This chapter seeks to tell you something more about action research and then how to do it.
It is not tempting at all to try and blind the reader with science. As you might guess a great deal of high falluting stuff has been written about Action Research [AR] yet when it comes to getting started much of this is not required. You should already have read chapters 1 and 3. The first gives you examples of the work of previous students. Chapter 3 gives you some idea about how to do qualitative research. As AR is considered in this publication to be a form of qualitative research, you should already have some notions of what needs to be done. Teachers are already doing a form of AR when they teach a lesson and feel it could be improved or changed. AR is simply about reflecting about your teaching and going through a cycle of change. Sometimes a group of colleagues do AR together.
In general, action researchers are expected to carry out their AR with greater rigour than indicated in the paragraph above. This book has already alerted the reader to the issues such as the ethical considerations which arise when data is collected, how theory is built and so on. This chapter should firm up your initial ideas. Ultimately, however, you will need to do it yourself while glancing through these papers and possibly other publications on AR. In a section below we will go through the steps you will need to take and explain, hopefully, in simple terms what needs to be done.
Rob McBride was with a group of teachers who were discussing steps they might take to improve the experience of 16-19 year olds. One commented that from his office he could, on most days, see a group of lads on an adjoining recreation ground who had left his school. They were unemployed, some were using drugs, some had already committed serious criminal offences. He felt he had failed them. Another teacher cited the economic situation and suggested that the teacher should not blame herself. A third believed that the rise of consumerism and the increasing gap between rich and poor was the cause. Already in this anecdote there is a rudamentary search for analysis and explanation. It follows the pattern outlined in the generalisation from cases in chapter 2 section I.
At this point there is no need to comment further on any of the arguments put forward and it could be accepted that an opportunity to let off steam and have a 'good moan' maybe something we should all do once in a while. Research that was solely about understanding situations might be content to pursue the contextual analyses indicated in the anecdote. Action researchers will also go through such a phase but ultimately they would have to ask themselves
What can I do, given that the situation does not look very promising?
AR begins when professionals begin to ask themselves this and similar questions. As you might imagine this has caused something of a split in those who claim to promote AR.
Carr and Kemmis  define AR as follows:
"Action Research is a form of self-reflective enquiry undertaken by the participants [teachers, students or principals, for example] in social [including educational] situations in order to improve the rationality and justice of [a] their own social or educational practices, [b] their understanding of these practices, and [c] the situations [and institutions] in which these practices are carried out."
Not all action researchers agree with this definition. There is widespread agreement that the participants undertake the research themselves. There is less agreement about the use of the terms 'rationality and justice' and about [c]. It is [c] which opens up the 'big' questions which have a tendency to divert research away from the work place and from the developmental possibilities as it did with the teachers I described above. One solution might be that action researchers should look at the 'big' questions and at their workplace. It has to be said that the central focus in AR is on improving practice.
Professor John Elliott, (at the time of writing) Dean of the School of Education at UEA, and a very influential figure, has defined AR as:
"the study of a social situation with a view to improving the quality of action within it." [Elliott, 1982]
Bridget Somekh [in McBride, 1989] has built upon this definition to derive a more inclusive one. She sees AR as:
"The study of a social situation, involving the participants themselves as researchers, with a view to improving the quality of action within it."
This last definition will inform this volume rather than any other. Definitions and uses of AR have been commented upon by a large number of academics in many fields. While it may appear to be a form of navel gazing, on this course we have to scratch surfaces more than we might normally, just as some academics have done with these definitions. We have to attempt to gain a deeper and more profound understanding of the uses of terms and ideas. It means that to some extent we have to make some terms, that are used in an everyday way, problematical - meanings have to be questioned. The same is true of some forms of authority, which are often exercised unreflectively. It is no coincidence that John Elliott has written widely about the problems of centralised control of education. Indeed, he has maintained that by engaging in AR teachers are taking an anti-bureaucratic position through which they become aware of the greater contextual framework. This is certainly a credible view though it would be surprising if some practitioners had not come to AR after being critical of national policy, i.e. concerned with 'big' questions first.
Action researchers are usually anxious to assert that their work seeks to narrow the gap between theory and practice. That is to say they have no interest in theory which is not grounded in practice nor practice which is not reflected upon and theorised about. In this way theory and practice interact. We can also conclude from Somekh's definition that AR is not carried out by the archetypal scientist in a white coat who is detached from the 'objects' he [as it usually is] is studying. To some extent the researcher is an 'insider' who changes the social situation by virtue of studying it. When conducting an interview, for example, it is more than likely that practitioners will have had thoughts and ideas which otherwise would not have emerged had an interview not taken place. Yet it sometimes becomes clear that an action researcher can become something of an 'outsider' simply by collecting data. In this sense knowledge is power and the knowledge is held by the researcher. It is likely that the action researcher will balance uncomfortably between the insider and outsider positions during the course of her project.
While the research is taking place it is not unusual for an action researcher to have a 'critical friend', i.e. someone who is often an academic and who acts as a sounding board for discussion, comments upon conclusions that might be drawn and/or action about to be taken. Those engaging in AR are expected to collect data more formally than classroom teachers would normally . This implies an understanding of the methodologies of data gathering - another role the critical friend might play. In general, the task of the critical friend is to support the work of the action researcher without taking over.
Figure 1, below, illustrates that AR is not only about research but about action too. For recruits to AR, the very thought of carrying out a formal interview can seem daunting. Taking action can be more frightening still and even act as a bar to trying AR at all. We might call the teacher starting out on AR as an enquiring teacher who does the research part of AR but puts the action part on hold. It has been argued that attempts to 'educate' without paying attention to action is not education at all. Readers will find that it maybe easier to begin acting in a small way in, say, your own classroom. More ambitious action may affect others or call upon others to make decisions about their own action. In this case it is important that these people have some stake and/or say in both the research and the action that follows. More will be said about this below.
We can now consider each of these steps to see what each entails.
For those new to AR this step can be the most difficult. It is likely that you will have to go through a period of reflection about your own practice or maybe even observe a colleague at work.
Some teachers may have a child that they feel they would like to understand better; there maybe an area where you think you are failing; or a part of the curriculum which feels inadequate; a topic about which you would like to know more; a topic which could add something to your teaching and so on.
Other professionals say in business, or the caring professions generally, may be concerned about the quality of decision maing in a particular department or team. There may be concerns about communication structures, particularly in multi-professional agencies, between agencies and between geographically dispersed sites. Doubts may be expressed concerning management structures, the distributions of responsibilities and duties leading to the possibility of role conflict and role strain. There may be suspicions that quality assurance procedures are inadequate or consultative procedures too narrow in focus. And so on.
We all, to some extent, relish the comfort of what we are used to and concluding that part of our work is not as good as it might be can create discomfort. One ploy for getting into posing some problems is to discuss your work with your colleagues and clients. A colleague may be prepared to run a critical eye over activities. Whether you try these steps or some others, you will have to think imaginatively. The first steps towards making changes are the worst and if you concentrate too much on the finer details you will rarely find any incentive to do anything at all. Talk yourself up and be perepared to fly by the seat of your pants a little. This is not to suggest that should be reckless rather that you should override the fears you might have and be prepared to sort out difficulties only if they arise. Many problems do not actually arise and when they do you can soon solve many of them. More often than not, you will find your colleagues remarkably resilient. Another alternative is to do nothing.
Of course to write all of this is a little risky. But in our experience change and improvement only comes about when people take risks. It is also our experience that as the AR process is worked through action researchers do not look back. Rather they develop a confidence about their study and, more generally, adopt a resourceful and flexible approach to their practice. In short they are empowered and independent.
In a sense problem [or interest] posing is, or can be, a first step in collecting dat a but the reader might feel that this step is more of a starting point. It may be added that a supportive colleague or 'buddy' in your place of work can be a great help in data collecting, as in other steps in process.
Let us first clarify what data is. Data can take any form. It can be:
Each of these have their associated strengths and weaknesses. Consequently, researchers do not simply depend upon one kind of data but try to obtain as great a variety as the situation will allow.
There are no hard and fast rules about how to collect data, but there are some guidelines which can be borne in mind. Although 'having an open mind' in the collection of data may seem to be some sort of 'objective' ideal, it is hardly possible. Every researcher is interested, or curious about something - otherwise why research? It is this curiosity through which questions are formed, and patterns are identified. In order to know more about something, one has to observe more closely, and collect more data. So,
We might advise you, then, to 'immerse' yourself in the situation, if you are not already. Once the guiding curiosities and interests are made sufficiently explicit. It becomes easier to know what data to collect and how. You may decide, for example:
The questions that have led you to carry out the research will guide you in becoming selective about what to record. For example, if you want to answer question two, you will exclude from your attention much of the 'irrelevant' descriptive material that would be necessary if you were answering question one. You would exclude anything that did not relate to 'children talking when at work'; that is, you would exclude things like teachers talking to each other, pupils talking when not in work situations, and so on.
What ever data you collect in the pursuit of your research focus will become part of your data base.
However, there is a danger in the collection of data that a researcher needs to recognise. It is simply that the very interests that led you to do the research may mislead you, may bias your findings in ways that are very subtle and hard to recognise.
Qualitative researchers, particularly those influenced by phenomenology, are aware of the need to suspend their own taken for granted values, interests and beliefs. No one can do research with an 'open mind', or clear of prejudice. That is not what is being asked for here. What is being asked for is the recognition that we all operate in everyday life with taken for granted beliefs about the world. What is taken for granted as real, true, holy and common sense in one culture and sub-culture will not be considered to be so in others. Research is largely about uncovering our own prejudices. As such it can be very threatening.
It is important that we be on guard to discriminate between judgements about 'data', interpretations of 'data', summaries of 'data' and 'data'. At a philosophical level it can be argued that all data is a form of judgement. Qualitative sociologists continually investigate the 'social construction' of 'data'. It is difficult to maintain the argument that some things are'facts' and are not socially constructed. For example, it can be said that a 'tree' is a 'fact' of nature. However, the concepts of 'tree' and 'nature' are not quite the same in mythological systems of thought as distinct from the systems of thought constructed by current scientific practices. Each concept takes its meaning within a given context of thought and practice; and these contexts can differ over time and between speakers. In this way the concepts are socially constructed by the speakers and practitioners within the context of their use.
A scientists may give priority to the scientif definitions of tree and nature. However, does this mean to say that the mythological definitions are 'wrong', 'inferior'? Our answers to this question can reveal our prejudices. The qualitative researcher will suspend judgement at this point and will inquire into the different kinds of meanings and practices associated with each context. If a judgement is to be made it will be made at the end of the research not at the point of data collection. This judgement, if it is made, will be accompanied by a rationale in defence of the judgement, based upon the data as an evidence base. For this reason it is very important that prejudices are not built into the data base itself otherwise the data-base will be biased and the findings falisified. The data base must be a fair or unbiased representation of the case itself. Thus the principles and procedures adopted by the researcher to do this must be made clear.
In developing the data base from which to make a representation you will be involved primarily in: observation, interviewing and the collection of documents and artifacts. These three will be briefly outlined in turn.
Observation is more than just looking and seeing. The task of observation is to be able to represent a social scene in a way which is recognisable to the actors involved, is considered valid and a true representation of their action. The problem here is the issue of what is meant by valid and true. It is a truism that different people make different judgements upon what they see. Observation cannot be separated from the different meanings that actors place upon their action, the actions of others, the 'stage' upon which they act and the 'props' they use. Thus, in making a representative description of observable characters, events, 'stages' and 'props', it is important to subject that description in some way to the interpretations of the actors involved. This can be done, for example, by:
By carrying out such 'tests' on a frequent basis, the researcher can ensure that his or her understandings of what is being observed do not differ significantly from those of the actors. However, rather than one interpretation, the researcher will collect many interpretations of a given observation, particularly when what is being observed involves an 'us' and 'them' situation. Thus, observation goes hand in hand with interviewing.
Much of your data will probably be collected by interview. The interview is a record of the other's voice. The voice is something very personal. When I speak, it is my voice and not someone else's, my words, expressing my feelings, my point of view. However, to what extent is this actually true?
How much of what I believe or know to be true is actually something that is second or third hand, bequeathed to me by television, parents or a school teacher? How much of what I speak is actually derived from the voices of others, powerful others, those who were significant in my life? Anyway, does my poor little voice actually count in this world? Many people feel they do not have a voice in their own organisation, family or life in general.
Interviewing involves wrestling with questions and issues like these. In representing the voices of others am I not also committing the same kind of theft of their voice by replacing it with my own. What am I doing when I speak on behalf of another? when I represent their voices? Am I not really contributing to the social construction of their silence, their voicelessness in society? How can I create a data record that truly represents the voices of others without censorship, without reinforcing the structures of power which continually deny a means of expression to their voices?
For this reason, questionnaires are not a good strategy to uncover the voices of participants. Questionnaires, in their search for 'objectivity' are pre-structured according to an agenda of interests closer to that of the researcher than the researched. Thus, if you are working in your own place of work it is probably not a good idea to be tempted to pass around questionnaires. Some people, doing AR for the first time, have been known to use questionnaires on their colleagues because they do not have the courage to carry out interviews. Resist the temptation. There is barely a better starting point than conducting an interview in order to identify the agenda of others. You might decide to use a questionnaire if you want to talk to a large number of people at a later stage to get a broad picture or to find out how widespread your conclusions apply. Bear in mind the issues associated with quantification [covered in chapter 2], and be careful about generalising.
Remembering the need for 'immersion' and the value of an early interview, start collecting data early on. Steps 1 and 2 as outlined here, should not be considered as necessarily distinct. When you begin you may feel that you have a poor grasp of the area you are studying but this should not prevent you conducting an open ended interview in which you may say little a learn a great deal.
Often we are asked how many interviews a researcher should carry out. There is no simple answer to this question. We could say that you need to carry out as many interviews as you need to in order to clarify or establish one case or another. You will probably find that in a remarkably short period you will have a substantial volume of data and this may begin to frighten you. Indeed, too much data can become very difficult to make sense of. Try and collect data from the various viewpoints that there appears to be and then have a pause and try and make sense of what you have. You can then decide in which direction your study might go. You might wish to reinterview some people or change the focus completely. As you acquire a sense of what is happening you may well be able to categorise data as you collect it, into 'important', 'peripheral', 'not sure' or whatever. In this way you can keep a finger on the pulse of your study.
One classification of interviewing style regards it as resting on a continuum between formal to informal. The most formal kind of interviewing would be similar to the researcher reading a questionnaire to the interviewee. The interviewer's task would be to ensure a correct interpretation by the interviewee of the interview schedule.
In a less formal interview the interviewer would have a list of broad questions but would follow up 'interesting' issues raised by the interviewee in reponse to the questions. Informal interviews try to engage the interviewee in 'conversations'. The intention is to allow the interviewees to address their own agenda of concerns and interests without imposition by the interviewer. The interviewer may begin with a simple "Can you tell me something about 'x'? Typically, the interviewee then makes the points perceived as significant, tying them together in a way he or she sees as 'rational'. Informal interviewing may take place under formally agreed conditions; or, it may be simply a passing conversation. In the latter case the researcher has to consider the ethical questions associated with using this information.
[a] The 'rules' of informal interviewing
Informal interviewing is more an art form than a 'science'. There are no hard and fast rules. However, there are possible problems that the interviewer needs to be aware of:
[b] Strategies for informal interviewing
Again there are no recipes. Everyone develops their own style according to their personality. Acting out of character is easily detected. Thus, it is best to learn by trial, error, discussion and reflection what works and what does not. The following describe a number of interviewing styles that we have seen colleagues adopt.
[i] The provocative style. Some people have what may be called a provocative style. It is almost an 'attack'. Its advantage is that it gets to the central issues very quickly by provokes responses. It can, however, also set up a distrust, a nervousness which closes down the interviewees responses.
[ii] The 'I'm on your side' style. Here the interviewer dresses and speaks to copy the image of the interviewee. This has the advantage of establishing an initial favourable response. However, an unconvincing interviewer may get 'found out'.
[iii] The 'laid back' style. The style is casual, 'cool', almost clumsy, clearly no threat to anyone. It requires a certain attitude of mind which is non-judgemental, does not indicate anger, or frustration but always indicates an interest in whatever is being said, no matter how boring or trivial or shocking. The danger is that the interviewee does not treat the interviewer seriously. But that too is its strength. It can lull the interviewee into making known what otherwise would have remained hidden.
[iv] The social worker/encounter therapy style. In this style the interviewer indicates 'that's interesting, tell me more'. This, can lead the interviewee into simply giving the interviewer what he or she wants [i.e., what is interesting]. It can also be unnerving - it may prompt a feeling of being secretly psycho-analysed. Nevertheless, it can build a trust. It provides an 'ever listening ear' which is useful in that people like to talk, especially about themselves.
These styles are not so much agreed techniques of interviewing, but rather caricatures of styles that can be seen in practice.
During interviews, it is sometimes helpful to test out understandings, or developing ideas or analyses. Thus, an interviewing may say, 'As I understand it at the moment I hear you saying ........ ' and then summarises the understanding he or she has of what the interviewee has been saying. Or, the interviewer may say, 'One argument/view/issue I've heard from other people is ....., What are your views?' In this way, ideas can be tested out or 'triangulated'[see F, below], to see to what extent they apply to other people.
It should also be borne in mind that an interview is not merely a question of words but of discourse, that is, communication of a broader kind than just words. Your dress does not have to be formal but appropriate. In general I find myself trying to fit in in some way. Often I find it useful to dress a little less formally than the person I hope to interview though this is not always true. It is sometimes useful to be 'less important' then the person you are interviewing. If you come across as 'more important' you may find the interviewee being a little overawed, restrained and not as forthcoming as you might hope.
Immerse yourself into your study as soon as you can. You may have to do some reading and some interviewing. But do whatever needs to be done, especially the interviewing, sooner rather than later. Interviewing requires some commitment from you. Get over any embarrassment or fear you might have at an early point and this usually means jumping in [but with sensitivity].
3. Documents and Artifacts
During the course of everyday action, documents and artifacts are often made or used. Documents include not only the official organisational papers/reports/brochures but also the more work-a-day memos, workplans, and materials. Even more transitory are the notes and instructions chalked onto the blackboard/whiteboard/flipchart which can be recorded as 'documentary' evidence of the work of a committee, or classroom. Then there are the artifacts [models, artwork, craftwork etc] and other 'props' [e.g., furniture, pictures and other background objects which can either be functional or have aesthetic or symbolic value] and tools of the day-to-day work of the organisation. Each of these hold a meaning for the actors which need to be discovered. In order to uncover the social meaning of documents and artifacts questions such as the following can be asked of oneself or of others to direct observation and analysis:
The researcher may also establish who wrote a document and what they intended to achieve by writing it. By exploring wording, illustration, examples used, general format and presentation, anticipated audience from the writer's point of view, important issues maybe established.
Remember to hoard and collect any paperwork which might be relevant and a small note which can serve as a diary can produce valuable data. Diaries have the added advantage of soon becoming reflective, that is you will find yourself writing down private thoughts which can be enormously illuminative about your own understandings and biases. It is not unusual to find that a diary will at least provide a useful record of your activities. At best it will become a reflective account of your activities and a major source of data.
As a diverse range of material begins to be collected - in the form of audio or video tape recordings, diaries, documents etc - the next major question is how to put it into an order which can be easily accessed and to make sense of it all.
Just as C and D above need not be separate, you may feel that step E is taking place while you are doing C and D. Note first of all that generating hypotheses takes place some time after you have begun to immerse yourself in your study. Central to your research will be trying to discern new understandings about your practice and other people are vital for this. Early hypotheses tend to lack understanding and familiarity. What might be called 'mature hypotheses' have a firmer basis. An example may help here.
One student set out to experiment with more participative classroom techniques. She had just returned to work after bringing up two children and, having had the opportunity to contemplate, she thought she would try what she had lacked the confidence to do before. Before long she found that while her classes were beginning to appreciate the new approach her major difficulty was the teacher in the class next door who was unhappy about children walking down the corridor during lessons on their way to the library and other places. Comments were being passed in the staffroom. Gradually the teacher's study became concerned not only with changing her own classroom but making these changes with the opposition of her colleague. Change in the school became as important as change in her class. The original hypothesis grew into a more relevant mature hypothesis.
Making sense of data is partly a matter of intuition and partly a matter of being systematic. The intuitive part is about seeing themes, patterns, make guesses, make arguments, ask further questions about the data. In this respect you may find that being familiar with the literature may help you decide what are the important themes, though you will need to test out your own ideas too. As patterns, arguments and questions arise the data begins to be transformed from a pile of notes, transcripts, documents and other recordings into 'evidence'
However, you do not have to wait for the inspired hunch or flash of genius. Much of the work of making sense of data requires a systematic ordering of data. This is largely a process of analysis which can be broken down into a remarkably simple set of routines:
If you store data upon a computer, there is an added bonus. Most wordprocessors will have a 'search' and 'find' facility. This can be used as a simple way of organizing data. By placing key words at the head of sections, the computer can be made to find these one after another, thus helping to quickly trace important passages.
Interview with a Houseteacher
*being an individual*
as a house master I'm expected to discipline kids who have misbehaved for other people ... and whom I might be getting on very well with ... apart from that. Now I have got to try and build a relationship which allows that situation to exist. I've got to try and ... kinds have got to understand - I try and teach kids that it's my, part of my job sometimes to be nasty to them even though they've not done anything personally to me. Um, and in that sense they've got to identify me with the institution. And when they hurt the institution they hurt me kind of situation. Having said that, I also personally want them to see me ... as somebody that they can trust and somebody that they can uh have relationship with as a person, as an individual. I don't want to be just part of the institution.
By instructing the computer to find the key word *trust* it will go to all sections, one after another, which have been labeled with the stars in this way. Thus it will not go to every single instance of the non-asterisked 'trust'. With more sophisticated data handling programmes such as hypertext, or hypercard, this process can be made very much more sophisticated.
While you are organising your data it is likely that you will find yourself deep in thought during odd moments, thinking about interviews or something said in your class. In short, you will have begun to analyse your data. If you have audio or video tapes, go through them without necessarily transcribing at first. Get a feel for them. If you have notes read them.Begin to tidy up your notes and use a highlighter pen to pick out the parts you believe are important. Do the same with documents and papers.
Gradually you will become aware of themes which begin to be emerging. By this is meant several teachers may allude to, say, the behaviour of a certain class or the learning of a group of children or the actions of the headteacher. These then become themes that you will pay special attention to. You may find yourself progressively focussing upon them while, of course, not ignoring other matters. As you go further you might see relationships between the themes. You will have to be careful that the relationships you see are not merely through your own eyes and biases. A second round of interviews maybe needed to ascertain that you are not misrepresenting those you have interviewed. During the course of your study, the relationships which cement your themes together will become your theories.
Take great care. It is neither essential nor good practice to create theories which are not accurate. Data used to support a theory should come from a range of sources, that is, it should be triangulated. Data used in this way is called evidence. Remember too, that it is not unusual for some data to be 'left over' as it were. In other words it may not fit neatly into a theory. This data should not automatically be discarded. It could be the most important data of all. You will have to judge, in association with your respondents. You will probably find out that people make their decisions according to slightly different logics and what is rational to you may not be to them and vice-versa.
A common notion in all of qualitative research is that if the planning of action is rooted in the data collected from those who have to take action, there will be a good chance that the action will actually take place. If you are doing AR in your own classroom there are a number of temptations. One is to do nothing and nobody else will know. This route, in my experience is rarely taken. Action researchers are often, by this stage, heavily involved in their work and cannot wait to try something. The questions then are whether to try and make great or small changes or, possibly, the action researcher will have a half hearted attempt.
Ultimately the sorts of changes that come about will depend upon the people and the issues that are involved and the on-the-ground judgement of the researcher will be a major factor of determining what happens.
When AR includes other people the decisions that have to be made are quite different. The importance of the first line of this section may well then be apparent. If you are trying to assess children differently as a department, for example, you may need to have several meetings. You may find it helpful to have representatives from other departments at your meetings so they can see what you are doing. Members of the Senior Management Team may have a role to play. I have known a small minority in a group who do not want to alter their practice. It is usually better to leave them alone than antagonise them. If the changes you plan work well, they may fall in at a later stage.
While we are formulating and planning the process of building theory is also going on. It is a many layered process. The issue of theory was alluded to in the previous section [D] and part 1, 'Building Theory', which follows, could just as easily have been included in D. But to demonstrate that it is ongoing, it is included here.
1. Building Theory
A theory is an explanation which links together elements of the data that you have collected. We all have and also make theories about the world all the time. Research allows us the chance to develop and then to test some of these theories in practice. If a teacher says "little Billy is just plain naughty' as a reason for bad behaviour, this can be tested out by examining the contexts in which Billy acts. We may find that he is always labeled naughty in some, sometimes naughty in others and never naughty elsewhere. This would be enough to suggest that the 'plain naughty' theory is at least insufficient and perhaps is wrong. Perhaps, the new theory, depending upon evidence would be 'Little Billy tends to be labeled naughty in excessively authoritarian situations', or alternatively, 'in insecure situations'. Such a reformulation no longer postulates the problem as residing in 'little Billy' but in an interaction between 'little Billy and certain social structures'. Furthermore, as repeatedly pointed out, to label someone as naughty is data not about the person so labeled but about the person who is doing the labelling. It is data bout the system of labels employed by a professional in given contexts of use.
Thus, one may go even further and begin to inquire into the meaning of 'naughtiness'. We may find that teachers A, B, C and D all have quite different definitions of 'naughtiness'. Or, we may find that all teachers in the school have broadly similar definitions, but the parents, social workers, and other social groups do not. We would then be led again to the view that whatever 'naughtiness' is, it is not an innate characteristic but is rather something that is defined by particular people. That is, it is socially constructed. It only exists when groups of people make it exist by continually referring to certain behaviours or to certain people as being 'naughty'.
Slowly, we are beginning to build a theory about naughtiness by constantly referring to our evidence of what people say it is, and what people do in different situations.
In building theories therefore we can ask ourselves such questions as:
These comparisons and contrasts between individuals, groups, and social contexts will begin to show what persists and what changes over time and in different situations and with different group composition. This process of constant contrast and comparison is often referred to by the term theoretical sampling.
In this process we have begun to examine not only the beliefs of others but also of ourselves. These beliefs and practices have been challenged and perhaps changed by listening to and observing a very much wider group of people than we would be able to do in ordinary conversation and observation. The principles outlined in the ethics section of this document have ensured that we accord each person equal status and rights to have a voice in the research. Typically, children have low status in a group which contains adults. Research, by listening to children seriously, can add a new and important dimension to the classroom and to our understanding of the world of children, the way they interpret and act upon their experiences of the classroom.
Theory and action are inter-related in that we act, develop theory, act and so on. When we build poor theory our actions often produce outcomes that are different or even contrary to what we expect. It is often said by scientists that there is nothing more practical than a good theory. Insofar as theory is a basis for action we have to develop it carefully and keep it constant review.
Social theory is often thought of as vague. However, the behaviours of millions of people are routinely co-ordinated by theoretical constructs everyday. Transport timetables, school timetables, calenders, dances, polite exchanges between strangers - all these in their different ways systematically co-ordinate the behaviour of people. We are very good and often very precise in our techniques of behaviour co-ordination. However, none of this is done in a mindless way. Unlike the materials manipulated by chemists or engineers, people can object to being manipulated, or can choose alternative actions. The timetables that organize our lives in schools, businesses and social life only work if we choose to make them work. Choice is the key.
By identifying and generating alternatives, choices can be made and action taken. The final major step in the process of action research is to formulate action plans.
2. Action Plans
We can distinguish two broadly different kinds of action plan:
[a] the first seeks only to change the details of existing ways of doing things; or, to solve certain problems in the execution of a plan without changing the overall plan.
[b] the second seeks to replace the existing way of doing things with a different way of doing things.
The first occurs when through discussion members of a group are convinced that their current practices are basically desirable but that certain problems still need to be ironed out.
The second occurs when through discussion members of a group become convinced that the current practices are less useful - or even no longer defensible - in comparison with another way of doing things identified through the research and observation carried out.
In either case the proposal for action should include the following:
By following guidelines such as these and adapting them to specific purposes, action research can be a continuous process of professional self and staff development.
As qualitative and indeed other forms of research have come to stress action there has been a developing and increasingly influential literature about change. Fullan , Holt , Fullan and Stiegebauer  and Fullan and Hargreaves  are some examples though the reader may be best advised to read Changing the Curriculum by MacDonald and Walker  for a good theoretical basis. Even a cursory read of these pages will suggest to the reader that our values suggest that change is more likely to be a process than an event. More importantly it is unlikely to be successful if teachers are told to implement somebody else's design.
For us, change occurs following consultation and debate and endeavours to incorporate the wishes of those involved. Of course this is not always possible and can be arduous but not always so. It is not necessary for managers, for example, to have long drawn out, formal debates with their colleagues. Sometimes a brief corridor word will suffice. People who work closely together find ways of discussing their work without too many formal meetings and develop an understanding for each others preferences and views. In collegial institutions teachers and others develop a culture which incorporates ways of understanding and working with each other. In such institutions chnage is decided by the body politic and while not everybody accepts all changes most professionals can find a way through. Dictatorial managers rarely have their fiats accepted in practice though people may give them the impression that all is well by being 'creatively compliant'. In qualitative research our focus is on what actually happens rather than what appears to be the case.
If you are working largely alone you are faced with the same sorts of choices outlined in 4, above. When all is said and done everything rests on a simple decision. "To do or not to do?" As I have suggested above, there has to come a time when you have to just go for it. This is the flying by the seat of your pants bit.
If you have colleagues involved, you will need to talk to each other but ultimately, as above, change rests upon somebody actually deciding to do something. There are often more reasons for not doing than doing. And it is in the nature of things that all change involves a risk. If you wait for no risk you will have a long wait. If you have collected and analysed your data carefully and attempted to ensure yourself that you have some people willing to participate, you at least have a chance.
The leading action researcher , enabler or critical friend will need to get colleagues to give accounts of what each is doing. We all learn a great deal from simple stories. And these have to be honest, warts and all. Those with doubts are often strengthened by hearing about the failures of others. This is not necessarily a question of being unkind but rather allowing others to realise that they are not the only one who is finding the path a rocky one. They can find fresh heart from commiserating with others, and having the confidence to share their own failures. Of course it does not have to be, and indeed, usually is not, a negative business. It is wonderfully exciting when colleagues express what they see to be success and show they have, then, the enthusiasm to see a new round of possibilities before the current ones are finished.
The first stage is to carry out what Elliott has called 'reconnaisance'. This involves the action researcher in first describing the situation following the first action steps and then beginning to explain the situation. Taking these actions implies that the action researcher is now setting out on another stage of the action research 'spiral'. Indeed the process is seen as a never ending spiral which looks like a coil.
Yet there is a variation in this rather neat picture. Sometimes during the course of carrying out action research the researcher comes across a piece of data, or has a new thought or for some other reason realises that there is a new and different perspective upon all that has been done. In this case the entire new coil is separate from the original. So, for example, an art teacher who was exploring whether children draw better from pictures or objects concluded that what was more important was the way she taught and the children learned drawing. She also found that a small group in her class could not and were not interested in drawing at all and she was better off finding different art activities for them to do. Having reconnoitred, the researcher may then be in a position to form a more mature hypothesis or at least to collect some data so that the study can proceed.
If you are feeling apprehensive, as many do, it is a good idea to jump in early to feel the water. There are usually difficulties in collecting data, and sometimes more substantial problems as you analyse, collect more data, put forward your findings and so on. Many practitioners return to study having not been on a long course for a number of years. The greatest difficulties are, in our experience, encountered by those who leave too much to the later stages. You cannot progress very far without data, so go and collect some.