Madrid, D. (2000): “Observation and research in the classroom”,
en Teaching English as a Foreign Language, Barcelona: The Australian
Institute. pp. 1-100.
INTRODUCCIÓN A LA INVESTIGACIÓN EN EL AULA
DE LENGUA EXTRANJERA
Daniel Madrid
(Universidad de Granada)
INTRODUCCIÓN
In this unit we will be examining how and why
teachers should be encouraged to reflect and analyse their own teaching and the classroom context. We will look at how
the notion of research in ELT has developed, and how a new paradigm of
reflective, classroom- based research has evolved.
It is assumed
that teachers can improve their knowledge and skills, not necessarily by
following a particular method or the findings provided by “official” research,
but with an enquiring attitude which allows him/her to find out the most
adequate techniques in each situation. Because each teaching situation is
different and needs to take a range of factors into account, from individual
students' characteristics to those of the teacher, the classroom context and
the community with its local educational and intellectual traditions. And every
language lesson is necessarily different and unique as it is jointly
constructed by the teacher and the learners (Allwright 1991).
Nowadays, the theory of constructivism is not
only applied to the learning processes, but also to the teaching situations. It
is widely accepted that each teacher
constructs his/her own 'theory' about teaching and learning. These ideas comes
from a variety of sources:
-
from
the teacher's personal beliefs and value system,
-
from
his/her experience as a learner,
-
from
training or education,
-
and
from his/her classroom practice and experience.
It is this latter area - that of classroom
practice - that we will be primarily focussing on in this unit. In fact, Stern
(1983) argues that thinking about classroom practice is an essential part of
learning to theorise. Thus a reflective approach to our teaching is very much a
product of the post-method condition of ELT (see Unit 5 for a discussion of the
end of method).
2.2 WHAT IS RESEARCH?
All of us often question ourselves about lots
of daily life events that we do not quite understand or seek answers to
questions that other people make. In this sense, we all become researchers of
our daily life when we try to find out explanations for the phenomena happening around us. In doing
this, we observe these events and analyse them. We make guesses and adventure
hypothesis that we later check till we can be sure that what we initially
assumed is true under certain conditions. So, the basic components of research
in our daily life can be summarised as
follows (see Seliger and Shohamy, 1989:7)
|
1.
There is a phenomenon that
is not clearly understood. 2.
To try to understand it, the phenomenon is observed and analysed. 3.
In this process we make
questions about the phenomenon: Why, how, when, ... does it happen?
What’s the connection with other events? Etc. 4.
Several guesses, answers and
possible explanations are anticipated intuitively. 5.
One (or some) of the anticipated explanations is/are considered that
most probably explain(s) the phenomenon (hypothesis). 6.
We check the hypothesis in other situations: collect more data and
test to what extent the hypothesis can be maintained. |
Even though we have seen a close connection
between scientific research and research in every day life, there are important
differences, as Seliger and Shohamy have pointed out (1989:10):
The differences between knowledge arrived at through common sense and
intuition on the one hand, and scientific research on the other, can be
expressed by concepts such as “organised”, “structured”, “methodical”,
“systematic”, “testable” and specifically by the notion of disciplinary
inquiry.
It is also important to differentiate the type
of research we decide to do:
a)
Basic or theoretical research: it aims to
construct abstract theoretical models which explain second language teaching
and learning.
b)
Applied research: it applies the theoretical models
provided by a) basic research to
different fields of study (e.g.
education).
c)
Practical research: it makes a practical utilization of
a) theoretical research and b) applied research. For example, by testing
classroom hypothesis, controlling
variables, providing pedagogical principles experimented in the classroom, etc.
It must be noticed that practical research
is based on the premises established by theoretical and applied research and it is influenced by them when
it is developed in classroom situations.
Classroom research is normally practical research which is centred on the classroom. It aims to
explains what actually happens inside the classroom, the direct and indirect influence of internal and external factors related to the student, the teacher
and the ELT curriculum. As van Lier points out (1988) we know very little about
what is going on in classrooms, so
classroom research becomes an important tool
to explain the relation between the diversity of variables that
continuously interact:
As yet we know too little about all the variables that play a role in
all the classrooms to be able to make rash recommendations about methods of
teaching and ways of learning (1988:7)
On
the other hand, classroom research has also been considered one of the most
difficult places to do research, hence its consideration as a “black box” as
far as the connection between the input received by the students and the output
they produce. In addition to that, and as we will see in the section about
research methods, the concept of research itself has not been unproblematic. Richterich, a Swiss scholar closely
involved in the Council of Europe Modern Languages Project (see Unit 4), and
himself involved in research, pointed out some of the piffalls involved:
Apart from these disadvantages, there was a
growing awareness among both researchers and teachers of the inadequacy of a
scientific, rationalist model of research, which we will examine later.
2.3 THE SCOPE OF CLASSROOM RESEARCH
When we aim to study what happens in our
English class we must take into consideration a great variety of variables that
may have a direct or indirect influence or relationship between one
another. According to van Lier
(1988:27) some of the central topics to research about are the following:
a)
The
nature and development of the students’
interlanguage.
b)
The
role of interaction between the
teacher and the students, that is, the type of communication which takes place
in the classroom
c)
The
use of learning strategies
d)
Cognitive styles
e)
Efficiency
of evaluation techniques
We think that a more complete picture of the
scope of classroom research is illustrated in
figure 1. Here, we present a model which includes the main groups of
variables that must be controlled in some way:
by qualitative, ethnographic and
naturalistic procedures or experimental
and quantitative ones (Madrid,
1995:60-62; 1998a, 1998b).
Fig. 1: A framework for L2
teaching analysis and research
PRESAGE AND CONTEXT VARIABLES
2. THE
TEACHER - Individual
characteristics - Beliefs,
preparation, personality 1. THE
LEARNER - Social
context -
Individual characteristics: cognitive styles, motivation, personality
PROCESS
VARIABLES
SLL
SLA
3. LEARNING
PROCESS ACQUISITION PROCESS - Mental
processes -
Learning strategies 4. THE
CLASSROOM - Educational
treatment - L2
curriculum - Tasks - Materials
PRODUCT
VARIABLES
5. RESULTS
/ ATTAINMENT / COMPETENCE / PROFICIENCY - Principles,
concepts: grammar,
functions, cultural aspects, ... - Procedures, skills: listening,
speaking, reading and writing - Attitudes and values
The model presented in this diagram
distinguishes five sets of variables that interact. At the top, are the learner and the teacher, the two main protagonists in the teaching and learning
processes.
If the learner acquires the L2
in a naturalistic setting, through communication in natural social situations,
without the teacher's guide and outside the classroom setting, then he/she
develops learning strategies and mental processes that lead to what is called second language acquisition (SLA)
(Krashen 1981). This sequence only involves the factors included in boxes 1, 4
& 5: the learner, his/her learning
process and his/her results.
An alternative sequence takes place
when the learner becomes a student
and is guided by a teacher in a
classroom formal setting. In this case the L2 learning takes place
through conscious study, with the help of the L2 teacher and some
teaching resources: books, recordings, etc. This educational treatment is
likely to develop different strategies and learning processes, and
consequently, different outcomes. This sequence, illustrated in boxes 1, 2, 3,
4 & 5 is said to produce second (or
foreign) language learning (SLL) (Krashen 1981).
Though, in many situations, it is
difficult (or impossible) to know if SLA or SLL takes place, this distinction
is useful, especially when we want to differentiate the type of context or
setting in which L2 learning takes place.
The model presented in the diagram
was first proposed by Dunkin and Biddle and adapted later by Stern (1983:338
and 500). Dunkin and Biddle used the following four terms for the set of
variables that we have identified in fig. 1(1975:38):
|
a) Presage variables ‑ Teacher
formative experiences and training ‑ Teacher
properties: skills, personality b) Context variables ‑ Pupil
formative experiences ‑ Pupil
properties: abilities, attitudes, etc. ‑ School
context c) Process variables ‑ Teacher
classroom behaviour (didactic performance) ‑ Pupil
classroom behaviour ‑ Learning
processes d) Product variables ‑
Outcomes/final competence/proficiency |
2.4 METHODS AND PARADIGMS FOR CLASSROOM
RESEARCH
As we will see, research questions and topics
can be looked into from many different perspectives, by following different
approaches and by using different procedures. The approach that we adopt needs
to be suitable for the kind of research we want to carry out, for the variables
we want to control. In some occasions, an observational process will be enough
because the data we want to collect cannot be quantified, but in others we may
need to illustrate our findings with figures and a statistic treatment may be
necessary. So, the approach we adopt depends on the nature of the research we
aim to do. Very often an eclectic position or a combination of paradigms may be
necessary. As classroom research is
defined primarily by its setting, the classroom provides the focal point for
the types of data collected. However, a wide variety of approaches are used to
obtain and analyse the data, and the choice of approach depends upon many
factors: the researcher's philosophy, the issue to be investigated, the
constraints inherent in the situation and so on.
First ELT research
studies
The beginnings of a research approach in
language teaching can be dated back to the late 19 century, linked to the
development of the language sciences and the scientific movement in education.
However, a truly consistent and deliberate research approach in ELT only became
apparent from the 1950s, and it is this that we will now look at.
One of the first attempts at systematic
research was carried out in 1948 by Agard and Dunkel at the University of
Chicago, in which 'new' and traditional methods of language teaching were
compared. As a result of this study, Dunkel put together a volume of all the
studies on language learning to date (1948). In the same year a journal with a
strong research orientation, "Language Learning", was published by
the English Language Institute at the University of Michigan in Ann Arbor. Its
director, Charles Fries, had been a key figure during the previous decade in
giving language pedagogy a basis in research (see Unit 2).
Nevertheless, despite these early attempts,
research into the field of ELT remained thin until the sixties, when a radical
change took place, and research began to influence policy issues and the method
debate in second language education. This was due to the fact that during the
fifties and sixties language centres with a strong research orientation had
been set up, such as the CREDIF (Centre
de Recherche et d'Etude pour Ia Diffusion du Français) in France in 1951,
or the CAL (Center for Applied
Linguistics) in Washington, D.C. in 1959, or the Centre for Information on
Language Teaching and Research in 1966, pIus a number of university centres in
applied linguistics, in Britain.
Thus when the Audiolingual method (see Unit 2)
appeared on the scene in the late 1950s, considerable interest was aroused, and
a number of major investigations were carried out in an attempt to resolve
controversies about its merits over the more traditional grammar translation
method.
Two other areas of language teaching which
began to be systematically researched from the 1960s on were those of language
teaching for younger children, and the effectiveness of immersion teaching
programmes.
The problem of
approach: Objectivity and subjectivity; quantitative versus qualitative
research
The problem of approach can be
addressed in terms of potentially opposing viewpoints on how research should be
conducted. The difference between these perspectives hinge primarily on
differing attitudes towards intervention and control. The issues of subjectivity and objectivity often arise when devising observation schedules for
recording classroom data. In constructing such instruments, researchers
distinguish between "low inference" and "high inference"
categories of analysis (Long, 1980a). Low inference categories include things
that can be counted or coded without the observer having to infer very much
(the number of times a certain student raises his or her hand, for example).
"High inference categories, on the other hand (learner's attention, for
example), demand that the observer make a judgement that goes well beyond what
is immediately visible. Generally speaking, in developing a coding procedure,
it makes sense to use low inference categories wherever possible, and try to
document all the evidence that leads us to make broader inferences. A universal
problem for classroom researchers, then, is one of finding low inference means
of investigating non-trivial aspects of what happens in language classes.
The terms quantitative and qualitative apply to both the data
collection and data analysis phases of an investigation. Any sort of
measurement that yields numerical information generates quantitative data. On
the other hand, some data are not the product of measurement or counting and
thus do not result in numerical information (prose descriptions, diaries, and
so on). This kind of data are qualitative.
Likewise, once we have collected
data we can analyse them by counting or measuring (quantitative analysis), or
by directly reflecting upon and trying to interpret them (a qualitative
analysis). Various combinations of quantitative and qualitative data collection
and analysis are possible.
Qualitative data, such as a set of
diaries kept by learners, would typically be considered subjective, a record of
opinions and perceptions, rather than "facts". For some researchers
learners' diaries are of interest not because they hold the "truth"
about something, but precisely because they are a record of opinions and
perceptions important to the learners -ideas which cannot easily be tapped in
other ways.
Quantification has a similar appeal
to objectivity in that if you can count things then you can subject them to
rigorous statistical analysis using procedures set out in statistical manuals.
If you get the procedures right, people can challenge you on your interpretation
but the figures themselves cannot be denied. The problem, here, of course, is
that not everything can be counted or measured adequately, and therefore
numbers cannot tell the whole story. Some vital element may be missing. Also,
there is the difficult problem that statistical procedures are themselves the
subject of endless controversy. Even deciding on a unit of analysis for
counting can be very tricky. Experimental research is objective and
quantitative, whereas action research tends to use subjective, qualitative and
ethnographic techniques.
Eclectic approaches to
research
Of course very few studies can be confidently
characterized as being purely experimental (+ structured, - controlled) or
purely naturalistic (- structured, -controlled). In fact, in recent years
observational procedures from naturalistic enquiry have been used to document
the process variables involved in implementing the treatment in
product-oriented experimental research.
While there are as yet relatively few examples
of this combined approach, the report by Bailey et al. (1990) demonstrates that
it is becoming increasingly clear in general educational research (though
research on language teaching and learning has lagged behind somewhat in this
area) that action research is often a viable alternative, and one which offers
immediate rewards to teachers and learners.
From all that we have said on the topic, it
should be clear that we see most value in investigations that combine objective
and subjective elements, that quantify only what can be usefully quantified,
and that utilise qualitative data collection and analysis procedures wherever
they are appropriate. This eclectic approach has been emphasied in several
occasions:
“Ideally experimental and interpretative research should be convergent
rather than parallel or divergent lines of enquiry” (va Lier, 1988/90:XIV).
“There is no need to oppose qualitative and quantitative research. Each
is capable of “critical thinking” and each has its place in IL (interlanguage)
studies. The danger is ... in failing to acknowledge the contribution that can
be made by “hybrid” research (i.e. research that employs both qualitative and
quantitative procedures) (Ellis, 1984:284).
“It should be clear that we see most value in investigations that combine
objective and subjective elements, that quantify only what can be usefully
quantified, and that utilise qualitative date collection and analysis
procedures wherever they are appropriate” (Allwright y Baley, 1991:67).
2.4.1 EXPERIMENTAL RESEARCH
Here the researcher exerts a high degree of
control and purposefully intervenes in the setting, to determine the effect on
intervention. It is important to note, however, that "intervention"
is by no means a negative thing. It is simply a technical terms which refers to
the "treatment" administered to some subjects (the "experimental
group") in order to test a hypothesis about a cause-and effect
relationship. This treatment is withheld from other subjects in the study
(typically called -the control group"). If the researchers are careful in
setting up the study, these two groups can be presumed to be virtually
identical in all respects, except that one gets the treatment (teaching method,
materials, for example) and the other does not. After the treatment has been
implemented, a test of some sort is usually administered to both groups and
their results are compared. Then various mathematical procedures are used to
determine whether or not there are statistically significant differences in the
test scores of the two groups. From the results, the paradigm claims, we can
infer that the treatment either did or did not cause a measurable change in
behaviour or learning (the hypothesis effect).
In
experimental research, there are some preparatory stages or phases that are
currently followed. These stages include the following steps (Seliger and
Shohamy, 1989):
|
1)
Formulating
the general question or the problem which will be solved. This usually
emerges from the researcher’s experience, from research paradigms or from
sources outside the second language field. 2) Feasibility of
the research work we aim to carry out to solve the problem. 3) Deciding on the
objectives that will be achieved. 4)
Formulating the research plan and hypothesis to be tested. |
In addition to that, research has to be
contextualized. According to Seliger and Shohamy it means (1989:85):
-
selecting
a research problem,
-
to
broaden the perspective of the research,
-
a
revision of the existing literature on the topic,
-
to
create a rationale for the study,
-
to help
the researcher narrow down the research question in preparation for conducting
the research.
-
to
describe the different sources for locating the literature:
-
references
to existing material, such as indices, computer searches, and bibliographies,
-
and the
actual material such as journal articles and reviews.
-
to
describe the criteria for determining the relevance of the material to the
research topic,
-
and
suggestions on ways of organizing and reporting the literature review .
Another key stage of research begins after the
research question or hypothesis has been developed. Planning the research takes place after the researcher has
identified the focus or objective of the research. In synthetic or analytic-deductive
research, planning requires the careful development of a plan in which
those factors to be controlled or manipulated are identified: the independent, dependent, subject, and
extraneous variables.
The dependent
variable is the means by which any changes are measured. The independent variable is the factor
the researcher manipulates in order to
see what effect the changes introduced will have. For example, if we want to
study the relationship between the students participation and their degree of
extroversion in their final results, we may start the research
plan by thinking that classroom
participation and extroversion encourage good results in SLL (hypothesis). The
degree of participation and extroversion are independent variables and the
student’s achievement is the dependent variable.
Research which focuses on variables, makes
predictions and tests hypothesis is primarily of the deductive type. Heuristic
research approaches the research context from a different perspective, without
preconceptions, with the aim of generating hypothesis but not to test them,
using qualitative methods. It does not attempt to control or manipulate
variables (see Seliger and Shohamy 1989).
The following checklist can be very
useful for research analysis (Seliger and Shohamy 1989: 80-81):
A
About the research topic
1 What is the main research area?
2 What is the research problem?
3 What are the major research questions or
hypotheses?
B
About the research context
1 What other research studies were
conducted in the same area?
2 What were their main findings?
3 What is the rationale of the research?
4 Why was it important to conduct the
research?
C
About the research method
1 What are the main variables of the
study?
2 Which research design was
used? (experimental, correlational, descriptive, multivariate, ethnographic?)
3 Description of the population, sample,
and selection procedures
4 The data collection
procedures - information about their development reliability, validity, pilot
study
5 Description of the data collected
D
About the data analysis
1 What are the specific data analysis
procedures used?
2 Were they quantitative or qualitative,
or both?
E
About the findings
1 What were the main findings?
2 What does the researcher conclude from
them?
3 How do the findings relate to the
research context and to the underlying theories?
4 What are the implications of the
findings?
5 What recommendations does the researcher
make based on the findings?
6 What recommendations are drawn from the
results?
F
Criticism of the research
Consideration of A-E
above and specifically:
1 the statement of the problem
2 the identification of the hypotheses
3 the description and definition of the
variables
4 the appropriacy of the design of the
study
5 the appropriacy of the instruments
6 the appropriacy of the data analysis procedures
7 the consistency of the results with the
analysis
8 whether the conclusion, implications,
and recommendations are warranted by the results.
Questioning the
scientific research paradigm
The success of modern science has had the
effect of imbuing anything that is 'scientific' with the flavour of absolute
truth. Thus research which is based on rational, 'scientific' experiment is
seen as unquestionable, objective, right. Why should this be so? According to
Winograd and Flores:
The rationalist orientation... is also regarded, perhaps because of the
prestige and success that modern science enjoys, as the very paradigm of what
it means to think and be intelligent... It is scarcely surprising, then, that
the rationalistic orientation pervades not only artificial intelligence and the
rest of computer science, but also much of llnguistics, management theory, and
cognitive science... rationalistic styles of discourse and thinking have
determined the questions that have been asked and the theories, methodologies,
and assumptions that have been adopted. (1986:16)
However, 'scientific' research has so far
failed to answer some of the most pressing questions in language teaching. The
issue of discipline in the classroom is a case in point. The scientific model
of research would try to use scientific knowledge to solve the problem of
maintaining discipline: thus the theorists Stones and Morris, writing in 1972,
claimed that "(this) important area of classroom and group management has
received detailed empirical study, and a body of theoretical and practical
information has been amassed which begins to put the problems of discipline on
a scientific footing..." (1972:14). The implication here is that
scientific study would soon provide a formula for dealing with classroom
discipline problems. As all practising teachers know, this has not been the
case. Thus scientific research seems to promise solutions to very complex
professional dilemmas, but rarely manages to deliver the goods.
Another weakness that is apparent in a
scientific research paradigm is that of its lack of grounding in specific
classroom practice. There has traditionally been a separation between theorists
who do research on the one hand, and practitioners - the teachers in the
classroom - on the other. Researchers tend to be working in university
departments where they have no contact with the daily realities of the language
classroom, particularly as represented in primary and secondary schooling.
Researchers and practitioners often have very different training, and there is
frequently a sense of status attached to the job of researcher which is not to
be found in the job of classroom teacher. In Schon's words:
It was to be the business of university based scientists and scholars to
create the fundamental theory which professionals and technicians would apply
to practice... But this division of labour reflected a hierarchy of kinds of
knowledge which was also a ladder of status. (1983:36)
Thus there has been a tendency in research
circles to downgrade the value of the classroom teacher's expertise which is
derived from experience, rather than from research.
At the same time, research findings are in a
sense imposed on teachers. Teachers are called upon to implement new ideas and
theories in language teaching in the form of teaching materials, rather than
asked to consider the ideas themselves and to evaluate them critically in the
light of their own experience and teaching context.
Ellis (1990) points to two reasons for the
growing scepticism with the conventional rational research paradigm: firstly,
the relationship between teaching and learning is extremely complex. It is not
linear (i.e. teaching does not automatically lead to learning), thus
experimental research can only provide us with an understanding of fragments of
the language learning process, not the whole process. Secondly, according to
Ellis, the findings from formal experiments conducted under laboratory
conditions are not necessarily applicable to the language classroom:
Innovation in the classroom can never be just a
question of implementing a recommendation derived from research. It is always a
process of negotiation, involving the teacher's overall educational ideology,
the learner's expectations and preferences and local constraints that determine what is feasible. There is no
single pedagogical solution which is applicable to all classrooms. (Ellis
1990:68)
Classroom research of any kind is
very likely to be a sensitive business, however carefully it is done, because
being observed in any way is anxiety-provoking, to say the least, and being
closely observed, recorded or analysed is enough to put anyone on the
defensive. In any given classroom study, the researcher must decide if the
learners are going to be treated like subjects in a classical psychology
experiment and kept uninformed about what the experimenter is investigating, or
-at the opposite end of the continuum, as in the action research paradigm- if
they are going to be involved as full collaborators from the start, helping to
define problems, deciding how to approach them, and so on.
2.4.2 CLASSROOM OBSERVATION
When the teaching and learning of a
foreign language takes place in a formal setting, it is necessary to observe
and analyse systematically what is going on in order to understand such
teaching and learning processes.
Classroom observation has been used with different purposes: to compare
teaching methods, to study the most efficient classroom techniques, to evaluate
teachers and materials, etc. All these topics are fundamental components of
classroom research. So, observation becomes one of the most important
techniques to study what actually happens inside the classroom in a systematic
way. In addition to direct observation, there are other procedures for classroom
research. These include surveys and self-reports. Self-report data are obtained
by conducting surveys, usually through interviews or written questionnaires.
Even though questionnaires are not always filled out truthfully and their
validity may be relative, the data obtained are important to form hypothesis
that can be tested later on. Another problem with observation checklists and
with surveys is that items have to be decided in advance and sometimes some
irrelevant aspects are not included and no feedback is obtained. This problem
can be solved with very open questions where the surveyed subjects can mention
anything that seems relevant to them. This desire not to prejudge the
importance of potential relevant events has led some researchers to explore the
procedures and techniques of ethnography as a viable approach to classroom
research (see van Lier 1988,
Watson-Gegeo, 1988, Erickson 1981, Green and Wallet 1981). For example,
van Lier (1988) considers ethnographic techniques the most suitable to study
classroom events. He states that to understand what happens in classrooms, we
must ask about the meanings that the participants give to the successive
classroom phenomena.
Observation in the classroom and ELT
research began to be used in the 1960s in connection with teacher training to
provide the trainees with feedback about their performance in class during
their teaching practice or after the short periods of microteaching with their
college peers. In most cases, the main objective was to investigate what
constituted effective teaching in order to select those reflective teaching
techniques and become competent to use them in the future, in similar
situations. But the teaching and learning processes have proved to be so
complex and unpredictable that this technological conception of education has
taken other directions, as we will see in the next section.
2.4.2.1 Interaction analysis
One
of the earliest and most popular observation model was developed by Flanders
(1970). He developed a list of categories of teacher and learner behaviour
associated with successful teaching. His taxonomy is called FIAC (Flanders’
Interaction Analysis Categories) and contained the following groups of categories (see Appendix 1.1):
|
TEACHER TALK Indirect influence
Accepts
feeling Praises
or encourages Accept
or uses ideas of student Asks
questions Direct influence
Lecturing Giving
directions Criticizing
or justifying authority |
|
STUDENT TALK Student-talk
response Student
talk-initiation |
|
Silence or
confusion |
To codify the behaviour used during
a teaching unit the observer marks every time a different category. When the
same category is repeated, the observer records this category every three
seconds. All the tallies are registered in a matrix which thus show a graphic
picture of the lesson. By analysing the matrix, some teaching patterns ca be
discovered.
Flanders model was soon criticized
by those who considered classroom behaviour too complex to be reduced to ten
categories and other models soon emerged. It was Moskowwitz (1976, 1971) who
produced the most widely used modification of Flanders’ model. She called her
taxonomy “FLint” (Foreign Language Interaction) and included the following
categories (see section 1.4.1):
|
TEACHER TALK Indirect influence
Deals
with feelings Praises
or encourages Jokes Uses
ideas of students Repeats
student response verbatim Asks
questions Direct influence
Gives information Corrects without
rejection Gives directions Directs pattern
Drills Criticizes
students behaviour Criticizes students
response |
|
STUDENT TALK Student
response, specific Student
response, choral Student
response, open-ended or student initiated Silence Confusion Laughter |
Moskowitz’s taxonomy was used as a
research tool and as a feedback instrument in teacher training. She trained
student teachers to analyse their teaching behaviour using her taxonomy for
them to have a more objective picture of themselves as potential teachers. It
was also used as a reference of what constitutes good and efficient language
teaching.
Fanselow (1977) made another
important contribution with his FOCUS (Foci for Observing Communication Used in
Settings). It was also an observation schedule developed for language teacher
training but it is applicable for classroom research. As we can see in the
following table, Fanselow’s model does not have separate categories for
teachers and learners, but instead has general categories which define five
characteristic of communication in settings (see section 1.4.1):
|
- Who communicates? - What is the pedagogical
purpose of the communication? - What mediums are
used to communicate? - How are the
mediums used to communicate areas of content? - What areas of
content are communicated? |
The model we propose (Madrid, 1998a, 1998b) is related to the communicative approach,
too, but it is primarily based on the communicative competence construct,
defined as the integration of five
basic sub-competences (see Canale, 1983; Savignon, 1983; Kohonen et al. (1985),
Bachman, 1990):
a)
Linguistic
or grammatical subcompetence
b)
Sociolinguistic
or illocutionary subcompetence
c)
Sociocultural
subcompetence
d)
Discourse
competence
e)
Strategic
(and procedural) competence
So, we aim to analyse the attention
paid to these subcompetences in the classroom; that is, the time that the
teacher and the students spend developing each one, in isolation or integrated
in different
teaching and learning situations.
These sub-competencies include the categories presented below that must be
observed and codified in a systematic way every minute (see instructions in
Appendix 1.8):
|
a)
Linguistic or
grammatical subcompetence: -
Grammar -
Lexicon (vocabulary) -
Phonetics |
|
b)
Sociolinguistic
or illocutionary subcompetence -
Communcative functions -
Appropriateness of grammatical forms |
|
c)
Sociocultural
subcompetence -
Cultural concepts, attitudes, values, ... |
|
d)
Discourse
subcompetence -
Working with texts: structure, coherence, cohesiveness, ... |
|
e)
Strategic
competence -
Learning
skills and procedures: listening, speaking, reading and writing -
Mental
operations which regulate the learning processes: cognitive, metacognitive,
affective, communicative, ... |
Obviously, the strategic
subcompetence is of a psychological nature and cannot be analysed without the
learner cooperation. That is, an additional information coming from the learner
is necessary. The following model is
suggested for data collection related to a) block 1 about the communicative
competence components and b) block 2
about the structural vs. the discursive paradigm (see Appendix 2.8 for a
fully description):
For the analysis of the classroom
teaching and learning activities related to the receptive and productive skills and procedures, we propose the
following model:
The model that we propose (fully
described in appendix 2.8) is also based on the process-product paradigm (see Section 2.3) which seeks direct and indirect relations between the teaching and
learning process and the final results that the student obtains (see Dunkin and Biddle, 1974). In
this sense, we also aim to analyse the English class to study the connection
between what is done in class and the effect on the students’ communicative
competence. In order to study that connection, the taxonomy presented in
Appendix 2.8 is proposed. In addition to that, suitable tests and evaluation
techniques must be administered to obtain the information needed about the
final outcomes and have thus some evidence about the relationship between the learning process and the final
product (outcomes).
2.4.2.2 Discourse analysis
The power emergence of discourse
analysis studies as a field of linguistic enquiry has led to classroom research
models also based on the analysis of classroom discourse. An example of these
types of contributions can be the COLT system proposed by Allen, Frölich and
Spada (1985). The COLT model (Communicative Orientation of Language Teaching)
was developed to distinguish communicative language teaching classrooms from
those more form-focused. The categories aimed to measure the extend to which an
instructional treatment was communicatively oriented and included the following
headings:
Part A: Classroom events
I.
Activity II. Participant
organization III. Content IV. Student modality V. Materials |
Part B: Communicative
features
I.
Use of target language II. Information gap III. Sustained speech IV. Reaction to code
or message V. Incorporation of
preceding utterances VI. Discourse
initiation VII. Relative
restriction of linguistic form |
Sinclair and Coulthard (1975) also
analysed the discourse of some L1 British elementary schools classrooms. In
their model, the discourse level is comprised of five ranks (lesson,
transation, exchange, move, act). The pedagogical move consists of various structures, which are realized by “acts”
having a specific discourse function. For example, an “opening” move consists
of a head act, either elicitation, directive, informative or check. These acts
resemble the concept of Searle’s speech acts (1969), illocutionary acts or
communicative functions, widely
employed to analyse the functional value of
language use. Sinclair and Coulthard include the following group of
categories (see Sinclair and Coulthard, 1975: 24-27):
|
Rank 1: Lesson Elements of structure,
structures, classes Rank 2:
Transaction Elements of structure,
structures, classes of exchange Rank 3: Exchange
(Boundary); Exchange (Teaching) Elements of structure,
structures, classes of move Rank 4: Move
(Opening) Elements of structure,
structures, classes of act |
For second language classroom
research Chaudron offers the following list of categories (1988:45):
|
STRUCTURAL UNITS utterance: a string of
speech by one speaker under a single intonation contour, arid preceded and
followed by another speaker's speech, or a pause of more than x seconds turn: any speaker's sequence of utterances bounded
by another speaker's speech T-unit: any syntactic main clause
and its associated subordinate clauses Communication
unit: an independent grammatical
predication; the same as a T-unit, except that in oral language, elliptical answers to questions also constitute
complete predications fragment: any utterance
which does not constitute a completed proposition (i.e., with explicit
subject and verb) FUNCTIONAL UNITS repetition: an exact
repeating of a previous string of speech (either partial or full, and either
a self- or other-repetition) expansion: a partial or full repetition which
modifies some portion of a previous string of speech by adding Syntactic or
semantic information clarification
request: a request for further information from an interlocutor about a
previous utterance comprehension
check: the speaker's query of the
interlocutor(s) as to whether or not they have understood the previous
speaker utterance(s) confirmation
check: the speaker's query as 10
whether or not the speaker's (expressed) understanding of the interlocutor's
meaning is correct repair: an attempt by a
speaker to alter or rectify a previous utterance which was in some way
lacking in clarity or correctness (either self- or other-directed); model: a type of prompt
by a speaker (usually a teacher) intended to elicit an exact imitation or to
serve as an exemplary response 10 an elicitation |
Discourse analysis has contributed to a better awareness of the internal
structure and functional purpose of the verbal classroom interaction and it has
been greatly influenced by the ethnographic research tradition
2.4.3 ETHNOGRAPHY
Given the shortcomings of scientific research,
which we have discussed above, it is not surprising to find that theorists
began to pose alternative research paradigms and used a naturalistic enquiry
for their research. In this case, the researcher tries not to intervene in the
research setting and does not try to control naturally occurring events. Quite
the contrary, they purposefully try not to influence the normally
occurring patterns of instruction and interaction because they wish to describe
and understand these processes rather than to test specific hypothesis about
cause-and effect relationships. While it may include comparison groups and can
involve the collection of both quantitative and qualitative data, it does not
use the experimental concept of "treatment". Nor does the researcher
create special groups for the purposes of experimentation and observation.
Instead, naturally occurring groups become the focal point (Watson-Gegeo,
1988). For this reason, regularly scheduled classes are likely sites in which
to use the naturalistic enquiry approach.
Hammersley and Atkinson present a summary of
views which can help us to understand the wide scope of ethnography (1983:
1-2):
|
Ethnography is: - the elicitation
of cultural knowledge (Spradley 1980) - the detailed
investigation of patterns of social interaction (Gumperz 1981) - holistic analysis
of societies (Lutz 1981) - essentially
descriptive, a form of story telling (Walker 1981) - the development
and testing of theory (Glaser and Strauss 1967; Denzin 1978) - one social
research method, drawing on a wide range of sources of information
(flammersley and Atkinson 1983) |
The growth of the discipline of ethnography has
had a profound influence on the acceptance of an alternative research paradigm.
Ethnography has its roots in anthropology and sociology, and it revolves around
two key ideas:
-
context is seen to have an important influence on
behaviour, therefore research needs to always be carried out in context, not
under 'laboratory' conditions.
-
the subjective
perceptions and belief systems of
both researchers and subjects need to be taken into account in any research.
This is because 'objectivity' is an impossibility, given that there are no
external 'Truths' or objective reality to be discovered. Thus research focuses
on the cultural meanings revealed by the behaviour of the subjects under study.
Thus ethnography involves "the study of
the culture/characteristics of a group in real-world rather than laboratory
settings. The researcher makes no attempt to isolate or manipulate the
phenomena under investigation, and insights and generalisations emerge from
close contact with the data rather than from a theory of language learning and
use." (Nunan 1992:55).
Obviously these ideas are akin to the idea of
the reflective practitioner, and have resulted in the development of a kind of
classroom research called "action research", which we will examine in
detail later in this unit. But many observational studies have also followed
the ethnographic methodology (see van Lier, 1988).
Mehan (1979) has also criticized
quantitative oriented surveys, coding and experimental studies and proposes what he calls a constitutive ethnography with the
following basic charecteristics:
|
1.
Retrievability of the data. Ethnograpic techniques such us video and
audio-tapes provide a record of the data studied that can be used to
re-examine and re-interpret them. 2.
Comprehensive data treatment. Its comprehensive approach to classroom
observation and analyses allows the study of each and every intance of
teacher-student interaction. 3.
Convergence between researcher’s and participants’ perspectives. The
structures and actions described reflect the participants view, how they perceive
classroom events. 4.
Interactional level of analysis. The goal is in Mehan’s words to ... locate
the organizing machinery of classroom lessons in the interaction, ... the
words in the actions of the participants, .. . |
A major procedure of ethnographic
work is the detailed analysis of recorded data. For any ethnographic study
based on classroom videotaped data, Erickson and Shultz’s stages are very
useful (1981:153-7).
|
Stage 1. Global viewing:
taking sparse notes to index the tapes and note transitions between occasions
of interest. Stage 2. Choice of specific occasions of interest for more detailed
analysis. Selected occasions are copied onto copy tapes. Occasions are timed,
and junctures described in detail. Participants may be asked to attend a viewing
session and give their emic views of the occasion. Stage 3. Specification of differences in the transitions or junctures,
with specific attention paid to nonverbal contextualization cues. Stage 4. Detailed description of the participation structures between
junctures. Stage 5. Construction of a model showing the principles of social
organization underlying the surface form of communication behavior in
interaction. Stage 6. Establishing the generalizability of the structures analysed.
(Summarized from: Erickson and Shultz 1981, p.153-7, in van Lier,
1988:65) |
2.4.4 ACTION RESEARCH
One of the new buzzwords in ELT is that of
"action research". Martin Parrott
provides the following definition:
Action research is not so much something that we do in addition to our
teaching as something that we integrate into it. In many ways it is a state of
mind - it is a scepticism about assumptions and a willingness to put everything
to the test. It is something which should take very little time and which does
not necessarily need to be made public. It is a way of ensuring that we
continue to learn even as we teach. It helps stave off staleness and routine. (1996:3)
Action research is thus research on the
smallest scale: that of the individual teacher in his/her classroom carrying
out investigations into teaching and learning in very specific contexts with
very specific groups of learners. Action research is what the reflective
practitioner actually does in the classroom. It is typically direct intervention with only limited possibilities
for control. Although it can take on many forms, action research in classrooms
basically involves taking an action and systematically observing what follows.
Action research has been defined as a form of
self-reflective enquiry undertaken by participants in social situations in
order to improve the rationality and justice of their own social or educational
practices, as well as their understanding of these practices and the situations
in which these practices are carried out" (Kemmis and Henry 1989:2). It is
characterised as being a participatory, self-reflective and collaborative
approach to research.
The fundamental characteristics of
action-research according to Elliott, one of its pioneers, are the following (1990):
-
It analyses
the human actions and social situations that students and teachers experience.
-
It
uses an exploratory approach
-
It
aims to explain what happens in the classroom in relation to specific teaching
contents.
-
It
interprets the different classroom events from the point of view of those
who take part in each situation; that
is, it involves teachers and students: their beliefs, values, intentions,
decisions, ...
-
It
uses a very direct and simple language to explain the classroom situations that
are analysed far from the technical and specialised language used by
conventional research.
Kemmis and McTaggart (1988) argue that action
research has three defining characteristics: firstly that it is carried out by classroom teachers rather by than
outside researchers; secondly, it is collaborative;
thirdly, it is aimed at changing
things:
A distinctive feature of action research is that those affected by
planned changes have the primary responsibility for deciding on courses of
critically informed action which seem likely to lead to improvement, and for
evaluating the results of strategies tried out in practice. Action research is
a group activity. (1988:6).
Thus for Kemmis and McTaggart the essential
impetus for carrying out action research is to change the system. Cohen and
Manion (1985) add that action research is primarily situational, as it is
concerned with the solution of problems in a specific context.
However, it is now widely accepted that action
research does not necessarily have to be collaborative, nor does it have to
result in change. A single teacher, working on identifying aspects of his/her
teaching practice in the classroom, can also be considered to be undertaking a
perfectly valid piece of action research.
2.4.4.1 Stages in action research
There are four classic developmental phases of
action research:
Phase 1:
Develop a plan of action to a)
improve what is already happening or
b) identify and examine a
"puzzle" or problem area in your teaching.
Phase 2: Act to implement the plan.
Phase 3: Observe the effects of action in the context in which it occurs.
Phase 4: Reflect on these effects.
Nevertheless, other authors establish the
following eight stages (Cohenand Manion, 1985):
|
Stage 1. The identification, evaluation and
formulation of the problem. Stage 2. Preliminary
discussion and negotiations amongst interested parties – teachers, advisers,
researchers, sponsors – culminating in a draft proposal. Stage 3. Review of research literature and
comparable studies. Stage 4. Restatement of the problem, or formulation
of a hypothesis; explicit discussion of the assumptions underlying the
project. Stage 5. Selection of research procedures,
allocation of resources, choice of materials and methods, etc. Stage 6. Choice of evaluation procedures - bearing
in mind that evaluation will be continuous. Stage 7. The implementation of the project itself,
including data collection and analysis, monitoring and feedback. Stage 8. The interpretation of the data; inferences
to be drawn; overall project evaluation. |
Nunan describes an action research project
which has a slightly broader scope (see Appendix 2). According to Nunan, the
key points to note about action research are that, firstly, the research is
initiated by the teacher him/herself and is derived from a real problem in the
classroom which needs to be confronted. Secondly, real objective data is
collected (in Nunan's project in Appendix 2, this data takes the form of
classroom interactions and learner language). Thirdly, the results of the
research project are disseminated. Finally, the project takes the form of an
ongoing cycle in which the teacher reflects on, returns to, and extends the
initial inquiry. (1992:18-19).
Action research is thus the putting into action
of the reflective model of research which we discussed earlier, and which is
summarised in the following diagram:
Action research can be used by the classroom
teacher for a variety of purposes: it can be used as a way of learning about
our learners, and about our teaching and its effectiveness, and as a way of
monitoring and evaluating innovation. We will look at each of these areas
separately.
A way of learning about our learners:
By asking about and taking learners'
attitudes and preferences into account, we can reflect on and evaluate our
teaching. Ways of doing this include: a simple show of hands (The teacher
simply poses a series of questions - eg. "Who wants to be corrected more?
Who wants only the teacher to correct them? Who likes mistakes to be corrected
by other students?"etc) - and students show preferences by putting up
their hands; questionnaires, in the mother tongue if necessary; teacher/learner
diaries or journals can be used for feedback on specific issues; brief
teacher/student interviews; and so on.
Reflecting and analysing our teaching:
There is a great deal we can learn
about the way we teach through simple observation of different kinds. For
example, we can record or video our lessons and listen to or watch the
recording later; peer observation; and so on. (See Subject *). Here a test-teach-test approach might be
appropriate. For example, the teacher might decide to 'test' the learners'
ability to recall vocabulary after teaching it in a variety of ways, with
various classes. The important point here is that evidence is unlikely to be
conclusive: an experiment of this type could well lead to the conclusion that
no one method is better than any other, and that certain students may learn
better than others depending on the approach taken. In this case, the teacher
may conclude that a variety of approaches is necessary to cater to a variety of
learner styles.
In a more general sense, action
research often has a specific and immediate outcome which can be directly
related to practice in the teacher's own context. As it does not make any
claims for universal relevance, the research methods employed in action
research can be more free-ranging and less rigorous than those employed in
conventional research. By allowing for discussion and the sharing of ideas and
experience among colleagues, it can open up dialogue on issues that worry
teachers, and it can help to kindle and maintain enthusiasm about our classroom
practice.
A way of monitoring and evaluating innovation: Simple but systematic evaluation of
innovation can help teachers to incorporate principled change into their
teaching. What is needed is to start evaluating before the innovation is introduced, and to make the evaluation
indirect.
2.4.4.2 Criticisms of action
research
Posing as it does a rather more flexible, reflective
model of research, action research has predictably come in for a certain amount
of criticism. This is inevitable, as it reflects an entirely different view of
the status of knowledge to the more traditional scientific research paradigm,
which we discussed earlier.
The issues of reliability and validity are key
in talking about and evaluating research. Reliability refers to the consistency
of results obtained from research
- how consistent is the collection of data
within a piece of research? Reliability also refers to replicability - can the
same piece of research be carried out with another group of learners and
provide similar results? Validity seeks to establish to what extent the
research really investigates what the researcher
wants it to - are the research tools used
consistent with what the researcher is trying to research?
Criticisms of action research have centred
around its lack of vigorous attention to both reliability and validity,
particularly external validity or replicability (the extent to which
independent researchers can reproduce a study and obtain results similar to
those obtained in the original study).
However, it has been pointed out that the very
nature of action research precludes such concern over issues such as external validity
because "in many cases practitioners are less concerned with generating
generalisable knowledge than with solving pressing problems associated with
their own particular workplace. While such (action research) activities
therefore fulfil a professional development function, I still believe that if
they address questions of interest to other practitioners, if they generate
data, and if they contain analysis and interpretation, then they qualify as
research." (Nunan 1992:18-19).
Another criticism of action research lies in
the fact that it is unable to provide 'facts' or universal truths about
language teaching, as its scope is limited. However, as this criticism lies in
a different perception of the status of knowledge to that proposed by action
research, as discussed earlier, it is enough to assert that "the aim of
action research is not to arrive at universal truths but only to learn more
about ourselves (at the moment), our teaching (at the moment), our learners (at
the moment) and their learning (at the moment)." (Parrott 1996:6).
2.5 TEACHERS' BELIEFS
Individual teachers exhibit certain typical
behaviour traits in the classroom. We all know that each teacher is different,
and will react differently in any given classroom situation, but what makes
that difference? MacLeod and MacIntyre (1977:260) suggest that individual
teachers are able to make very fast, complex professional decisions on the job
due to having a limited number of deep underlying "conceptual
schemata" or 'constructs' of what appropriate professional action might
be. We could define 'construct' as a broad term used to include a group of
related concepts such as beliefs, ideas, attitudes, etc. all of which shape our
behaviour in various typical or consistent ways.
Research studies about teachers’ beliefs have
increased dramatically for the last twenty years. The general aim has been to
find out what happens in the classroom in order to improve the teaching and
learning situations. This type of research provides ideas and principles which
can be undertaken as hypothesis to be checked in other contexts and also gives
evidence about the variety of pedagogical ideas and beliefs that teachers
have and their connection with
different modes of performing in the classroom.
Floden and Klinzing (1990) have
found a very valuable relationship between the teacher’s beliefs and his/her
initial education, which becomes an important source of ideas, beliefs and models that are applied
later to different classroom situations. De Vicente has also highlighted the
importance of this paradigm in order to train teachers within the
reflective-in-action-paradigm. Given
that the research studies that conform this paradigm are based on case studies
and report about very personal and sometimes unique situations, they often
derive into reflection-in-action processes and promote a critic reflection of
teaching. De Vicente has differentiated three basic stages:
a)
Cognitive stage: when these studies focus on the use of the
cognitive skills that teachers make when preparing their curricular activities,
implementing them and making decisions in
the classroom.
b)
Critical stage: when teachers report about their beliefs,
opinions, values and experiences critically.
c)
Interpretative stage: when teachers interpret what happens
in the classroom and narrate classroom events.
Nunan (1991:50) gives a clear example of how
mental constructs could operate in a specific classroom situation:
Thus, for example, language teachers who have detected a learner making
a pronunciation error in the target language will usually deal with it in a way
that is at least partly determined by mental constructs... (ie. in the light of
their views on the importance of fluency as opposed to accuracy, on the nature
of the cause of the error; appropriate ways of dealing with such errors, the
importance of pronunciation errors as opposed to other kinds of errors, and so
on). Of course their reaction may also be influenced by aspects of the incident
specific to this particular situation: their relationship with a
particular learner; whether they are
feeling particulariy tired or irritable at the time, etc. But, in general
terms, if they are professionals, they will exhibit an overall consistency of
response in this area which will relate to the repertoire of mental constructs
which they have developed.
Wittrock (1990) has designed a model to explain
the relationship between teachers’ thought and its effect on the student’s
behaviour and achievement that we present in fig. 2 with some changes:
Fig. 3: Components of the teacher’s thought
paradigm (Wittrock)
As we can see in fig. 3, the teacher’s
thought/beliefs operate following three stages which cyclically conform each
teacher’s mental system:
a)
Pre-action thought / beliefs, which include the previous ideas,
preconceptions and initial opinions we have at the moment we are preparing and
planning the teaching units.
b)
Interactive thought / beliefs, which operates while implementing
what we have planned, when we teach what we have prepared in class and a
variety of interactive processes take place.
c)
Post-action thought / beliefs, these include the ideas we have
formed after we have finished our teaching, as a result of the unpredictable
events that may have happened in class, after we have observed the students’
reaction and the outcomes obtained. In this 3rd stage, our
preconceptions and initial ideas may have changed when confronted with reality
or may be confirmed and consolidated.
Stage 1: Planning
stage: lesson planning, preconceptions, pre-action thought
What curricular ideas underlie this lesson
planning? What kind of preconceptions are taken for granted when this teacher
designed the unit in this way?
Stage 2: Implementation: teaching performance, learning processes, interactive thought
If a group of students is available, try to
implement this lesson planning. If you do not have any students, think about
the students’ reaction when receiving this teaching and the ideas and concepts
that you would question and revise as a result of the teaching and learning
process.
Stage 3: Post-action thought / beliefs: after the students’ reaction and
the questions that emerged and were raised in the second stage, what initial
ideas have you changed? What modifications are necessary for the future? What
theoretical aspects do you think are still weak and need more research and
practical verification to be fully accepteded?
______________________________________________________________________________
The sources of mental constructs are many, and come
from both received and experiential knowledge, as well as from personality
factors, social factors and cultural factors. Even novice teachers, with no
direct classroom experience or formal training will have mental constructs
about teaching, derived primarily from what Gutiérrez Almarza calls their
"apprenticeship of observation" (1996:51). By this she means their
experiences of being taught and of watching teachers teach.
Obviously action research can provide a tool
for teachers to discover and examine their own individual mental constructs.
2.6 DATA COLLECTION
Other steps in classroom research are to decide
what we want to investigate, in which way and how to approach it. On the other
hands, the sort of data we can use to analyse classroom language can be
obtained from two different sources:
-
Using
reports of earlier classroom research
-
Watching
language learners
With
the first approach we can read different types of research and replicate those
studies that seem more interesting for our purposes. In this case, it is acceptable to use research plans,
questionnaires or observation instruments from published documents to conduct
our own research, provided we give credit to the original authors. From the
point of view of research, the replication of existing studies is an important
way to go a step further as it is part of the refinement process.
In any case it is important to bear in mind, if
we want to adopt this view, that some studies are not probably as well designed
as we might believe. Accordingly it is usually more appropriate to talk of
follow-up studies rather than of exact replications. This way of approaching
and analysing language data can be advisable if we are not yet working with a
group of learners.
Watching language learners in a classroom
context can definitely be a good source of ideas for investigations or be based
on previous research and/or theory. Regarding this latter point, Allright and
Bailey (1991) have underlined what there is a dynamic tension between these two
opposing points of view on how to determine the topic of an investigation. The
first position, which is associated with experimental science holds that a
researcher should decide in advance what to investigate, on the basis of
predictions generated by theory. In the second view, that more commonly
associated, the research questions and hypotheses arise from the data that are
collected.
Some theorists would argue that any
hypothesis or research question ought to come directly from a theory which
makes predictions that can be empirically tested by some sort of classroom
investigation.
However, as it has been suggested, there are
two problems with this way of approaching research. On the one hand, the
theories of language learning do not always lend themselves to making directly
testable predictions. In addition, some researchers (van Lier, 1988) consider
that classroom lessons are such complex affairs that it is virtually impossible
ever to control the number of different variables that could bias the results
of any attempt to test a particular theory-driven prediction.
On the other hand, the second problem with
putting theory first is that it misses the point that theories themselves have
to come from somewhere.
But, as Allwright and Bailey have put it,
classroom research does not always have to concern itself so directly with
theories at all, whether to test them or to illuminate them. Instead, classroom
research can be directed at trying to understand and deal with the immediate
practical problerns facing teachers and learners. The term for this sort of
work, aimed as it is at investigating and dealing with immediate practical
problems, is "action research".
In any case, whether we begin from the
data-first or the theory first position, it is important to have a specific
issue in mind, a particular problem to think about, because there is a strong
risk of wasting a lot of our own and everyone else's time if we begin a
research project with no clear idea of what we are going to do.
Given these opposing viewpoints on the research
sequence (data-first versus theory-first), it would probably be best to start
off with at least a general issue we want to investigate, and to use your
thought about that issue to help you to decide what sorts of data you will
need.
Another aspect of the decision about what to
investigate is the understandable temptation to look at the most visible things
only, the things that are easiest to observe, to record, and to count.
Two problems arise in this area of deciding
what to investigate. First, the
overall picture we have of classroom language learning from research so far is
already distorted by this bias towards the visible.
Second, the bias towards the easy things to
investigate is a luxury that action researchers at least cannot usually afford.
2.6.1
Interviews
The objective of interviews is to obtain
information by actually talking to the subjects under study. Sometimes, it is
also necessary to ask the students questions about the classrooms events either
in an individual face to face situation or to the whole group of students.
Perhaps the two main disadvatages are that individual interviews are a) time
consuming and b) they often introduce subjective and biased information, given
that interviewees often say what they
think that will please the interviewer (see Appendix 2.9).
Besides that, certain covert variables (e.g.
attitudes, prejudices, interests, needs analysis, learning strategies,
motivation, etc.) cannot be fully studied unless we interview the subjects
involved in the research work and collect information about their beliefs,
feelings and opinions. According to the degree of explicitness and structure,
interviews can be “open” and “semi-open or “semi-structured”.
a)
Open interviews provide the interviewee with very open
questions which allow a great freedom of expression to give the answers. Very
often, they are carried out through informal talks about the topics under
study.
b)
Semi-open interviews provide some core questions
predetermined in advance but the subjects
interviewed still feel quite free to answer them.
c)
Semi-structured interviews consists of specific and defined
questions determined beforehand but tyhey allow some ellaboration in the
questions and answers.
d)
Structured interviews include very specific close questions that require very
specific answers.
In general, semi-structured and structured
interviews need some kind of interview schedule, checklist or questionnaire
which presents the questions to be asked and the topics to be discussed with
some space for the interviewer to write down the answers.
2.6.2 Questionnaires
Questionnaires are printed and used for data
collection. They include questions or statements that are responded in an
anonymous way (see Appendix 2.10). They are similar to interviews in the type
of data that are provided but in questionnaires the questions/statements are in
a written form whereas in interviews they are oral. The use of questionnaires
has also some advantages:
a)
They
can be administered to large groups of subjects
b)
The
data provided tend to be quite uniform and standard
c)
If
they are applied to groups of subjects at the same time, the data collected can
be very accurate.
But there are also some disadvantages:
a)
If
they are sent home to be answered the response rate is usually very low and
that may affect validity and reliability.
b)
Sometimes,
the data provided can be very subjective and need to be contrasted and checked
in other situations
Questionnaires, the same as interviews, can also be open, semi-open, semi-structured
and structured. Quite frequently, structured questionnaires use the Likert scale to grade the statements
from 1 to 5 (e.g.: 1 = never, 2 = seldom 3 = sometimes, 4 = often, 5 = always).
The semantic differential is another
technique which grades the items with a bipolar scale (e.g. bad/good; high/low,
...)
2.6.3 Case study
This type of research focuses on one or a few individuals. In most cases, these studies
are longitudinal, that is, they follow the individuals for a long period of
time. Many case studies have centred, for example, on the development of one subject’s (or a few ) interlanguage for
a few years in order to describe the process. For example, Ellis work on the
grammatical and semantic development of three subjects (1984) and his
recommendations about the importance of the learner’s initiative in interaction
for second language acquisition. This type of research is very important to
draw up hypothesis that need to be tested later on with more representative
samples.
2.6.4 Diaries
Even though diaries can be anecdotal and
subjective, they can provide very important clues about what the learners feel
and how they process information. There are many mental operations and
strategies that play a crucial role in language learning and that cannot be
observed and studied unless the student reports about them. This technique is
very useful to explore the learning strategies that the students use in
different situations, especially when they receive metacognitive instruction and we want to know what
kind of effect has it got on the student
learning.
2.7 PROBLEMS IN COLLECTING DATA
In the early days of classroom
research observation instruments were used which focused primarily on the
teacher's behaviour. But as language classroom research has become more deeply
involved with issues in language learning, a clearer focus on the learners, and
on the interaction among learners and teachers, has superseded the earlier
emphasis on teacher behaviour. One result of this shift in focus has been a
decrease in the use of observation schedules and an increase in the use of
discourse analysis of transcribed data.
2.7.1 Discourse analysis and transcription
Discourse analysis" refers to a
variety of procedures for examining chunks of language, whether spoken or
written (Allwright and Bailey, 1991).
In the case of classroom research,
discourse analysis usually involves the analysis of spoken language as it is
used in classrooms among teachers and learners. Van Lier (1988:122) describes
it as "an analysis of the processes of interaction by means of a close
examination of audiovisual records of interaction".
The focus on lengths of oral
discourse in classroom interaction leads us naturally to units of analysis
which are different from the concepts of sentence, clause, or phrase, as these
terms are used in syntactic analysis. Instead, discourse analysts, who are
interested in the way talk is structured, have investigated concepts such as
utterances, repair strategies, topic nomination and turns.
Discourse analysts typically use
transcripts and audiotaped or videotaped interactions as their data base. Some
use transcripts and accompanying videotapes in order to document the nonverbal
channel of communications (see Appendix 2.6).
Transcripts vary widely in their
level of technical complexity. They may use standard ortography or detailed
phonetic transcription of speech, depending on the research goal.
2.7.2 Triangulation: the value of multiple perspectives
Anthropologists have borrowed this
term from land surveying to suggest that at least two perspectives are
necessary if an accurate picture of a particular phenomenon is to be obtained. Triangulation can take several different
forms (Denzin, 1970:472).
a)
One of
them is data triangulation, which
means using a variety of sampling strategies, for example, data related to the
time and social situations collected in different occasions to ensure that the
objectives proposed are studied in more
than one way.
b)
Another
is investigator triangulation, in
which more than one observer contributes to the findings to gain more
reliability.
c)
Methodological triangulation refers to using different methods
(for example, observation, analysis of transcripts and self-report surveys) to
collect the data.
d)
Finally,
theoretical triangulation demands
that the researchers approach the data analysis with more than one perspective
on possible interpretations.
The combination of multiple methods, data types, observers and
theories in the same research study is called multiple triangulation (Denzin 1970:472).
2.7.3 Reliability
This term refers to the fact that the research
procedures must be consistent, both over time and across a variety of people
who might use them. It applies to both the data collection and data analysis of
classroom research.
One area of obvious concern about reliability
is in situations where more than one observer is involved in trying to count or
code the same things.
The degree of reliability between observers can
in fact be calculated, and of course steps can be taken to train observers to
improve their "inter-observer agreement" figures. In general, before
going ahead with the coding of a large corpus of data, classroom researchers
strive for at least 85% agreement among observers and raters. This figure is
called "inter-observer agreement" or "inter-rater
reliability". Another matter for concern is the percentage of "intra-observer
agreement"- that is, the extent to which a single observer or coder,
working with the same data, codes or categorises the data consistently after a
lapse or time. The figures range from 0 to 1.00, with decimals nearer to 1.00
representing greater consistency in the observers' use of the categories.
2.7.4 Validity
As a technical construct in experimental
studies, validity takes two important forms. First, there is the notion of
"internal validity". A study is said to have internal validity if the
outcomes of the experiment can be directly and unambiguously attributed to the
treatment applied to the experimental group, rather than to uncontrolled
factors.
Internal validity relates to the extent to
which the results of an experimental study can be reliably and unambiguously
related to the treatment which was implemented.
Chaudron (1 988b) has discussed 3 types of
validation which relate to classroom research:
-
Construct validation: the construct is verifiable and
can be "captured" through various measurement procedures.
-
Criterion-related validity: some form of measurement is used
to measure a trait along with another form.
-
Treatment validation: the treatment was in fact
implemented and that is identifiably different from whatever it was being
compared with.
According to Seliger and Shohamy
(1989:190) the information needed for determining the quality of data
collection procedures is the following:
Technique The information it provides
Reliability whether the scores are
accurate
Test-retest whether the scores are stable overtime
Inter-rater whether
there is agreement among judges about the score assigned
Intra-rater whether
a rater will assign the same score after some time has elapsed
Parallel form whether
two similar instruments supposed to measure the same thing actually do
Internal whether
the test items are related to one another and
consistency measure the same thing
Validity whether it measures
what it is supposed to measure
Content whether
the procedure represents accurately the content it is supposed to measure
Concurrent whether
it correlates well with a different type of instrument which is suppose to measure the same thing
Predictive whether
the measure can predict accurately a certain future behavior
Construct whether it represents accurately
the theory of the variable which it measures
Item analysis whether the items and
questions which appear on the instrument are difficult or easy, and whether
they discriminate among the subjects of the research.
2.7.5 Generalisability
The other form of validity in experimental research,
to set against the internal kind, is of course "external validity" or
generalisability. As the name implies, the issue involves the extent to which
the findings of a study can be generalised, or applied, to other (external)
situations.
Generalisability is connected to the concepts
of "population" and "sample". In experimental science, the
population is the entire group of subjects of interest -in our case, typical
learners-. The "sample" is that smaller group which is actually
studied by the researcher.
It is important to note that reliability and
validity are intertwined1 and importantly so. If the results of an
experiment are not reliable they cannot be applied to other contexts.
Furthermore, in experimental research terms, there can be no "external
validity" (generalisability) without "internal validity". The
reasoning behind these maxims is that the findings of an experiment (if there
are problems in measurement, for example) then there is no point in trying to
apply that treatment in other settings.
We will naturally want to know whether or not
the findings would hold true with our own learners. For this reason, it is
important to know something about the background of the sample of learners
investigated in the study.
When the results of the study are presented in
numerical form, it is common to see "mean scores", which are simply
mathematical averages, and "standard deviations". The standard
deviation is a statistic measure which indicates how much variety there is in a
group of scores. It is a measure of the average distance of the scores in a
group of scores from the average score of that group (see Brown, 1988).
Experimental research aims at
investigating a situation without changing it in any way other than the
treatment, while action research deliberately aims at bringing about changes
for the better in the specific situation being investigated. Another major
difference between experimental research and action research -by definition-
can afford to confine itself to looking for local situations to local problems,
whereas experimental researchers are committed to trying to develop the sort of
understanding that will apply much more globally.
For this reason, the issue of generalisability
is not so crucial in the tradition of action research. In naturalistic enquiry
similarly, and again in contrast to experimental research, generalisability is
not always such a primary focus.
In language classroom research, Van
Lier has argued that generalisability cannot be a major goal because -the first
concern must be to analyse the data as they
are rather than to compare them to other data to see how similar they
are" (1988:2)
But then the whole issue of
generalisabilty looks very different to anyone doing research in naturalistic
tradition. Instead of claiming that whatever has been discovered must be true
of people in general, a naturalistic enquirer will claim that whatever
understanding has been gained by an in-depth study of a real-life classroom may
illuminate issues for other people.
2.8 ANALYSING THE DATA
Data analysis is the final stage of research.
It implies organising the data collected in order to study to what extent the objectives that we predetermined have
been achieved, to check if the
hypothesis formulated can be confirmed or simply to draw conclusions from an
ethnographic study. This analysis leads to the final conclusions of the
research.
In this stage, as with the previous phases,
there is a variety of techniques that
we can use and the results will depend on the type of analysis. But it is
important to notice that there must be a relation between the nature of the
research problem, the research method, the variables we want to control, the
tools used to control the variables and the procedures chosen to analyse the data collected. Some data analysis
techniques will be more appropriate for quantitative research, while others
will be more appropriate for qualitative research. Some of the well known handbooks which offer a gool deal of practical information for data analysis are Seliger
and Shohami, (1989); Brown, (1988); Tuckman, (1978), just to give a few
examples.
2.8.1 Qualitative research
data
In qualitative research, where the
data have been collected through observations, interviews, diaries, or any
other qualitative procedure, the information is gathered in recordings or
written reports. Then the researcher has to identify the most relevant segments
of the text according to an organised scheme. Quite often, some categories
emerge from the data, without having to apply a fix taxonomy. Sometimes, the
researcher does the opposite: (s)he applies a predetermined classification.
In short, the two main types of
techniques that can be identified in analysing qualitative data a the
following:
a)
deriving
a set of categories for dealing with text segments from the text itself
(inductive procedure)
b)
applying
a system of categories or predetermined classification to the data .
According to Tesch (1987), there are some features that are common to
all qualitative research analysis:
|
-
The analysis of qualitative data is systematic, but not rigid. -
The main procedures used are comparison, a search for likeness and
differences. -
In order to be compared and contrasted, the raw data need to be
summarised and condensed. -
As a result of summarising and organising the information, some
preliminary and tentative classifications of categories emerge. -
The analysis is not the final phase of the research project. The
results of each analytical session
point to other questions that need new data. -
Qualitative analysis is a process that demands deep involvement on the
part of the researcher. -
Finally, there is not an exclusive and right way of analysing
qualitative data. It is possible to analyse the information in different
ways. |
2.8.2 Analysing descriptive research data
Descriptive research is generally
analysed by means of descriptive statistics. Some of the most common
descriptive statistical procedures are: frequencies,
central tendencies and variabilities.
a)
Frequencies are used to indicate how often a phenomenon
occurs
b)
Central tendency measures provide information about the average
and the typical behaviour of subjects.
-
The mean is the sum of all scores of all
subjects in the group divided by the number of subjects.
-
The mode is the score which has been
obtained by the largest number of subjects, i.e. the most frequent score in the
group.
-
The median is the score which divides the
group into two parts, so that half of the score are above it and half are below
it.
c)
Variability provides information on the differences
or spread of the behaviours. It indicates how homogeneous (or heterogeneous)
the groups are. The most common variability measure is the standard deviation: the
higher the standard deviation, the more heterogeneous a group is. Another
measure of variability used in statistical analyses is the variance, which is the standard deviation squared.
2.8.3 Correlational research data
Correlation techniques are used to
explore existing relations between variables. For example, if we have data
about the students’ achievement in the EFL class and in the L1 class, we can
correlate data and study the relationship between achievement in L1 and in the
FL class. If a positive correlation
is obtained
it means that there is a close
relationship between both variables. A negative
correlation would indicate the opposite view.
Correlations are indicated by means of the correlation coefficient, which ranges
from – 1.00, indicating perfect negative correlation to 1.00, which indicates
perfect positive correlation. The significant level obtained is very important
since it relates directly to whether the null hypothesis is rejected or not.
The conventional level of rejecting the null hypothesis is p < .05 or p <
.01. When reporting correlations the researcher needs to specify all these
data: the sample size (n) the correlations were based on, as well as the level
of significance (p).
2.8.4 Multivariate research data
There are three well known
multivariate procedures (see Seliger and Shohami, 1989:222-231)
1)
Multiple regression, which is used to examine the relationship and
predictive power of independent variables. In the case of the relationship
between L1 and L2, regression would indicate the prediction of L2 achievement
under the influence of L1.
2)
Discriminant analysis indicates which combination of
independent variables distinguish
between two or more categories of the dependent variable. For example, a
researcher may want to study which combination of variables, L1, motivation,
aptitude, etc. can best distinguish between two types of second language
learners (males/females; learning in formal/informal contexts).
3)
Factor analysis. In this case the interrelationships between
and among the variables of the data are examined in an attempt to find out how
many independent dimensions can be identified in the data. Factor analysis is a
procedure frequently used to validate language tests, for example, to check if
the items of the Cultural tests really measure the cultural competence.
2.8.5 Experimental research data
When a control and experimental
group are used in experimental research, other procedures are commonly used to
compare results: The t-test, ANOVA and Chi-square.
At present , most of the data
analysis techniques described in the previous sections can be performed with
the computer. Some common statistical packages are: Statgraphics and SPSS
(Statistical Package for Social Science).
2.9
CONCLUSION
In this brief overview on classroom observation
and research, we have defined the main key concepts of the field and have
presented a concise survey of the scope, methods and paradigms as well as their
corresponding research techniques. We hope that our readers appreciate the
variety of options they have if they intend to do classroom research (CL).
As we could see, CL on language teaching and
learning has moved from the early simple attempts to identify “good” language
teaching towards the study of classroom
interaction processes. This move has brought specialised researchers and
teachers much nearer in order to study, sometimes in a co-operative way, what
actually happens in the classroom and not to test or apply what other people
think that should happen.
We have seen that , in instructed second
language acquisition, what happens in the classroom is crucial if we want to
understand how language learning takes place. We should not forget that the
learning opportunities that learners have are confined to the classroom in most
contexts. That is why the way teachers and learners interact determine the way
that the language is learnt. Learners
do not seem to learn from the lesson plan but from the way it is implemented:
from the input provided, the practice opportunities and the intake that they
experience.
One basic goal of this unit has been to present
several alternatives for classroom observation and analysis in order to study
what happens in class and draw practical conclusions to improve the teaching
and learning process and the final outcomes. Several models and procedures for
CR have been introduced which involve varying attitudes on variables control
and the intervention in the research study. We think that each method and
paradigm has its place and it is useful for different purposes. The researcher
must select the most appropriate approach according to the type of study that
(s)he aims to carry out. The usefulness of particular research instruments
depend on the purposes of the research, the limits of generalizability and
other factors that we have mentioned in the previous pages.
We have given to observation a very high
status, since it has often been considered the single most important component
of classroom research. In this sense, classroom research is contextual
research. It begins with detailed observation, recording and transcription.
Such techniques include conversational analysis, linguistic variation, etc.
They involve the use of intuitive judgements and decisions but also the
establishment of a coherent research plan which inform about what is going to be studied and how
we aim to proceed (methodology).
Nevertheless, the most important conclusion may
be to recognise the importance of classroom research as a basic tool for
innovation in ELT. When classroom teachers practise research they use their
knowledge and experience to explore specific classroom phenomena in order to
gain a better understanding and find out their right meaning. In this sense, CR
may also contribute to the classroom teacher professional development. This
enquiring practices promote the teacher’s personal construction of his/her knowledge
and help to improve the quality of classroom teaching and learning.
The consideration of teachers as researchers of
their own practice has been fully discussed when describing the action research and reflection-in-action paradigms. These two paradigms advocate a new
philosophy which gives teachers a more active role in the classroom when they
reflect about classroom events and draw personal conclusions to solve the daily
pedagogic problems, retreating from the standardised ready-made techniques supplied
by official research.
But CR is not easy. We must admit that an
enquiring attitude in education demands certain qualities from teachers:
-
A
general background about education and, in our case, in ELT methodology.
-
An
enquiring and diagnostic ability
-
Identifying
the implied theories, approaches and paradigms that lay the foundations of the
classroom techniques that are applied every day.
-
Designing
strategies for future decision-making occasions.
-
A
critic evaluation and reflection which is shared by other colleagues for the
classroom research process to become co-operative.
We hope that the theoretical principles, models
and project work that we have introduced in this unit also encourage our
readers to become researcher of their own practice in order to know themselves
better and improve the results of their
work.
2.10
FURTHER READING
ALLWRIGHT, D AND BAILEY, K. M. (1991): Focus on the language classroom: An introduction to classroom research
for language teachers. Cambridge: Cambridge University Press.
VAN LIER, L. (1988): The Classroom
and the Language Learner. London: Longman
CHAUDRON, C. (1988): Second
Language Classrooms. Cambridge: Cambridge University Press.
WAJNRYB, R. (1992): Classroom
Observation Tasks: A Resource Book for Language Teachers and Trainers.
Cambridge: Cambridge University Press.
SELIGER, H. W.
(1989):
Second Language Research Method. Oxford:
Oxford University Press.
BROWN, J. D. (1988): Understanding Research in Second Language
Learning: A Teacher’s Guide to Statistics and Research Design. Cambridge:
Cambridge University Press.
MALAMAH-THOMAS, A. (1987): Classroom
Interaction. Oxford: Oxford University Press.
ALLWRIGHT, D. (1988): Observation
in the Language Classroom. London: Longman
McLAREN, N. , MADRID, D. (1996): A
Handbook for TEFL. Alcoy: Marfil
MADRID, D., N. McLAREN (1995): Didactic
Procedures for TEFL. Valladolid: La Calesa.