Glossing inadequacies: problems with definitions of key concepts in some methodology texts

Abstract

It is argued in this paper that a significant number of definitions of terms provided in currently available methodological texts are inadequate or positively misleading. Some do not cover the main meaning of the term but focus instead on a non-standard one. Others fail to pick out the distinctive features of what is being defined. There are also those that fail to indicate relationships to related terms or those with similar meanings. Finally, there are some definitions that are simply incorrect in significant respects. It is argued that defective definitions are worse than useless and are a disservice to students.

Keywords:

• Definitions of key methodological terms
• semantic failings of methodology texts

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Clearly what I have outlined here is a longer and more complex definition of ‘symbolic interactionism’. But while simplification is often necessary in defining terms, so that some features may not be mentioned and none can be fully explicated, a definition is useless unless it identifies some of the distinctive features of the phenomenon the term refers to.

2. In fact, ‘triangulation’ was first introduced in the quantitative methodology literature, though today it tends to be most used by qualitative researchers.

3. And this is by no means always picked up in definitions of ‘logical positivism’, as for example in the following formulation: ‘a view that logic is crucial in understanding cause-and-effect relationships’. Furthermore, given that many logical positivists rejected the notion of cause as metaphysical, this definition fails comprehensively.

4. A similar problem arises with this definition: ‘Cause and effect (causal, causation) refers to the process of establishing a causal link between a “treatment” and a “research outcome”’. Moreover, these terms do not refer to a ‘process of establishing’ but to a process in the world. We should also note that this definition is circular: the term ‘causal’ is used to define ‘cause’. This is another occasional fault in definitions offered in methodology texts and glossaries. For example, in one glossary we have ‘randomization’ defined as ‘a process of assigning participants into groups randomly’. Of course, circularity can be a problem to varying degrees, a circular definition may still convey some information, but in this case ‘groups’ is probably too vague: mention of treatment and control groups, or of groups receiving different treatments in an experiment, would be necessary to give the definition any informational value.

5. It is also true that Popper rejected the idea that statements can only have meaning if they can be subjected to direct empirical or logical test.

6. The entry on the correspondence theory of truth referred to in the definition of ‘realism’ quoted claims that this: ‘assumes that truth corresponds to a fact or facts, i.e. it is established in relation to an observable reality in the world’. There are difficulties, hard to avoid, concerning how ‘fact’ can be defined without employing the concept of truth, thereby rendering this definition circular. Furthermore, the term ‘corresponds’ appears in the definition, so that the definition is circular in this respect too. Even more problematic is the requirement that reality be observable. Commitment to the correspondence theory of truth does not entail this, as is made clear by the case of Aquinas and others, see Prior, (1967).

7. Defining ‘ethnomethodology’ is a contentious matter because there are significant disagreements among those who are affiliated to this approach; and the same applies in some other cases. Thus, the outline definition I provide here might well be criticised by some as inaccurate or misleading in some respect. But I doubt that it could be charged with failings on anything like the same scale as the definitions I have quoted in this section. Inaccuracy is a matter of degree, and can be a contested issue, and we can all be guilty of errors; but this does not alter the fact that there are deficiencies that are below any threshold of adequacy, and that these are more frequent than they should be in texts on research methodology.