23 Generalisation (3)
Speaking only for myself ...
We've been talking about generalisability. But what is it?
In experimental and quasi-experimental research I think it is relatively easy to define. Suppose the research identifies a relationship between two variables. Then generalisability is the extent to which that relationship applies to the same variables in situations. In other words, the relationship can now be regarded as approaching universality.
But think a little more closely than that. How often, in fact, will the studied variables be the only ones operating in a given situation? If the answer is yes, then it seems to me that generalisability has real and useful meaning.
Otherwise, I have my doubts.
Now consider the situation which applies to action research and similar approaches. Very often it's hard to know what is true about the situation until it is studied.
For example, I may have learned that in a given situation, certain actions were followed by certain outcomes. As situations differ it would be rash of me to assume that the same actions will always be followed by the same outcomes. I assume this is why it is often said that you can't generalise from action research.
But consider the two situations. Let's say I've conducted an experiment in which I have demonstrated that a certain treatment resulted consistently in a given change in some variable. And let's say I've carried out an action research study in which certain actions were followed by certain outcomes.
The key question ... What happens when I now take those two sets of findings into a different situation? Can I have faith in either of them?
It seems to me that if it is a typically complex field situation involving people, the answer is a qualified "no".
In short, the generalisability which experiments provide does have some universality. But only when the researched variables are the only variables operating. In the field settings where action research is typically conducted, that is seldom likely to be true.
Copyright Bob Dick 1999-2000. May be copied if it is not included in any
material sold at a profit, and this and the following notice are shown
This may be cited as: Bob Dick (1999) Generalisation (3). Occasional
pieces in action research methodology, # 23. Available online at
Version 1.2; last modified 20000101
These "occasional pieces" form part of a substantial action research site at
Southern Cross University. To access the "front page" of the site, click here