The development of political research methodology: the relationship between political research and theory
How can we be sure that voting behaviour may be linked to economic prosperity or that political interest is related to educational attainment? The political research we carry out helps us to formulate, and if appropriate, revise political theories. Mannheim and Rich (1995, p. 21) claim that: 'theories make facts useful by providing us with a framework for interpreting them and seeing their relationships to one another… [they are] sets of logically related symbols that represent what we think happens in the world.'
To reiterate a point made earlier, research data are means to an end. They are of little use if we are unable to relate them to structured, logical explanations. Therefore, theories help us to relate distinct pieces of research to each other. Our research may not always be intrinsically comparative (see below), but we do make use of theoretical foundations in order to assess whether our research findings are unique or whether they fit into a more regular pattern of findings. There are two basic means by which we emphasise the relationship between research and theory. If we take the deductive approach we develop our claims, or hypotheses, from existing theory, or to put it in simpler terms, 'we might consider a general picture of social life and then research a particular aspect of it to test the strength of our theories' (May, 1997, p. 30). For example, our theory might state that political party membership is positively related to trade union membership, and when we examine party members we would expect to find that they are also members of a trade union. In contrast, we can take the inductive approach in which we build our theory as a result of our empirical findings. For example, we might look at the political affiliations of
members of trade unions and then make claims about whether they are members of any (or a particular) political party; or perhaps we can identify a different variable which influences political affiliation, such as membership of particular types of trade unions. In both cases, however, we are relating our findings to an assumption (or theory) of political behaviour - that is, that membership of employment-related organisations and political identification are related.
As you may have noticed, we tend to talk of research methods rather than a singular method. As political science has grown as an academic subject so has the interest in, and attempt to fine tune, the methods of research employed. For example, Blondel's text Thinking Politicallydevotes a single chapter (just seventeen pages) to the issue of methodology. In contrast, the text Theory and Methods in Political Science(Marsh and Stoker, 1995) concerns itself solely with the relationship between theory and research, and just as political scientists do not always agree over what it is they study, they also disagree about 'how to do it' (ibid., p. 7). As we shall see in later chapters on survey methods and interview techniques, academics are constantly engaged in assessing the process of information collection as much as the actual findings of research itself. We are, therefore, not only seeking the most accurate answers to our political conundrums, but the 'right' methodological approach to achieving this aim. Appropriate research methodology is much more than a mechanical process which can be attained by ticking off a check-list of 'actions to be taken'. Being able to defend our research methodology is as important as being able to defend our findings (in fact, the two go hand in hand). As such: 'By methodology, one does not mean a technique, but an analytical process of defining problems, setting their limits, and finding some means of relating these problems to others already solved' (Blondel, 1976, p. 8). This is something we should bear in mind when reading the research of others too.
Marsh and Stoker (1995) assess the relationship between political research methods and theory in terms of evolving approaches. This began with the normative approach (which is concerned with moral notions of what 'ought to be'), followed by institutional studies (which consider rules, procedures and formal organisations); the behavioural approach (examining political behaviour and attitudes via individual and large-scale studies); rational choice theory (which explores the way in which political decisions are taken); feminism (which attempts to redefine politics from a non-patriarchal perspective); and discourse analysis (where we consider the structuring of social meaning, language and symbols in political debate) (see Marsh and Stoker, 1995, for a greater elaboration of these approaches).
In contrast, Blondel (1976, p. 13) identifies the development of political analysis as a history of the 'three main battlefields', each dominating political analysis at different times. One 'battlefield' has been represented by the distinction between 'normative and descriptive political science' - that is, the study of what ought to be versus the study of what actually occurs.
Second, Blondel identifies the 'battlefield' between 'law and reality, the problem of structures', in which a legal approach is taken towards the study of politics. Examples of this approach would be analyses which focus upon constitutional law, public law and administrative law, when 'the problem rotates around the question of implementation of rules, and the role which the notion of rules plays in political life' (ibid., p. 21). The problem here is that political behaviour is influenced by a range of
structures and procedures (for example, the family or membership of a political party) which lie beyond the remit of legal rules. Blondel's third 'battlefield' is that between 'the unique and the general', in which we have witnessed a move towards the quantification of political analysis and the development of 'behaviourism' (which first developed in the USA). It is this approach which marks the difference between political studiesand political science. Hence, what is clear is that we can take different approaches to the study of political life; and a strong relationship exists between the theoretical beliefs and underpinnings that prompt our research and the methods which we find most 'appropriate'. Rather than consider the historically developed approaches to political research, this text focuses upon 'types' of empirical research methods. The main distinction made here is between the use of quantitative and qualitative forms of data and the analytical frameworks in which such information is employed. So, for example, we shall consider research which is operationalised via case studies and research which is operationalised via comparative analysis. How do these two research designs differ? Rose (1991, p. 446) claims that comparison is 'one of the oldest forms of the study of politics, tracing its roots back to Aristotle', which can be defined by its 'focus and method'. For Rose (ibid., p. 447), comparative analysis includes the following criteria:
Comparative studies begin with what Rose terms a 'matrix' - rather like a table in which we can 'tick off' characteristics as they appear. The beginning of comparative analysis involves two decisions: which countries and which concepts, and 'the more countries examined, the fewer the concepts; the fewer the countries, the more detail, approaching holistic comparison' (Rose, 1991, p. 453). Yet it may be difficult to identifytrue comparative analysis as 'even the most insular types of works frequently contain references to other countries' (Page, 1990, p. 438).
'The Comparative Method' is explored in some detail by Mackie and Marsh (1995). Its major strength is that it avoids 'ethnocentrism' (ethnocentrism occurs when we are insensitive towards the social context of what we are researching, or when we are unaware of historical and cultural differences - that is, we view what we are studying from our own personal and cultural perspective). However, it would probably be more appropriate to refer to comparative methods rather than to a singular method. This is because comparative analysis can take different forms. It can be used to retest earlier studies - a comparison of individual case studies over a period of time. It can also be used to compare across types - for example, parliamentary
procedures in different countries, voting behaviour in different regions, economic development across continents.
The development of comparative analysis has grown rapidly since the 1960s, aided by the popularity of quantitative analysis and the development of data archives, although such research tends to focus upon the industrialised nations (and more recently upon the newly democratised nations of Eastern Europe), and upon studies in the field of international relations (particularly as globalisation has been marked by a rise in the numbers and importance of supranational political organisations). Its popularity and importance as a research approach is highlighted by the fact that the European Consortium of Political Research has been established to promote the comparative study of politics.