How can I conduct worthwhile research into effective homework in my primary school?

This paper considers how to plan and implement an educational inquiry in a primary school. The subject of this planned inquiry is the identification and application of effective homework practices in the small school where I am head teacher.

The paper is in two parts.

Worthwhile research: I discover useful lessons from the natural sciences. I declare myself a reluctant convert to the methodology of practitioner-action-research .

Effective homework: I indicate how I will apply the practitioner-action-research methodology to an inquiry into effective homework.

My account may be of interest to others undertaking school-based inquiries, particularly those for whom the action research approach is unfamiliar. It may also be of interest to established action researchers as a reminder of how very unfamiliar their methods can appear to prospective researchers.

Worthwhile research

Hopes, expectations, limitations

To call the product of my inquiry 'worthwhile research' I was hopeful that I would do more than merely plan, implement and evaluate a school-based initiative. I hoped to learn and apply trusted inquiry techniques to create a lucid, novel account that would be:

These are the marks of a broadly scientific, or positivistic, methodology. I was not naive enough to suppose this process would be straightforward. I certainly anticipated difficulties in satisfying these criteria within a small-scale inquiry and, more fundamentally, was suspicious of their application to the complex social setting of a school.

My training allows me to claim an understanding of the powerful and convincing methods employed by natural scientists. Fully aware that these methods could not be applied directly to children's learning, I was keen to discover how social scientists generate knowledge. I expected to tread what I fondly imagined would be a well-worn path from the natural sciences to the social sciences. Little did I realise how quickly that path becomes indistinct...

Seeking clarity I consulted a number of texts, the insights from which are incorporated in what follows. Yet all the time I have harboured the strong belief that I do not have a free choice about how to conduct the inquiry, so circumscribed are the conditions under which it is to be completed. I am acutely aware that, amongst other limitations, I:

These restrictions are not entirely surmountable if I am to remain in post. I believe this makes an action research methodology inevitable. Consequently, I have directed my efforts to clarifying two features of such a study, namely:

The goal of the analysis that follows is an explicit understanding of how these two entwined features might show themselves in examples of small-scale practitioner research, such as I might realistically undertake. To attain this goal it has been necessary, in the first instance, to broaden the theoretical field.


Consider the product of research in general. Practical inquiry results in new opinions, new attitudes, decisions and actions, but research creates knowledge.

What counts as knowledge is by no means obvious - there are certainly different ways of knowing - but in the present context a good working definition is that knowledge is 'justified true belief'. This derives from Plato's suggestion in the Thaetetus where Socrates indicates that knowledge is 'true belief accompanied by a rational account' (p.115). Beliefs are two-a-penny. What counts is that they are true and rationally justified.

Lessons from science

The dominant paradigm for propositional knowledge of this kind is provided by natural scientists, pre-eminently by mathematical physicists. The life sciences and positivistic social sciences aspire to the same, or a similar, form of knowledge, expressed causally in theories. If such theories did follow logically by induction from the observations that engendered them then the scientific method would indeed give rise to a very special sort of knowledge. For then, assuming observations were reliable, scientific theories would not only be true and justified; they would be unassailable. However, reasoning from the particular to the general (science's characteristic induction) is emphatically not the simple opposite of reasoning from the general to the particular (the logically infallible deduction) so scientific theories can never be proven. 'How can one derive a factual proposition from a fact?' asks Lakatos (1999), picking up the same point in one of his lectures on scientific method. Without hesitation he responds: 'One cannot' (p.37). So scientific knowledge is not fundamentally different from other learning and is as much the product of intuition and imagination as other human ways of knowing.

Nevertheless, understanding that scientific knowledge is not created uniquely cannot detract from the fact that its products possess distinctive strengths. Actually or potentially scientific theories possess:

Whether or not the degree to which they possess these qualities permits a clear demarcation between science and other theories, consideration of these qualities makes clear that educational inquiries are, at best, not very scientific:

Precision: High precision is a consequence of quantitative analysis and concise mathematical expression. Quantitative analysis of a social setting, like a school, is possible though crude. A key question is to what extent numerical data (e.g. attainment statistics, questionnaire responses) can be held to measure important features of the processes of education. What is incontestable is that it cannot capture them all.

General applicability: This is not a feature of particular theories but an underpinning assumption of the whole scientific enterprise. If the universe behaved capriciously then science would be impossible. Scientists must assume that there is regularity in the objects of their study. It is manifestly not an assumption that educational researchers can share in their work: one school, one class, one pupil, is patently not like others. Trying to frame a conclusion with even wide applicability is a challenge in an educational inquiry.

Authority: A theory is all the more believable when its evidence base is broad and strong. When the body of experimental evidence available to the natural scientist is orders of magnitude larger than that available to the social scientist (and often repeatable to boot) it is little wonder that the latter's conclusions appear tentative by comparison. It makes the position of the practitioner researcher, relying on so-called case studies, positively precarious.

Predictive power: The assumptions of regularity and causality, coupled with precision, make predictions possible in scientific work. Predictions are possible only in the very broadest terms in a social inquiry, and even then the margin of 'error' will be very great.

Falsifiability:'...The criterion of the scientific status of a theory is its falsifiability, or refutability, or testability' (Popper, 1963: 48). If a theory is irrefutable, Popper declares, it is simply unscientific. This places most social 'science' well outside the scientific fold, for two reasons. Firstly, a great deal of social science is descriptive and thereby unsuited to testing at all. Secondly, where testing would be conceivable the ability to plead special circumstances makes the propositional content so slippery that it would be hard to decide whether a test had been passed or not. Lakatos (1999) is forthright on this point: 'The social sciences are on a par with astrology, it is no use beating about the bush' (p.107).

Thus to allow the possibility of educational research at all the field of inquiry must be broadened, since the preceding discussion makes clear that science as practised by natural scientists is not suited to the task.

Though natural scientists study nothing less than the whole universe this does not place within their remit everything that is. Popper (1963) defines science's rightful field of activity as 'systems which (are) well isolated, stationary and recurrent,' adding, 'these systems are very rare in nature, and modern society is surely not one of them' (p.457). Coming at the same issue from the opposite direction Nagel (1986) is at pains to admit to the world more than natural scientists can handle:

'For many philosophers the exemplary case of reality is the world described by physics, the science in which we have achieved our greatest detachment from a specifically human perspective on the world. But for precisely that reason physics is bound to leave undescribed the irreducibly subjective character of conscious mental processes... The subjectivity of consciousness is an irreducible feature of reality... and it must occupy as fundamental a place in any credible world view as matter, energy, space, time and numbers' (pp.7-8).

Education is not a natural phenomenon but a human activity based entirely on mental processes (conscious or not). Irrespective of the high regard in which I hold the work of natural scientists, to have supposed that the hallmarks of their work could characterise the product of a small-scale educational inquiry was folly. For reasons both practical and epistemological I do not believe that any educational inquiry I undertake can deserve to be called scientific.

This conclusion does not entirely negate my original aspirations. In place of science's exacting qualities I propose more modest alternatives appropriate to small-scale practitioner research, as below:

General applicability
Predictive power
Suggestive possibilities

Thus, a good quality account of an educational inquiry will possess:

Clarity: It will be concise, readily understood and hard to misinterpret. Technical terms will be employed, though always serving to illuminate meaning rather than obscure it.

Transferability: Accounts must be written for others. They must be sufficiently general to teach others something useful. Autobiographical elements, if present, will always serve the didactic purpose of the text.

Credibility: A range of evidence will be presented and limitations acknowledged. Practical and theoretical contexts within which findings are interpreted will be explicit. The temptation to 'over-report' change will be vigorously resisted.

Suggestive possibilities: Implications of research will be offered and not merely implied. Thinking about the application of findings is the author's task, as well as the readers'. Possible unintended outcomes of the application of findings to novel situations ('side-effects') will be pre-empted, where possible.

Humility: Findings and practical suggestions will be offered provisionally.


On the basis of the above, 'the scientific method' would appear to have little to commend it in the conduct of a school-based inquiry. However, the distinction between actually doing science and any other critical inquiry is not that great. Following Popper, Medawar (1968) - one of the few working scientists who has written on the subject - says of science's underlying approach (what he calls the hypothetico-deductive process) that there is 'nothing distinctively scientific' about it, and that 'It is not even distinctively intellectual' (p.54). If it is understood that no unwarranted carelessness is implied this hypothetico-deductive process could be termed the 'guess-and-check' approach, an approach as suited to understanding how schools work as anything else. In any inquiry imaginative consideration of why things are as they are and how they might be different (guessing) alternates and interacts with critical testing and observation (checking).


It is not easy to describe the imaginative/intuitive phase of the inquiry process nor ensure that it happens successfully. Medawar (1968) emphasises the creativity that underlies acts of making-sense-and-imagining-things-otherwise when he describes this as 'the invention of a fragment of a possible world' (p.56). He is prepared to sketch four types of intuition though is not so bold as to indicate how to manufacture them. That notwithstanding, 'guessing' is certainly not an accidental process. It cannot occur without fuel for thought or the experience that structures observations coherently.

Structuring observations coherently: Perhaps the lack of an explicit theoretical framework within which to interpret observations is no great handicap when practical engagement over a period of time offers a strong basis on which to understand the situation being studied (e.g. as a teacher in a school). Action research possesses a healthy respect for the embodied professional knowledge of practitioners. This places the onus on practitioner-researchers to consciously apply their implicit knowledge and articulate this act (of application) clearly.

Fuel for thought: The risk of parochialism attends the recognition of practitioners' everyday expertise as a valid basis for critical inquiry. This can be obviated. Medawar (1968) says: 'We can put ourselves in the way of having ideas, by reading and discussion and by acquiring the habit of reflection' (p.57).


The critical phase - in particular its need of empirical evidence - hints at a fruitful metaphor. If it is impossible to prove a proposition in an educational inquiry then it is appropriate to require that at least a convincing case be presented, as in a court of law. Here, if an argument is to be proved true beyond reasonable doubt, a mosaic of evidence in a variety of forms will be eloquently and persuasively presented. This will include witness statements, expert testimony, scientific observations and more: a diverse mix chosen simply because it is available and it supports the case.

Available evidence: Judgements about the efficacy of particular educational practices will also be supported by evidence in a range of forms. The practitioner action researcher will not be bound by the methods of a particular school (e.g. a sociological or psychological interpretation) but will gather data eclectically and opportunistically.

Supporting the case: Quantitative data may be useful, though its statistical significance will need close scrutiny in a small-scale inquiry.

Expert opinion will probably be available since few research subjects could claim to be entirely novel and there will often be evidence of existing practices to both confirm and challenge new findings.

Action researchers will be seeking to effect changes for the better and will gather testimony from those most affected by their actions (pupils, colleagues and others). In each case practical techniques will inform the gathering of evidence, from statistical analysis to ethical guidelines to questionnaire design. There is a need for careful attention to technical methods since by their skilful use researchers help ensure that the evidence they amass is admissible and convincing. Conversely, inattention to established methods will weaken research. This is a danger, as practitioner-researchers are unlikely to be experts in such methods.

Action research

The 'guess-and-check' approach sketched above is too general to deserve the title of a methodology, though importantly it does give the imaginative process due emphasis.

A scientific methodology has been shown to be inappropriate for my educational inquiry.

A naturalistic, ethnographic approach is equally inadmissible since I intend to change school practice through the inquiry. (It would be difficult to justify the commitment of time and energy if improvements were not sought and I believe it is implicit in the Teacher Training Agency funding arrangements that allow me to complete the inquiry at all.)

Consequently I will conduct the various activities of my substantive inquiry within the structure of an action research project.

Several mechanistic descriptions of the action-research process exist, such as The Action Research Planner by Kemmis and McTaggart (1982). This is described as a 'procedural guide' and methodically describes a four-stage cycle of planning, acting, observing and reflecting, the intention being that this sequence is repeated to effect practical improvements and increase knowledge simultaneously. Notwithstanding the authors' insistence that their guide be worked through systematically and completely, it is not to be supposed that the slavish application of even a well-defined methodology will guarantee a fruitful inquiry.

Winter (1996) fights shy of declaring a programme but advances six key principles for the conduct of action research. Application of these principles, he suggests, will ensure that action research methods:

There are numerous methodological challenges to be faced. Some are considered below as the general processes of 'guessing' and 'checking' within an action research framework are applied to the conduct of my inquiry into effective homework.

Effective homework


The school that I lead is a small primary school serving a village community. It currently comprises 104 pupils aged 4 to 11 years old, in four classes. Children enter the school with broadly average attainment. Most parents/guardians are very supportive of the school.

The school was inspected by the Office for Standards in Education (Ofsted) in January 2001. The inspection report indicated that there were significant shortcomings and the school was declared to be 'underachieving'. A large number of substantial key issues were identified for action.

In February 2003 the school was re-inspected. Very good improvement was noted. The inspection report stated that all key issues from 2001 had been addressed. Two new key issues were identified. These have since been addressed.

Motivation for the planned inquiry

One minor issue was declared in the inspection report of 2003, namely: 'Ensure... that homework is consistently set in all classes, that its collation is systematic, and that its return to parents is informative.' (Ofsted, 2003: 21)

This issue was prompted by parents' and carers' replies to a pre-inspection questionnaire. When asked to comment on the statement 'My child gets the right amount of work to do at home' a significant minority expressed dissatisfaction. Inspectors commented: 'Whilst homework is used to a satisfactory extent, there is scope for the school to adopt a more consistent approach to its use.' (Ofsted, 2003: 10)

With the major challenges that once faced the school largely resolved it is appropriate that governors and teachers, in partnership with pupils and their parents, consider secondary issues that hold the promise of further improving children's attainment and attitudes to learning. Hence the planned inquiry.

Activities in the inquiry


An important early activity will be to read relevant research to place our practices in context and to guide developments. Summary reviews of a wide, though inconclusive, literature exist. These include theoretical models of the subject, which will help structure the inquiry.


The principal activities of both the imaginative and critical phases of my inquiry will be listening to others. I will talk with the obvious stakeholders within my school, namely teachers, governors, pupils and pupils' parents. At different times discussions will furnish information about current practices, ideas for improvements and evaluations of initiatives. The responses of these various groups will be communicated to each other, through the conversations themselves and in school newsletters.

Teachers: Formal conversations will take place in staff meetings. We will employ common techniques for structuring discussions, e.g. nominal group technique, SWOT analysis, forcefield analysis. I will encourage colleagues to share thoughts and observations informally as they arise. From time to time we will briefly consider the subject together to maintain a running dialogue as the theme develops and as new information becomes available. The school's general primary adviser will bring an outside perspective to one staff meeting.

Governors: Conversations will take place in the governors' Curriculum Committee and, later, with the full governing body. Committee meetings take place once per term, but more frequently if necessary. Discussions are usually informal but some structured discussion techniques may be employed.

Pupils: The oldest pupils (aged 9 - 11 years) will consider the subject in lessons. They will gather evidence of their own opinions by creating and completing a questionnaire. Younger children may also complete this. They will analyse findings and discuss these, and other thoughts, in established ways (e.g. 'Circle Time'). I teach the class of oldest children and will complete this work with them, making notes of class discussions.

School Council members will seek views from their classmates and discuss the matter in School Council meetings. Children minute these, but I will attend as an observer and make notes too. Involving children will furnish some interesting information and raise the profile of homework in their minds as we seek to make improvements in school practice.

Pupils' parents: Gathering parents' views will require careful attention, as this is the first time that the school has consulted the whole parent body in this way. A questionnaire will be devised. This will be discussed with teachers and governors before use, trialled in other schools (I have secured offers of help from other head teachers), modified where necessary, and distributed locally. To shortcut the rigorous trialling necessary for the development of a good questionnaire some questions will be based on those used in a larger, published study. Sentence-completion questions will be employed in early trials. The final questionnaire will probably comprise a mixture of multiple-choice answers and sentence-completion questions.

Opportunities for face-to-face conversations with parents will be created. These will take the form of unstructured depth interviews in groups. Because of my authoritative position within the school it would be preferable for someone else to conduct these, but this is impractical. Though I have characterised the interviews as 'conversations' I will try to listen rather than talk.

Reaching understanding

Conversing and reading are similar in that both are marked by a determination to understand what others have to say.

Reaching an understanding with others is carefully analysed by the social theorist Jurgen Habermas. He seeks to uncover the processes that all competent language users employ intuitively and identifies the conditions that they must meet to bring about understanding. His analysis (Habermas, 1979) emphasises that reaching an understanding involves a balance between saying something clearly ('cognitive use of language') and allowing it to be heard ('interactive use of language'). Attending to both these aspects will be necessary when conducting all the consultations described. Creating a questionnaire that no-one feels inclined to return will hardly serve the purposes of the inquiry, after all.

Another powerful notion contained within Habermas's analysis is that of the 'illocutionary force' of an utterance. In the setting of my inquiry I take this to mean the intention that lies behind any of my consultations. I believe it will be worthwhile to make this explicit, in my own preparations and to others. It would prove unhelpful to imply that all parents' suggestions will be acted upon, for example, but, equally, contributors will want to know that their opinions are valued. I will attempt to address these issues consciously rather than intuitively in order to forestall any possible misunderstandings.

Winter (1996) offers 'collaborative resource' as one of his six important principle to apply to action research. Careful attention to the views of teachers, governors, pupils and their parents, as described above, plus an explicit consideration of my role in the consultative process will be an appropriate way to apply this principle to my inquiry.

Quantitative data

I will seek to gauge whether new practices affect:

Questionnaire responses may be summarised quantitatively but statistical analysis will be inappropriate due to the tiny sample size involved and the strong likelihood of bias, through self-selection in the return of questionnaires. The views expressed will be representative only of the respondents: there can be no pretence that replies will be typical of the whole school community, let alone a wider constituency.

The question of whether homework affects children's standards of achievement will not be answerable. Analysis of numerical assessment data occurs as part of the school's normal cycle of self-evaluation but the contribution of homework will not be identifiable.

Ethical concerns

All conversations will be conducted openly and findings will be communicated between groups as necessary. In the final report permission will be sought from individuals for the inclusion of remarks that may be directly attributable to them.

The forensic analogy (above) might suggest that my role will be to cross-examine witnesses regarding current and future practices. It is better to characterise these interactions as conversations rather than anything more probing. Happily, the relations between members of the school community are friendly so any methods that seem to deny this fact would be counter-productive and will be avoided. Partly for this reason I do not propose to record conversations on audiotape or videotape because of the distorting effect I believe this would have on the conduct of the discussion and on the relationship between contributors. In addition most, if not all, conversations will be group discussions where the technical problems of ensuring a successful recording would be great. Finally, transcription would be impractical. All formal meetings are minuted and I will make informal notes of my own for future reference. This will be sufficient.

Success criteria

Criteria for the success of any interventions will be negotiated during the course of the study. The limited quantitative data to be collected may help determine the success of what is done. Alternatively it may be that the very act of establishing what we mean by 'successful' will itself count as one of the more important outcomes. I take this sort of introverted analysis to be an example of what Winter (1996) calls 'reflexive critique' (another of his six principles), a process whereby claims (e.g. 'This way of doing homework is effective') are questioned rather than asserted. The basis on which the judgement is made is itself questioned, and always understood to be one of several options. Thus: 'claims can be transformed into questions, and a range of alternatives suggested, where previously particular interpretations would have been taken for granted.' (Winter, 1996: 43)


I hope to write an account that will be useful to primary school teachers who share an interest in effective homework. I will share a draft of my report with colleagues involved in its creation, including the head teachers of several schools who trial the parents' (and possibly the pupils') questionnaire. I will invite their responses, incorporating these in the final account (with their permission). My account and supporting notes will be freely available online.

Concluding unscientific postscript

It is not obvious how I should conduct an educational inquiry, but there is much at stake in finding the right way. For in answering the question 'How?' I simultaneously delimit what I can discover. Methods and outcomes are truly and inextricably entwined.

My original goal was to do 'worthwhile research'. Under the hegemony of the positivist tradition this seemed to imply doing something that looked like science but in my setting this is no more practical than it is appropriate (and I will not call my efforts 'scientific' just because I execute them thoughtfully).

The rejection of science as a method is necessarily the rejection of science as a product and with it the loss of those very things that make it so attractive. Fortunately some desirable qualities (clarity, transferability etc.) can be salvaged.

I am adopting what I believe to be the only method fitted to my situation, namely action research. Any reluctance stems from uncertainties about the status and quality of the knowledge that will be created. I take heart from the fact that professional researchers also struggle to define what 'good' action research is. In editorial documents for the journal Action Research, Reason declares 'there really can be no pre-established rules for validity based on firm foundations' (at: and his considered discussion of the same point can therefore offer only plural, speculative conclusions ripe for revision (Reason, 2003). My uncertainties fit well with an insistence that 'action research (is) an orientation to inquiry rather thanů a methodology' (Reason and McArdle, 2003). So, reassuringly, not everything has to be sewn up before I can begin.

If, as I start my prosaic look at homework, the utopian calls to arms of some action research evangelists seem misplaced I can, equally, glimpse that an educational community whose corporate activity is marked by self-criticism and self-interpretation would be an exciting place to belong. And the hint that a reconciliation might be effected between 'theory' and 'practice' is a vision both large enough to inspire effort and modest enough to seem attainable.


Habermas, J. (1979) Communication and the Evolution of Society. London: Heinemann

Kemmis, S. and McTaggart, R. (1982) The Action Research Planner. Victoria: Deakin University Press

Lakatos, M. (1999) Lectures on Scientific Method. In: Lakatos, M and Feyerabend, P. For and Against Method. pp.19-112.Chicago: University of Chicago Press

Medawar, P. (1968) Induction and Intuition in Scientific Thought. Philadelphia: American Philosophical Society

Nagel, T. (1986) The View from Nowhere. Oxford: Oxford University Press

Ofsted (2003) Inspection Report: St. Nicholas CE (VC) Primary School, Bromham, Chippenham [Online] Available at: (accessed February 18 2004)

Plato. Thaetetus. Translated by Waterfield, R. London: Penguin

Popper, K. (1963) Conjectures and Refutations: The Growth of Scientific Knowledge. London: Routledge and Kegan Paul

Reason, P. (2003) Choice and Quality in Action Research Practice, Keynote Address, World Congress of Participatory Action Research, Pretoria, South Africa [Online] Available at: (accessed February 17 2004)

Reason, P. & McArdle, K. (2003?) Brief Notes on the Theory and Practice of Action Research [Online] Available at: BriefNotesAR.htm (accessed February 17 2004)

Winter, R. (1989) Learning from Experience: Principles and Practice in Action-Research. London: The Falmer Press