The art of systematic analysis: the development of political science
We can identify within the study of politics a trend away from the term political studies in favour of the label political science. The first American university chair in political science was established in 1858 at Columbia (Dahl, 1961, p. 764), and since then we have seen a progression and growth of different approaches to political research. What does the term political science, as opposed to political studies, mean exactly? Marsh and Stoker (1995, p. 7) say that: 'by science we mean the organised production of knowledge that demands of its practitioners certain intellectual disciplines, in particular, logical coherence and adequate evidence.'
While the real world of politics and politicians may seem chaotic, as political researchers we cannot afford to ignore the rules of science. This will be elaborated upon in more detail in the next chapter when we examine terminology such as reliability, validity and causality. If we adopt Marsh and Stoker's approach we are claiming that the study of politics involves more than merely accounting for singular findings and events. Rather, we are employing a recognisable process of analysis which includes logic and maybe even prediction.
The term political science is often (although not exclusively) associated with the growth of the behavioural approach. The behavioural approach became a prominent focus of political research in the early part of the twentieth century, and its establishment is associated with the University of Chicago in the United States during the 1920s and 1930s (although the 'Chicago School', as it is commonly known, was heavily influenced by European political scientists). However, the development of the behavioural approach was not just driven by ideas, but also by methodological tools - particularly the availability of surveys for studying political choices and attitudes (Dahl, 1961, p. 765). The label political science is important from a funding level. For example, the Economic and Social Research Council (ESRC) is one of the largest independent funding agencies involved in funding and promoting 'social science research'. Created in 1965, the ESRC has a current annual budget of over £60 million. Without such funding (from the ESRC and many other organisations) we would certainly be severely restricted in the breadth and depth of our political knowledge.
Much of the research we can carry out only begins to have any value when we can use it to support a particular theory. Research data are merely a device we employ to illustrate and reinforce theoretical foundations. Do not be misled into believing that lots of data can be a sufficient substitute for appropriate explanation. Research data should be illustrative, not and end in themselves. We are scientists in as much as we are trying to 'make sense' of a political world which is defined by human thoughts and actions.
In some ways politics is very different to other social science subjects, and in others there is a great deal of similarity. The obvious difference is that we are concerned with political behaviour and institutions, rather than purely historical, sociological or psychological concepts. Yet despite this, these subjects often share the same basic research principles. This is partly because the distinction between the subjects is not always clear cut (one only has to look at a range of book titles - Dowse and Hughes's Political Sociology, Childs's Britain since 1945: A Political History,
Samuels's The Political Psyche - to see that this is the case). Many existing text books focus upon general social science research methods. Indeed, the principles relevant to carrying out an appropriate survey or interview will be broadly similar for any social science. There are some important differences between the subjects though. 'Observation' is a technique common to sociology and anthropology, but it is employed relatively rarely in political research. This is not to say that it cannot be used - it is just not feasible in many cases. For example, we may find it useful to 'observe' Cabinet meetings, but are unlikely to find access easy!
Within political science, we rarely take the psychological approach, which employs experiments to test hypotheses. Any student of psychology will invariably be introduced to the principles of behavioural conditioning in which animals have been trained to respond to particular stimuli (as exemplified by Skinner's rats and Pavlov's dogs). However: 'Social scientists cannot create laboratory conditions and then intervene in the experimental process to see what changes on existing relationships are induced as a result' (Broughton, 1995, p. 26). In the human world, people receive their political cues from a whole range of sources (such as the family, the workplace as well as the media) and it would be unethical for us to control these environments to any great extent. Hence, the main reason for avoiding experimental techniques in political research are similar to the justifications we put forward in relation to observation: it is very difficult to create a totally experimental environment in politics (although not altogether impossible). When might the experimental approach prove to be useful?
In order to assess whether the media can influence voting choices, an experimental study of television news was carried out during the 1997 general election campaign (a more substantial account of this study is provided in Chapter 6). The study tested four hypotheses which were based upon existing theories of media-agenda setting (Norris et al., 1999). Respondents were asked to complete two questionnaires, in between which they were shown a thirty-minute selection of video news, in order to assess whether the news exposure influenced responses. Yet the authors behind this study recognise that while 'experiments provide a more satisfactory way to examine the short-term impact of media messages… In contrast, many different factors other than the news media could influence public concerns' (ibid. p. 129). It is simply not possible to create a totally experimental situation in politics, living in the safe hope that, were our scenario to occur in the real political world, our political actors would behave in exactly the same way.
Whilst this text will focus upon the main methods used in empirical political research, examples of less conventional methodologies will also be identified and examined. We have stated that a scientific approach to the study of politics is an attempt to produce a 'logically coherent' and empirically supported knowledge, yet in the drive to develop explanations and theories we have to recognise that much of what we research rotates around attitudes, beliefs and interpretations. We can state the turnout in an election with a high degree of certainty, but we are unlikely to be as confident about the precise reasons why some people did not vote. In this sense, the political research method(s) we adopt is often central to our overall findings. It is easy enough to read a book or article and say, 'but the question you really should have asked is…', or 'why did you look at institution x and not institution y…', or 'it's all right asking people who did vote what they think about New Labour, but what
about those who didn't vote…'. Any published research worth its salt should have a clear explanation of methods, and I suggest you read this section of any research carefully, and use it as a reference point when reading the rest of the work. As such, there may be some disagreement as to whether an 'ideal' research method exists. Rather, there are methods which are more appropriate in different circumstances, enabling us to provide more accurate explanations and theories, and this text will predominantly focus upon these. There are several questions we can ask about all political research: