A Dissertation Submitted to the Faculty of English

of The University of Birmingham

for the degree of



This dissertation consists of 14,302 words

School of English

Centre for English Language Studies

University of Birmingham

Westmere, Edgbaston

Birmingham B15 2TT


December 1996


Personal Construct Repertory Grids, often referred to as "rep grids," have a long and distinguished history as a research instrument in psychological, educational and cross-cultural studies. However, they have seen only limited use in EFL-based research. This paper presents the background, theory and terminology associated with Repertory Grids. The research tradition, reliability and validity of repertory grids will also be discussed. For EFL teachers unfamiliar with the procedure, a step-by-step description is provided on one way to use repertory grids for cross-cultural research in the second language classroom. The remainder of this paper covers the first documented use of repertory grids in a Japanese EFL setting.


Many thanks to the following:

Dr. Chris Evans, Senior Lecturer in Psychiatry at the University of London and expert in the field of PCP, for his invaluable help with analyzing the data for this study.

Professor A. Charles Muller of Toyo Gakuen University in Chiba, Japan, an internationally known Confucian scholar, for graciously allowing the use of his recent translation of Confucius' "Five Books" (Muller, forthcoming, Five Chinese Classics).

Professor Yukihiko Ushiyama of Niigata University, whose help with Internet access was vital for the completion of this research.

James B. Brown of Keiwa College and Dr. Malcolm Benson of Hiroshima Shudo University, for reviewing earlier versions of this manuscript, and for their many helpful comments.

Professors Yukio Yamazaki and Kazuo Fukuda of Niigata University, whose discussions of shushigaku over lunch paved the way for interpreting this research.

Hiromi Hadley, my spouse, for her help in translating the data from this research into English.




1.0 Introduction: Cultures, Classrooms and Research Instruments

1.1 Purpose

2.0 George Kelly and Personal Construct Psychology

2.1 Kelly's Philosophy

2.2 Elements, Constructs and Bipolar Constructs

2.3 The Personal Construct Repertory Grid

3.0 Research Tradition

3.1 Reliability and Validity

3.2 Reliability of Repertory Grids

3.3 Validity of Repertory Grids

4.0 A Repertory Grid Procedure

4.1 Define Your Purpose and Area of Research

4.2 List the Elements

4.3 Eliciting Constructs

4.4 Linking the Elements

4.5 Analyzing the Data

5.0 Cross-Cultural Research in Japan

5.1 Research Question

5.2 Subjects

5.3 Research Procedure

5.4 The First Session: Element Elicitation

5.5 Results and Translation of Elements

5.6 Justification for Selected Elements

5.7 Second Session: Repertory Grid Procedure

6.0 Presentation and Discussion of Data

6.1 Interpretation of the Elements

6.2 Interpretation of the Constructs

6.3 A Piece of the Puzzle: Confucianism

6.4 Qualities of the Confucian Superior Man

6.5 Neo-Confucianism

6.6 Confucian and Neo-Confucian Influence in Japan

6.7 From the Meiji Era to World War II

6.8 Reform and Tradition: 1945 to Present

6.9 Teacher as Superior Parent

7.0 Implications for Teachers

8.0 Final Considerations: Strengths and Weaknesses

8.1 Future Research

9.0 Conclusion






Zan Yu was driving for the Master on a trip to Wei. Confucius said, "How populous it is here." Zan Yu said, "Once there are so many people, what should be done?" "Enrich them," said the Master. "Once they are enriched, what next?" "Educate them." (Analects 13:9)

1.0 Introduction: Cultures, Classrooms and Research Instruments

Like Confucius, many EFL teachers often express wonder and astonishment as they enter their first Asian classroom. Looking over the multitude of students packed into their classrooms, they may ask themselves, "How do I enrich them? How can I educate them?"

The search for answers to these questions will lead many on a journey. During that time a few will develop a deepening awareness of something they already knew: Their schools and classrooms, while appearing deceptively Western, are actually unique cultures within the foreign culture itself. Greene and Hunter (1993) are among those who define the EFL classroom as a special culture:

If 'culture' can be loosely defined as a particular system of actions and assumptions within a social setting, the oral language learning and teaching which occur within an EFL classroom is a culture (p. 9).

Typically the EFL classroom culture is a synthesis of two distinct cultures, that of the teacher and of the learners. Richards and Lockhart (1995) note that the teachers' culture finds its strength in well-articulated axioms which converge on concepts such as "success" and "failure":

Teachers' belief systems are founded on the goals, values and beliefs teachers hold in relation to the content and process of teaching, and their understanding of the systems in which they work and their roles within it. These beliefs and values serve as the background to much of the teacher's decision making and action . . . (p. 30).

Unfortunately, few EFL teachers know as much about their students' educational beliefs and values. Schumann (1978) and Anderson (1981) both explain that if language teachers fail to gain deeper insight into their students' culture of learning, social and psychological distance develops between them and the learners, and also between the learners and the target language. It is for this reason Shimizu (1995) urges language teachers to find out what their learners look for in a good teacher:

. . . research on this topic is doubly important because not only are student attitudes towards them (as teachers) important, but also because they are representatives of the culture which speak the target language. Viewed in this manner, attitudes towards foreign teachers could adversely affect student motivation not only in the classroom, but also in terms of a student's desire to continue learning the language (p. 6).

Teacher-researchers in English language teaching normally use questionnaires to study their students' opinions and cultural beliefs. But while surveys are an economical method of data collection, the designers of questionnaires must constantly guard against "researcher-bias." Researcher-bias crops up when questions focus on issues pertaining to the researchers' worldview. For example, in Japan there are several excellent reports on student motivation (Greer 1996; Kobayashi et al. 1992; Kirk 1992). These studies are of great interest to teachers, but it is uncertain whether motivation is such an all-consuming concern for students. Rather than shedding light on the students' culture of learning, such research runs the risk of reinforcing values and beliefs already held by teachers.

However, there is another research instrument that has potential for lessening these cross-cultural challenges to ELT research. Called the Personal Construct Repertory Grid, it was designed " . . . to avoid, as far as possible, the limitations of a particular time and culture" (Bannister and Fransella, 1971:16). According to McCoy (1983), Personal Construct Repertory Grids, or "rep grids":

. . . allow subjects to reveal their own judgements in their own vocabulary regarding some important set of elements in their own experiences (p. 175).

If the rep grid technique truly gets around some of the problems associated with researcher-bias, it would be of great value to the teacher-researcher interested in cross-cultural studies.

1.1 Purpose

This paper is divided into two parts. The first part presents the background, theory and terminology associated with rep grids. The research tradition, reliability and validity of repertory grids will also be discussed. For EFL teachers unfamiliar with the procedure, a step-by-step description is provided on one way to use repertory grids for cross-cultural research. The remainder of this paper will study a research project which used repertory grids in a Japanese EFL setting. After interpreting this study's substantial qualitative findings, we will conclude by answering the following questions: 1. Are rep grids useful as a cross-cultural research tool? 2. What value do they have for teachers and students in an EFL context? 3. What strengths and weaknesses are inherent to this research method?

2.0 George Kelly and Personal Construct Psychology

Repertory grids were developed forty years ago by George Kelly, an educational psychologist who developed a cognitive science called Personal Construct Psychology (PCP) (Cohen and Manion 1987). Kelly passed away more than twenty years ago, but his theories continue to influence not only the field of psychology, but also other disciplines such as management, logic and artificial intelligence (Shaw and Gaines 1992). Every year the PCP academic community produces numerous books, journals, and newsletters. Besides holding a biennial international conference, there are excellent World Wide Web sites in Canada, Australia, Spain and the United Kingdom which keep researchers informed about new developments in the field.

2.1 Kelly's Philosophy

Personal construct theory is systematically organized in Kelly's two-volume work: Principles of Personal Construct Psychology (1955). It consists of a fundamental postulate and eleven corollaries (see Figure One). While it is not within the scope of this paper to cover all the aspects of Kelly's philosophy, a brief description will suffice.

Kelly's fundamental postulate states that " . . . a person's processes are psychologically channelized by the way in which he anticipates events" (Kelly 1955, p. 46). People experience events and interpret them. Their future actions are based upon their previous experience. If a student is scolded by a teacher for talking in class, based upon what she has previously experienced with the teacher, the student may behave differently in anticipation of what the teacher will do (which may include construing that the school administration will ultimately subvert the teacher's efforts). This is further explained in Kelly's Construction Corollary, which describes humanity as "man-the-scientist" (1955, p. 4):

What I think this view of man as the paradigm of a scientist - and vice versa - does mean is that the ultimate expression of human behavior lies in examining man's undertakings, the questions he asks, and the strategies he employs, rather than in analyzing the logical pattern and impact of the events with which he collides (Kelly 1969b, p. 16).

Theorizing that "a person anticipates events by construing their replications" (1955, p. 50), Kelly explained that humans develop coping mechanisms in order to deal with the world around them:

A person's processes, psychologically speaking, slip into the grooves which are cut out by the mechanisms he adopts for realizing his objectives (Kelly 1955, p. 49).

These "grooves" become a person's individual framework for understanding reality, which is what Kelly called personal constructs:

Man looks at his world through transparent templets which he creates and then attempts to fit over the realities of how the world is composed (Kelly 1955, p. 8-9).

Kelly's Individuality Corollary sets forth that people differ in the way they perceive the world around them, while his Commonality Corollary allows for the possibility of similarity between certain groups of people in the way they perceive certain events. The Sociality Corollary discusses the possibility that group members may have similarities with each other, but also the ability to reinforce and/or change other group member's existing constructs in order to build a consensus. These last two corollaries are especially important when we consider using the repertory grid technique in cross-cultural research.


2.2 Elements, Constructs and Bipolar Constructs

Fundamental to Personal Construct Psychology are the terms elements, constructs and bipolar constructs. To use the analogy of an iceberg at sea (see Figure Two), elements are the conscious associations that first come to an individual's mind when asked to think about a particular subject. For example, if a person were asked, "What comes to your mind when you think about 'vacation'?", some possible elements could be fun, relaxation, sunshine, palm trees, and so on. Elements can be "people, events, objects, ideas, institutions and so on" (Cohen and Manion 1987, p. 316), but they should be " . . . well-known and personally meaningful" to the subject (Shaw 1980, p. 10). Easterby-Smith (1981) explains that it is helpful " . . . to think of elements as being the objects of people's thoughts, and constructs as the qualities that people attribute to these objects" (p. 11).

This definition of construct should not be confused with psychological constructs for testing construct validity. A psychological construct is " . . . a theoretical label that is given to some human attribute or ability that cannot be seen or touched because it goes on in the brain." (Brown 1988, p. 103). Kelly's constructs are a mental category, but they need not be an attribute or ability (Bannister and Fransella 1971, p. 8).

Returning to our iceberg analogy, constructs are typically "under the surface." They are unconscious assumptions of the world through which we understand reality. Constructs are unique to each person. Two people looking at the same elements could articulate completely different constructs. Constructs can also change depending upon a person's mental development, mood, and contrasting perceptions of the world:

Constructs are used for predictions [that can] be either correct or misleading. This fact provides the basis for revision of constructs and, eventually, of whole construct systems (Kelly 1955, p. 14).

Personal constructs are finite, fragmented projections onto reality (Kelly 1969, 1970, Shaw and Gaines 1992). They form a person's basic, albeit unconscious, assumptions of what he or she believes to be real or unreal, true or false, right or wrong, and good or evil. This also implies that constructs center on sets of discernible opposites, which Kelly called bipolar constructs (see Figure Three). Bannister and Fransella (1971) define bipolar constructs as:

. . . having two poles, a pole of affirmation and a negative pole, rather than . . . as concepts or categories of a unipolar type . . . Most people recognize bipolarity where it has an explicit verbal label to cover it -- black versus white, up versus down, nice versus nasty, here versus there, concrete versus abstract, noisy versus quiet, intelligent versus stupid, male versus female, and so forth (p. 24).

If constructs and bipolar constructs function as an unconscious framework for understanding reality, how then does one become aware of them? For the moment, we must remember that elements emerge from personal constructs (see Figure Three). Kelly felt that when a person focused upon his or her stated elements, he or she would begin to discern a certain relationship between some elements, and there would be occasions when some elements would seem to be less related to each other:

In its minimum context [a construct] is a way in which two elements are similar and contrast with a third (Kelly, 1955:61).

Let us return to three of the elements concerning the subject of vacation: sunshine, palm trees and fun. Depending upon the way one understands reality, a person might feel that sunshine and palm trees have more in common than other possible combinations, (although another person might feel that fun and sunshine have more in common). If the individual were asked to describe what it is that sunshine and palm trees have in relation to each other over the element of fun, the person might respond with the construct of "on the beach in Hawaii." If asked why fun is less related in his or her mind to the other two elements, the person might answer, "Playing computer games at home." The point is that constructs relate to issues or events which have significantly influenced the way the person perceives present, and possibly, future events. This is why Shaw (1980) maintains that ". . . thoughts and feelings, objective and subjective descriptions, attitudes and prejudices all constitute valid constructs" (p. 10). While we cannot discount the possibility for some of these thoughts and attitudes to be superficial, most often constructs and bipolar constructs point to issues and perceptions which come from the very core of the individual.

2.3 The Personal Construct Repertory Grid

Kelly originally designed rep grids as a psychoanalytic technique to " . . . investigate the role relationships between patients and their families, friends, etc., and for assessing the relationships between a patient's constructs about people" (Pope and Keen 1981, p. 36). Variations of Kelly's rep grid technique are still used by psychologists, but recently educators, management specialists, and social scientists are finding that rep grids are useful in their research as well (Feixas and Cornejo-Alvarez 1996).

The repertory grid procedure helps people to express their opinions and beliefs in the form of bipolar constructs, but it should be emphasized that rep grids will not "drag skeletons out of the closet" or trick people into revealing something they would prefer to keep private. Rep grids merely provide the framework within which individuals can focus their thoughts, without being "fed" the researcher's predetermined areas of interest in the form of questions or multiple-choice answers.

Because of the technique's flexibility, Pope and Keen state that there is really "no such creature as 'The Grid'" (p. 37). However, all grid techniques share the common goal of studying a person or group's bipolar constructs (Bannister and Fransella 1971, p. 70). All rep grids consist of elements, personal constructs, and often some form of rating system (Feixas and Cornejo-Alvarez 1996, p. 5). All highlight the subjects' perception of reality without forcing them to conform to an outsider's point-of-view. All elicit data in a way that recurring patterns can be interpreted by the researcher either quantitatively or qualitatively. The resulting analysis can then be presented in a form that is understandable to a larger audience.


Element 1

Element 2

Element 3

Element 4

Element 5

Element 6

Element 7

Element 8

Element 9


Construct 1










Construct 1

Construct 2










Construct 2

Construct 3










Construct 3

Construct 4










Construct 4

Construct 5










Construct 5

Construct 6










Construct 6

Construct 7










Construct 7

Construct 8










Construct 8

Construct 9










Construct 9

Table 1 A Sample Repertory Grid Sheet

3.0 Research Tradition

If rep grids can be interpreted either quantitatively or qualitatively, into what research tradition do they fit? How reliable and valid is the procedure as a research instrument? This section offers answers to these questions, and considers ways to protect the grid's reliability and validity.

Chaudron (1995:13) highlights four well-documented traditions in L2 research: Psychometric, Interaction Analysis, Discourse Analysis, and Ethnographic. With regard to these traditions, Nunan (1989) says, "each of these research methodologies has a different focus and function, and it is not possible to say that one is necessarily better than another without knowing what it is that the researcher is trying to find out . . ." (p. 4). The two main traditions commonly associated with L2 research are the psychometric and ethnographic traditions.

Psychometric research is concerned with quantitative results. Psychometric research methods use experiments, statistical tests and control groups to " . . . collect facts and studies the relationship of one set of facts to another. They measure, using scientific techniques that are likely to produce quantified and, if possible, generalizable conclusions" (Bell, 1995:1). Psychometric research is often used to statistically measure the improvement of students' test scores after being exposed to a different teaching method, technique or teaching materials.

Ethnographic research methods deal with qualitative (or interpretive) results. They use surveys, systematic observation, and journals to " . . . seek insight rather than statistical analysis" (Bell, 1995:2). Ethnographic research " . . . observes, describes and interprets the classroom in ways similar to those employed by anthropologists when they study unfamiliar cultures and societies" (Nunan 1989, p. 4). Ethnographic research methods are often used to study the classroom "culture" or some other aspect of the students' learning experience.

Psychometric research then, to use Long's terminology (1983a, cited in Ellis, 1986), is a theory-then-research approach interested in data, while ethnographic research is a research-then-theory approach interested in insight. While there is some debate about which is more valuable to L2 research, Grotjahn (1987) explains that most research techniques employ a mixture of psychometric and ethnographic research methods. For that reason, Nunan's (1989) advice is to " . . . acknowledge that different projects will have different purposes and audiences, and that it is these purposes and audiences which should determine the research methodology and design" (p. 10). This is a wise suggestion, because repertory grid techniques, while flexible, cannot bend to the needs of every research project. Rep grids are an excellent method for discovering an individual or group's opinion on a certain issue, or as an action research, problem- solving tool (Shaw 1980, Shaw and Gaines 1995). But if one wishes to objectively measure students' retest scores on a cloze test, rep grids are obviously not the best choice.

Grotjahn's (1987, p. 60) research paradigms (see Table Two) place repertory grids as either a Paradigm 6 (Exploratory-Quantitative-Statistical) or Paradigm 7 (Exploratory-Quantitative-Interpretive) research instrument. Rep grids are nonexperimental in design, and produce qualitative data which can be analyzed either statistically or interpretively. Which way one chooses to interpret the grid data of course hinges on how the researcher designs his or her particular grid. The data from some grid experiments may be more amenable to a quantitative analysis, while others would be better analyzed qualitatively.

3.1 Reliability and Validity

Bell (1995:64-65) defines reliability as " . . . the extent to which a test or procedure produces similar results under constant conditions on all occasions," and validity as " . . . whether an item measures or describes what it is supposed to measure or describe." The two are commonly separated into internal reliability, external reliability, internal validity and external validity. Nunan (1992) defines internal reliability as "the consistency of data collection, analysis and interpretation," and external reliability as "the extent to which independent researchers can reproduce a study and obtain results similar to those obtained in the original study" (p. 14). He defines internal validity as "the interpretability of research," and external validity as "the extent to which the results can be generalized from samples to populations" (p. 15).

3.2 Reliability of Repertory Grids

Are rep grids reliable? One would wonder after reading the ambiguous responses of PCP experts. For example, Bannister and Fransella (1971) answer this question by writing:

Kelly is reported as referring to reliability as 'a measure of the extent to which a test is insensitive to change'. This is no facetious comment but a logical deduction from his theory which sees man as a form of motion. Our aim should be to understand the meaning of change, not to regard it as an irritating interference with the 'reliability' of our tests by an irresponsible subject - to be looked on as 'error variance' (p. 76).

While this is certainly an interesting point of view, they skirt the issue by saying there is " . . . no such thing as the reliability of repertory grid technique, since there is not one form of the grid" (p. 76). While Bannister and Mair (1969) confront the subject of grid reliability head on, they come to a tepid conclusion:

One practical rule must be that if the reliability of a particular grid in a particular context needs to be known, for either theoretical or practical reasons, then it will have to be specifically assessed as part of the experimental venture . . . 'reliability' is, in itself, a target for experimental investigation (p. 175).

Nevertheless, there have been promising experiments designed to test the reliability of repertory grids. Bannister and Fransella were able to demonstrate a high degree of retest consistency in a grid technique constructed to identify thought disorder in schizophrenics (Bannister and Fransella, 1966). Other more recent studies also show that reliability in rep grid techniques can be strengthened if it is a necessary factor for the research project (cf. Bavelas et al. 1976, Lansdown 1975 and Gathercole et al. 1970).

Still it must be admitted that most ELT researchers will lack the time and/or resources to bolster rep grid reliability in a psychometric sense, especially if they plan to use the technique with a group or individual only once. People change. Rep grid techniques are most reliable in discovering the change which has occurred in a person's construct system. For example, McCoy's cross-cultural study (1983) used rep grids to measure the progressive stages of culture shock and resolution in an Australian subject living in Hong Kong. Her research showed how properly administered rep grids can reliably measure changes in an individual's construct system over time.

Yet even if a researcher uses a rep grid technique with a group or individual only once, LeCompte and Goetz (1982) say the internal reliability of ethnographic research can be strengthened if more than one researcher works on the project, others are conducting the same research in similar settings, and if the data is systematically recorded. The external reliability of rep grids can be strengthened if the researcher carefully details the subjects, process, conditions, form of analysis, and form of data collection. Nunan (1992) summarizes LeCompte and Goetz's research by saying that protecting against threats to reliability " . . . can be summarized in two words - care and explicitness . . . if one is careful in the collection and analysis of one's data, and if one is explicit about the way the data were collected and analyzed, then one can reasonably claim reliability for one's investigation" (p. 62).

3.3 Validity of Repertory Grids

Nunan (1992:59) adds that if one is not looking for causal relationships between variables and is not concerned with generalizing the results to represent a larger group, then validity will not be a pressing issue. However, the issue of validity will be a concern for most EFL teacher-researchers, since many will wish to apply the data gathered from rep grids to their classroom teaching, and also share their findings with other colleagues.

Fortunately the validity of rep grid techniques for discovering the constructs of both individuals and groups has been well-proven in a number of studies. Fransella and Bannister (1967) showed that it was possible to predict how people would vote from an analysis of the relationship of their constructs to the elements, which were stated as party affiliations. Hill (1976) demonstrated that his rep grid technique was valid for measuring various forms of thought disorder. Leitner (1981) demonstrated the construct validity of a rep grid that measured personality styles. Munby (1982) presented evidence that his rep grid technique was valid in measuring teachers' attitudes toward classroom decisions.

As long as one has not "stacked the deck" by choosing subjects that one feels will validate his or her research, repertory grid procedure by design has a high level of internal validity. The technique is specifically designed to identify the bipolar constructs of groups and individuals. Ethnographic research such as that found in the repertory grid procedure can make reasonable claims to external validity not through generalization, but through comparability and translatability:

Comparability requires that the ethnographer delineate the characteristics of the group studies or constructs generated so clearly that they can serve as a basis for comparison with other like and unlike groups . . . Translatability assumes that research methods, analytic categories, and characteristics of phenomena and groups are identified so explicitly that comparisons can be conducted confidently. Assuming comparability and translatability provides the foundation upon which comparisons are made (LeCompte and Goetz, 1982:34).

If other teachers identify similarities in their own students after reading a published report, reasonable claims to external validity can be made. We can conclude that repertory grids fit squarely in the ethnographic tradition. Provided that an appropriate level of rigor is applied, research shows that rep grids can be both reliable and valid as a research instrument.

4.0 A Repertory Grid Procedure

This section is provided for readers who need a better understanding of how rep grids work. The repertory procedure is composed of several stages, beginning with a development stage, where the parameters of the research and the grid are decided. Next is the administration stage, where elements, constructs and a rating system are used by the subjects. The final stage includes the analysis and interpretation of the grid data.

4.1 Define Your Purpose and Area of Research

It is important to have a clear reason for why one wishes to use rep grids over other research instruments, as well as what area, problem or issue one wishes to learn more about. Otherwise the elements and constructs elicited from the students will be unfocused, thereby ruining the entire procedure. Teacher-researchers need to ask several question prior to administering the grid. Will the grid be used in an action research project? Will it be used with individuals or with groups? Will the grids be administered by the teacher, another researcher or both? How much time will the subjects have for filling out the grids? How many elements and constructs does one plan to use in the grid? Will a rating system be used? Will the data be analyzed quantitatively, qualitatively or both?

4.2 List the Elements

Elements define the focus and outcome of the grid data, so it is important that special attention is given to their selection. There are several ways to elicit elements, but Easterby-Smith (1981 p. 10-12) suggests that, whatever method one uses, elements should be as specific as possible. Elements should also be homogeneous, meaning they must be related to the research question. "Dependable", "Honest" and "Fair" might be appropriate elements when thinking about one's ideal manager, but "Chocolate Ice Cream" is unrelated to the task at hand (unless the manager runs an ice cream shop!). Homogeneous elements are important because elements from an unrelated subject might elicit constructs that are significant to the individual, but unconnected to the research project.

Easterby-Smith (1981) also suggests that elements in a grid need to provide representative coverage of the research subject. A grid about the current strengths in an ESL teaching program which omitted the names of the classroom teachers would naturally miss major areas of discovery. If a group is working on the same grid, they should feel free to elicit any elements they feel are connected to the research question, but:

. . . it is important to ensure that all the people are able to relate directly to the elements specified: a research chemist asked to name five subordinates may not actually have any subordinates; a graduate trainee asked to rate his reactions to 'Chairing Meetings' may have no direct experience of chairing meetings (Easterby-Smith 1981 p. 10).

Pope and Keen (1981) feel eight to fifteen elements are usually sufficient for a well-balanced grid. Easterby-Smith (1981) suggests no less than six or seven, but no more than twelve.

Elements can be elicited through brainstorming, discussions with the subjects, providing a list of possible elements to choose from, or even by deciding the elements in advance. The subjects then write the elements on the top of the grid sheet. If groups work together, it may be necessary to gather all the elements in the form of an open-item survey and then provide a representative list of the most frequent elements. But Pope and Keen (1981 p. 41) stress that a researcher must do extensive work to assure the elements truly exemplify what the subjects have expressed.

4.3 Eliciting Constructs

Pope and Keen (1981:47) advise that "prior to the elicitation of a first grid from an individual or group of individuals, it is a sensible idea to . . . complete a small 'dummy' grid in order to come to grips with some of the 'mechanics' of completing a grid." This is a good practice, since most people have never participated in a repertory grid task before. If the subjects do not clearly understand what to do from the beginning, the whole process may be destabilized. It is best to walk the subjects through a grid based on a question totally unrelated to the intended research project. This way the subjects can grasp the basics of the repertory grid without running the risk of being unwittingly influenced by the researcher's views on the real subject to be studied. While there is no limit to the number of constructs that one should be allowed to express, most researchers (Shaw 1980, Easterby-Smith 1980, Pope and Keen 1981) feel that eight to fifteen constructs are adequate:

One is not aiming to encapsulate the whole of an individual's construct system but that part of it which is relevant to the defined purpose (Pope and Keen 1981 p. 44-45).

There are several ways to formulate bipolar constructs, the simplest being to ask the subjects to write them on the grid sheet. Pope and Keen and Easterby-Smith all say that this is the best way if the subjects can spontaneously produce constructs. But for most people, constructs do not come to mind automatically.

The traditional way researchers use to elicit constructs is Kelly's triadic method (1955). This technique asks an individual to choose three elements on a grid and then decide which two of the three have something in common with each other. The subject describes in what way the two elements are alike, and writes this description on the first row of the "likeness" (or emergent) pole, (often on the left side of the grid). The subject then defines how the third element is unlike the other two, and writes this description on the first row of the "contrast" (or implicit) pole, (usually on the right side of the grid). Sometimes defining the way a third element is unlike two other elements may be too difficult for the subject(s). Pope and Keen advise that "it is important to remember triadic elicitation is an aid to formulating constructs but can often be abolished altogether when constructs 'come to the respondent' freely" (1981:45).

Another option is the Opposite Method, developed by Epting et al. (1971). This allows subjects to write what they perceive to be the opposite of the relationship between the two similar elements. An advantage of this approach is that it usually ensures the constructs are bipolar in nature. Other methods include supplying example constructs in the grid to help guide subjects into what areas the researcher is interested. Dyads (Epting, et al. 1971) ask subjects to work with only two elements. They write constructs based on how the two elements are alike or different. The most complex method is Laddering, which involves asking an individual a steady progression of questions why the subject stated his or her elements.

Regardless of how one chooses to elicit constructs, Pope and Keen emphasize that, ". . . the issue of contrast should be stressed because, whichever method is used for construct elicitation, one occasionally encounters a situation where the construct poles given by a person represent two ends of what are, in fact, two dimensions" (pg. 51).

4.4 Linking the Elements

After eliciting bipolar constructs, the subjects usually implement a scaling system that shows the relationship between the constructs and the elements. There are three basic methods: Dichotomizing, Rating and Ranking (see Figure Four). In dichotomous methods, the subjects rate each element they perceive as related to either the construct on the "+" side or the "-" side. Ranking consists of grading the elements from one to six, from the one which is perceived to be more closely identified to the emergent pole (1), to the element thought to be closer to the implicit pole (6). Of course, this system depends upon the number of elements in the grid. If there are nine elements, each element would be ranked from one to nine. Rating involves using a graded scale of 5, 7, or 11 points. Each element is then graded on a scale of one to five, one to seven, etc., in relation to the construct on the emergent pole (1) or the implicit pole (5, 7, or 11).

Shaw (1980) reports that at the time of her research, more than 70% of the published reports using repertory grids implemented some form of rating method. However, while linking elements to bipolar constructs is advantageous, it is secondary to elicitation. If studying the interrelationships between constructs and elements is not a concern to teachers, they should feel free to forego any rating or scaling method in their research.

The process of selecting an emergent construct, implicit construct and rating the elements continues until the grid is completely filled or until the subject runs out of ideas. Both Easterby-Smith (1980) and Pope and Keen (1981) warn that rep grid procedures can take a long time for subjects to complete. The time needed to finish a grid of nine elements and nine bipolar constructs can take up to two hours. For this reason, teachers wishing to use the repertory grid technique as a research tool with their students should carefully consider how much their students can handle in a limited amount of time.

4.5 Analyzing the Data

There are several ways to analyze rep grid data. Probably the most practical way for busy EFL/ESL teachers is to count up the frequency of constructs and, (if a rating system has been chosen) study the connection of the constructs to the elements. One can gain a surprising amount of insight from this fast and simple method.

Alternative ways to analyze the grids includes the correlation matrix, which is described in Easterby-Smith (1981). This is an easy way of parametrising correlations by counting up the numbers of matching pairs of elements or constructs for each row. This method has the main advantage of allowing a researcher to see how similarly the subjects perceived the elements or constructs in relation to each other. But this is only one of many ways to work out correlations. Those wanting a statistical method of analysis will be interested in clustering, a method that builds on correlation matrices. This technique calculates the matching scores between the elements and constructs. Afterwards the elements and constructs with the highest frequency of matching scores are clustered together. Not only does this method show the interrelationship of several concepts within a grid, the graphic display is easy to read and gives a better "feel" for the data than some other methods. Another means for analyzing the data is through focusing. This method rearranges the rows and columns of the grid so that similar constructs and/or elements are situated together. Researchers either interested in the focusing or clustering method can find step-by-step instructions in Shaw and McKnight (1981).

There are several computer programs developed to analyze repertory grid data. For Macintosh computer users, RepGrid (Shaw 1994, CPCS, 1993) analyzes individual or group grids, and statistically analyzes the data through focusing, clustering or a number of other methods not covered in this paper. G-Pack (Bell, 1987) is another program which offers similar features as RepGrid, but is significantly cheaper. It is available in both Macintosh and IBM formats. IBM users will find CIRCUMGRID (Chambers and Grice, 1986), OMNIGRID (Sewell 1989, Sewell et al. 1991) and INGRID 96 (Legg 1996) as freeware available on the World Wide Web. Probably the best program currently available for IBM is GRIDCOR, an English version of the Spanish program RECORD, which is produced at the University of Barcelona. GRIDCOR is more flexible than some of the older programs, and also can use multidimensional scaling to analyze the data (Feixas and Cornejo-Alvarez, 1996).

A sample of the kind of data produced by rep grid analysis programs can be seen in Figure Five. An EFL teacher in Japan completed a repertory grid using OMNIGRID. Like many of these "homegrown" programs, OMNIGRID's printing capabilities were very limited, so what is seen in Figure Seven is a much-improved display for a research presentation (Hadley 1996a). Programs such as these are superb in helping individuals elicit personal constructs. The statistical analysis in many of these programs often helps subjects identify areas of consensus and conflict in their construct systems, which suggests the potential of rep grids for action research projects centered on a number of pedagogic issues (Shaw and Gaines 1995).

All of these programs are excellent but none is a complete solution to all the possible rep grid designs. Rep grid-based research deals with a large number of variables, and the programs mentioned above may or may not match the needs of a specific research project. If one wishes to use a computer program to analyze the data elicited from his or her students, it is best to find out first what the program can do, then plan the research project around the program. Doing it the other way around - conducting the research and then finding a program - is almost certain to bring disappointment. To date there is no computer program that one can get "off the shelf" that is able to take into consideration all the variables in rep grid research.

Most language teachers will find that studying their students' constructs and interpreting them based on simple frequency yields the quickest, simplest and most effective method of analysis. But for those teachers skilled in statistics, Bell (1994, 1996) suggests a number of practical ways that rep grids can be analyzed with standard computer statistics packages. But however one decides to analyze the grid results, Shaw and McKnight (1981) remind us that while the methods and tools described above have been used in many studies, there is no one best way to interpret grid data. An eclectic approach to grid analysis is welcome so long as it is well-informed and consistent with rep grid theory. As Easterby-Smith (1980) says, ". . . the interpretation of grid data is very much an art and not a technology" (p. 17). Let us now turn our attention to an actual research project which used repertory grids to study the constructs of Japanese English language learners.

5.0 Cross-Cultural Research in Japan

Repertory grids have had a long history of use in educational research (Diamond 1984, Diamond 1983, Munby 1982, Olson 1981). Most of these research projects have studied only the attitudes and practices of teachers, while a few have investigated the learning preferences of native English-speaking students.

Repertory grids have seen only limited use in Japan. Takagi and Sakamoto (1991) used repertory grids with Japanese college students to measure their changes in perception of different countries after the 1988 Seoul Olympics. And while there have been some recent studies using rep grids with Japanese students (Sakamoto 1996, 1993; Sakamoto and Numazaki 1989), the focus of these studies has largely been to determine the validity of the rep grid technique itself, or to measure psychological concepts such as cognitive complexity. Currently there have been no published reports on the use of rep grids in Japanese EFL research.

5.1 Research Question

The research question was formulated as: "What is a good teacher?" The purpose of this research was to find out what general attitudes and values Japanese college students held about good teachers. It was hoped that this research would contribute to the efforts of many teachers in Japan struggling to find ways to better communicate with their students. There is a considerable amount of "angst literature" in Japan describing the very real trials and tribulations that teachers suffer in encouraging their students to talk with them on even the most basic level. Providing those teachers with further insight as to where their students are coming from culturally could prove to be a valuable pedagogic resource.

5.2 Subjects

A total of 165 Japanese college students (ninety-nine males and sixty-six females) was randomly chosen to take part in the project. The subjects included eighty-nine first-year students from Niigata University, a high-ranking national university and seventy-six third year students from Keiwa College, a private, four-year, liberal arts school. Both institutions are located in Niigata Prefecture, on the Northwest coast of Japan. More than half of the students came from Niigata Prefecture. The rest were from various prefectures on the main island of Honshu, and the northern island of Hokkaido. All of the subjects came from one of the following departments: International Studies (Keiwa College), Pre-Medical, Agriculture, Elementary Education, and Economics (Niigata University). All the classes were composed of strong and weak learners. Their relative strengths and weaknesses as language learners were determined from attendance, test/homework scores, and a successful system that measures active class participation (Hadley, 1995a).

5.3 Research Procedure

Each class was taught how to perform the rep grid procedure in the same manner. The procedure was first explained in Japanese and English, mainly because the research was during class time.

All instructions were verified with the stronger students before starting each step of the research. The stronger students would then instruct the weaker students. I monitored the stronger student's explanations to confirm there were no misunderstandings. We would begin only after everyone had reached a consensus on what they were expected to do. The subjects' behavior during the research was observed and recorded in a journal for later reference.

During this research project, most of the students asked to work in groups with close friends. I allowed group work so long as they remained in the same group for the duration of the project. The procedure was divided into two sessions. The first involved the discovery of elements, and the second used the repertory grid procedure to elicit bipolar constructs.

5.4 The First Session: Element Elicitation

Each class was told they would participate in a research project, the results of which would be read at a larger forum. The research question, "What is a good teacher?", was written in English on the blackboard. Students were asked to brainstorm on eight or more items which, in their opinion, described a good teacher. They were assured that no part of this research was a test consisting of right or wrong answers, and that no efforts would be made to track the responses back to any particular individual.

The subjects were asked not to misinterpret the question by writing what they thought of the teacher conducting the research. They were neither to "read between the lines" and write elements which they perceive the researcher might like, nor were they to write down qualities of only good foreign teachers.

After it was clear that all students understood what they were supposed to do, I did not wander around the room. When students asked for clarification, the process would be explained again, but requests for example elements were declined so as not to contaminate the data. All the elements were then collected and translated into English.

5.5 Results and Translation of Elements

Out of all the elements elicited (see Appendix One), the eight elements which were used in the rep grid are in Figure Six. While the supplied elements may seem similar to each other in English, they are more clearly delineated in Japanese.  

Kind (yasashii) means one who is gentle, calm and mild in his or her dealings with people. Fun/Entertaining (omoshiroii/tanoshii) is a literal translation of the Japanese. Caring/Understanding (omoiyari no aru) denotes one who is able to feel the pain of another. It carries with it the feeling of one who is genuinely sympathetic and can put oneself in the other's place with ease. Friendly (Shitashimi-yasui) means one who is warm, open, outgoing, easy to feel close to and easy to talk with. Fair/Impartial (kohei) comes from two Chinese characters, one meaning "public", and the other meaning "normal", "standard" or "regular". Together the characters mean a sense of detached impartiality. Understandable (wakari-yasui) means literally, "easy to understand". Interesting Lesson/Story (kyomi-bukai kogi/hanashi) is also a literal translation which means, "A lecture or story that is deeply interesting." Enthusiastic (nesshin) means zealous or passionate. It comes from two Chinese characters, one meaning "heat" and the other meaning "heart" ("heart" in the metaphoric sense of soul or spirit).

5.6 Justification for Selected Elements

In order to protect the validity of grid-based research, Yorke (1985, p. 386-88) states that the elements should be coherent next to each other, representative of the area to be studied, and understood by the subjects. Keeping this in mind, I chose three of the elements: Kind, Friendly, and Fair/Impartial because of their high numeric frequency. The elements of Understandable, Caring/Understanding, Enthusiastic, Entertaining/Fun and Interesting Lesson/Story were based on the frequency of interrelated elements (see Appendix One). I determined that these eight elements represented the subjects' dominant concerns. Pope and Keen (1981) defend this method of supplying elements:

. . . if one is exploring the nature and sharing of construing within a group, it is often the case that a common set of elements are selected and provided for each individual - this could be followed by either provision or elicitation of constructs or a combination of both. It should be emphasized, however, that if one decides to provide elements/constructs then adequate groundwork should be done in order to obtain what one hopes are representative elements and constructs (p. 40).

Before administering the grids, I showed these eight elements to all the subjects to verify whether or not any elements seemed inappropriate. None found the elements to be unsuitable or a distortion of their views. The elements were written both in Japanese and English on a repertory grid sheet (see Figure Seven). The grid sheet was enlarged and photocopied on A3 paper to allow room for students to write their bipolar constructs.

5.7 Second Session: Repertory Grid Procedure

As with the first session, the research question was again written on the blackboard. Each class was reminded that they would participate in the second part of a special research project. They were again asked not to misinterpret the research question as stipulated in the first session. The students were allowed to work either individually or in the groups they joined during the first session. Students choosing to work alone were asked not to confer with others during this part of the research. Groups were asked as well not to seek a consensus with others outside their group.

A small number of students who participated in the element elicitation were not present for the grid procedure, and in some classes there was a student or two present who was not part of the element elicitation. In most cases the new students were members of existing social groups, and had already been briefed about the research project. Three students with sporadic attendance and no group ties were given a complete briefing and encouraged to fill out their rep grid sheets individually.

Pope and Keen (1981:44) state that "in many circumstances there may be time limit constraints on . . . conducting the grid interview which well may impose a limitation on the number of constructs which are elicited for any one session." I felt that the first-time challenge of the repertory grid procedure would allow individuals or groups only enough time to complete a rep grid sheet which contained the maximum potential of eight bipolar constructs. A large repertory grid identical to the one the subjects were to receive was drawn on the blackboard. The concept of bipolar constructs was briefly explained on the blackboard, using the visual analogy of elements being the tip of an iceberg, and constructs being the part of the iceberg submerged under water (see again Figure Two).

Each individual or group was then given one rep grid sheet and a set of cards numbered one through eight. I explained that each number corresponded to the numbered element on the grid (see again Figure Seven). A modified form of the Triadic Method was demonstrated on the blackboard with large sample cards before allowing the elicitation process to begin. The Triadic Method was conducted in the following manner:

*   Each group or individual would turn the element cards (which were numbered from one to eight) face down on their desk and shuffle them.

*   They would then turn over three of the cards and mark on the grid with a circle or an "X" which elements were drawn.

*   Each group or individual would then decide in their own opinions which two of the three elements had something in common with each other. They would then connect these with a line.

*   The subjects were instructed to write the emergent construct on the plus side, and the implicit construct on the minus side. They were instructed to fill out the grid row-by-row, only moving to the next row only after completely eliciting a bipolar construct.

*   In the case of drawing the same three elements as in a previous turn, the students were instructed to reshuffle the cards and try again.

*   This process was repeated until the rep grid sheet was completely filled. As before, students were asked to write their answers first in Japanese, and later in English if time allowed.

The students were asked to avoid as much as possible making elements into constructs. However, if the subjects could not come up with any other way to express their constructs, writing an element label in the construct box was allowed. But the students understood the futility of writing something like "a kind teacher is kind", so most did not label elements as constructs.

Anticipating that attempts to elicit why the third element was different from the chosen two would be too challenging for some subjects, I modified Epting et al.'s (1971) Opposite Method by asking the subjects to write in the implicit construct box what they perceived to be the opposite of the emergent construct. The subjects found this to be an easier framework for defining their bipolar constructs. They were instructed before beginning the grid experiment that their bipolar constructs did not have to be worded in stark "black and white" terminology. "Black and Good", "Coal and Ice Cream", or any other perceived opposite would be acceptable so long as it represented their point-of-view.

Explaining this stage took an average of twenty minutes, leaving seventy minutes for grid elicitation. After the students began, I monitored the classes to verify that the procedure was being followed. Out of the six classes, only three groups in one class were discovered to be following a different procedure. I stopped these groups, re-explained the process, and gave them a new grid sheet. Most subjects finished their grids within fifty-five minutes, while others took the entire seventy minutes.

Pope and Keen (1981:47) feel that a simple grid without any ranking system works best with children, teenagers or adults who would prefer to not use a lot of numbers. Because I was more interested in what types of bipolar constructs would be elicited from the subjects, and from observing the cognitive strain the procedure was putting on a majority of the subjects, I opted for a dichotomous link of two elements to the emergent pole, with the third tacitly linked to the implicit pole. The rest of the elements were not ranked or rated. Samples of actual grid data elicited from two different groups can be found in Appendix Two.

6.0 Presentation and Discussion of Data

Even with a minimal link of the three elements to constructs, the design of this grid experiment could have been studied with statistical techniques such as multidimensional scaling. But it is doubtful that most language teachers would benefit from such an analysis. The findings were instead studied qualitatively, which is more in line with the way that most EFL teachers would interpret the data elicited from their grid research. Research such as this is typically data rich. Space does not allow for all the data to be interpreted, so only the major aspects of the elements and constructs will be discussed.

6.1 Interpretation of the Elements

A detailed review of these elements is found in Hadley and Yoshioka-Hadley (1996). But even a peripheral study the elements found in both Figure Eight and Appendix One reveal much about the qualities Japanese students associate with good teachers.

From the elements we can see that the subjects' general portrait of a good teacher is that of a kind-hearted, friendly individual who is open-minded, sympathetic but impartial in student relations and class decisions. He never resorts to physical violence or forces his opinion on an issue. He is punctual for class and is fun to be with. He should not only be understandable, but understanding as well. He focuses on the needs of his students, not on tests or homework. The good teacher is knowledgeable and experienced, but humble. Whatever other teaching methods that he uses, the good teacher is a storyteller who shares real-life anecdotes of interest with his students. His enthusiasm for teaching, sense of humor and cheerfulness will encourage students to participate with him in class. In effect, the good teacher is a person students look up to, believe in, and depend on.

This depiction reflects the Japanese sempai approach to leadership. The bond between sempai, literally "companion ahead" and kohai, meaning "companion behind" is important in Japan. Rohlen (1974) describes the relationship by writing:

Ideally, the sempai will represent, advise, console, teach and discipline their kohai. Kohai, in return, will confide in, listen to, depend upon, follow, and respect their sempai . . . there is an implication that leadership should be as sympathetic, protective and unselfish as good sempai (p. 23).

The sempai-kohai relationship is essential for consensus-building and the reinforcement of cultural values in Japanese companies, organizations or institutions. To be successful in Japanese society, it is important to have a wise and influential individual who can teach you the proper way to think and act. Japanese students seem to desire these qualities in their teachers. Perhaps this is why students recognize storytelling as a significant good teacher practice. Wright (1995) says that storytelling fosters warmth and closeness between the speaker and listener. The data from this sample supports Wright's recent plea for language teachers in Japan to give more time for storytelling in their classes. However, the issue of nonviolence as one of the top ten responses came as a surprise. There has been a long history in Japan of bullying by teachers (Murakami, 1985 and Horio, 1988). This sample may imply that much more violence goes on behind closed doors than many are aware of.

6.2 Interpretation of the Constructs

While each grid sheet used with the subjects (see again Figure Seven) contained the same supplied elements, and allowed for only a maximum of eight bipolar constructs, afterwards the data from all 62 individual grid sheets were joined together into a "megagrid" containing the eight elements and all 496 bipolar constructs elicited from the subjects. A weakness in many studies using repertory grids is that they work with only a limited number of subjects. The purpose behind joining all the individual grids into one large pool of data was to create a valid sample from which to work with. In this way language teachers in Japan and abroad could be assured that the data in this study came from a representative group of Japanese university students, not merely from a few individuals. Table Three contains the most commonly-occurring constructs.

Top Constructs by Frequency





- Boring


- Not Serious


- Cold


- Narrow-minded


- Selfish


+ Attention to Detail


+ Cheerful


+ Earnest


- Gloomy


+ Warm-hearted


- Lazy


+ Attention to Students


- Half-hearted


+ Understands Students


+ Parental Kindness


+ Good Speaker


Table 3 "+" Signifies an Emergent Construct, "-" Signifies an Implicit Construct.

One is struck by the high frequency of implicit, negative constructs. Whatever else this may mean, it does suggest that the sample has a clearer image in their minds as to what qualities a good teacher does not have. Another way to present as much data as possible is in the form of Concept Mapping. Developed by Novak in the 1960's, Concept Mapping is a technique used as a means of visually representing thoughts and ideas (Novak 1977, Novak and Gowen 1984). The concept map in Figure Nine features the most-frequent constructs and connects them to their converse implicit or emergent constructs. This allows for a more balanced presentation of the data, and visibly demonstrates the interconnected nature of this sample's construct system.

Each box in the map represents a construct label. It shows how many times it occurred in this research, and whether or not it was labeled as an emergent construct ("+") or implicit construct ("-"). For example, if we look at Gloomy, we see that it occurred nine times, and was always labeled as a negative, implicit construct. The nine emergent constructs which occurred opposite to Gloomy are, Sociable, Always Smiling, Pleasant, Nice Character, and five occurrences of Cheerful. Cheerful occurred a total of ten times, and the implicit constructs connected to it were: Gloomy (5), Silent, Repetitive, Uptight, Selfish, and Cold-hearted. Following the lines to Selfish and Cold/Cold-Blooded/Cold-Hearted reveal connections to other constructs.

This is only a portion of the data, but the cognitive map provides a fair representation of the construct findings. No attempt was made to combine constructs into groups, with the exception of the Cold and Parental Kindness groups, since they used the same Japanese characters with only minor variations.

Translating the constructs from Japanese into English was surprisingly straightforward, with the exception of the construct label Half-Hearted and Unprepared. This comes from the Japanese word iikagen, which is difficult to translate directly into English. Iikagen roughly means a devil-may-care attitude that results in haphazard, sloppy work. When used with regard to speech, it means the vague, scatterbrained response of someone who is ad-libbing in the place of conscientious preparation.

Why would 18 to 19-year-old adults express constructs such as these? Through studying a number of historic influences in Japan's educational system, we can better understand how to interpret this data.

6.3 A Piece of the Puzzle: Confucianism

Confucius, or K'ung-fu-tsu (551-479 BC), taught that most of life's problems occurred when common people abandoned their stations in life, and rulers left the path of virtue. He developed a socio-political philosophy which stressed that one should know his place in society, be satisfied with it, work hard, and study hard. Confucius felt that Heaven had established five relationships (wu-lun), which must be maintained:

The relationships are those between ruler and minister, father and son, husband and wife, older brother and younger brother, and between friends. (Doctrine of the Mean: V. 20).

The parent-child relationship had priority even over one's obligation to the ruler. Upholding it kept all other relationships, and ultimately, the society, in perfect accord. The key to maintaining harmony was through proper outward conduct (li) and humane benevolence (jen). Jen was by far the most important virtue for Confucius, yet it was useless without li.  Jen (Japanese: Jin) also appeared as an emergent construct label in this study (See Appendix Two). When jen and li were unified, children loved their parents (hsiao), citizens obeyed their rulers (chung), and friends or business partners treated each other fairly (yi) (Hooker 1995, Chung 1995, Hall and Ames 1987).

According to Confucius, a man (women were ignored by Chinese philosophers) who loved his parents, maintained honest relationships with friends, practiced benevolent justice (jen) with those below him, and respected those above him might become a chün-tsu, (lit. "son of a prince"), usually translated in English as a "Superior Man" (Muller, forthcoming p. 4).

6.4 Qualities of the Confucian Superior Man

Although Confucius' Superior Man was first and foremost a teacher, not all teachers were Superior Men. Confucius said, "Someone may have profound theories - but is he a Superior Man? Or is he only superficially impressive?" (Analects 11:20). The Superior Man had the authority to teach others not only because he was well-educated, but because his life was characterized by jen:

Tzu Chang asked Confucius about jen. Confucius said, 'If you can practice these five things with all people, you can be called jen.' Tzu Chang asked what they were. Confucius said, 'Courtesy, generosity, honesty, persistence, and kindness. If you are courteous, you will not be disrespected; if you are generous, you will gain everything. If you are honest, people will rely on you. If you are persistent, you will get results. If you are kind, you can employ people (Analects 17:6-7).

In all things, the Superior Man was fair and impartial in his dealings with others:

Confucius said, 'When the Superior Man deals with the world he is not prejudiced for or against anything. He does what is Just.' (Analects 4:10). There has never been a case where a person was able to teach others the things he had learned, if he lacked fairness (Great Learning, V. 9).

The Superior Man was open-minded:

Men of high office who are narrow-minded . . . how can I bear to look on such things?! (Analects 3:26). To be broad-minded and gentle in teaching and not rashly punish wrongdoing . . . The Superior Man abides in this. (Doctrine of the Mean, V. 10).

He also was careful to do all things tastefully. This involved being prepared and paying close attention to detail:

Confucius said, 'If raw substance dominates refinement, you will be coarse. If refinement dominates raw substance, you will be clerical. When refinement and raw qualities are well blended, you will be a Superior Man.' (Analects 6:16) In all affairs, if you plan ahead you can be successful, and if you don't plan ahead, you will fail . . . If you are prepared before teaching, you won't run out of material (Doctrine of the Mean, V. 18).

When teaching others, Confucius taught that the Superior Man should not be boring and inflexible (Analects 15:36). Instead, a Superior Man's instruction should bring out the best in people:

The Superior Man develops people's good points, not their bad points. The inferior man does the opposite. (Analects 12:16).

The Superior Man could do all of this because he had purged from himself the following faults: "imposing his will, arbitrariness, stubbornness and egotism" (Analects 9:4).

From this brief review of Confucius' teachings we can clearly see that much of the qualities the student sample associates with good teachers are also found in Confucius' Superior Man. However, the subjects' do not want their ideal teacher simply to be a Superior Man. The Good Teacher for them is warmer and much closer to the heart than the monolithic figure found in Confucianism. A good teacher should also nurture and entertain students because he or she cares about their comfort and happiness. To understand this facet of the Japanese Good Teacher, we must look to later developments in Confucian thought which were first initiated by the Chinese reformer Chu Hsi, and later refined by Japanese Neo-Confucian scholars.

6.5 Neo-Confucianism

During his era, Chu Hsi (1130-1200), felt that many aspects of popular Confucianism had fallen into a state of disrepair (Ichii 1974). Chu Hsi and his disciples collected and codified the extant writings of Confucius, and emphasized conformity to various social and religious rituals (li). By paying serious attention to the details of these rites, and by guarding against a lazy, apathetic heart, internal virtue would ultimately be manifested in a person's life:

...Refinement is substance; substance is refinement! When the hair is taken off the hide of a tiger or leopard, it looks the same as the hide of a dog or a sheep (Analects: 12:8).

The core of Chu Hsi's teachings emphasized sections of Confucius' book, The Great Learning (Ta Hsüeh), which admonishes those in authority to study hard, live a moral life, and lead by example:

When things are investigated, knowledge is extended. When knowledge is extended, the will becomes sincere. When the will is sincere, the mind is correct. When the mind is correct, the self is cultivated. When the self is cultivated, the clan is harmonized. When the clan is harmonized, the country is well-governed. When the country is well-governed, there will be peace throughout the land (Great Learning: v. 1).

Other Confucian classics frequently quoted by Neo-Confucians called on those in authority to treat their subjects as if they were their own children (Doctrine of the Mean V. 20). While this "ruler-child" concept was never taken seriously in China, it flourished later in Japan.

6.6 Confucian and Neo-Confucian Influence in Japan

Confucianism entered Japan as early as the fourth century. Its influence can be seen in many ancient political writings of the time. By the seventh century, the main Confucian concepts of jen and li had inspired the Japanese educated classes.

Introduced to Japan by Zen priests in the late 12th century, the Neo-Confucianism of Chu Hsi (Japanese: shushigaku) was destined to have a greater impact on the whole of Japanese society and education. Shushigaku was taught in monasteries and schools such as in Mito, where the Tokugawa clan ruled. Eventually a Japanized form of shushigaku, unique from the actual teaching of Chu Hsi, was developed by Japanese scholars (Tomikura 1981).

In Japanese shushigaku, devotion to one's lord, teacher or other superior was taught to be the same as (and later, more important than) love for one's parents. One was to unquestioningly love and obey one's higher parents over the will and needs of their biological parents (van Wolferen 1989, Anesaki 1980). The Emperor was believed to be the supreme spiritual parent of all Japanese. Senior government officials, teachers and samurai were upheld not only as Superior Men, but also sometimes as elder brothers, sometimes as parental figures serving under the Emperor to take care of the national family (kazoku kokka). Through government-sanctioned private schools, (juku), the common people were taught to maintain this natural harmony by staying in their place, and obeying the will of their superiors. (Bellah 1957, Anesaki 1980, Tomikura 1981).

This contributed to the formation of an interconnected family-like society based upon a system of mutual interdependence (amae kozo). In business, politics and education, the young were taught to depend upon their sempai for guidance, while sempai took full responsibility for the success or failure of their kohai. To a great extent, this aspect of Japanese culture continues to be prevalent in various forms up to the present (Doi 1985).

6.7 From the Meiji Era to World War II

In the wake of the Meiji restoration of 1867, the new Imperial government enacted several policies aimed at making Japan a modern nation that could catch up technologically with the West (Fujita 1985, Aso 1982, Amano 1982). This required the creation of a national compulsory school system. Disenfranchised in the new social order, but well-educated, many from the samurai class became teachers in these new schools. Through their efforts a Neo-Confucian national education system was formed:

School education became an institutional nexus among family, community and polity, and in turn served to diffuse various formal doctrines, promoting diligence at work and in school, advocating the virtues of loyalty and filial piety, and strengthening the Confucian moral order and the emperor system (Fujita, 1985:150).

Further support for this educational theory came through the Imperial Rescript on Education of 1890, which reaffirmed the five relations of Confucianism and the importance of their inclusion into every aspect of Japanese education (Fujita 1985).

The result was an education system that promoted teacher-student relationships patterned after the shushigaku parent-child model. An important function of schools was to instill these Neo-Confucian values and assure that everyone conformed to the social norms of Japanese society.

6.8 Reform and Tradition: 1945 to Present

Japan's national ideology and educational theory were largely called into question after the end of World War II. Nevertheless, the attempt by American occupation administrators to erase Japan's prewar cultural ideology can be likened to a gardener pulling a weed, but leaving its roots behind. Vestiges of the old system still remain (Fukue 1991, Tomikura 1981). Despite their new freedom from direct government control, educators have found it difficult to discard over a millennium of Japanese Neo-Confucian practices, traditions and teaching styles. Fujita (1985) notes that, except for cosmetic changes in the national curriculum, Japanese schools remain essentially unchanged in form, function and purpose since the Meiji Period.

The present function of the teacher is still compared to that of a parent. Japanese teachers frequently affirm this parental role in the research literature. For example, Yamazaki (1996) states, "My job is not simply to teach English well, but to help students to develop well-rounded personalities by continuously encouraging them" (p. 14). Shimahara (1986) notes that the Ministry of Education's policy still considers teachers, not parents, responsible for training students in issues of social morality:

It should be a basic principle that moral education in the school be provided throughout all educational activities of the school . . . In carrying out moral education at school, due consideration should be given to establishing closer human relationships between teacher and students . . . (p. 79-80).

Okuyama (1996) reveals that a parent-child relationship between teachers and students is strongly encouraged, especially during the impressionable high school years:

It is perhaps less well known that Japanese high schools are also family-like and teachers often play more important roles in students' lives than their own parents do (p. 89).

This is supported by the community as well. According to Okuyama, if a student gets into trouble by breaking the law, it is the school, not the parents who are contacted:

. . . Japanese people tend to think of youngsters as students who belong to a school rather than children who belong to a family . . . In Japan, both crimes and personal problems are considered to be the school's responsibility. This demonstrates how deeply Japanese teachers are involved in students' lives (p.91-92).

This relationship often continues for a lifetime, depending upon how well the teacher has cared for the students. Okuyama explains that, over time:

. . . the parent-children-like ties with their students become stronger, and these ties may last for many years, with invitations to weddings and regular visit from former students . . . [because] the students depend more upon their teachers than on their parents, who typically have little knowledge regarding jobs or colleges (p. 92, italics mine).

6.9 Teacher as Superior Parent

This review of Confucianism, Neo-Confucianism and shushigaku allows for an interpretation of the construct findings. The subjects seem to feel that good teachers should be dynamic, morally superior, benevolent figures who will prepare them for success in Japanese society. The construct data reveals that the sample does not merely want teachers to be entertaining, permissive parents. Neither, as stated earlier, do they want their teachers to be only Confucian Superior Men (or Women) who can be looked up to as role models. The good teacher is best expressed in the role of Teacher as Superior Parent.

Seeing the teacher's role as similar to that of a parent in itself is nothing special. To a certain degree Western teachers are also viewed in a semi-parental role by younger students. What may be significant is the cross-cultural conflict that can occur around the construct label of "parent." Idealistically in the West, parents and teachers sometimes allow their children or students to fail at a certain task so they can "learn their lesson." Some feel that failure is able motivate some to try harder in the future. If the student fails again after receiving adequate instruction, usually it is the student, not the teacher, who is ultimately held responsible. Not so in Japan. Generally speaking, most parents and teachers prepare their children in order that they do not fail. Failure brings shame upon parents, who have historically been held responsible for their children's actions. Daidoji Yozan, writer of the ancient Code of the Samurai (Budo Shoshinsh), says this about parents and education:

And children are not to blame for lack of education either. It is entirely due to the neglect and incompetence of their parents who really do not know the way of affection for their children (Code of the Samurai, p. 19).

Teachers in Japan, as Superior Parents, are expected to take full responsibility for the success or failure of their learners. Only a "bad" teacher would hold the student responsible for his actions. Throughout the education process, teachers must guide students to the right class, the right school and the right job where they can succeed if they try hard. As Superior Parents, they demonstrate their compassion for the students entrusted to them by being the prime mover in all of these endeavors, for as Confucius said, "Can you love someone without exerting yourself for them? Can you be sincere to someone without teaching them?" (Analects 14:8).

Since the students in this sample were all new first-year college students, it is reasonable to assume that they were construing about good high school teachers, not college or university teachers. One would suppose that their expectations of good teachers would change as they grow older. But a couple of experiences during the writing of this study have led me to believe that these assumptions may be open for debate.

Recently in my private children's English class, my elementary school second-graders came storming into the house. They were quite upset with a certain teacher who had disciplined them earlier that day. Sensing an opportunity, I asked them to tell me what they thought a good teacher was like. They told me that a good teacher should be kind, intelligent, understandable and sensitive to their feelings - elements strikingly similar to those elicited by the college student sample in Appendix One.

A couple of weeks later, I finished teaching an "English for Citizens" course at the Niigata Municipal Education Center. In Japan, when students finish a course, it is sometimes traditional to make a shuryo bunsho, an informal collection of essays by the students on their final impressions of the course. One gentleman in his late fifties wrote this unsolicited comment:

I thought about a good language teacher. I think he (or she) should have qualities of patience, [sic] generousness, humor, etc. And he (or she) should be like [sic] a actor (-ress) and a newscaster.

While these experiences are admittedly subjective, they allude to the possibility that the role of Teacher as Superior Parent may be a cultural construct formed early in life, and remains essentially unchanged even years after finishing one's formal education. More research is needed to investigate whether this observation is valid.

It is not the intention of this study to offer an overly simplistic interpretation of the sample's constructs. Japanese Buddhism, American Pop Culture, Television, mangabon (phone directory-sized comic books read by Japanese of all ages), and Shintoism are just a few of the many influences that have also affected the lives of Japanese students. The Neo-Confucian influence identified in this study is only one of many cultural patterns in a vast cognitive tapestry. Yet when construing about teachers, a much-Japanized form of Neo-Confucianism appears to be a cultural pattern which stands out in vivid color. Nozaki (1992) notes how, despite the Western veneer of today's Japanese students, their "core values remain traditionally Japanese" (p. 27). This research suggests that Nozaki may be correct - that the core values of Japanese college students appear to have changed little from those of their parents or grandparents, especially when it comes to their demands on teachers (Nozaki 1992 p. 27-28).

7.0 Implications for Teachers

Japanese colleagues frequently voice their concern that today's students are becoming "too Westernized." The findings of this study should give them both relief and pause. Relief because there is little evidence that younger generations are losing their Japanese Neo-Confucian values. Pause because the expectations of their students remain blindingly high. C.S. Lewis (1960, p. 137) once quipped that the teachings of Plato, Aristotle and Confucius, while excellent, have yet to be followed by anyone:

The moralities accepted among men may differ - though not, at bottom, so widely as is often claimed - but they all agree in prescribing a behavior which their adherents fail to practise (1962, p. 21).

Japanese teachers yearning for "the good old days" should ask themselves if they can really be Superior Parents to even a fraction of the hundreds of students they teach in a week.

Some Western teachers will be dismayed by Japanese students' expectations in this study. While many would find a relationship containing the qualities described earlier as fulfilling, few would desire this level of intimacy with their students. And yet the shocking fact of the matter is that this sample wishes to receive such warmth, affection and nurture from their teachers in the classroom. Western teachers might also agree that some of the sample's elicited qualities are important for a good teacher, but most would want to add more items to the list. For example, the elements of the teacher who tested the OMNIGRID program (Hadley 1996a, see again Figure Five) suggest that good teachers should effectively impart knowledge of the subject to the learners. Yet this teaching aspect appears to be one for which a majority of Japanese students have little regard. Their constructs focus almost entirely upon feelings and relationships, not on teaching and learning (in the Western sense of the words). This points to a major reason why so many Western teachers report difficulty in teaching Japanese students: They and their students are on different cultural wavelengths. The negative classroom atmosphere sensed by teachers and students alike might result from cross-cultural misunderstanding around the role expectations for teachers and the implicit goals of education.

A question that begs answering (and I am speaking from a decidedly Western point-of-view here) is whether or not we as teachers should live up to our students' expectations. Shouldn't we rather show our Japanese students that good teachers might be kind, fair and student-centered, but they also may be individuals who have much more to share than simply good feelings? Shouldn't we help them to reconstrue what constitutes a good teacher?

8.0 Final Considerations: Strengths and Weaknesses

There are several admitted weaknesses in this study and the rep grid technique in general, the main one being the unfamiliarity and difficulty of the procedure for both new researchers and students alike. Another is the necessity of having bilingual researchers to oversee the research project. Pope and Keen (1981:44) warn that the rep grid procedure can be exhausting for both the subjects and the researchers. This was certainly the case in this project. Many students commented how mentally drained they felt after finishing the rep grid procedure. Time limitations on element and construct elicitation provided focus for many students, but others felt rushed.

The unavoidable decision to divide the project into two sessions resulted in a slight change in the number of original subjects, thus weakening its external validity. This is an unavoidable fact of life in Japanese college classes, where students are allowed up to 33% total unexcused absences, and even more if they are involved in a school club or searching for a job.

Some may disagree with the rationale used for selecting the grid's elements. If time and resources were available, it would be interesting to see if a different choice of elements from Appendix One would have resulted in significantly different constructs (but I suspect that it would not). Some as well might fault me for risking the halo effect by telling the subjects they would be participating in important research. However, the Japanese colleagues who have seen this data assure me that the constructs elicited by the sample appear to be similar to the expectations they had for their teachers during their high school and college years.

Others may be concerned that by allowing groups to form constructs I risked the creation of a self-selection variable (Brown 1993, p. 31), which could have allowed any strong-willed students to stifle the opinions of the more reserved group members. Related to this is the concern that by creating a megagrid of all the constructs, the unique understanding of construct labels elicited by each group or individual might have been ignored.

However, Japanese strongly affirm their homogeneity in terms of race, education and cultural values. This research set out to respect Japanese Consensus rather than to impose Western Individuality. More studies from Middle and Far Eastern countries (cf. Jones 1995, Barnett and Jordan 1991) call for Western researchers to show more sensitivity to Asian values and concerns. I allowed group work for the following reasons:

Repertory grids were developed in the West where individuality is, idealistically speaking, more valued. However, when embarking upon something new, many Japanese would prefer to work in a group.  Asking the students to construe about good teachers in a "pop quiz" atmosphere could have negatively affected the data. Social groups in Japan are formed through a consensus. Individuals normally join a group where their ideas are accepted. I am convinced that the data produced by the groups represents the cooperative effort of like minds, construing together in a manner consistent with their cultural norms.

Despite these weaknesses, it appears that the rep grid technique gets around the cultural bias of the researchers by allowing subjects to communicate from their cultural nexus. The flexibility of the rep grid technique allows for adjustments according to the needs of the researcher and the subjects. This is a major advantage over fixed techniques such as questionnaires and interviews, which must be piloted first to avoid disaster. The data from grids can be used for the creation of valid surveys or any number of action research projects. Repertory grids also go one step further than questionnaires by identifying the deeper mental constructs which form the subjects' "surface opinions", as we saw in this research project. Some of the elements elicited by the subjects in the survey section of this research mitigated against an interpretation of Confucian/Neo-Confucian influence in the subjects' understanding of the Good Teacher (for example, "Friendly" or "Entertaining"). If was only after going under the surface by studying the constructs elicited through the rep grid technique that we discovered that Japanese Neo-Confucian thought is still a significant influence when the subjects think about good teachers and good teaching practices.

When properly designed, the repertory grid technique is a powerful tool for cross-cultural research. Provided one is willing to make the investment of time, EFL teachers will find that the repertory grid technique is excellent for obtaining data not normally found in other research instruments. As long as efforts are made to protect the reliability and validity of the procedure, conscientious teachers involved in similar EFL environments may find the results of grid-based research interesting and directly applicable to their classroom teaching.

8.1 Future Research

This study offers tantalizing leads for further research. As hinted earlier, an interesting project would be to find out whether or not the good teacher constructs of Japanese college students change after four years of university education. Are good teacher constructs of Japanese University teachers different from those of High School teachers? What are the areas of consensus and conflict in construct labels for expatriate teachers and Japanese teachers? Would cross-cultural comparisons of the constructs elicited by this sample differ from their counterparts in American or British schools?

9.0 Conclusion

This study has had a number of implications for my teaching. The insights gained from this research have motivated me to adjust some of my previous teaching practices. My basic goals, to lower affective barriers and encourage increased L2 use and fluency, have not changed. What has changed however is the way in which I work toward those goals. I have moved from a structural syllabus-driven approach to a process syllabus approach which allows for more humanistic forms of learning (cf. Moskowitz 1978). The warmth and caring this change has stimulated has encouraged many in my classes to make greater efforts to express themselves in English. Whereas before I was performing a one-man show (complete with a pedagogic whip and chair), now in some classes the students and I are cooperating in the task of studying English. In these classes the students are bringing English materials to class that we incorporate into the lesson. Several now speak with me after class on a regular basis, wishing to discuss topics ranging from advice when traveling abroad, to how to get an English magazine subscription, to Western philosophy.

Regardless of how much or little EFL/ESL teachers of Japanese English language learners adjust their teaching as a result of this study, it is hoped that it has provided them with a deeper understanding about the perceptions their students have about good teachers. Japanese students understand their ideal teacher in terms of a Superior Parent. They are not nearly so concerned about what their teacher does as they are about who their teacher is. Character issues far outweigh any skills or abilities the Good Teacher might have. This does not mean that language teachers should discard their pedagogic goals and theories. Nor does it mean they should interpret these findings as justification for a glorified popularity contest. What this paper does propose however is that enriching and educating our Japanese learners has a better chance for success when teachers understand Japanese Neo-Confucian thought, and appreciate the humanistic concerns of the Japanese educational system.

In conclusion, language teachers must continue the process of cross-cultural discovery in their EFL classrooms. This will require willingness on the part of both teachers and learners to try new approaches in an effort to bridge their cross-cultural differences. The repertory grid procedure, while complex, opens doors to an enhanced understanding of our students' cultural perceptions. This technique provides opportunities not only for personal enrichment, but by entering into our students' culture, we equip ourselves with insight on how to better teach and reach our learners.



Sample Grid Data

The following are two example grid sheets selected from the 62 grid sheets received. Sample Grid One was filled out by a group of men, while Sample Grid Two was filled out by a group of women.

Translation of Constructs in Sample Grid One:



Easy to get along with

Hard to get along with

Teaches eagerly

Teaches with neglect

A good impression

A bad impression







Shows no discrimination

Shows discrimination

Humane (Jin)

Human garbage


Translation of Sample Grid Two:



A good speaker

A poor speaker







Easy to ask questions to

Difficult to ask questions to

Understands students' feelings

Ignores student's feelings

Doesn't get angry


Has many things to talk about

Not knowledgeable


(1996). Corel WordPerfect User Manual 6.1. Corel Corporation Limited: Dublin, Ireland.

(1993). RepGrid 2 Manual. Centre for Person Computer Studies: 3019 Underhill Drive NW, Calgary, AB, Canada T2N 4E4.

Amano, I. (1982). Kyiku to senbatsu (Education and Selection) Tokyo: Dai-ichi Hoki Shuppan.

Anesaki, M. (1980). History of Japanese Religion. Tokyo: Charles E. Tuttle Company.

Anderson, R. (ed.)(1981). New Dimensions in Second Language Acquisition Research. Rowley, MA: Newbury House.

Aso, M. (1982). Kindaika to kyiku (Modernization and Education) Tokyo: Dai-ichi Hoki Shuppan.

Bannister, D. (ed.)(1970). Perspectives in Personal Construct Theory. London: Academic Press.

Bannister, D. and Mair, J.M.M. (1968). The Evaluation of Personal Constructs. London: Academic Press.

Bannister, D. and Fransella, F. (1966). 'A repertory grid test of schizophrenic thought disorder', in British Journal of Social and Clinical Psychology. 2:95-102.

Bannister, D. and Fransella, F. (1971). Inquiring Man: The Theory of Personal Constructs. London: Penguin.

Barnett, L. and Jordan, G. (1991). 'Self-access facilities: What are they for?', in ELT Journal. 45:4.305-12.

Bavelas, J.B., Chan, J.A., and Guthrie, J.A. (1976). 'Reliability and validity of traits measured by Kelly's repertory grid', in Canadian Journal of Behavior Science. 8:1.23-38.

Bell, J. (1995). Doing Your Own Research Project: A Guide for First Time Researchers in Education and Social Science. (2nd edition), Philadelphia: Open University Press.

Bell, R.C. (1987). G-Pack: A Computer Program for the Elicitation and Analysis of Repertory Grids, Version 3.0. University of Wollongong: Personal Construct Group, Department of Psychology.

Bell, R.C. (1994). 'Using SPSS to analyse repertory grid data'. Unpublished report. Department of Psychology, University of Melbourne.

Bell, R.C. (1996, May). 'Analyzing repertory grid data'. Paper presented at the University of Huddersfield, UK.

Bellah, R. (1957). Tokugawa Religion. Glencoe, Illinois: The Falcon's Wing Press.

Beversluis, J. (ed.)(1995). A Source Book for Earth's Community of Religions. Grand Rapids: CoNexus Press and New York: Global Education Associates. Available HTTP: [1996, June 25].

Brown, J.D. (1988). Understanding Research in Second Language Learning: A Teacher's Guide to Statistics and Research Design. New York: Cambridge University Press.

Chambers, W.V. and Grice, J.W. (1986). 'Circumgrids: A repertory grid package for personal computers', in Behavior Research Methods, Instruments and Computers, 18:5.468.

Chaudron, C. (1995). Second Language Classrooms: Research on Teaching and Learning. New York: Cambridge University Press.

Chung, D. (1995). 'Confucianism', in Beversluis, J. (ed.) (1995). A Source Book for Earth's Community of Religions. Pg. 31-32. Grand Rapids: CoNexus Press and New York: Global Education Associates. Also available HTTP: [1996, June 25].

Cohen, L. and Manion, L. (1987). Research Methods in Education. London: Croom Helm.

Diamond, C.P. (1983). 'The use of fixed role treatment in teaching', in Psychology in the Schools. 20:1.74-82.

Diamond, C.P. (1984). 'The repertory grid in teacher education', in Bulletin of British Psychological Society. 37:57-58.

Doi, T. (1985). The Anatomy of Self: The Individual Versus Society. Tokyo: Kodansha International Ltd.

Easterby, M. (1981). 'The design, analysis, and interpretation of repertory grids', in Shaw, M.L.G. (ed.) Recent Advances in Personal Construct Technology. Pp. 9-30. London: Academic Press.

Epting, F.R., Schuman, D.I, and Nickeson, K.J. (1971). 'An evaluation of elicitation procedures for personal constructs', in British Journal of Psychology. 62:513-17.

Feixas, G., and Cornejo-Alvarez, J.M. (1996). A Manual for the Repertory Grid: Using the GRIDCOR Programme (Version 2.0). Barcelona: Paidós.

Feixas, G. y Cornejo-Alvarez, J. M. (1996). Manual de la técnica de rejilla mediante el programa RECORD v. 2.0 (2 ed.). Barcelona: Paidós.

Fransella, F. and Bannister, D. (1967). 'A validation of repertory grid technique as a measure of political construing', in Acta Psychologica. 26:97-106.

Finkelstein, B., Imamura, A., Tobin, J., (eds.). (1991). Transcending Stereotypes: Discovering Japanese Culture and Education. Yarmouth, Maine: Intercultural Press.

Fujita, H. (1985). 'Educational policy dilemmas as historic constructs', in Finkelstein et al. (eds.). (1991). Transcending Stereotypes: Discovering Japanese Culture and Education. Pp. 147-61. Yarmouth, Maine: Intercultural Press.

Fukue, H. (1991). 'The persistence of IE in the light of Japan's modernization', in Finkelstein, et al. (eds.) (1991). Transcending Stereotypes: Discovering Japanese Culture and Education. Pp. 68-73. Yarmouth, Maine: Intercultural Press.

Gathercole, C., Bumble, E., and Ashcroft, J. (1970). 'The reliability of repertory grids', in Journal of Clinical Psychology. 26:4.513-516.

Greene, D. and Hunter, L. (1993) 'The acculturation of oral language learners and instructors in EFL', in The Language Teacher. 17:11.9-15, 47.

Greer, D. (1996). 'Gardener and Lambert in the classroom', in The Language Teacher. 20:1.10-14.

Grotjahn, R. (1987). 'On the methodological basis of introspective methods', in Faerch, C. and Kasper, G. (eds.). Introspection in Second Language Research. Pp. 54-81. Clevedon, Avon: Multilingual Matters.

Hadley, G. (1995a). 'Class participation: A solution for Japanese and Korean university English courses', in The Korea TESOL Journal. 3:3.124-25.

Hadley, G. (1996a, May). 'New ways to answer old questions: An introduction to repertory grids'. Paper presented at the Niigata University General Education and Language Research Group. Niigata, Japan.

Hadley, G. and Yoshioka-Hadley, H. (1996). 'The culture of learning and the good teacher in Japan', in The LanguageTeacher. 20:9.53-55.

Hall, D.L. and Ames, R.T. (1987). Thinking Through Confucius. New York: State University of New York Press.

Hill, A.B. (1976). 'Validity and clinical utility of grid test of thought disorder', in British Journal of Psychiatry. 128:251-254.

Hooker, R. (1995). Jen. University of Washington. [Online]. Available HTTP: [1996, October 28].

Hori, I., Ikado, F., Wakimoto, T. , Yanagawa, K. (eds.). (1981). Japanese Religion: A Survey by the Agency for Cultural Affairs. Tokyo: Kodansha.

Horio, T. (1988). Educational Thought and Ideology in Modern Japan: State Authority and Intellectual Freedom. edited and translated by Platzer, S. Tokyo: University of Tokyo Press.

Ichii, S. (1974). 'On Tradition and Innovation', in Japan Quarterly. 21:3.273-283.

Lewis, C.S. (1960). Mere Christianity. New York: Macmillian Publishing Company.

Lewis, C.S. (1962). The Problem of Pain: How Human Suffering Raises Almost Intolerable Intellectual Problems. New York: Macmillian Publishing Company.

Pope, M. and Keen, T.R. (1981). Personal Construct Theory and Education. Ontario: Academic Press.

Jones, J. (1995). 'Self-access and culture: Retreating from autonomy', in ELT Journal. 49:3.228-34.

Kelly, G. A. (1955). The Psychology of Personal Constructs. (Vol. 1), New York: W.W. Norton & Co.

Kelly, G. A. (1969b). 'Strategy of psychological research', in Maher, B.A. (ed.) (1969). Clinical Psychology and Personality: The Selected Papers of George Kelly. Pp. 114- 132. New York: Wiley.

Kelly, G. A. (1970). 'A brief introduction to personal construct theory', in Bannister, D. (ed.) Perspectives in Personal Construct Theory. Pp. 1-29. London: Academic Press.

Kirk, D. (1992). 'Motivators in English Conversation Schools', in The Language Teacher. 16:2.22-25.

Kobayashi, S., Redekop, B., and Porter, R. (1992). 'Motivation of college English students', in The Language Teacher. 16:1. 7-9, 15.

Lansdown, R. (1975). 'Reliability study of the 8*8 repertory grid', in Association of Educational Psychologists Journal. 3:10.24-25.

LeCompte, M., and Goetz, J. (1982). 'Problems of reliability and validity in ethnographic research', in Review of Educational Research. 52:1.31-60.

Legg, J. (1996). Ingrid Thought Processor V6.0b1. New Zealand: Newton, Auckland. Also available HTTP: [1996, November 9].

Leitner, L. (1981). 'Construct validity of a repertory grid measure of personality styles', in Journal of Personality Assessment. 45:5.539-544.

Long, M. (1983a). 'Input and second language acquisition theory'.  Paper presented at the Tenth University of Michigan Conference on Applied Linguistics. Cited in Ellis, R. (1986). 'Theories of Second Language Acquisition'. Pp. 248- 276, in Ellis, R. Understanding Second Language Acquisition. Oxford: Oxford University Press.

McCoy, M. (1983). 'Personal construct theory and methodology in intercultural research' in Adams-Webber, J. and Manusco, J.C. (eds). (1983). Pp. 173-186, Applications of Personal Construct Theory. Ontario: Academic Press Canada.

Moskowitz, G. (1978). Caring and Sharing in the Foreign Language Classroom. Boston: Heinle & Heinle Publishers.

Muller, A.C. (forthcoming). Five Chinese Classics: New Translations of the Basic Texts of East Asian Indigenous Thought. Taipei: Jin Luen Publishing.

Munby, H. (1982). 'The place of teacher beliefs in research on teacher thinking and decision making, and an alternative methodology', in Instructional Science. 11:201-25.

Murakami, Y. (1985). 'Bullies in the classroom', in Finkelstein, et al. (eds.). (1991). Transcending Stereotypes: Discovering Japanese Culture and Education. Pp. 190-96. Yarmouth, Maine: Intercultural Press.

Novak, J.D. (1977). A Theory of Education. Ithaca, Illinois: Cornell University Press.

Novak, J.D. and Gowin, D.B. (1984). Learning How to Learn. New York: Cambridge University Press.

Nozaki, K.N. (1992). 'The Japanese student and the foreign teacher', in Wadden, P. (ed.)(1992). A Handbook for Teaching English at Japanese Colleges and Universities. Pp. 27-33. New York: Oxford University Press.

Nunan, D. (1992). Research Methods in Language Learning. Cambridge University Press.

Nunan, D. (1989). Understanding Language Classrooms: A Guide for Teacher-Initiated Action. Prentice Hall.

Okuyama, M. (1996). 'Teacher-student relationships in Japanese high schools', in Wada, M and Cominos, A. (eds.). (1996). Japanese Schools: Reflections and Insights. Pp. 89-93. Kyoto: Shugakusha.

Olson, J. (1981). 'Teacher influence in the classroom', in Instructional Science. 10:259-75.

Richards, J. and Lockhart, C. (1995). Reflective Teaching in Second Language Classrooms. Cambridge University Press.

Rohlen, T. (1974). 'Up and down', in Finkenstein, et al. (eds.). (1991). Transcending Stereotypes: Discovering Japanese Culture and Education. Pp. 20-25. Yarmouth, Maine: Intercultural Press.

Sadler, A.L. (1988). The Code of the Samurai: A Translation of Daidji Yzan's Bud Shoshinsh. Tokyo: Charles E. Tuttle Co.

Sakamoto, A. (1996, Aug.). 'Cognitive complexity and the number of points for rating'. Paper presented at the 103rd APA Convention. Toronto, Canada.

Sakamoto, A. (1993). Cognitive Complexity and the Evolution of Sociocognitive Systems. Tokyo: Kazama Shbo (in Japanese).

Sakamoto, A. and Numazaki, M. (1989). 'The effect of time for rating on a cognitive complexity scores of the Rep test' in The Japanese Journal of Psychology. 60:5.316-319. (in Japanese with English abstract).

Schumann. J. (1978). 'The acculturation model for second language acquisition', in Gingras, R. (ed.) Second Language Acquisition and Foreign Language Teaching. Pp. 27-50. Arlington, VA: Center for Applied Linguistics.

Sewell, K.W., Mitterer, J.O., Adams-Webber, J., and Cromwell, R.L. (1991). 'OMNIGRID-PC: A new development in computerized repertory grids', in International Journal of Personal Construct Psychology. 4:175-192.

Sewell, K. (1989). OMNIGRID-PC, Version 1.5. University of Kansas: Department of Psychology.

Shaw, M.L.G. (1994). 'Methodology for sharing personal construct systems', in Journal Of Constructivist Psychology. 7:35-52.

Shaw, M.L.G. (ed.)(1981). Recent Advances in Personal Construct Technology. London: Academic Press.

Shaw, M.L.G. and Gaines, B.R. (1995, October). 'Comparing construction through the web'. A paper presented at the Proceedings of CSCL '95: Computer-Supported Cooperative

Learning. Bloomington, USA. Also available HTTP: [1996, October 30].

Shaw, M.L.G. and Gaines, B.R. (1992). 'Kelly's 'geometry of psychological space' and its significance for cognitive modeling', in The New Psychologist. 9:23-31. Also available HTTP: [1996, May 11].

Shaw, M.L.G. and McKnight, C. (eds.). (1981). Think Again: Personal Problem-Solving and Decision-Making. Englewood Cliffs, N.J.: Prentice-Hall.

Shaw, M.L.G. (1980). On Becoming a Personal Scientist: Interactive Computer Elicitation of Personal Models of the World. London: Academic Press.

Shimahara, N. (1986). 'The cultural basis of student achievement' in Comparative Education. 9:1.19-26.

Shimizu, K. (1995). 'Japanese college student attitudes towards English teachers: a survey', in The Language Teacher. 19:10.5-8.

Takagi, E. and Sakamoto, A. (1991). 'Changes in national images due to Seoul Olympic: Panel survey using college students', in Shakai Shinrigaku Kenkyu (Social Psychology Research). 6:2.98-111. (in Japanese with English abstract).

Tomikura, M. (1981). 'Confucianism', in Hori, et al. (eds.). (1981). Japanese Religion: A Survey by the Agency for Cultural Affairs. Pp. 103-20. Tokyo: Kodansha.

Wadden, P. (ed.) (1992). A Handbook for Teaching English at Japanese Colleges and Universities. New York: Oxford University Press.

Wright, A. (1995). 'A traveling storyteller', in The Language Teacher. 19:10.16-19, 26.

van Wolferen, K. (1989). The Enigma of Japanese Power: People and Politics in a Stateless Nation. New York: Alfred A. Knopf.

Wada, M. and Cominos, A. (eds.). (1996). Japanese Schools: Reflections and Insights. Kyoto: Shugakusha.

Yamazaki, Y. (1996). 'What are the students there for?', in Wada, M. and Cominos, A. (eds.). (1996). Japanese Schools: Reflections and Insights. Pp. 10-14. Kyoto: Shugakusha.

Yorke, D.M. (1985). 'Administration, analysis and assumptions: Some aspects of validity', in Beail, N. (ed.) Repertory Grid Technique and Personal Constructs: Applications in Clinical and Educational Settings. Pp. 383-399. London: Croom Helm.