HOW WE TEACH AND WHY: THE IMPLEMENTATION OF AN ACTION RESEARCH MODEL FOR IN-SERVICE TRAINING

Lily Belleli
Abstract
This study investigates the extent to which teachers’ reflection on their teaching affects their perception of what happens in their classrooms. The research was based on the deployment of an Action Research model of counselling with a group of seven in-service teachers, once a week, for a period of six months. The effects of the reflection-contact period were investigated through the use of a Curriculum Perception Instrument based on Kelly’s Personal Construct Theory. The instrument identifies the constructs underlying teachers’ practice before and after the contact period. A semi-structured interview was also applied as a measure for changes in teaching behaviour. Findings reveal interesting differences in teachers’ perceptions of their teaching and of curriculum materials before and after the contact period.

Theoretical background
The main objective of the in-service project was to help teachers understand and identify their approaches to teaching and teaching materials. The project also aimed at developing teachers’ ability to revise and adapt curriculum materials to the needs of diverse student populations as they shared experiences with other colleagues in team work (Ebbut & Elliot, 1985).
The underlying assumption was that instructional improvement is brought about through getting teachers to reflect on their own teaching (Easen, 1985; Fullan, 1982). Within this context, reflection comprised two levels: (a) analysis of understanding, and (b) monitoring of performance.
A teacher’s capacity to identify and become aware of the implicit theories guiding his/her practice is a matter of analysis. Bringing to a conscious level the structures operative in one’s teaching is a way of discovering one’s implicit theories. Implicit theories can also be revealed by investigating the criteria by which teachers select, adopt and interpret curriculum materials available to them. The criteria by which these are collected and interpreted reveal the constructs or theories inherent in teachers’ practices; which in turn determine their professional activities in the planning and implementation of lessons (Brumfit & Rossner, 1982).
The level of performance monitoring entails the revision and reformulation of assumptions in the light of practice, as well as adjusting and developing curriculum materials as a result of these reformulations.

Kelly’s theory of personal constructs
The main instrument developed for evaluation in the project, the Curriculum Perception Instrument, has its theoretical underpinning in Personal Construct Theory (Kelly, 1955. See note at end of paper.). According to Kelly, people strive to make sense of new experiences by developing a set of personal constructs.
Constructs cluster in a system that represents an individual’s knowledge and view of the world. The system’s boundaries are defined as a result of one’s interaction with the environment. Consequently, no two systems are the same; although they may be similar when people inhabit similar internal, as well as external, worlds (Bannister, 1970). This conceptual framework provides a tool for investigating the unique processes of development of individual teachers and groups of teachers.
Teachers spend much of their time interpreting and transforming materials into lesson plans. This would mean, in Kelly’s terms, using personal constructs in order to make sense of their teaching environment. Since constructs differ from individual to individual, every teacher will make predictions and construe activities from curriculum materials in his/her unique manner.

The research
I conducted regular counselling sessions (once a week for two hours, over a period of six months) with a team of seven in-service teachers, from four different schools of a deprived neighbourhood in the Sharon area in Israel.
The team systematically engaged in analysing practice by reporting their work to one another and by evaluating and developing curriculum materials. As a tutor, my role was to facilitate the interaction among the members of the team and help them clarify their own thinking. The point of departure was the problems of teachers as defined by the teachers themselves. Thus, counselling sessions were devoted to discovering the pertinent practical personal knowledge that the teachers possessed to help solve pedagogical problems. This knowledge was applied by the team to working out methods and preparing materials suitable to the special needs of their classes.
The theoretical framework for the design of the project followed an ‘action-research’ cycle based in the model of action-research described by Kemmis & McTaggart (1982). In practice, the cycle of activities included:
1. identifying the main idea or statement of the situation one wishes to change or improve;
2. deciding on a field of action – describing the facts of the situation and the context in which it takes place;
3. constructing a general plan of the actions to be undertaken;
4. monitoring the process of implementation and its effects on action;
5. observing the effects of action; and
6. revising the general plan through reflection on the effects of action as a basis for future planning and subsequent action.
The study was designed to assess changes in the teachers’ behaviour and in their conceptualisations of teaching while they were taking part in the sessions as well as after they had completed the training.
I will describe changes in the relationships among the members of the group and in teaching behaviours that could be observed during the sessions. In addition, I will report on significant changes in how teachers construed curriculum materials after the six-month period. In this connection I attempted to answer the following questions:
1. How did teachers’ constructs develop in the course of the contact period?
2. Could any prevalent constructs, common to all teachers, be identified after a six-month contact period?
3. To what extent did changes in teachers’ constructs reveal patterns of teaching behaviour?

Procedures
The first session was devoted to open discussion of various pedagogical issues in order to release tension and create a non-threatening atmosphere in which teachers would experience the benefits of sharing experiences with each other. My role as a co-ordinator at this point was to provide the team with a basic framework within which they would organise their thoughts, share insights and consequently arrive at a definition of the problem they wished to explore and try to solve.
During the six month contact-period the group completed four cycles, each with a planning, action, observation and reflection phase. Some of the problems that were analysed through the four cycles were:
(a) weak learners in mixed ability groups (the need for a more structured rationale and methodology);
(b) discrepancies between stated aims and procedures (teachers reported on different tasks they had devised that seemed suitable in terms of content, gradation and clarity, yet the results in class were not satisfying);
(c) teaching or testing (by investigating different questions that teachers had posed during their lessons, the balance between teaching and testing activities was discussed).
The most interesting feature that characterised the four cycles completed by the group was the way in which plans changed and developed in the light of observation and reflection. In relating to materials development there was a drastic shift from devising specific tasks for specific levels to devising frameworks suitable for weaker or stronger learners in general. In terms of group work, there was a change in the perception of collaborative team work, from a view of team work as a product facilitator (i.e. in order to produce materials) to a view of collaborative team work as a process facilitator (i.e. in order to help reflect on and analyse practice).
Throughout the six-month period, teachers engaged in continuous, concrete and precise talk about teaching practice. Thus, they built up a register of professional language which they could share during the workshops. They planned, designed, and evaluated teaching materials together. Workshops were characterised by a climate of interaction, collegiality and exchange of ideas. Teachers experienced a sense of meaningfulness and practicality relatively early in the process, because they tried the materials in their classrooms and reported their findings, impressions and insights to the group.
They were constantly encouraged to analyse discrepancies between ‘aim’ and ‘procedure’ as a basis for the further development of curriculum materials. It is important to point out that the adaptation and development of material for their own schools constituted the basis for experiencing the meaning and practicality of the project among team members. Furthermore, it generated the need for collective ‘reflection on practice’.

Factors affecting implementation:
Fullan (1982) identifies major factors influencing the implementation of any kind of educational change, several of which are of salient importance to the implementation of the project reported on here. We must view proposed changes in terms of: need, relevance, clarity, complexity, quality and practicality. Our project focused on the identification of special pedagogical problems perceived as ‘priority needs’ by the team. The relevance of change to a solution was thoroughly discussed during the ‘reflection and developing’ action phases and much time was devoted to clarifying the problem(s) and identifying ‘goals and means’. The project was implemented in an incremental manner and teachers were not expected from the beginning to interact in a collaborative manner. Gradually, as a result of recognising the benefits of team work, collaboration developed. Materials development was used as a starting point and as a springboard to reflection on practice. Materials prepared by the team were relevant to their teaching at that particular time and workshops enabled participants to spend time on preparation. This was a major feature of the project.
Three factors, according to Fullan, relate to the social conditions: the history of innovative attempts in the district, the adoption process, staff development and participation. Participating teachers had had no previous experience with implementation attempts and some of them were sceptical at the beginning about the purposes of the project, and the potential for implementing change. The project was taken very seriously by principals and teachers and there was full collaboration. However, participation in decisions during the implementation and ‘planning ahead’ phase was solely a team process; other members of the school staff were not involved in decision making of any kind.
One of the members of the team, a veteran teacher, was responsible for providing constant support to two members of the team who were inexperienced. This was done in the form of regular meetings during and after school sessions. The purpose of these meetings was to discuss practical classroom problems. All team members took equal part in decisions concerning project operations and modifications, regardless of the extent of their teaching experience.
A major drawback was the fact that there was almost no possibility for paired observation, due to the constraints of the schools’ scheduling. Most of the teachers taught on the same days, at the same hours. It was, therefore very difficult to release them on a permanent basis for paired or group observation in addition to the two hour workshop.

Instruments
In order to monitor the project the team relied on several techniques. Observation sheets with criteria devised by the group were used as tools for observing lessons. Field notes and protocols were kept by members of the team during the workshop sessions. In addition, observations, interpretations, reactions and feelings were noted in a diary kept throughout the six-month contact period.
The main instrument used for the evaluation of the project was a curriculum item repertory grid (CIR) (Ben-Peretz and Katz, 1980). This is an extension of the ‘role repertory grid’ used in psychology, and is based on the methodology of triadic sorting (Bannister and Mair, 1968:54). Triadic sorting methodology requires respondents to relate to random triads from a set of items, and to decide for each triad, which of the three items differs from the other two, according to a self-generated criterion. By establishing the criteria for comparing and contrasting items, the subject is expressing a personal interpretation of the materials in question as a dichotomous construct.
The instrument (CIR) consisted of a set of fifteen curriculum items, selected from a set of available EFL materials. Items judged to represent diversity of form and content were included. The items included paragraphs, illustrations, unit headings and exercise types extracted from units of a textbook.
Every participant received a copy of the fifteen items numbered from one to fifteen and a grid form to be completed during the investigation. Teachers were asked to take random triads and establish a self-generated criterion by which one of the items differed from the other two. The criterion and the number of the item which ‘differs’ was then indicated on the grid form. The procedure was repeated ten times with the same fifteen items. When respondents were asked to take random triads, it was possible for the same item to be reviewed more than once; nevertheless, every time the item reappeared, it would be in a different environment/context. Hence, it would generate new possibilities for construal (establishing constructs).
The constructs generated by participants were then classified by investigators. The investigator and two teachers who were not taking part in the procedure classified the constructs generated under superordinate categories. Inter-rater reliability was above 70%.

The semi-structured interview
To evaluate the long-range effects of the contact-period upon teachers’ constructs, a semi-structured interview was carried out four months after the end of the contact period.
It was agreed by investigators that teachers’ attitudes in regard to collaborative team work and the particular framework developed for the project would be explored in as informal a manner as possible. Instead of asking predetermined questions, we opted for free discussion. Points for discussion were extracted from the CIR findings. The interview was tried out with two typical respondents, commented on, and revised in the light of such comments.
Seven participants were interviewed on an individual basis. Each interview lasted for about three quarters of an hour. The interview was carried out by a trained teacher colleague who had not met the participants before. She met the teachers in their own schools.
The interviewer aimed at creating a balance between friendliness and objectivity. At no point was there any attempt to guide the participants or to bias the direction of their responses. On the contrary, the interviewer repeatedly questioned the relevance and validity of the project. The transcribed interviews were double-checked by two readers and areas of disagreement were analysed and discussed.

Findings – Repertory Grid
Examination of the criteria and their distribution among the superordinate categories disclosed interesting predominant constructs in relation to curriculum materials among all seven respondents.
At the beginning of the project the category of highest frequency was: text/task type/content of materials and format of materials. The next category of highest frequency was ‘pupils’ tasks specified by materials’. The category of lowest frequency was: ‘the methodology advocated by materials’.
At that point, teachers were concerned mostly with the ‘form’ and ‘content’ of materials used for teaching in terms of language, exercise, type, topics, grammatical points and skills. Little emphasis was given to differences in the way in which ‘items’ could be presented or practised in the classroom (i.e. methodology implied by materials).
In the category, ‘pupils’ tasks specified by materials’, teachers were mainly concerned with the type of task the pupil is required to perform rather than with the extent to which the task takes the learner’s needs/lacks or desires into consideration. Here, again the emphasis was on the ‘requirement’ of the task, in terms of ‘content’ rather than on the learner’s capacity to perform a particular task.
Participants’ personal construct systems indicated a high degree of similarity in their relation to curriculum materials. The predominant categories occurring in all seven grids emphasised the ‘What’ rather than the ‘How’ implicit in the items. At the end of the project the two categories of highest frequency were ability level implied by the materials (71%) and methodology advocated by materials (50%) – format of materials’ occurred in only two grids.

Findings – interview
Most of the constructs disclosed by the CIR were also elicited in the interview. They were evident in teachers’ accounts of the different activities they had tried in their classrooms. Teachers referred to the aims they had set as well as the procedures used. Moreover, constructs relating to ‘methodology’ and ‘students’ knowledge’ which were not manifested in a number of grids were elicited in the interview.
It was not possible to identify patterns of teacher behaviour in terms of teacher/pupil interaction. However, most of the participants expressed concern for the degree of student involvement (and the implication for classroom organisation) as advocated by the activity they were describing.
In the interviews, teachers also described classroom practices. Those cited included:
a. eliciting information from students (through brainstorming for example) as a way of introducing material or as a follow-up activity;
b. introducing new language material through known concepts;
c. adapting activities involving the production of language to activities involving the recognition of language in weak classes;
d. using tasks that require non-linguistic response in weak classes;
e. grading tasks and the amount of teachers’ input/students’ output according to the level of the class;
f. integrating the four language skills through a common topic and different types of classroom management for different types of activities (pair and group work used for information gap activities).
Most of the interviewees showed concern for students in terms of the world knowledge they bring to the English lesson. The interview revealed how a number of teachers used this concept of ‘student world knowledge’ in the planning and implementation of their lessons.
General attitudes to the project’s framework were disclosed by the interview. Teachers raised a number of points in relation to the project. They felt that a great deal of time was spent on defining the problem, developing action, monitoring, and revising the plan. They also felt that feedback sessions after monitoring the materials in the classroom were productive. Feedback sessions provided the teachers with the opportunity of seeing how the same topic could be realised through different teaching styles. Teachers mentioned that there was a constant pooling of ideas and the workshops provided the teachers with a forum for discussing day-to-day problems.
They felt that the interaction among them was productive and that decisions made in the workshops were their responsibility. Contact with teachers from different schools was felt to be an encouraging experience. Interviewees were, however, quite honest in pointing out the weaknesses they were aware of. They recognised that personal biases caused certain innovations to fail at the level of procedure. Moreover, they felt that there was a need for a tutor to guide the group. However, they all stressed that the tutor should be an outsider and not a member of the group.

Conclusions
From analysis of the findings it was possible to draw several conclusions:
1. Teachers learned to construe elements differently by relating to possibilities of developing different teaching strategies for the same activity item. The criteria they developed did not specify exactly how to go about an activity, but rather revealed a general awareness of the existence of different teaching strategies applicable to a single item.
2. A major common concern amongst the seven participants after the contact period was the extent to which an item demanded the use of recognition or production skills. Here again, constructs revealed teachers’ awareness of the possibilities of modifying an ‘activity’ so as to cater to different ability levels.
3. Criteria related to ‘content of materials’ put little emphasis on the skills, vocabulary, or grammatical points inherent to an item. Rather, teachers’ criteria stressed the students’ world knowledge and production/recognition skills implicit in the content.
4. An interesting discovery was the finding that constructs suggested by all seven participants were predominantly complex and it was possible to extract many ‘categories’ from a single criterion. Teachers tried to interrelate form, content, teaching strategies and cognitive demands in one criterion.
In comparing the findings at the end of the project with those obtained at the beginning, we found that teachers came to show more concern for the ability level implied by a specific task. Criteria that considered the difficulty of the item (in relation to a student’s age and/or the complexity of material) were seen from a different perspective. The emphasis was on the ‘ability level’ of the activity and on whether the task involved the recognition or production of language. There was less emphasis on the categories of ‘content of materials’ and ‘text type’.
In their concern with the demands of pupils’ tasks, teachers emphasised not the task’s content/or format, but rather the field of knowledge incorporated as well as the ability level it subsumed.
Teachers’ overall focus was on the teacher and his/her ability to manipulate different teaching strategies. Wherever a concern for the student was manifested, it was usually through constructs which related to the interaction between ‘student’s ability level’ and ‘teaching strategy’.
It is possible to evaluate the effectiveness of the action-research model in terms of the definition of ineffective in-service training as provided by Hopkins (1986). He claimed that most current in-service work is ineffective because it is frequently based on single shot workshops involving large or undifferentiated groups of teachers, providing limited time for teachers to learn, and enabling little evaluation or practical follow-up support. Furthermore it is not usually linked to a particular classroom or school problem (1986:268). In evaluating the project described here according to these points, it is possible to say that:
• workshops were held on a regular basis;
• teachers participated as planners, decision-makers, and evaluators of in-service activities relevant to their immediate practice;
• teachers had the opportunity to apply and assess specific techniques in their classroom;
• teachers interacted, shared ideas, and provided assistance to each other during and after the workshops.
In Hopkins’ account of the problems to be resolved in school focused in-service, he mentions the danger of ‘becoming unnecessarily limited to immediate problems and attached only to the teaching role’ (1986:56). He therefore urges institutions adopting INSET (in-service training) models to focus on a ‘balanced approach’, i.e., a balance of action and reflection, which provides teachers with the opportunity to ‘conceptualise more complex approaches to teaching’ as well as ‘to observe specific practices and practice specific techniques’ (1986:54).
This balance of ‘action and reflection’ was a major factor taken into consideration in the design and implementation of our project. The cyclical nature of the model provided teachers with the opportunity to reflect ‘in’ and ‘on’ action. Nevertheless, since the programme concentrated on identifying and solving specific immediate problems as defined by the group, ‘reflection in action’ occurred mainly on the level of analysis of specific practices rather than on the level of conceptualising broader aspects of teaching/learning.
Despite the fact that the project could be defined as ‘effective’ in Hopkins’ terms, it would seem appropriate to evaluate its ‘degree of effectiveness’ as far as the project’s overall objectives were concerned. The programme was effective in terms of the extent to which it increased self-confidence, reduced anxiety and developed an awareness of individual differences and a tolerance to others. Teachers also developed an ability to transfer concepts, principles and skills from one medium to another and to bring intuitive practice to a conscious level.
However, the programme was far less effective in fostering an awareness of processes and perceptions of how learners at various levels of development respond. Although the CIR and interview findings point to an increase in teachers’ concerns with the ‘ability level implied by the task’ and ‘methodology implied by materials’, the increase was manifested only on the level of skills and application. The shift from intuitive action to ‘conscious action’ was not apparent on the level of understanding students’ learning strategies. This might have been achieved through the adoption of a different focus. A framework that would focus on systematic individual student observation could provide insights into students’ construct systems and their patterns and behaviour. We did not account for this dimension in a fully systematic manner mainly because teachers could not engage in paired or group observation beyond the two-hour workshop.

Personal constructs and the in-service framework
The impact of the framework upon teachers’ constructs can also be evaluated by comparing the way in which the four cycles of action research developed and the way in which teacher construct systems developed. Collaborative work, mutual reinforcement and support had strong implications beyond the surface level of experiencing a new model of interaction.
Positive attitudes and predispositions are crucial for the development of cognitive skills (Stern, 1983). Hence, teachers’ ability to apprehend their teaching and curriculum materials in a more divergent and complex manner, (as reflected through their personal constructs) bore a direct relationship to the particular framework of collaborative team work developed. By sharing insights, learning about others’ behaviour and supporting one another, teachers’ motivation and self confidence were enhanced. This led to the development of a greater capacity for diversification of their approaches to teaching.
If constructs are dynamic and influenced by past and present experience (Kelly, 1955), then it is anticipated that teachers’ future practices will operate less at a level of surface structure (i.e. focusing on the appearance of materials) and more at a level of deep structure (comprehending the underlying purpose of materials and teaching activities) as a result of reflection. In the four cycles of action research the teachers’ constructs were characterised by a shift from ‘What’ (i.e. what was done in the classroom) to an emphasis on ‘How’ and ‘Why’ (i.e. the rationale behind action). In much the same manner, constructs developed from a focus on the content and format of the item (i.e. What) to the methodology implied by the task (i.e. How).
Thus, in terms of the project’s overall outcome, the development of teachers’ interpretative abilities was more important than the development of materials.
From this study, I conclude that it is possible to draw on personal construct theory not only as a diagnostic tool but also as an integral component of teacher education programmes.
A methodology based on personal constructs would encourage teachers to analyse their patterns of construing in relation to different curriculum materials during the course of the programme (and not only at its termination). This process might further sensitise them to learners’ perceptions of curriculum material and to perceptions of how these are similar/different to their own perceptions.
Learners and their construing patterns is an area that needs to be fully accounted for in the design of future in-service programmes.

A NOTE ON PERSONAL CONSTRUCT THEORY
Each individual can be seen as perceiving the world in terms of important elements and relations among those elements. This individual set of personal constructs is the basis on which we each build our lives.
Triadic sorting is an investigative technique for exploring such constructs. We can best explain it in terms of doing it. Choose a topic. Let it be discipline problems. On separate pieces of paper, write down several important issues inside this topic. Some of them might be: social background, drug abuse, motivation, student-student relations, outdated school policies, insensitive teaching. Some of these items might never have occurred to you, but that is part of the point: different issues will be important to different individuals at different times.
Turn your pieces of paper face down and select three at random. Let us take social background, student-student relations and insensitive teaching as examples. Divide these three into a pair, which have something in common, and the odd one. Now say what it is which the pair have in common, and/or what it is which makes the third one different from the other two. These two descriptions of similarity and difference make up two poles of a construct.
To continue our example above, one person might say that social background and insensitive teaching belong together because these are external influences on the group, while student-student relations are internal to the group. So, this construct has the two poles of external to group and internal to group. Your construct may well be different.
Having established this construct, we can now go back to the other elements we listed and place them along the continuum between the two poles. What about drug abuse? Is that more like the first pole or the second? As well as placing our other elements on a scale between the poles of this construct, we can select another three items at random and make explicit another of our constructs. Once again, we can place the other elements somewhere between the poles of this construct.
As we continue in this fashion, we build up an intermeshing grid of our personal constructs. This is the repertory grid, or repgrid.
These techniques are extremely powerful in helping a person become more aware of thoughts and feelings about issues under investigation. In the hands of a sensitive investigator, the accompanying talk and clarification can be even more illuminating.