You are on page 1of 16

Paedagogica Historica

Vol. 44, No. 6, December 2008, 691705

Qualitative and quantitative research methods: old wine in new bottles?

On understanding and interpreting educational phenomena
Paul Smeyers*
Ghent University and K.U. Leuven, Belgium

Generally educational research is grounded in the empirical traditions of the social

sciences (commonly called quantitative and qualitative methods) and is as such
distinguished from other forms of scholarship such as theoretical, conceptual or
methodological essays, critiques of research traditions and practices and those studies
grounded in the humanities (e.g. history, philosophy, literary analysis, arts-based
inquiry). Since the early twentieth century, mainstream educational research is of an
empirical nature. In quantitative research, one typically looks for a distribution of
variables (how many are there with this or that characteristic) and for explanations,
which can be of a deductive-nomological kind, incorporating universal laws, or be of an
inductive nature, which employ statistics. Due to being subsumed under its own set of
laws, quantitative research can offer an explanation either in terms of an argument (a
logical structure with premises and conclusions governed by some rule of acceptance),
or as a presentation of the conditions relevant to the occurrence of the event and a
statement of the degree of probability of the event given these conditions. Using
Polkinghornes distinction between an analysis of narratives and narrative analysis
one can further differentiate between two kinds of qualitative research. One may be
interested in common features in different cases. Here the purpose is not only to describe
categories, but also to deal with the relationships between different categories. In many
cases this kind of research is generally analogous to a quantitative design (including
hypotheses), with the exception that qualitative data are gathered, i.e. they refer to what
people feel about, or what their experience is with, particular things, what they say that
their reasons, desires and intentions are. To be distinguished from this is a second kind
where the researcher arranges events and actions by showing how they contribute to the
evolution of a plot. The plot is the thematic line of the narrative, the narrative structure
that shows how different events contribute to a narrative. This interpretive research thus
goes beyond research as the accumulation of knowledge and comes close to those areas
of scholarship (see above) that were distinguished from educational research grounded
in the empirical traditions of the social sciences. In other words, an interpretation is
offered. In this paper various problems relative to the different types of research will be
dealt with. It will be argued that educational research (the study of education) should be
characterised by various modes of explanation depending on the kind of theoretical
interest one is pursuing. That is does not give us fixed and universal knowledge of the
social world as such, but rather that it contributes to the task of improving upon our
practical knowledge of ongoing social life. It presupposes dialogue between all those
involved and furthermore invokes a normative stance. Finally, it should be seen as a case
of positive slowness that prevents us from being absorbed in the chaos of unmediated
Keywords: quantitative educational research; qualitative educational research;
interpretative educational research; method; science

ISSN 0030-9230 print/ISSN 1477-674X online
2008 Stichting Paedagogica Historica
DOI: 10.1080/00309230802486168


P. Smeyers

The debate: an introduction

An educational research problem is an issue, topic or question, so it is said, that may be
theoretical, practical or a combination thereof. However, in most cases, the starting-point is
a particular educational reality that is unsatisfactory to the parties involved. Examples are
language learning, participation in higher education of particular groups of society, implementation of educational policies, but also bullying in primary schools, burn-out of teachers,
empowerment of parents etc. Generally1 educational research grounded in the empirical
traditions of the social sciences (commonly called quantitative and qualitative methods) is
distinguished from other forms of scholarship such as theoretical, conceptual or methodological essays, critiques of research traditions and practices and those studies grounded in
the humanities (e.g. history, philosophy, literary analysis, arts-based inquiry).
This has not always been the case. According to Kant, science is always a system:
knowledge built on the basis of principles. What is expressed has to emerge as necessary
for the mind. A science whose foundations and principles are only empirical can produce
only false knowledge. Therefore, a proper educational science has to be based on a priori
ideas. Our reason is practical: it should give the norms which bind our acting unconditionally. Ones realisation as a being of pure practical reason is actualised in so far as one
commits ones acts to the law of ones own reason. In order to be able to do that an education that is orientated by ethics is required. The main task of a human is to become moral.
A historical-teleological orientation will express the way in which the human being is
shaped. Kants interest is thus not focused on the individual, but on humanity and on how
it actualises, and gradually realises the perfection of a human being. The educational theory
that emerges is a science of principles, a philosophy. And though he was not indifferent to
how society develops, as a scientist of education he was interested neither in empirical falsification, nor in outcomes or in prediction. In some sense, the debate initiated by Kant has
continued ever since. Herbarts reformulation of the issue, outlining that philosophy should
give us the aims, and psychology the means to achieve them, reconceived the place of
empirical facts, but left the overall design almost untouched. Such is also clear in later
developments as for instance in Brezinkas technological stance, where it is argued that the
science of education can only deal with the means in relation to achieving certain ends, not
with the ends themselves.
Since the early twentieth century mainstream educational research has been of an empirical nature. Though quantitative methods are still very much used, qualitative research has
gradually become more and more important. In quantitative research, one typically looks for
a distribution of variables (how many cases are there with this or that characteristic) and for
explanations which can be of a deductive-nomological kind (incorporating universal laws)
or be of an inductive nature, which employs statistics. A subsumption under laws can also
offer an explanation not in terms of an argument (a logical structure with premises and
conclusions governed by some rule of acceptance), but as a presentation of the conditions
relevant to the occurrence of the event and a statement of the degree of probability of the
event given these conditions. It has been doubted by many whether it is possible to find
universal laws within the context of the social sciences, but even if one accepts the more
moderate version that degrees of probabilities of an event can be given, most scholars argue
that contextualisation of theoretical insights is necessary.
1See for instance: Standards for reporting on empirical social science research in AERA
publications. American Educational Research Association, Educational Researcher 35, no. 6
(2006): 3340.

Paedagogica Historica


Contextualisation and generalisation of quantitative and of qualitative

research findings
An interesting example to illustrate this is the Tennessee Studies of Class Size, known as
project STAR (Student/Teacher Achievement Ratio).2 Project STAR is seen as an experiment that starts from the idea that in smaller classes teachers have more time to give to
individual children. In the experimental classes, the class size was reduced from around 23
to 15, by approximately one-third, in kindergarten, first, second and third grades (ages 5
8); the children moved into regular-size classes in the fourth grade. There were three kinds
of groups: classes one-third smaller than regular-size classes, regular-size classes without a
teacher aide and regular-size classes with a teacher aide. The experiment was carried out in
79 schools in the first year; both children and teachers were randomly assigned to the
classes. In the second year it included 76 schools with 331 classes including 6572 children
in inner-city urban, suburban and rural schools. It was continued for four years (1985
1989). After this period there was a second phase, the Lasting Benefits Study, which
followed participating children into later grades and recorded their academic progress. The
major findings on class size are, first, that smaller classes did bring substantial improvement to early learning in cognitive subjects such as reading and arithmetic; second, that the
effects persisted into grades 4, 5, 6 and 7, after pupils moved to regular-size classes, and
finally that students who had originally been enrolled in smaller classes continued to
perform better than their peers who had started in larger classes. Incidentally, minority
students gained twice as much as the rest during the first two years before settling to about
the same gain as the rest.
In their discussion Mosteller et al. indicate that there are many issues involved when a
well-designed and implemented study comes out with a definite finding. Serious consideration has to be given to all the available alternatives, and to the costs and social consequences of implementing the new policy suggested by the findings. They stress that these
findings do not automatically mean that reducing class size is the best way to improve
schooling this has to be compared with other measures (for instance, one-to-one tutoring
by qualified teachers, peer tutoring or cooperative group learning). Though the results of
the STAR project have not generally been disputed,3 some critics have pointed out other
elements as well that have to be considered, such as the preparation time for teachers,
which is supposed to be higher for larger classes; whether larger classes are given to more
experienced (or possibly better) teachers; and the views of pupils themselves (whether
they feel happier, believe they are less likely to be bullied and are more confident about
speaking up for themselves and participating in practical activities). Other more general
issues have also to be taken into account: the relationship between class size, teaching
2See F. Mosteller, R.J. Light and J.A. Sachs, Sustained inquiry in education: Lessons from skill
grouping and class size, Harvard Educational Review 66 (1996): 797842.
3Goldstein and Blatchford draw attention to several technical problems, which may arise because
researchers have ignored the problematic aspects of measuring or defining certain concepts such as:
the sample population may differ from the target population; reduction of class sizes within a large
school may not be the same as an equivalent change in a small school; the institutions or populations
which are most accessible for study are often atypical; a design where randomisation occurs only at
the school level may not be representative of the real world where typically differential sizes do
exist within schools; teachers may alter their style of teaching (they might tend to use more wholeclass teaching methods and concentrate more on a narrower range of basic topics), and consequently
compensate in a number of ways with larger classes. See H. Goldstein and P. Blatchford, Class size
and educational achievement: a review of methodology with particular reference to study design,
British Educational Research Journal 24 (1998): 255268.


P. Smeyers

methods and the age of the pupils. Therefore, it is suggested that the effects of class size
may be different at various ages, a matter which will interact again with the kinds of
teaching and instruction that are offered. For instance, it could be the case that class-size
reductions will be more effective in the first years of school when children are more
dependent on adult help, whereas peer tutoring and computer-assisted learning are likely
to be more effective once pupils have been in school for a few years; it may also be the
case that reduced class sizes can prevent problems but are not sufficient to remedy
problems later on, and so forth.
Such studies also give rise to comments concerning this kind of research (randomised
field trials) itself. First of all, it seems that the benefits of reducing class size are determined in terms of factors (independent and dependent) that can be measured and manipulated in their constituent parts. What does not fit into this experimental pattern is mostly
simply left out, in any case in the experimental design (such as the well-being of pupils,
and teacher workloads). It is true that most of the researchers working in this area accept
that the higher cost of smaller classes is a relevant consideration; however, they are much
more concerned with establishing whether or not there is an effect than with considering
the strength of the effect that would justify higher spending on education. This generates a
picture which suggests that once the facts have been determined, the conclusion (i.e. to
decrease class size or not) follows on of its own accord. Second, it is difficult to see how
long-term studies can accommodate situational/historical change. It is not only impossible
to foresee which new elements have to be taken into account, but what is ignored is the
different elements which, in their interaction with each other, create something new (which
is not just the result of addition or subtraction of variables seen as factors). Problems of
discipline for instance may disrupt the interactions to such an extent that regularly
observed relations between variables no longer hold. Third, and less technically but
perhaps even more importantly, the favoured design seems to ignore the fact that teachers
deal with class situations (or learning situations) in a creative manner. It comes as no
surprise to find in many studies that it is not so much class size that is important, but the
way the teacher deals with it, i.e. varies his/her teaching to accommodate optimal student
learning. Teachers will look for opportunities for students to learn and thus act more in the
spirit of making the most of it, rather than carefully following regularities or causal inferences. They realise that there are many roads to Rome, and also that it may not be the only
place worth going to. All three of these conclusions could be seen as strengthening the case
for a more holistic approach, where the relation of the elements that are involved is given a
more prominent place. It seems that in educational contexts it is not so much factors or
elements that have to be studied as such, but the complex relationships between them. Here
the presence or absence of something may change the whole picture and, consequently, the
conclusions that can be drawn from a particular setting. Yet from the position that is generally embraced, such studies are seen as irrelevant due to their lack of potential for generalisation. Quantitative empirical research belongs to the paradigm of causality, which cannot
(or only at great pains and by changing the meaning of causality, i.e. by incorporating
reasons) give a place to the reasons human beings invoke for doing what they are doing.
Or, it is so piecemeal that it is hardly relevant given all other kinds of factors. Does this
rule out experimental or even empirical research? For some that is the conclusion, but this
seems wrong to me. I will return to this later.
There is of course a strand of criticism against the use of cause, of determinism but
even of indeterminism (as a descriptive statistical category) in the sphere of human
explanations. This is not to deny that human beings are exempted from causal processes
generally, but that behaviour itself can exhaustively be made clear in such a way. It is

Paedagogica Historica


argued that human beings give meaning to their life (among others by Ricoeur,4 Gadamer,5
Wittgenstein,6 Winch7 and Taylor8), to be understood as something that is different from
things which just happen to them (understood as the law-like explanations and predictions
of the natural sciences) and that research should focus on this. For example, for some (e.g.
Winch) to understand human conduct comes down to comprehending that the reasons for
ones actions and the understanding that is offered should be of the same kind as the understanding involved in the practice in question (using descriptions in terms of everyday
language, often by verbatim expressions of the practitioners themselves). This does not
imply that technical concepts cannot be used, but if so their meaning will rely on everyday
language. Various qualitative methods and techniques have originated from this interest
including case studies, participatory observation, interviews, analysis of policy documents,
content analysis and so on and so forth; even a technical vocabulary has developed which
includes terms such as horizontal and vertical analysis, thick concepts and triangulation.
The researcher brings to the forefront what was as yet not fully realised by the participants
or she may re-conceptualise the problem through her interpretation and in this way solve
the problem. As in many cases such as multicultural issues, feminist research but also
studies about teachers and teaching, narrative data are gathered through interviews and selfdescriptions for instance.
A closer look at a typical example of qualitative research may be helpful to understand
the nature of this kind of study. Some time ago the professional development of Belgian
primary school teachers was investigated.9 The study deals with recent graduates who for
the first time have their own class. It is an important year for their professional development
conceived as a lifelong and complex process of learning, which leads to qualitative changes
in the way they act and function, i.e. interact with the professional context. Obviously, they
still have a lot to learn, most importantly matters which could not be taught at college. The
leading research question of the study is: How does the personal frame of reference and their
micro-political learnedness evolve and change during the first year of their career? A
qualitative research approach was chosen, more particularly biographical research. After a
questionnaire collecting some background information, there was an in-take interview and
three more interviews with in total eight primary school teachers. The interviews were
recorded and typed, subdivided into fragments and then coded to prepare for a content analysis. A synthesis of the interview was composed and presented for final approval to each of
the teachers. All the time two researchers were involved who checked each others interpretations, for instance concerning the codes that were given and the syntheses that were made.
A so-called vertical (or in-depth) analysis was conducted for each teacher and a horizontal
for the different aspects across the respondents. Again checks were made between the
researchers who moreover kept a logbook of everything that they did or that happened
within the context of this one-and-a-half-year period in which the data were gathered. The
4P. Ricoeur, Hermeneutics and the Human Sciences: Essays on Language, Action and
Interpretation, ed. and trans. J.B. Thompson (Cambridge: Cambridge University Press, 1981).
5H.G. Gadamer, Truth and Method, trans. G. Barden (London: Sheed & Ward, 1975).
6L. Wittgenstein, Philosophical Investigations/Philosophische Untersuchungen, trans. G.E.M.
Anscombe (Oxford: Basil Blackwell, 1953).
7P. Winch, The Idea of a Social Science (London: Routledge & Kegan Paul, 1958).
8C. Taylor, Philosophical Papers (Cambridge: Cambridge University Press, 1985), 2 vols.
9S. Mahieu and R. Vanderlinde, De professionele ontwikkeling van beginnende leerkrachten vanuit
het micropolitiek perspectief [The professional development of recently graduated teachers from the
perspective of micro politics] (Unpublished masters thesis, Katholieke Universiteit Leuven,
Belgium, 2002).


P. Smeyers

researchers were conscious of the part they played themselves in doing this research, which
could also influence their findings. They tried to be as honest as they could and as methodologically pure as possible. The result is a piece of research that is detailed, sophisticated,
aware of what could go wrong and constantly involved in anticipating possible future
methodological problems. Because it uses the lived experience of the teachers as expressed
in the interviews it has a high degree of what is technically labelled validity, concerning
what is studied. As indicated, the researchers were anxious also to have reliable data and
constantly monitored the quality of what they were doing themselves.
So much for the research question, the design and the method that was followed. I will
now briefly deal with some of the results evidently a selection has to be made in which I
focus on the elements that were used in the comparative analysis of the syntheses (the
horizontal analysis). It was concluded that: the first year is a year of intense learning, particularly relational being responsible for her/his own class, the teacher needs to develop
relationships with other actors within the school, this has its own problems, is a big challenge and an important task (teachers say they know too little about this on the basis of their
own college education); when they applied for a job teachers really felt the micro-politics
of reality, the power of the local school governors; they all regard it important that there is
a colleague who is in charge of a parallel class (someone they can or would like to rely on),
and they also mention the crucial importance of their dealings with parents; the relation with
the head too was felt to be very important, and so was the experience of functioning in a
team; on a more basic level they were all concerned with securing a job the following year,
and furthermore they all experienced the importance of practical things in the context of the
day-to-day functioning in the school. There are a few more results the researchers report,
but let me end with a final and interesting conclusion: All teachers experienced themselves
as insecure, badly prepared, inexperienced and not fully qualified.
I have deliberately gone into some detail about this research and have tried to give a
good picture of it. Moreover, I want to acknowledge that in my opinion this research has
been carried out according to the highest standards appropriate for qualitative research. Yet,
one cannot but wonder whether what was found, i.e. what we know now, after so much
work and with such evidence of expertise, is really something that was worth the effort. In
what sense is it different from what one knew beforehand? But even if there is something
we would not have come up with in our armchair thinking, is this kind of knowledge useful
then, and for whom, in what circumstances, and to what extent? Surely, if one is not familiar
with teaching and teachers all of this will open up the sphere of the experiences of primary
school teachers during their first year of teaching. But would not most people who are teaching in a school be able to come to the same conclusions without engaging in this kind of
research? And in so far as detailed information is concerned, this will necessarily be different from case to case. So there is not much point in gathering this with a view to using the
results for other settings. I do not disagree that the insights and the well-written research
report may be interesting for heads, and even for lecturers teaching in teacher training
colleges, but let me be blunt about this: Is it really more than truisms, more than common
sense that draws on a particular context the researchers have come up with? Their general
conclusions are very similar to the relationship between speeding and accidents, drinking
and driving, being depressive and suicidal, and so on and so forth. And though there is value
in seeking empirical evidence for so-called generally held beliefs, it does not go very far.
Of course, sometimes people need to be reminded of these general facts of human nature,
but one may wonder whether and in what sense this kind of research is helpful at all. There
are also more fundamental criticisms of this kind of research. The last decades have seen
a growing interest in qualitative research and recently narrative analysis has become very

Paedagogica Historica


popular. This has generated heated discussions between the quantitative and qualitative
camps, but also more in general among educational researchers. The question is whether the
use of stories in narrative analysis is reconcilable with the focus of generalisation characteristic of most of the research. According to Phillips10 epistemology is being swept away and
replaced by politics. Since the acceptance of a narrative can have consequences, the narrative ought to be true, features like an enticing plot, adequacy or plausibility are not
epistemically relevant. Carter and Doyle11 also ask whether students personal narratives
of past educational experiences can be the core of the curriculum in their own training. It
seems that in general, although qualitative educational research aspires to true knowledge,
its insights are seriously questioned.
The conceptual framework and the value-ladenness of educational research: the
reconstruction of the researcher
Educational researchers are interested in how things are (what the facts are, how those
who are involved feel about particular things), and in this sense they are interested to understand what they are presented with. In some areas this implies descriptions or reconstructions of the participants experiences, in others being able to make predictions. It is clear
that this presupposes a particular conceptual framework (sometimes also a theory) or at least
a set of concepts in order to make sense of the multitude of phenomena one is confronted
with. Clearly, it is generally accepted that one is part of an intersubjective reality that may
be characterised in various ways (what is considered to be a fact, what we value, how we
situate ourselves as human beings). But research is at least potentially also nearly always
interested in change, in making improvements (either to prevent particular problems or to
address these). Thus it is interested in understanding ways to manipulate certain elements
with a view to certain outcomes. In this sense the value-ladenness and maybe even the
utopian dimension (how one could conceive of things differently) come unavoidably
forward. This presses the point about the nature of what the researcher is really doing (or is
allowed to do or should have to do).
The distinction made by Polkinghorne12 between an analysis of narratives and narrative analysis is particularly illuminating. In an analysis of narratives one looks for
common features in different cases in order to define them within a broader category. By
pointing at features that different experiences have in common, one can construct cognitive
conceptual frameworks. The purpose of the paradigmatic analysis is not only to discover
and to describe categories, but also to describe the relationships between categories. In
many cases this kind of research is generally analogous to a quantitative design (including
hypotheses), with the exception that qualitative data are gathered, i.e. they refer to what
people feel about, or what their experience is with, particular things, what they say that their
reasons, desires and intentions are. In narrative analysis, on the other hand, the data are
mostly not in a narrative form. The information comes from different sources: the researcher
arranges events and actions by showing how they contribute to the evolution of a plot. The
10D.C. Phillips, Gone with the wind? Evidence, rigor and warrants in educational research, in
Papers of the Annual Conference of the Philosophy of Education Society of Great Britain, ed.
J. Tooley (Oxford: Philosophy of Education Society of Great Britain, 1993).
11K. Carter and W. Doyle, Personal narrative and life history in learning to teach, in Handbook of
Research on Teacher Education. Second Edition, ed. J. Sikula et al. (New York: Macmillan, 1996),
12D. Polkinghorne, Narrative configuration in qualitative analysis, International Journal of
Qualitative Studies in Education 8 (1995): 523.


P. Smeyers

plot is the thematic line of the narrative, the narrative structure that shows how different
events contribute to a narrative. The writing of it involves an analytical development, a
dialectic between the data and the plot. The resulting narrative must not only fit the data but
also bring out an order and a significance not apparent in the data as such. The result is not
so much an account of the actual happening of events from an objective (i.e. something we
agree about intersubjectively) point of view as the result of a series of constructions; it is
instead a particular reconstruction of that researcher. Where in the analysis of narratives
the narratives (gathered from the participants) are the source of knowledge, the narrative in
narrative analysis is the result of the research, i.e. the creation or interpretation the
researcher comes up with. Not only in the conclusion that is offered is the researcher
present, but she is involved all through the process (though different compared with the
practitioners involvement). This kind of interpretive research comes close to those areas
of scholarship (see above) that were distinguished from educational research grounded in
the empirical traditions of the social sciences. It seems analogous to history and philosophy
of education where, in other words, an interpretation is offered.
If it is accepted that to study educational problems one needs not only a quantitative
approach but also one or other kind of qualitative stance, it is not clear whether the kind of
qualitative research that is merely a use of qualitative data within an overall quantitative
design is not contradicting its own presuppositions (because it is likely to betray the holistic
nature of the meaning-giving process as a consequence of generalisation). The question
can be raised whether, in other words, it is not trying to do something which cannot be done,
at least in so far as it accepts seriously that one should not strip words of any context in
which they might be used for saying something in particular. Second, and perhaps even
more important, is not all empirical educational research guilty of trying to help to escape
from the particularities of a situation one finds oneself in, which closes one off from being
responsive to the situation one finds oneself in? Granted, there is a sense in which knowing
certain facts, being acquainted with the experiences people are likely to have in certain situations, is helpful to understand educational problems. But the issue is how far this type of
research goes, in other words whether it offers more than a starting point. What occupies
me is the nature of research that does justice to the particularities of the situation and what
its characteristics are. These reflections on the nature of empirical educational research are
highly relevant in a time when research is used as a quality label almost equivalent to
sound thinking, and experts are solicited to give advice on all kinds of issues belonging
to the educational context.
What we seem to need is an idea of research that gives up a number of the old distinctions such as values/facts, objective/engaged, researcher/practitioner, concept/fact and qualitative/quantitative/interpretive. Not that these issues do not matter, but they seem to be
dated given certain developments in philosophy itself. The disdain directed at some of the
achievements of so-called postmodern philosophy is hardly an issue here. I will therefore
characterise educational research differently and avoid the so-called philosophical/empirical
dichotomy. So if we start afresh, what is it that we should bear in mind when we gather data
that are relevant and come up with insights that are valuable for theory and practice, again
not to be seen as different domains, but rather as different ways of dealing with the problems
envisaged? Many of the aforementioned oppositions or dichotomies tend, unfortunately, to
lead a life of their own. Instead of being helpful they obscure the full picture of what is at
stake. A more fruitful metaphor may be to see these as two sides of the same coin. The
fairytale that it is possible to isolate (observable) factors and still deal with something that
is relevant in future cases should once and for all be buried. The most a researcher can come
up with is a new angle that might be helpful for particular problems. And it is not clear

Paedagogica Historica


beforehand which problems may be solved and which others may be generated through
particular interventions. Moreover, this is no different for a philosopher than it is for a socalled empirical researcher.
Overcoming the dichotomy of reasons and causes
It is not so important to identify the fact that there are many factors at work. Rather, we
should consider which of these factors are relevant and the extent of that relevance. It is
precisely here that feminism and postmodernism can play their part and put a finger in the
wound. Empirical research may claim to deal with multiculturalism but, to reverse this
scenario, a different stance may equally find a place for particular kinds of empirical
research. What we do as researchers is constitutive of the reality, which we shape together
and which is unavoidably value-laden. If empirical research does not rise above the level of
description, there is a danger it will degenerate into a kind of empiricism. It often plays it
too safe and engenders more of the same the Matthew effect, more details of what is in
the end irrelevant. Instead, to make real progress empirical research should take risks and
play a more imaginative, possibly dangerous game. Evidently judgement goes far beyond
this, involving experience and the development of wisdom through participation in a
community of practice (practitioners and scholars). Weighing the evidence and foremost
arguments, i.e. judgement as Richard Smith13 uses that term, may seem more crucial than
so-called empirically based know-how.
In the background of what we do there are always normative or value-laden elements
operative, besides distinctions between the meanings of concepts, which constitute the frame
of reference of our activities, i.e. how we understand human reality. It has been the aspiration
of science to bracket or eliminate these, and thus to be able to deal with reality as it really
is. The pitfalls of this positivist stance have for long become clear. Our move should therefore be in the opposite direction, that is in accepting that concepts, theories, reasons etc.
always presuppose a background in order to make sense. It is similar but broader than the
position which argues that all sciences are theory-laden. If this is correct there is no longer
a need to rule out causes or observed regularities when explaining human action. Do they
too not presuppose a meaningful context? To give a causal explanation of human behaviour,
then, only refers to the fact that it is described in certain terms, in the same sense as an explanation in terms of reasons presupposes a background of shared understanding. Some human
actions may thus be characterised in terms of causes and effects, but it may also be possible
to give descriptions in terms of regularities (how antecedent variables go together with
subsequent conditions) or to refer to reasons. Some activities may almost exclusively be
understood by using one type of explanation, while in other cases several will be possible.
Thus whether something is really explained, or whether reality here is merely a matter of
not being fictitious, should not necessarily invoke a correspondence theory of truth where
sense data are the exclusive building-blocks. Instead, as Winch rightly argued, it is always
about what is real for us. It goes without saying that answering a research question in terms
of causes and effects will not generate an answer in terms of the understanding of those
involved. But this kind of circularity is not to be regretted as it is characteristic of all explanation. Science, like for that matter any kind of explanation, will always take the data which
are to be interpreted to the next higher step of abstraction, thus invoking a particular theo13R. Smith, Technical difficulties: The workings of practical judgement, in Educational
Research: Why What works Doesnt Work, ed. P. Smeyers and M. Depaepe (Dordrecht: Springer,
2006), 159170.


P. Smeyers

retical construction which makes sense. This is a circular process in which each level is taken
to account for, to derive from or to elaborate on the other. Thus instances are explained by
patterns and patterns by instances. Clearly, here it is not prediction that may exclusively
provide us with a point of reference, nor is the method of the natural sciences the only way
to come to valid conclusions. But even if the possibility of prediction is what one is interested
in, even then a meaningful background cannot be absent. How could it possibly be doubted
that we always start from making distinctions in terms of what makes sense for us? For
example and at a slightly higher stage of abstraction, there seems at first sight to be a socalled objective point of reference in medicine (being healthy, living longer etc.). But it is
not only the case that surgery or medication can have side effects which may prompt one to
give up on a particular treatment; there is also the debate about quality of life which plays
a role in this decision process. And, as is clear for instance from the area of environmental
ethics, there are in many cases conflicting interests which have to be dealt with. We have to
decide for example whether to safeguard a particular wildlife area or to build in that location
a new airport, thus relieving thousands of people of aircraft noise.
An adequate methodology of the social sciences should therefore combine causal explanation with intentional understanding. Following Bohman14 I too accept that hermeneutic
philosophers of social science he mentions Taylor and MacIntyre, but the same goes for
Winch are in danger of ignoring this due to their one-sided arguments, ruling out every
kind of causal explanation in this context, and also due to their claim that all explanations
are supervenient on actors own interpretation of their actions. Understanding the reality we
live in demands an understanding of nature itself but also of the kind of beings we are. And
if one accepts that human beings are of a different kind than physical objects, it follows that
they have to be studied at least partly in a different manner. But Winch too seems to go to
the other extreme. He does not deny that human beings are also bodies (in Merleau-Pontys
sense), but in so far as they are human, this aspect does not seem to count for much. He
attacks the Humean view of causality as no more than constant conjunction, a view that
leads to a neo-positivist sociology where the embeddedness of what we do, in a horizon of
inherited meaning for instance how we understand ourselves vanishes. But if there are
reasons to accept that human behaviour may at least partly be explained in terms of the
regularities one observes and grants for these other techniques and methods, the problems
with seeing human behaviour as dualistic, either in causal terms or in terms of reasons,
should be overcome. Thus Bohman argues that macrosociological and functionalist
explanations supervene upon actors intentions, a thesis which I believe is very plausible.
A pertinent example of this might be a university climate in which constraints of a macroinstitutional nature may be observed on scientific production.
Thus a place has been found for instrumental behaviour but it has also become clear how
different methods or ways of explanation can be used. It goes without saying that in the end
the explanation that is offered will have to be subjected to the community of researchers and
scholars but also to that of practitioners, to be judged on the basis of whether the position
argued for makes sense to say. And clearly it is at least in principle always possible to
offer another explanation referring to reasons, causes or regularities and to apply theoretical
insights from philosophy, history, aesthetics etc. But the fact remains that the acceptability
of it will rely heavily on whether others can agree with this particular interpretation.
Educational research should come to terms with Winchs suggestion that it must be
philosophical in character. Here philosophy is not just concerned with eliminating linguistic
14J. Bohman, Pluralism, indeterminacy and the social sciences: Reply to Ingram and Meehan,
Human Studies. A Journal for Philosophy and the Social Sciences 97 (1997): 441458.

Paedagogica Historica


confusions, although there is a proper place for conceptual analysis, but is concerned
foremost with what makes sense for us. This implies that the starting point is the selfunderstanding of those involved: educators and parents, students and children. It does not
make sense therefore to break down teaching into atomic skills and sub-skills; nor to conceive
of education almost entirely in terms of examination results; nor to regard moral psychology
as a series of points on the scale of self-esteem; nor to talk of parenting in terms of developing
particular skills and quality time to empower children. In empirical educational research
nowadays there is typically a starting point (topic, research question etc.), then a procedure
(the gathering of data, etc.) and then an analysis and discussion. A lot of attention is paid to
the middle stage, concealing the fact that the really big questions concern the values at the
heart of, or taken for granted in, the identification of the topic and the question, and of course
in the ensuing discussion. Winchs position implies that the discussion has to start from a
particular social intercourse or practice. It follows that the empirical observational methods
(and statistical techniques) cannot possibly be the only yardstick. Instead, the human situatedness of the phenomena being researched requires that all our observations, arguments and
considerations must be based in our practices. Normative and value-laden elements have to
play a crucial role throughout educational research and not just in the first or final stages.
Pre-understandings, ethical sensibility and language: interpretative research
In trying to be objective, and in identifying objective with free of bias the fact is
concealed that we always and inevitably bring our pre-understandings with us into any situation. There is a standing temptation to embrace the notion of objectivity that consists in
imagining that the explanations of phenomena are waiting out there for us to discover
with our instruments and techniques. There is no need to embrace the claims of a crude
constructivism to the effect that all meaning is created ex nihilo, but only to argue that
whenever we conceptualise a particular part of reality, this necessarily occurs within the
boundaries of what already makes sense for us. Ideas about what is worthwhile, about the
nature of a human being, necessarily enter into the picture. We are no longer aware of things
in terms of their multifarious meanings and interconnections, but isolate them in order to
study their reality and then worry about how they are linked and connected to each other.
We do this because we are unwilling to live with complexity. To replace scientism and
doctrinaire empiricism with a more modest view of science is perhaps the first step towards
wisdom. Of course this does not mean that scientific method has no role at all to play within
social science, but that it must make out the case for its relevance in each particular instance
against other approaches that also offer insight and understanding, whether in conjunction
with measurement and statistics or apart from them.
The conceptualisation of social (and political) problems demands an ever-renewed
rethinking of reality with similar instruments. To think again can only mean to think from
a different point of view what one is trying to understand (perhaps change). It will be clear
that an investigation of what exists is only a starting-point. What is at stake shifts to what is
at stake for someone (again for the other and for myself), where the other is recognised in
her personal struggle as an emotional being unstructured justice. Rigid approaches to
social (and political) problems will have to be complemented by a more flexible ethical
sensibility. To see the other is to look for the way in which she expresses herself, gives
shape to herself in the struggle with herself. But to touch the other is also to confront the
other with ones own struggle by means of the evocative instruments which are at my
disposal. That we inevitably violate the other is clear enough. After all, the understanding
of the other is at the same time a negation and a constitutive affirmation. We understand the


P. Smeyers

other as an intentional object which we crave to understand. We want to read the story of
the other, too often without recognising the illegibility of her story. This does not necessarily imply that we would not be able to understand her or do not want to do justice to her.
The reading of the story of the other is, however, at the same time a reading which interferes
with my own story. What rests for us is to surrender to the intersection of this reading with
its reader, and to what this does to us.
By investigating how the subject is part of the intersubjective level Wittgenstein shows
the value of our freedom, of our autonomy, not as the exercise of an arbitrary choice, but as
the result of the way in which nature as well as artistic products and moral responsibilities
are taken seriously and even seen as necessary. He mocks those who are seduced by the
promise to be able to control the cultural and who think they are able to represent our
thoughts and concepts as necessary. To write is for him to surrender to certain readings
(words seen as what is given to us) and philosophy as the result of a play of reading and
writing on the basis of ones own authority. Or we are able to rethink a thought that comes
up on our way, to possess it and to judge it, or we have to let it go, it belongs not to us. We
find ourselves and in the answer to the way we see ourselves we find a place to begin.
Cavell15 refers to the consolations of the word; to this meaning for the other; as a song; as
sharing in the case of food and drink, to have in some sense the same experience. Here,
to write becomes a means to fight the struggle with oneself (with ones own language), and
poetry a means to make a bridge. The written word, the poem is a weapon in this struggle.
It requires no other material presence; it does not want to explain; it only suggests seeing
things in a particular way. It seems a means to be at home for a moment, for the lonely
individual, for the subject-with-the-others.
In a paper called The Investigations: Everyday Aesthetics of Itself16 Stanley Cavell
argues that Wittgenstein claims for the ordinary its own possibility of perspicuousness. This
ordinary does not refer to the everyday, but it insists on giving up the disappointment
with criteria. Hence the importance of the literary, which of all forms of texts have in
common that they do not only use language to express certain contents, but also direct
readers attention to the workings of language. Moreover, by telling stories novels may allow
us to describe uncommon situations and develop a new perspective on everyday situations.
Thus they provide the context necessary for exploring not only the grammar of our language,
but also the limits of our form of life. Poets may provide short, carefully crafted texts that
are particularly apt for minute and acute analyses and critique of single expressions and their
roles in language. Thus new metaphors may shed light on the limitations of ordinary
language to express certain situations or feelings.
Thinking about the nature of a story, not only in educational research as the raw data one
starts from but of educational research itself, may be a way to do justice to the study of
education. On the one hand a story can be conceived as what joins people together, on the
other hand as what can only show. Here education may be seen as what starts from an
initiation into what is groundless. An educator and a student of this area can do no more
than give expression to her story, and appeal to the educandus by what she holds to be valuable and constructive, taking into account the childs or students response and thus being
responsive to her. If educational research can be heterogenous and produce different (kinds
of) results and moreover can be presented in various ways, if different stories can be told,
this will eventually endorse the classical insight that education is about instilling and evoking a good disposition, and educational research as another way to express this showing,

Cavell, In Quest of the Ordinary (Chicago: University of Chicago Press, 1988).

J. Gibson and W. Huemer, eds, The Literary Wittgenstein (London: Routledge, 2004).

Paedagogica Historica


as a mode of the will to join in this kind of dialogical speaking and doing. But this implies
understanding educational research no longer as research along the lines of the positions
described above as quantitative or quantitative using qualitative data, but opting radically
for an interpretative stance.
Indeed, the fateful mistake would therefore be to attempt to say something about what
must be happening in us when we see (or otherwise experience) one thing or another, to be
so and so. This characterises an attitude not even of keepers of truth but of registrars of truth,
something that we cannot become without losing ourselves, as Baz claims.17 This is, as
Cavell holds, not to deny the fact and the significance of the background of pervasive and
systematic agreements among us, but for him seeing the point of our own words or the
words of others is just as much a condition of human speech. Following Cavell, Baz argues:
And when one of those rare occasions does occur, and something speaks to us, what speaks to
us is never simply those objective properties that anyone who speaks the language and has eyes
in her or his head can see. What speaks to us in things always goes beyond, comes from
beyond, what anyone can just plainly see. These would be the moments in which it might make
sense for us to give expression to what we see, for no other reason but that it is true. What we
then would say might perhaps be unwarranted there would be no obvious way for convincing
the other of its truth, or for fitting it into a world-view. But in spite of that, and in fact precisely
by virtue of that, it would be an expression of our faithfulness to what we see. 18

Cavell reminds us that saying something in particular has its conditions. Only by acknowledging that, can we avoid overlooking the question of the point of what we say. Following
Cavell, Baz argues that what we say cannot be specified independently of why we say it.
And he continues, if Cavell is correct in his recurrent reminders to the effect that the
intelligibility of our words is a matter of their being found to be worth saying here and
now in this way then whoever wishes to understand our experience of the world on the
model of saying something about the world would have to take into account the question
of what calls for the words with which we give voice to our experience of the world.19 To
say that educational research has to be interpretive echoes Mulhall claim that interpreting
things into practical life has no distinctive structure or principles because it is fundamentally
not based on the following of some pre-given set of rules; it depends upon imagination, the
ability to see connections, the creative shaping of ones sense of how aspects of human
experience hang together or fail to do so.20
Educational research: joining the ongoing debate
Affeldt21 argues that we should be hesitant about the use of criteria because they may be
seen as marks and features which tell us when the application of a concept is licensed, and
by articulating grammatical relations among our concepts tell us what exactly we have
said in any particular instance of applying a concept. They should not be seen as determining, that is, what our concepts mean and to what else we (must) have committed ourselves
17A. Baz, On when words are called for: Cavell, McDowell, and the wording of the world,
Inquiry 46: 478479.
18Baz, On when words are called for: Cavell, McDowell, and the wording of the world, 496.
19Baz, On when words are called for: Cavell, McDowell, and the wording of the world, 476.
20S. Mulhall, Misplacing freedom, displacing the imagination: Cavell and Murdoch on the fact/
value distinction, Philosophy 47 (Suppl. Vol., 2000): 264.
21S. Affeldt, The ground of mutuality: criteria, judgment, and intelligibility in Stephen Mulhall and
Stanley Cavell, European Journal of Philosophy 6 (1998): 31.


P. Smeyers

or made ourselves responsible in employing a particular concept.22 In a similar vein content

analysis of interview protocols, horizontal and vertical analyses of case studies, and the use
of observation categories (stipulating a neutral description in behavioural terms) diffuse us
from the particular in the name of objectivity and generalisability. There is something what
may be called a first-person aspect that is present in the story the researcher presents. Yet
in another sense he seems to claim to speak for others. His endeavour therefore involves a
peculiar mixture of self-reliance and vulnerability. What he does is only reminding his audience of the very possibility of a different scenario.
Gustafson interprets Cavell as arguing that the epistemologist starts from familiar
language and makes a projection, i.e. in importing words into not so familiar contexts. The
pervasive significance of our unregularisable projective imagination manifests, he says, the
extent to which keeping language alive and the world in view is not a matter of passive
conformity, but a continuous undertaking which requires the employment of those interests,
feelings, modes of response, sense of humor and significance and fulfilment. This undertaking is our task, as language using creatures; we are burdened with this responsibility. It
cannot be transferred, he says, to a machinery of rules the application of which is fixed independently of human modes of response.23 If the researcher claims to sense something others
fail to see, there may be no agreed procedure by means of which the issue can be solved.
One cannot decide in advance what projections are tolerable. But this is not different from
how we are situated in a moral debate. Mulhall argues that to present your position in a
moral debate is to present it by defining your sense of its relation to other positions, to place
yourself in a particular space of moral options, and so to place yourself in relation to those
who would plot that space differently or would have you place yourself differently in it. And
he continues:
But this placing is yours alone to do: the logic of moral argument offers no impersonal background on to which ones responsibility not only for the choices one makes but for the range
of choices one regards as available can be sloughed off. What it does provide, however, is the
possibility of accounting for ones choices, by engaging in modes of explanation and defence
which not only make reasoned agreement on that choice a real possibility, but also ensure that
a sense of mutual respect, of mutual moral intelligibility, might survive eventual disagreement
over the rectitude of a given choice.24

Such a kind of respect should characterise educational research which aspires to do justice
to the nature of education. It leads inescapably to the interpretive kind.
I will conclude with a number of characteristics that in my opinion mark the nature of
educational research. In the study of educational phenomena various modes of explanation
may find their place in trying to understand what is involved in teaching pupils and students,
in child-rearing, in continuing education and in educational policy and evaluation and so on.
There is indeed no need for a single method or to prioritise one; much will depend on the
problem that is studied, but also on the kind of theoretical interest one is pursuing. Since
various modes of life are involved (the religious, the social, etc.) and various principles (for
example rights and freedoms) must be taken into account, an educational context has to be
dealt with from a variety of theoretical stances. This moreover points to something outlined
22Affeldt, The ground of mutuality: Criteria, judgment, and intelligibility in Stephen Mulhall and
Stanley Cavell, 5.
23M. Gustafson, Perfect pitch and Austinian examples: Cavell, McDowell, Wittgenstein, and the
philosophical significance of ordinary language, Inquiry 48 (2005): 377.
24Mulhall, Misplacing freedom, displacing the imagination: Cavell and Murdoch on the fact/value
distinction, 272.

Paedagogica Historica


at various points in this paper: that social research does not give us fixed and universal
knowledge of the social world as such, but rather that it contributes to the task of improving
upon our practical knowledge of ongoing social life. That this presupposes dialogue
between all those involved goes without saying. But when we realise that there are many
and often highly contested versions of participants self-interpretation, we will also see that
though the latter are the only plausible starting place, more is needed for good dialogical
and social scientific practice. Here science is no longer seen as disinterested and value-free:
instead there do not seem to be strict boundaries between science and society. In her
contribution the researcher, the interpretive pluralist, will among other things explore the
operation of many different practical norms, thus through her interpretation making implicit
norms explicit. But she will also and in my opinion necessarily invoke a normative stance.
Here facts are no longer seen as exclusively made to refer to objective things in the world
or things in themselves; neither are values seen as subjective states of the mind. In avoiding
these and other conceptual confusions science reveals itself instead as a performative intervention. Though the researchers work is in this sense also of a political nature, it does not
coincide with that of the practitioner or the politician. The writing of research may be seen
as what prevents us from being absorbed in the chaos of unmediated complexity. It allows
us time to think and is performed at some distance in the interests of perspective and justice.
Clearly, the characterisation of educational research as interpretive resembles the stance
that philosophy and history of education embrace, without, though, coinciding with it. Starting from the life-world it transgresses the hypothesis testing design and makes sense of
all the aspects of what is relevant in and for a particular case. Thus, and only thus, is it truly
Notes on contributor
Paul Smeyers is Research Professor of Philosophy of Education at the University of Ghent and parttime at the Katholieke Universiteit Leuven, both in Belgium. He taught philosophy of education and
methodology of the Geisteswissenschaften (Qualitative Research Methods). He holds, or has held
several positions in the International Network of Philosophers of Education (President since 2006).
He chairs the Research Community Philosophy and History of the Discipline of Education:
Evaluation and Evolution of the Criteria for Educational Research established by the Research
Foundation Flanders, Belgium (Fonds voor Wetenschappelijk Onderzoek Vlaanderen). Together
with Nigel Blake, Richard Smith and Paul Standish he co-authored Thinking Again. Education after
Postmodernism (1998), Education in an Age of Nihilism (2000), and The Therapy of Education
(2007) and co-edited The Blackwell Guide to the Philosophy of Education (2003). Together with
Marc Depaepe he co-edited Beyond Empiricism: On Criteria for Educational Research (2003),
Educational Research: Why What works doesnt work (2006), Educational Research: Networks and
Technologies (2007), and Educational Research: The Educationalisation of Social Problems (2008).
With Michael Peters and Nick Burbules he co-authored Showing and Doing: Wittgenstein as a
Pedagogical Philosopher (2008).