22     Generalisation (2)



Speaking only for myself ...


A detour.  I was about to address the commonly-held fiction that action research usually does not permit generalisation.  In the meantime some interesting communications from Vic Wooddell and DP Dash have suggested that there are some other issues to address first.

In addressing the issue of generalisation in its various forms, and its relationship to action research, I think it is useful to differentiate two forms of action research.  I might characterise them as "action research" and "action research".  I think they have different requirements for generalisation.

By "action research" I mean the form where action is obligatory, or at least the primary reason for conducting the activity.  The research is desirable but can be jettisoned if necessary.  Programs for community or organisational change are often like this.  A client or client group is paying for certain results and expects to get them.

In situations like this the researcher will probably achieve some research outcomes.  At the least, it is likely to take the form of an improved understanding of what was done.  Beyond that, the client may or may not develop better understanding.  The client may or may not have better understanding as an explicit goal.

Then there is "action research".  The research outcomes may well be crucial -- for thesis purposes, for example.  It may be good to achieve some action outcomes too.  That may even be an explicit goal.  But if necessary the action can be abandoned.

It seems to me that in action research, generalisability may not be a great concern.  There is likely to be some, at least for the researcher.  Without generalisation each situation is so different that no learning can carry from one situation to another.  However, the main emphasis is likely to be on learning how to deal with specific situations.

Now consider action research.  Here, unless there is a contribution to knowledge, there is no research outcomes.  If it is a thesis study the examiners may decide there is no thesis.  Generalisability is a requirement.



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 (2).  Occasional
pieces in action research methodology, # 22.  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