Policy Process and Large-N Comparative Studies', in P. A. Sabatier (ed.) Theories of the Policy Process, Boulder, CO: Westview Press, pp. 201–232.
3. Cohen, M.D., March, J.G. and Olson, J.P. (1972) 'A Garbage Can Model of Organisational Choice', Administrative Science Quarterly 17(1): 1–25. | Article | ISI |
4. Collier, D. and Levitsky, S. (1997) 'Democracy with adjectives. Conceptual innovation in comparative research', World Politics 49(April): 430–451. | Article |
5. Collier, D., Mahoney, J. and Seawright, J. (2004) 'Claiming to Much: Warnings about Selection Bias', in H. E. Brady and D. Collier (eds.) Rethinking Social Inquiry. Diverse Tools, Shared Standards, Oxford: Rowman and Littlefield, pp. 85–102.
6. Dogan, M. (1994) 'Use and Misuse of Statistics in Comparative Research. Limits to Quantification in Comparative Politics: The Gap between Substance and Method', in M. Dogan and A. Kazancigil (eds.) Comparing Nations. Concepts, Strategies, Substance, Oxford, UK and Cambridge, USA: Blackwell, pp. 35–71.
7. Dogan, M. and Kazancigil, A. (1994) 'Introduction. Strategies in Comparative Research', in M. Dogan and A. Kazancigil (eds.) Comparing Nations. Concepts, Strategies, Substance, Oxford, UK and Cambridge, USA: Blackwell, pp. 1–13.
8. Easton, D. (1953) The Political System, New York: Knopf.
9. Gerring, J. (2004) 'What is a case study and what is it good for?', American Political Science Review 98(2): 341–354. | Article |
10. Hague, R., Harrop, M. and Breslin, S. (1993) Comparative Government and Politics. An Introduction, 3rd edn, Houndmills and London: Macmillan Press.
11. Hofferbert, R.I. (1990) The Reach and Grasp of Policy Analysis: Comparative Views of the Craft, Tuscalloosa: University of Alabama Press.
12. Huntington, S. (1993) 'The clash of civilisations', Foreign Affairs 72(1): 22–49.
13. Immergut, E. (1992) Health Politics: Interests and Institutions in Western Europe, Cambridge New York: Cambridge University Press.
14. Katz, R.S. and Mair, P. (eds.) (1992) Party Organisations. A Data Handbook, London: Sage.
15. Katz, R.S. and Mair, P. (eds.) (1994) How Parties Organise. Change and Adaptation in Party Organisations in Western Democracies, London: Sage.
16. Katzenstein, P.J. (1984) Corporatism and Change: Austria, Switzerland and the Politics of Industry, Ithaca, NY: Cornell University Press.
17. Katzenstein, P.J. (1985) Small States in World Markets: Industrial Policy in Europe, Ithaca, NY: Cornell University Press.
18. King, G., Keohane, R.O. and Verba, S. (1994) Designing Social Inquiry. Scientific Interference in Qualitative Research, Princeton, NJ: Princeton University Press.
19. Klingemann, H.-D., Hofferbert, R.I. and Budge, I. (1994) Parties, Policies and Democracy, Boulder: Westview Press.
20. Klingemann, H.-D. and Hofferbert, R.I. (2000) 'The Capacity of New Party Systems to Channel Discontent', in H.-D. Klingemann and F. Neidhardt (eds.) Zur Zukunft der Demokratie (Future of Democracy), Berlin: Sigma, pp. 411–438.
21. Laakso, M. and Taagepera, R. (1979) 'Effective number of parties: a measure with application to West Europe', Comparative Political Studies 12(1): 3–27. | ISI |
22. Lieberson, S. (1992) 'Small N's and Big Conclusions: an Examination of the Reasoning in Comparative Studies Based on a Small Number of Cases', in C. C. Ragin and H. S. Becker (eds.) What is a Case? Exploring the Foundations of Social Inquiry, Cambridge: Cambridge University Press, pp. 105–118.
23. Lijphart, A. (1971) 'Comparative politics and the comparative method', American Political Science Review 65(September): 682–693. | Article |
24. Lijphart, A. (1975) 'The comparable-cases strategy in comparative research', Comparative Political Studies 8(2): 158–177.
25. Lijphart, A. (1999) Patterns of Democracy. Goverment Forms and Performance in Thirty-Six Countries, New Haven: Yale University Press.
26. Lijphart, A. and Crepaz, M.M.L. (1991) 'Corporatism and consensus democracy in eighteen countries: conceptual and empirical linkages', British Journal of Political Science 21: 235–246.
27. Lowi, T.J. (1972) 'Four systems of policy, politics and choice', Public Administration Review 33: 298–310. | Article |
28. Norris, P. and Inglehart, R. (2004) Sacred and Secular: Religion and Politics Worldwide, Cambridge: Cambridge University Press.
29. Ostrom, E. (1999) 'Institutional Rational Choice: An Assessment of the Institutional Analysis and Development of Framework', in P. A. Sabatier (ed.) Theories of the Policy Process, Boulder, CO: Westview Press, pp. 35–71.
30. Peters, B.G. (1998) Comparative Politics. Theory and Methods, New York, NY: New York University Press.
31. Ragin, C.C. (1987) The Comparative Method: Moving beyond Qualitative and Quantitative Strategies, Berkeley, CA: University of California Press.
32. Ragin, C.C. (2000) Fuzzy-Set Social Science, Chicago: University of Chicago Press.
33. Ragin, C.C. (2004) 'Turning the Tables: How Case-Oriented Research Challenges Variable-oriented Research', in H. E. Brady and D. Collier (eds.) Rethinking Social Inquiry. Diverse Tools, Shared Standards, Oxford: Rowman and Littlefield, pp. 123–138.
34. Riggs, F.W. (1994) 'Conceptural Homogenisation of a Heterogeneous Field', in M. Dogan and A. Kazancigil (eds.) Comparing Nations. Concepts, Strategies, Substance, Oxford, UK and Cambridge, USA: Blackwell, pp. 72–152.
35. Sabatier, P.S. (1999) 'The Need for Better Theories', in P. A. Sabatier (ed.) Theories of the Policy Process, Boulder, CO: Westview Press, pp. 3–17.
36. Sartori, G. (1994) 'Compare Why and How. Comparing, Miscomparing and the Comparative Method', in M. Dogan and A. Kazancigil (eds.) Comparing Nations. Concepts, Strategies, Substance, Oxford, UK and Cambridge, USA: Blackwell, pp. 14–34.
37. Sassen, S. (2000) Cities in the World Economy, 2nd edn, Thousand Oaks, CA: Pine Forge Press.
38. Sassen, S. (2001) The Global City, New York, London, Tokyo, 2nd edn, Princeton, NJ: University Press.
39. Schmitter, P.C. (1981) 'Interest Intermediation and Regime Governmentability in Western Europe and North America', in S. Berger (ed.) Organizing Interests in Western Europe, Cambridge/New York: Cambridge University Press, pp. 287–330.
40. Siaroff, A. (1999) 'Corporatism in 24 industrial democracies: meaning and measurement', European Journal of Political Research 36: 175–205.
41. Siaroff, A. (2003) 'Varieties of parliamentarism in the advanced industrial democracies', International Political Science Review 24(4): 445–464. | Article |
42. Strøm, K. (2000) 'Delegation and accountability in parliamentary democracies', European Journal of Political Research 37: 261–289.
43. Tsebelis, G. (1999) 'Veto players and law production in parliamentary democracies: an empirical analysis', American Political Science Review 93(3): 591–605. | Article |
44. Walton, J. (1992) 'Making the Theoretical Case', in C. C. Ragin and H. S. Becker (eds.) What is a Case? Exploring the Foundations of Social Inquiry, Cambridge: Cambridge University Press, pp. 121–137.
45. Waarden, F. van (2002) 'Dutch consociationalism and corporatism: a case of institutional persistence', Acta Politica Special, International Journal of Political Science 37(1–2): 44–67.
46. Wieviorka, M. (1992) 'Case Studies: History and Sociology?', in C. C. Ragin and H. S. Becker (eds.) What is a Case? Exploring the Foundations of Social Inquiry, Cambridge: Cambridge University Press, pp. 117–159.

About the Author

Hiltrud Nassmacher has carried out empirical studies on decision-making and changes of governance structures as well as on participation in specific policy fields in a comparative perspective, at different levels of political systems. In addition to publishing many articles on these themes, two books have drawn the findings together in a systematic way. She has worked at the universities of Konstanz, Göttingen, Trier, Münster, Siegen and Oldenburg.