This guest post is by Rod Pitcher. Rod is a PhD candidate in Education at The Centre for Higher Education, Learning and Teaching at the Australian National University, Canberra, Australia. The focus of his study is the metaphors that researchers use when describing their research. Rod can be contacted at Rod.Pitcher@anu.edu.au
Deemter (Not Exactly, 2010) shows that vagueness is inherent in any research. This is partly due to our use of language in describing the results, but also because of the nature of our research itself and the person doing the research.
Many words in English have rather vague meanings. For instance, what does ‘tall’ mean? Does it mean bigger than a certain height, say 2 metres? That would mean that all buildings are ‘tall’. But if we choose a measurement for ‘tall’ that suits a building, such as 100 meters, then we wouldn’t be able to call any person ‘tall’. The terms are vague unless we specify the value that ‘tall’ refers to. Usually we don’t specify a measurement because ‘tall’ gets its meaning from the context of the situation in which it is used. But if we see the word ‘tall’ written down, with no other information, it is so vague as to be meaningless.
In reports of research one comes across words like ‘sometimes’, ‘in this case’, at present’, and so on, which actually specify that they only apply to the research described. Their application to other areas becomes problematic.
We go even further, and narrowly specify the group of people we interview, the questions we ask, and our method of interpretation. We do this because if we change any of those things the results will most likely be different. In other words, our results are narrowly applicable to one real life situation and have a large degree of vagueness when applied to other situations.
Not even the ‘hard’ sciences are completely free of vagueness. For instance, if we want to measure the time for a certain reaction to take place we need a measuring device. That, in itself, will have built in tolerances that affect its accuracy. If the time measured by a stop watch is 20 seconds, how accurate is that? Could it be 19.99 seconds or 20.01 seconds? The results must always have a degree of tolerance and thus vagueness.
Then there’s the subjective response of the researcher. No-one can possibly be completely objective, so the observations will be subject to the vagaries of the act of observation itself. Human nature being what it is, the vagueness can be quite large, not only due to the researchers physical characteristics such as hearing and vision, but also will depend on his or her psychological characteristics.
What can we do to reduce vagueness in our results? According to Deemter, very little. It is part of life. Unless we can demonstrate that our procedures are 100% accurate and that they apply in all situations then we cannot avoid vagueness. Since that is physically and psychologically impossible we are stuck with a certain degree of vagueness. The only thing we can do is to reduce it as much as possible, by explaining the results accurately, specifying the method and situation, and adding a warning that it only applies in that situation, only when that method is applied to the specific data. Only if the research is duplicated with extreme accuracy will the results be the same. Then we run into the problem of how accurately the research could be duplicated. Again, the problem of tolerances in the procedure arises.
The answer is that we can never entirely remove a degree of vagueness, we just have to live with it. The best we can do is reduce it as much as possible by being careful with our language and methods.
van Deemter, K. (2010). Not Exactly: In Praise of Vagueness. USA: Oxford University Press.