Classifying Methods

How many scholarly methods are there? There is no obvious way to answer this question deductively by identifying some logical typology of ‘types of method.’ It is possible, however, to inductively identify some dozen methods used in the scholarly enterprise:

· Experiments (including natural or quasi-experiments)

· Surveys

· Interviews

· Mathematical models (and simulations), including game theoretic models

· Statistical analysis (often, but far from always, associated with models)

-including secondary [that is, collected by others] data analysis

· Ethnographic/observational analysis (‘Participant Observation,’ in which the investigator interacts with those under observation, is most common, but discreet observation is also possible).

· Experience/ intuition [some would treat this as an important subset of observational analysis, since we are in effect ‘observing’ ourselves here]

· Textual (content, discourse) analysis

· Classification (including evolutionary analysis)

· Mapmaking (both representational and conceptual)

· Hermeneutics/ semiotics (the study of symbols and their meaning)

· Physical traces (as in archaeology)

· Some would treat ‘evaluation’ of programs as distinct, though it can be seen as a combination of some of the above methods. Similar arguments can be made with respect to demography, case study, feminism, and perhaps also hermeneutics.

Having inductively identified these dozen methods, it is then possible to ask the Who, What, Where, When, and Why questions of these {These questions were outlined in Classifying Theories). This allows one to see if there are methods well suited to the investigation of each type of theory. In asking these questions, some subsidiary questions emerge:

• • How many agents can a method investigate?

  • Does the method identify any/all of the criteria for specifying a causal relationship? These are: that the cause and effect are correlated, that the cause occurs before the effect (at least this is the general case), that intermediate causal factors be identified, and that alternative explanations of correlation be excluded.

  • Does the method allow scope for induction (or is it entirely deductive)?

  • Does the method allow movements through time to be tracked?

  • Does the method allow movements through space to be tracked?

The table which follows shows how ten of the above methods answer these five as well as the original five questions. The subsidiary questions are listed immediately after the question that they build upon. Thus the first row indicates which type(s) of agent a particular method can investigate, while the second captures how many agents that method can investigate.

The table which follows establishes that there is no ‘one’ scientific method, but rather a dozen such methods with different strengths and weaknesses. In trying to comprehend the full complexity of human life we can usefully have recourse to each (while remaining cognizant of each method’s limitations).

Typology of Strengths and Limitations of Methods

Since methods are classified here along similar dimensions as Classifying Theories, it is possible to establish empirically that disciplines choose mutually compatible sets of theory and method (as would be expected from our discussion of Defining Disciplinary Perspective. The fact that disciplines choose methods that are generally biased in favor of their theories is a powerful motive for applying multiple methods [See Performing Mixed Methods Research].

Source: Rick Szostak, Classifying Science, 2004, 138-9.