We have essentially adapted information written by Dr. Valerie Stewart of Enquire Withintm as the basis for this technique. Dr. Stewart has facilitated the technique for over thirty years and obviously has a great deal of information to offer. It is a sign of the times that her primary consultations are with business and not human services. This technique is uniquely suited to the community assessment process although it can be productively used with individuals and with groups. It is our hope that we can bring this aspect of assessment without bias to the human services arena.
ENQUIRE WITHINTM – Trademark Registered in New Zealand
Copyright © 2000 by Dr Valerie Stewart and John Mayes
NOTE: The Repertory Grid Technique is an extraordinarily versatile tool. While we have chosen to list it as an assessment tool because of our own desire to make assessments of children more goal oriented than ‘need’ oriented, it could easily have been listed as a management tool or a practitioner tool. It is emphasized throughout that the technique is a conversation – however, it is a structured conversation, which may be documented or not. As a documented conversation, it can be used to identify and direct change of a child-serving organization, or to cognitively map a client for counseling. As an undocumented conversation, it can be used in situ by the counselor to explore with the client his/her thoughts, beliefs and values about a given domain of interest. Because of this versatility, we have included many of the examples in the original materials which describe use of the grid with business organizations. For the most part, we have left those references in because the examples can be applicable to other organized groups [families or schools] and practitioners may want to use the technique in that way.
Repertory Grid Technique CBAT#06
The Repertory Grid technique is a structured conversation which should be used in a relaxed atmosphere.
Record: Document all information as it is divulged.
01. define domain to be explored
02. define concrete and discrete elements. This may be planned by the facilitator or done with the subject(s).
03. define qualifiers – this should be done by the facilitator.
04. start conversation – select [planned or random] three  elements for comparison
05. subject identifies a similarity that exist between two  of the selected elements
06. subject identifies the difference [contrasting aspect] of the third element – this constitutes a two poled construct
07. change selected elements [planned or random] – may change as few as one selected element or as many as three
08. repeat steps five  and six  09. repeat step seven  as many times as comfortable
10. have subject(s) consider and group the constructs that appear to be similar [NOTE: new constructs may emerge in this process
11. select the constructs to be ‘laddered’ – How important is this construct to the subject(s) and what pole if the most valued?
12. organize the list of constructs with a Likert Scale [1 to 5] between the poles
13. ask the subject(s) to rate each element from 1 to 5 on each construct
NOTE: Give the opportunity for feedback throughout – e.g., if facilitator has selected the elements – give the subject(s) the opportunity to comment on the validity of the select and suggest others – constantly give the subject(s) time to reflect upon and reconsider their responses.
Repertory Grid Technique CBAT#06
The Repertory Grid is an instrument designed to capture the dimensions and structure of personal meaning. Its aim is to describe the ways in which people give meaning to their experiences in their own terms. It is a structured interview designed to make explicit those constructs with which persons organize their world. The way in which we get to know and interpret our environment, our understanding of ourselves and others is guided by an implicit and self created theory of meaning which is the result of conclusions drawn from our experiences. The repertory grid is a method used to explore the structure and content of these implicit theories (personal meanings) through which we perceive and act in our day-to-day existence.
Personal Construct Theory [PCT] is considered to be a predecessor of the cognitive approach that is currently dominating the field of clinical and social psychology. Developed by George Kelly, PCT is predicated on one axiom: man is a scientist. In other words, from first consciousness each of us tries to make sense of the world as we experience it, and we do this by constantly forming and testing hypotheses about the world. By the time we are between four to seven years of age, we will have formed a naive, but complete ‘Theory of Meaning’ and will have developed a very complex model of the world and our place in it. This model is, according to Kelly, our personality.
Kelly’s theory of personal constructs develops this principle further – for example, by considering whether and how we modify our constructs when faced with contradictory information, what are our ‘core constructs’ – that is, the deeply held values and principles which are unlikely to change, etc.
At that time, Kelly was employed as a school psychologist and was interested in those students described as ‘problem children’. What he wanted to develop was some way of identifying how the teachers, who referred the children to him, ‘construed’ the children’s problems themselves.
The fundamental postulate of the theory is that: a person’s processes are psychologically channeled by the ways in which s/he anticipates events. The theory is based on the notion that human beings create their own network pathways [cognitive structures/processes]. The individual’s major motivation is his/her need to predict and thus personalize constructs. These are the dimensions that are used to conceptualize aspects of our everyday world. Kelly’s theory suggests that these dimensions are continually being developed, and that those that prove good ‘predictors’ are kept and elaborated, while those that are not reinforced by successful application are ultimately discarded.
The term construct is particularly well-chosen, because it reflects the concept’s dual role. On the one hand, your constructs represent the view you have constructed (designed and built) about the world as you experienced it. On the other hand, your constructs indicate how you are likely to construe (perceive, interpret) the world as you continue to experience it. Your construct system is your personal history and your predisposition to perceive it in set ways.
Constructs are the symbols we use to describe our judgements about things, events and people. Because constructs represent some form of judgement or evaluation, by definition they can fit on a scale or are balanced or bipolar: that is, the concept good can only exist in contrast to the concept bad, the concept gentle can only exist as a contrast to the concept harsh. Any evaluation we make – when we describe a car as sporty, or a politician as right-wing, or a sore toe as painful – could reasonably be answered with the question ‘Compared with what?’ Constructs are bi-polar and always identify a preference. Subjects are therefore asked to select which end of the bi-polar construct they favor and why. This process called ‘laddering’ can be used to elicit deep construction.
A person is guided by an entire network of meaning which consists of a network of personal constructs analogously arranged concentrically with the most central or ‘core’ constructs being those that define the person’s identity. Most of these core constructs are involved in construing those significant others with whom the person interacts and the nature of one’s ‘role relationship’ to them. In addition, there are more peripheral constructs that, although subordinate to these core constructs, are actively involved in construing events and further actions. The repertory grid is a structured procedure designed to elicit a grouping of constructs around a specific domain and to explore their structure and interrelations.
Later theorists have developed these constructs and models into schema theory and posit the concepts of schemata, scenarios and mental maps. Our interest here, however, is to explore how to use these notions as a means of assessing individuals and/or groups (e.g., family or classroom) to discover the major constructs they use in specific domains. For example, you and I may talk about motherhood and both agree that a person we both know is an ideal mother. What is not immediately apparent, however, is whether we are each making the same estimation on the basis of the same or different criteria. In assessment, it is important that we understand the mental maps of our clients: what criteria [constructs] do they use to make decisions in a specific domain [e.g., family or school] and how important is each construct to the final decision [e.g., what are the values of the constructs organized around the domain family; what will the person change relatively easily and what core constructs will take more evidence].
Kelly developed the Repertory Grid interview as a means of getting people to show their construct systems or mental map. He had some very important methodological concerns about the standard of interviewing. His major concerns were:
• Interviewer Bias. He had seen that the interviewer often contributes more to the diagnosis than the person being interviewed. In a recent book on cognitive behavioral therapies, for example, a clinician is recommending that parents be asked: “What is the problem with your thoughts?” – implying, of course, that the parent’s thoughts were wrong. Such bias is difficult to overcome, but causes the subject to interpret the conversation quite differently than they might if the language were ‘clean’.
• Specificity in measuring, and where possible predicting, the characteristics of individual people and small groups. Many psychologists were seeking ‘laws’ of human behavior. So there was the ‘rats, cats, and stats’ approach to studying behavior, and large-scale studies showing the correlations between different aspects of personality and behavior; but these were no good to Kelly, nor are they to any other clinician, because they see people one at a time or in small groups.
• Over-dependence on the expert. Clinical services psychology can too often be described as perceiving the client as a non-autonomous agent who needs the ‘expert’ to tell him/her what is wrong. Kelly took the view that most people can take responsibility for understanding their behavior and, where necessary, adapting it; and that the more useful role of the clinician would be as a ‘skilled mirror.’
The Essence of Repertory Grid Technique
In order to circumvent these obstacles in his quest for the client’s construct system or mental map, Kelly invented the Repertory Grid interviewing technique. This manual outlines the procedures involved in the repertory grid techniques as well as the main associated areas, making it easier to adapt the technique to the user’s area of interest.
The process consists of four  basic steps:
1. design phase: given the multiple applications [the process has been used for marketing and management consultation, as well as for clinical applications] the design phase is where the parameters that define the area of application are set out;
2. administration or implementation phase: is where the type of structured interview for the elicitation of information and the resulting matrix is defined;
3. analysis phase: is where the data is analyzed [mathematical analyses are possible through readily available pre-developed software such as EnquirewithinTM]; and the
4. response phase: in which the information is used to decide how to intervene in a manner that will help the person with problems in living to progress.
NOTE: Because the learning process is overlapping among the first three phases, we will repeat section titles as necessary.
The Design Phase
Basically, a repertory grid consists of a) a series of elements that are representative of the area under study, b) a set of personal constructs that the subject uses to compare and contrast the elements, and c) a rating system that evaluates the elements based on the bipolar arrangement of each construct. The essence of Grid technique therefore, is very simple:
• Identify clearly the domain that you wish to explore. The domains can be anything (such as family, school classroom, friends, etc.).
• Select a set of elements. The elements are concrete examples of the domain you wish to explore – for example, working with a client who has had problems in making satisfactory relationships, the elements would be people with whom the client had relationships.
• In a presentation, the interviewer takes the elements in groups of three, and ask the question:
‘Can you tell me a way in which any two of these people are different from the third, in terms of ……?’
(The ‘in terms of’ phrase, called a qualifier, directs the client to consider the elements in a way appropriate to the purpose. So in this case the qualifiers might be: ‘… in terms of how you feel about them, …. in terms of how they felt about you, …. in terms of how you behaved to each other,’ etc.). This two-against-one question produces a bipolar scale – for example had a sense of humor – I never saw him smile. This scale is a construct that comes entirely from the interviewee. The interviewer has set up the parameters for the conversation, but has suggested none of the content.
Kelly suggested that the triads be presented sequentially by changing only one element at a time. If at least one ‘self’ element is retained in each presentation, the personal relevance of the elicited constructs is reasonably assured.
There are ways of exploring the constructs in more depth and detail that we will get into later, but at some point in a full Grid interview the constructs are turned into Likert scales, usually 1 to 5 with 1 being the preferred pole (See CBAT- #06-001) and the interviewee is asked to rate every element on every construct. This gives the facilitator a matrix which can be analyzed statistically in order to pursue the progress of the discussion with the client. The statistical analysis answers Kelly’s need to measure people individually; for example, you could compare the person’s before-and-after perceptions. There are several computer analysis programs available, but it should also be stressed that it is not always necessary to use them. A great deal of information can be learned about a person simply by listening to how they use their constructs and an analysis of data can be done without letting yourself become dependent on a computer program to do the work for you.
That is the essence of Grid. It is a powerful and content-free procedure. The interviewer sets up the session in order to meet the purpose, but provides none of the content. There is simply no other interviewing technique which will allow you to cover the breadth and depth of the client’s cognitive map of their world. Its complete freedom from interviewer bias, and the transparency of the process, means that the interviewer can say to the client ‘All I’m doing is playing your own perceptions back to you.’
Because the Grid is a standardized protocol, any interviewer can pick up the work of any other and understand what happened in the interview. If you are conducting a large project [a total school population] with a number of interviewers, you don’t have to have reconciliation meetings where everyone has taken notes in their own way and you spend time explaining to one another.
The Administration Phase
Keeping Free From Interviewer Bias
What does all this mean for a new interviewer utilizing the Grid? At its simplest, it means that we have the means of knowing when we have influenced the discussion. Because the Grid can be completely free from interviewer bias, it should be kept that way until, or unless, the purpose requires us to intervene. Some practical examples:
• Don’t supply the contrast pole yourself. If the client says ‘These two had a sense of humor,’ don’t say ‘And the other one didn’t?’ Say ‘And how would you contrast the other?’
• Don’t summarize the client’s constructs, either verbally or when writing them down. If the client says ‘S/he could almost always find two or three new ways of looking at a problem’, that’s what you write. You don’t write ‘Creative problem-solver.’ Try to avoid asking the interviewee to summarize a construct, even if there are lots of words, because s/he may say ‘Creative problem-solver.’ What you have in the first phrase is a detailed, behavioral description of the element. You will probably get a similarly specific description of the contrast pole. Later on you could find yourself needing that specificity.
• Don’t imply that you’re judging the client’s constructs or waiting for a particular type of construct to appear. You do want to elicit constructs that are relevant to your purpose (that’s what the qualifiers are for) but in the early stages of a Grid interview you should respect the fact that the two-against-one process is not how most people are used to thinking and your first goal is to get them comfortable with it. Once they are comfortable with it, you can remind them of the qualifiers (for example ‘Could you look at these three in terms of ……’), or you can ask ‘Does that construct relate to the purpose?’
• Don’t judge the importance of someone’s constructs. For example, a facilitator interviewed a number of people and wrote ‘We stopped the interviews after eliciting twelve meaningful constructs.’ Who decided that they were meaningful – you, or the client? And what exactly was meant by ‘meaningful?’ It’s not the facilitator’s job to decide that. It is much better to elicit the constructs until the client runs out, do some ‘laddering’ to change the focus, and then ask if that has prompted any more constructs. Ask the client to sort the constructs into high, medium, and low priority.
One core value for a good Grid interviewer is: don’t interfere with other people’s construing. Don’t judge. Ask open questions: for example, if you can see a pattern in someone’s constructs it’s better to ask ‘Can you see any patterns, or groups, in what you’ve said so far?’ than to say ‘You’ve got a lot of constructs about humor.’
However, the interviewer should not be completely passive. There will be occasions in any interview when questions or input from the interviewer are appropriate. It is important to make as much use as you can of the unique opportunity. The Grid helps you to understand the client’s world in their own terms. If you interpose yourself in the process, you’ll never get that state again. Be open about the process and what you’re recording. Make the interview a joint endeavor. Most interviewees will quickly find the process interesting and many become almost self-managing. Then the interview becomes a cooperative process in which you offer techniques and they offer answers and insights.
Designing a Session
There are lots of cross-references to other sections of this material, because some of the issues (such as choosing an element set, and methods of analysis) are treated in greater detail there.
Example: The Assignment
Suppose that you’re studying Modern American History and your tutor gives you an assignment: ‘Discuss the factors which characterize the most effective Presidents of the 20th century.’ How will you proceed with the task?
First you assemble a list of the 20th century Presidents. Second, who should you consider all of them, or a selection. You will want to make sure that your sample includes the extremes and maybe some about whom opinions differed. While you are doing this, another part of your brain is probably asking questions like ‘Effective in whose eyes? Isn’t effectiveness dependent to some extent on the situations they had to confront? Is it my opinion or should I do a literature search as well?’
What you are doing with this internal dialogue is only a very short step away from designing a Repertory Grid session. The Grid is simply a way of formalizing this quest and then playing back your perceptions so that you can see how much ground you have covered, what patterns have emerged and whether you are content with them, and encouraging you to be thorough.
Configuring the Session
So how would you configure the session? The building blocks of any Repertory Grid session are: the purpose, the elements, and the qualifying questions. The purpose is a statement of why you are doing the interview; the elements are concrete examples of the domain you want to explore; and the qualifying questions direct you to think about the elements in a way that is relevant to the purpose. In this case, your purpose statement would be something like ‘To explore my perceptions of 20th Century American Presidents’. Your elements would be a selection of Presidents, chosen according to the sampling method outlined above.
In choosing the qualifying questions you need to strike a balance between the general and the specific; for example, quoting from the internal dialogue you might have started thus: ‘Effective in whose eyes? What about those where history has changed how they are evaluated? Isn’t effectiveness dependent to some extent on the situations they had to confront? How do we define effectiveness anyway? Is it my opinion that the tutor wants, or should I do a literature search as well?’ This internal dialogue is prompting some qualifiers and some constructs. A good rule of thumb is that you should make your qualifying questions fairly general, because you’ll probably need only three or four. Therefore you might have: ‘In terms of their personal attributes ……, In terms of the situations they faced ……, and ‘In terms of public opinion……’.
Then you’re ready to take the elements [selected Presidents] in groups of three and ask yourself ‘In what way are any two of these Presidents similar to each other and different from the third?’ and now you’re into construct elicitation.
You will also want to discover the selected elements of your clients domain, being careful to understand that in doing so, you will want the client’s decisions about elements, not your own. For example, suppose you are concerned with the child’s family relationships. You would first want the child to identify the ‘family members’ [elements of the domain] which may include or exclude people that are surprising to you. It is the child’s construct of family that you are concerned about so the inclusion or exclusion of certain people would not, at least at this point, be questioned.
Your next task would be to list these elements of the family domain on cards so that they can be compared.
A Note on Qualifiers
The qualifiers direct the client to think about the elements in ways that are relevant to the purpose. They can be very important at the start of the interview, because they set the scene, but mostly they fade out of the discussion once the interviewee has hit their stride. A couple of examples may clarify:
• An interviewer who was using the Grid for career counseling reported that she had difficulties getting ‘personal’ constructs (her words) from her clients, using careers as elements. The constructs were mostly what are called ‘propositional’ – that is, constructs that described objective properties of the elements, such as regular hours – irregular hours, large firm – small firm – and she wanted her clients to express their ‘feelings’ as a first step in counseling. The suggestion was that she ask the clients to think about the jobs ‘in terms of the skills they’d need, in terms of how I feel about them, in terms of what it would be like to work there ….’ and presto! she started to get ‘personal’ constructs.
• One application of Grid is for mentoring newly-qualified teachers. The elements used are the children in the class. The first two qualifiers are ‘In terms of how they behave in class,’ and ‘In terms of their home and family circumstances.’ However, the third qualifier is ‘In terms of how I behave toward them,’ that radically changes the person’s viewpoint – but in a way which will enable a discussion of how the worker’s behavior influences the children’s, and vice versa.
So the qualifiers can be really useful at the beginning of the session, to steer the interviewee in the general direction of the purpose; but after a while you might not need them if the interview has become self-managing. It’s very rarely the case that you need to record which construct goes with which qualifier.
The point of this note on qualifiers is to say three things:
• you need them at the start;
• as long as they’re in the right area, don’t agonize about the wording, and
• if all goes well, you’ll forget them as the interview progresses.
Back now to the session on American Presidents. The purpose, elements, and qualifiers suggested themselves easily. But – how are you going to analyze it? Given your assigned task of discussing the factors that characterize the most effective Presidents of the 20th century, the essence of your quest could be summarized as: ‘List those constructs that are most closely associated with the construct effective – less effective.’ Of course, you will also want to group the Presidents themselves in terms of their effectiveness. There are other refinements which you could apply, but the basic question is to explore the constructs associated with effectiveness.
This means that you might want to put your matrix of elements rated on constructs through a statistical analysis. Please again note: a statistical analysis is not necessary to analyze the data. The information is included only for those who choose to use it.
Statistical analyses of Grid data fall into two camps: multivariate analysis and dendritic analysis. Each approach has its adherents. A multivariate analysis manipulates the matrix of elements and constructs so that the relationships between them can be plotted on a two- or three-dimensional chart. A dendritic analysis calculates which two elements are most closely correlated, places them next to each other in the matrix and makes them into a ‘virtual’ element, and goes on doing that until all the elements are shown grouped in ‘families’ according to their degree of correlation; and then it does the same thing for the constructs. We are committed to the dendritic analysis approach because it does not lose any of the information – as is inevitable with a multivariate analysis – and because the presentation of the analysis makes it easy to go on growing the Grid by looking for the areas where more clarity is needed.
The detailed exposition of analyses can be found elsewhere. What’s important in this session on American Presidents is that before you start the process, you know what you want the analysis to tell you – in this case, which constructs are correlated with effectiveness, and how many types of President do we seem to have – and make sure that your method will support these questions. It is simply no good waiting until later and asking ‘How do I analyze this Grid?’ because the answer you’ll get is ‘What do you want the analysis to tell you?’
In the family domain, you may want to know how the child’s constructs are related to child management or to child development; e.g., either who helps manage the child most effectively and what is effectiveness, or who the child’s turns to for help and support.
The Administration Phase
The Nature of the Interview
The example using American Presidents may have looked obvious, but with a little practice and feedback you should find it easy also. The rest of this section is not concerned with the details, but with the overall planning of a Grid study.
There are two questions that you should ask yourself before planning a Grid session. The first is: where does the project fall on the construct extractive – reflective, and the second is: where do we expect to find the information that makes a difference?
Some Grid interviews are designed for an assessor to entice information out of the client: market research, for example, some approaches to developing management competencies, etc. These are ‘extractive’ interviews. In other interviews, and these are more pertinent to our major purposes. We are concerned with counseling, conflict resolution, helping someone draw up a person-specification – you are there to serve the purposes of someone who has asked for your help. These are ‘reflective’ interviews.
Knowing where you stand on this scale is helpful. For example, if you were planning a project which would involve ‘extractive’ interviews with a large number of interviewees, you might invoke the 80/20 rule and put a time limit on each interview; whereas when you’re doing someone a service you are there until they no longer need you, but you may be committed to a series of visits with time for reflection in between. Again, in an ‘extractive’ interview you have rather more license to get the participant to keep to the point (politely, of course, and first having listened for whether what’s said is useful); in a ‘reflective’ interview you are more likely to follow the client’s lead.
Both kinds of interview – in fact all kinds of interview – demand feedback, of course.
Where to find the gems
Knowing or guessing the answer to this question is a great help when deciding what form of analysis you should consider. By ‘the gems’ we mean those parts of the interview where you expect to discover the information that is relevant to your purpose. These could range from a simple frequency count of the number of elements, or constructs; content analysis of elements or constructs; analyzing how just a few elements or constructs are used (for example, in the session about American Presidents your first priority is to examine the constructs which are closely correlated with the construct effective – less effective, though you may go on to examine others); looking at the whole picture; seeing what happens when you introduce a new element or construct, or delete an existing one; and so on. Knowing where and how to look for the useful information will guide your choice of analysis.
Some final points to bear in mind
• Pilot your interview design (including analysis) before committing to it.
• Bear in mind that there may be two or three different configurations that will achieve the same purpose.
• The most common mistake that new Grid interviewers make is to have the elements too abstract. You will hardly ever go wrong by making your elements more concrete.
• Remember that the Grid lets you see the other person’s world as s/he has learned to understand it. Even if the client has asked for counseling (i.e. asked your help in reframing their world), much of it has worked for them until now. Don’t judge it – listen to it, understand it, be a skilled mirror before anything else.
The Design Phase
This is an absolutely fundamental skill for Grid practitioners: get it wrong, and the rest of the process will never right itself.
Elements Must be Concrete and Discrete
The important thing to remember is that elements must be concrete and discrete. Also, your element set should be homogeneous – that is, each element should carry the same ‘weight’, have the same right to be in the element set. However, if you have a nonhomogeneous element set you will soon notice it when the interview starts and you can correct yourself. Break the ‘concrete and discrete’ rule, though, and you won’t be able to recover.
Don’t Use One End of a Construct as an Element
For example, someone recently asked for help with a project intended to clarify the characteristics of effective managers as the interviewee saw them. The element set s/he was proposing to use was: MOTIVATION, STRESS, COMMUNICATION, LEADERSHIP, RISK, TEAMWORK, PROBLEM-SOLVING, and AMBITION. S/he was not getting anywhere with this set, as you can see for yourself if you try asking the question ‘What have MOTIVATION and STRESS in common which makes them different from AMBITION?’ It was pointed out that s/he had fallen into the trap of using as elements concepts which were really one pole of a set of constructs: low motivation – high motivation, prone to stress – not subject to stress, etc. What he needed to do was to give his interviewee a set of real managers as elements, and then see whether these constructs emerged in the Grid conversation and how the interviewee used them. With the elements he had originally proposed, the interview would have meandered into meaninglessness.
A good analogy is to think about how you might write the analytical essay about American presidents. You would probably have some ideas about the dimensions you would want to uncover or explore, but once you started to be systematic in your background research you would have no choice other than to list a sample of presidents and then, as you read about them, make notes on the important characteristics which emerged. And when you start to write your essay, you would have to introduce individual presidents to illustrate your points. A Grid interview is exactly the same: start with concrete representations of the domain you want to explore (your elements), and then use construct elicitation to derive the important dimensions within your field of inquiry.
Elements Can’t Overlap or Contain One Another
Another analogy will illustrate the point about the elements being discrete. Suppose that your essay will be about battles. Again, you would start by drawing up a list of battles that give a good spread over your area of interest – begin with concrete examples. Suppose further that one of your battles was the D-Day landing, and another was Omaha Beach – one of the subsidiary battles within the D-Day landing. You’d find that you needed to treat Omaha Beach as a subject on its own, because although it was part of the D-Day landings what happened there was very different from what happened on some of the other invasion beaches. In Grid terms, you’ve got a non-discrete element set: OMAHA BEACH as an element is a subset of D-DAY LANDING as an element. This is a potential problem whenever you use an element set composed of events or activities. The best way to guard against it is to try to be sure that your elements don’t overlap, and perhaps to think about events or activities which are sufficiently bound in time and place that they could be captured on a video clip.
To change the subject-matter, if you were counseling someone with a difficult relationship, then QUARRELLING WITH SYBIL is not as good an element as THE LAST QUARREL WITH SYBIL, and/or THE WORST QUARREL WITH SYBIL, etc.).
To summarize: if your Grid interview seems never to get off the ground, you’re getting a lot of generalities which are difficult to refine and/or ladder, your interviewee shows signs of impatience with what feel like meaningless questions – go back and ask yourself whether your element set is truly concrete. Abstract concepts, half-poles of constructs, large-scale events, don’t work.
Methods of Selecting Elements
This section is about the three different ways you can derive an element set, with the pros and cons of each approach.
‘Offered’ Element Sets
This is the term for when you as the interviewer determine the element set in advance of the interview, with no input from the interviewee. Use it when you are certain that it is these elements, and these alone, which you want to start the interview. For example, if you were doing market research to see how people construed eight different brands of soap powder, you would use those eight brands as your elements. Or if you were doing separate Grids with all the members of a family or class about their perceptions of the other members, then the family or class members must be the elements for every interview.
The advantage of using offered elements is simply that of control: you as the interviewer determine what the Grid interview will be about. The disadvantage is that your interviewee may not be familiar with some of the elements, and so you need to check that the interviewee does recognize all of them (and perhaps keep a few spare in reserve).
Offered Category – Elected Elements
With this strategy, you would name the category into which the elements should fall, but leave it up to the interviewee to name the actual elements: for example “Think of eight brands of soap powder’, or think of the four best managers you know … and four of the least effective”, and so on. The advantage of this is that you can be certain that the elements are known to the interviewee, but you might get a slight bias toward those which are more familiar.
This strategy has you, the interviewer, prepare a list of questions to which the answers will be the elements: for example ‘Tell me the career you would most prefer … and one you would never consider … and your best friend’s career … and another which is appealing … and another which you wouldn’t like …’ etc. There are several advantages in this process: it makes sure that you have a good scatter over the domain you are exploring, you know that the elements are familiar to the interviewee, there’s a stronger feeling of ownership, and if you are doing a project which involves getting Grids from several people then the collated answers to the element questions are themselves informative. The price you pay is that this kind of element set takes longer to elicit, but in many applications of Grid it’s worth it.
Make sure you cover both sides of the boundary
Whatever strategy you use, if you are using the Grid to help define a boundary then you need to have elements from both sides of the boundary. In other words, if you are using the Grid to uncover how the interviewee perceives the characteristics of good team members, then you must have in your elements some good team members and some not-so-good, otherwise you won’t get the contrast. If you want to help someone explore occasions when they have successfully been assertive, you need in the element set some occasions which were successful and some which weren’t.
You can of course use a mixture of strategies. You might use element creation questions to begin with, and then supply some elements yourself if you want to be sure that those elements are included. In that case, it’s probably best to begin with the element creation questions because you will then know whether a given element is there as a response to a particular question.
And remember: elements should be concrete, discrete, and homogenous.
‘Ideal’ elements are a very good way to explore scenarios, ask ‘what if’ questions, and be more specific in your inquiry. An ‘ideal’ element is not a real one, but one which you and the interviewee invent and then put into the Grid part-way through.
Example: Relationships at School
For example, in a Grid about relationships at school, you might get a sense that your interviewee has a strained relationship with his/her teacher. So you might suggest that you create one or two new imaginary elements and rate them: for example, MYSELF AS MY TEACHER WOULD LIKE ME TO BE, or MY TEACHER AS I WOULD PREFER HIM/HER TO BE. Together you then look at the differences between the first of these elements and MYSELF, the differences between the second of these elements and MY TEACHER, and maybe some other comparisons which the process suggests. You can then get into a discussion about the size and nature of the problem, who owns the problem, what would have to happen to change it, etc.
• in a market research Grid, you could introduce MY IDEAL CAR;
• setting up a person-specification, THE IDEAL PERFORMER or THE IDEAL FATHER;
• examining stressful situations, THE MOST STRESSFUL SITUATION I CAN IMAGINE and THE LEAST STRESSFUL SITUATION I CAN IMAGINE.
The idea is to name an imaginary element which gives expression to the question you want to explore when contrasted with the real elements. They can be very instructive in getting to the core of an issue quickly. The difference between MYSELF and MYSELF AS MY MOTHER WOULD LIKE ME TO BE might need to be taken in bite-size chunks.
Introducing the ‘Ideal’ Element
For most Grid purposes and practitioners, it is best to introduce the ‘ideal’ element part-way through the session rather than at the beginning. Get a good number of constructs out through consideration of real elements (people, cars, stressful situations, etc.) and then introduce the ‘ideal’. There are two connected reasons for this:
• if you include your ‘ideal’ element from the beginning, you have a non-homogeneous element set and it could be difficult to work with; and
• the ‘ideal’ element may not be clearly defined in the client’s mind at the start of the process. Using real elements to give you a rich selection of constructs from which the client can come to their own definition of the ideal.
Where’s the difference that makes a difference?
One of the most important considerations when you are planning and later analyzing a Grid interview is where in the process you are going to find your most useful and insightful information – the difference that makes a difference (Bateson). Quite often, this is not in the final Grid analysis; it may happen much earlier. Some examples:
The language people use for their constructs
For some purposes, all you need to know is the pattern of language people use when describing the elements, and a simple content analysis is all the analysis you need. The classic example of this is using construct elicitation to get people to describe their colleagues at work, as a preliminary to a study of competencies or an organizational change intervention. It can be quite enough to know that (for example) 30% of their constructs have to do with knowing the right way to communicate with the Main Office, or that 40% have to do with conflict management.
The responses to an unrehearsed element set
If you do a Grid with someone where the elements are ‘situations in my life where I learned something’, or ‘times when I tried to be assertive,’ it’s unlikely that the responses will come tripping off the client’s tongue. You’re asking questions that most people may not formally have asked themselves before. You and they can learn a lot through the process of eliciting them.
Analyzing how just one or two elements, or constructs, are used
You may be doing a counseling interview where the elements are key relationships. Obviously you use a wide range of relationships in order to elicit the full domain of constructs, but you may find that the bulk of your work can be done by simply comparing MYSELF and MYSELF AS I WOULD LIKE TO BE, or MYSELF AS MY MOTHER [father, teacher, etc.] WOULD LIKE ME TO BE … depending on where the agenda of the interview seems to lie.
Exploring a Grid about experience of service in stores and other businesses, you may find that most of the work can be done by looking at the constructs related to one such as ‘made me decide never to shop there again – not as important as that,’ and looking at the elements that rate highly on the first pole of the construct.
If you are aware of all these choices – and the others that are open to you – then you can plan how to do the interview, how much time to use, how much technology you will need, etc. And always remember that you can do a simple ‘once over lightly’ with a simple protocol that will guide you toward deciding how to plan a larger investigation
Note: It is important to see the Repertory Grid as a conversation; the importance of feedback; and not collecting masses of data which you then don’t know how to analyze. So it might be useful to share some very simple, low-tech, cost-effective applications of the Grid, and trust that you can extract the principles so that you can adapt the example if it’s appropriate.
At least half the serious problems people experience with Repertory Grid are due to failure to include the method of analysis into the projected plan. Regrettably, many of these problems can’t be fixed, and the phrase ‘If I were you I wouldn’t start from here’ applies. Take note of the following:
• There are many different and valid ways of analyzing Grid data. Some methods need a computer; others don’t. If you ask ‘Is there a computer analysis program for Repertory Grid?’ the answer will be ‘Yes, several; they do different things; and the choice is dictated by your purpose, the type of data you have, and the questions you want the data to answer.’
• For almost any work that uses the Grid, it is possible to think of at least two or three protocols for designing the session. There could therefore be two or three different ways of analyzing your data, provided that you incorporate your chosen form of analysis into the design of the session.
• No method of analysis exempts you from looking at the results and deciding what they mean. If you keep a spreadsheet of your family finances, it will show you where the money goes; but it is up to you to decide what this means in terms of turning off the lights, buying a new car, and thinking of the next holiday. Analysis of Grid is exactly the same.
• Don’t think of ‘analysis’ as a singular activity – there are many purposes – especially but not exclusively in the ‘reflective’ interviews – where you and the interviewee should stop, do an intermediate analysis, and then move on to the next stage. You will never get a complete picture of the client’s cognitive map on the first sweep – people are simply not that superficial. For some ‘extractive’ purposes you can compensate for this by taking a sample and relying on the 80/20 rule, but the closer you come to helping an individual person reflect, the more you need to be aware that analysis is part of the process, not its endpoint.
• The Repertory Grid technique is a method for structuring a conversation. It is not a rush to complete a matrix which you enter into a computer program. For many purposes, especially the ‘reflective’ uses of Grid, the journey matters more than the arrival: meaning that the insights garnered in the course of eliciting elements, constructs, laddering, rating, and looking at the analysis so far may be much more useful than the final presentation of the data by your chosen methodology.
Bearing in mind that there is advice on analysis elsewhere, we can explain the basics of the different analysis methods without going into great detail. Analysis can be greatly enhanced with he aid of computer software packages and web applications defined specifically for that purpose . Here is a general overview of the options:
• Frequency count, usually of the constructs but sometimes of the elements as well. The rationale for doing a frequency count is that people have more constructs about topics of which they have more experience. In a simple example of using frequency counts before and after a training course on psychological tests – before the training trainees were asked to name as many tests as s/he could think of, and give as many constructs as possible; and then repeated the Grid afterwards, which showed that s/he knew more tests and more ways to differentiate between them. A frequency count is a very rough guide; you can rely on mega-trends only, not small differences; you must be sure that your interviewee has had every chance to give as many as possible; and it’s best used when making a before-and-after comparison of the same person, rather than comparisons between people.
• Content analysis, which may be combined with frequency counts. Like any other form of content analysis, you look at the data (usually the constructs), see what themes suggest themselves, and sort into those themes. Again, the presumption is that people have more constructs about issues they know well; so your analysis is likely to focus on the relative proportions of different themes – such as the preponderance of constructs about conflict management in the way managers described one another in the example above. Whatever your subject-matter, there isn’t likely to be a benchmark against which you can compare results – for example, it would be difficult to say what the ideal proportion of constructs about conflict management should be – but a combination of common sense and input from the client is usually enough to get you started. For example, if you were interviewing a client in order to help him understand why many of his relationships were unsuccessful, and there were a lot of constructs about trust, and when you laddered up it seemed to be a core construct, you might ask the question: ‘Can you see any major themes in your constructs so far?’ and if he didn’t spot it for himself you might then offer a comment like: ‘It looks as if you have a lot of constructs about trust – does this seem important to you?’
The most difficult aspect of content analysis is seeing what is not there – for example, my seeing that there were no constructs about innovation in my conflict-ridden client. Experience is the best teacher here, combined with your general knowledge of the client’s circumstances. Content analysis is very useful when you are comparing the constructs produced by two or three groups of people about the same subject, especially because you don’t need any external benchmarking to draw conclusions – for example, in a study in the Public Service in which senior managers, executives, and control agencies all contributed constructs about effectiveness at senior management level – one outstanding finding was that about half the managers’ constructs had to do with managing their departments, but this didn’t figure at all in the way executives construed them – which was very interesting when viewed in the light of the performance contracts between managers and executives.
• Examination of just one or two elements, without statistical analysis. For example, if you were doing a career counseling interview, using different careers as elements, you could then ask the client to think of the ideal job and rate it on all the constructs. What you’re doing here is using the real life elements as the means for generating constructs about careers – so you have grounded the interview in your client’s real experience – and then used this information to generate a profile of the ideal job. Obviously, it’s useful to put the constructs into some kind of priority order as well.
• Statistical analysis using multivariate analysis. There are many different analysis packages available, commercially and non-commercially. They are all based on analyzing the matrix you get from rating the elements on the constructs, by searching for the smallest number of independent variables which could account for the relationships in the matrix. Most programs will then present this information visually, which restricts them to using two (three at most) variables; the variables appear as the x and y axis, with a z axis if three are used. The position of each element (and sometimes the constructs) are plotted on the visual diagram, so that ones which are similar are close to one another, and so on. At some point the axes have to be named (ideally by the interviewee, as part of the feedback process) and then, depending on your purpose, you look at what the visual plot tells you – for example, if your Grid is about relationships at work and there is a great distance between the elements MYSELF and MY BOSS you would want to explore this: does it matter? How do you feel about it? How does your boss feel about it? Does anything need to change, and if so, what? If this discussion results in an action plan you may then do a new Grid interview and look at what has happened, if anything, to the distances.
• Statistical analysis using dendritic analysis. In this type of analysis, the calculations are made by first looking at the elements to find which two are most closely correlated. So if you have ten elements in your Grid and numbers 2 and 8 are most closely correlated, the program will re-sort the visual matrix so that it places them next to each other, and will make a ‘virtual’ element number 11. It then drops 2 and 8 from the analysis, replaces them will number 11, and looks for the next two closest combinations, and re-sorts the Grid again until all the correlations have been calculated. Above the matrix, with the elements on the top row of the Grid, it draws a set of ‘trees’ which show the strength of the correlations. When you look at a dendritic analysis you usually see the elements grouped in ‘families’ of closely-correlated elements. The analysis will then do the same process for the constructs, putting together those which are most closely correlated (and taking into account that some constructs need to be reversed). The interpretation of the results is based on the axiom that elements (and constructs) which are very closely correlated have very similar meanings, and so the first stage is known as differentiation – you look at the elements which are closely correlated and ask whether that degree of correlation actually represents the truth, as the interviewee sees it. For example, if you were doing a Grid about characters in Shakespeare, and the first dendritic analysis showed a 98% correlation between LEAR and HAMLET, the question is: ‘Are those characters as similar as they seem?’ If the answer’s Yes, you go on to look at the next correlation, but if the answer’s No the program will then ask you for a new construct on which LEAR rates at one end and HAMLET at the other. You then rate all the elements on the new construct, and the Grid is recalculated. Going through the differentiation process for the constructs is slightly more complex, because you have three choices – to combine the two constructs into one, to offer a new element which will be rated at one extreme on one construct and the other on the second, or to treat the correlation as an important insight which you want to leave in place. For example, if the interviewee gave the constructs tragic character – comic character and make great demands on the actor – easier for the actor and they were correlated at the 95% level, the question posed would be: ‘Almost always you describe tragic characters as making great demands on the actor, and comic characters as making fewer demands – is this a true representation of how you see things?’ If the answer is No, then the next question is: ‘in that case, can you think of a tragic character which makes fewer demands on the actor? Or a comic part which makes great demands on the actor?’ Maybe the interviewee can think of an example or two, but s/he may decide to treat this information as an important insight to leave in place for further thought. This process is a very effective way of highlighting and challenging the interviewee’s stereotypes and prejudices. Dendritic analysis is a dynamic process, in which the first calculation serves as a starting-point for building and testing the interviewee’s perception of the subject until s/he is satisfied that it is clear and complete.
The difference between these two approaches to statistical analysis can be summarized as: Multivariate analysis condenses the information in the Grid, and loses some of the detail in the process, whereas dendritic analysis expands the Grid and loses none of the detail. If you really want to go into as much detail as possible then (i) dendritic analysis is the only choice, and (ii) the differentiation process gets people ‘hooked’ and you can leave the session with them to carry on alone. However, let us leave this session as we began, by reiterating two of the Golden Rules:
• Build your method of analysis into your project plan. Pilot it so that you can be sure that it will tell you what you want. And remember that you will still have to evaluate what the analysis tells you.
• The Grid is a structured conversation, of which the matrix and its analysis is only a part. The journey may matter more than the arrival. The map is not the territory.
The Repertory Grid Interview
The learning curve for being an effective Grid interviewer is very steep, provided that you practice your first few interviews in a safe place, with a tolerant friend, on noncontroversial topics. If you have access to an experienced interviewer, so that you can ‘shadow’ each other for the first two or three times you go live, so much the better.
In your first few interviews, you’re learning several different skills at the same time – that’s one reason why it can feel confusing. You’re learning:
• how to phrase and, where appropriate, rephrase the two against one question;
• how to manage the presentation of the elements, which means keeping control of the cards they’re written on;
• how to manage recording the constructs, which also means keeping control of the cards they’re written on;
• how to monitor the process – how is the interviewee reacting, are you getting information which is relevant to the purpose, etc.; and
• how to control the use of time, if time is a constraint.
One consistent feature is that people learn these diverse activities at different rates – nothing to do with their intelligence or speed of comprehension, it’s just one of those things. So if you decide to teach yourself Grid in a group, which is a good idea because you can observe and get feedback, please don’t be surprised if you get different rates of acceleration and pause. In this guide, we assume that you have access to at least one friend who’ll let you practice, and if you have a third who can observe then so much the better.
Start by assembling some small index cards (which you’ll use for recording the elements) and larger cards (which you’ll use for recording the constructs). You’re going to start with some easy elements, so ask your interviewee to name six well known public figures, or television programs, or models of car – something you know they are familiar with. Don’t bother with a purpose statement, but do think of a couple of qualifiers, such as ‘…. in terms of how I feel about them, ….. in terms of their appeal…..’ Number the element cards.
Now practice asking the ‘magic question’ in two of three different ways. The basic question is ‘Tell me something that any two of these have in common that makes them different from the third,’ but it’s worth having some alternative phrases, such as ‘In what way are any two of these similar but different from the third,’ or ‘Can you put any two of them together and the third one is different?’ As you’re asking the question, lay down three element cards at random; not in a straight line, but at random and shuffle them around on the flat surface.
Note: write the names of the elements in big writing so they occupy most of the card, because some people will try to write the two poles of the construct on the element cards. Also, it’s really important to present the triad of elements so that the interviewer can see two against one – whether it’s on the table or into a computer. Giving people a printed list of all the elements and asking them to concentrate on just three is very difficult.
It’s quite likely that the first two or three constructs will be ‘propositional’ – that is, objective properties of the elements, like male – female, young – old, entertainment – documentary, sports car – family car. Don’t worry about this for the moment, because your job is to write down the construct. Use the bigger index cards, one card per construct. It’s a good idea to write the construct about two-thirds from the bottom, leaving space to record any laddering up and down you may want to do later. A useful discipline is to write the characteristic of the pair on the left, and the singleton on the right, and to note the numbers of the elements in the pair and the singleton. If the interviewee names one end on the pole only, don’t offer the answer: ask ‘How would you describe the other(s) by contrast?’
Note the exact wording of that phrase – ‘the other(s) by contrast’. Don’t use the word ‘opposite’ because the other pole may not be a dictionary opposite. Tagging your question to the other element(s) helps the other pole be distinct. Best practice is that each end of the pole should carry equal ‘weight’; so if the first pole is ‘shows leadership’ you need something other than ‘doesn’t show leadership’ – the other pole might be ‘easily led,’ or ‘sabotages others’ leadership,’ or ‘stopped being a leader,’ or a whole variety of other poles, depending on how the interviewee sees the elements.
After you’ve practiced writing down two or three constructs, if the first constructs were propositional try introducing one of the qualifiers, so that you ask ‘Again, can you tell me a way or ways in which any two of these are similar to each other and different from the third, in terms of the way you feel about them?’ That should shift the emphasis. Also, the constructs are likely to be longer and more personal, and this will give you a sense of how long it takes for the interviewee to think, and for you to write.
Note: people vary greatly in the length of time it takes them to think of a construct. It depends on their knowledge of the topic, how they feel about the process, etc. But at some point you will learn to present the next triad while you’re finishing writing the existing one. Some other hints:
• Don’t be afraid of the silence – it means that the interviewee is thinking;
• try to write down everything that’s said, and not condense or summarize the construct – but if you absolutely have to, then ask the interviewee to summarize it, don’t do it yourself;
• if you simply must put in some words of your own, to explain or summarize, find a convention for remembering that they’re yours – for example – put yours in square brackets;
• some interviewees will give you a long list of single poles, if they can see a number of constructs – if this happens, write each pole on a fresh card and then go back and ask for the contrast pole.
Managing the Cards and the Recording
By this stage you should be reasonably comfortable with presenting triads, asking the question, and recording. You’ve probably learned that you need to give yourself plenty of physical space – a good horizontal surface at the appropriate height for writing. It is also a good idea if you can sit so that the interviewee can see what you’re doing – it symbolizes the fact that this is a joint exploration. The next stage is to learn how to impose some order on the process, especially managing the cards.
A useful hint: if there are no other pressing reasons dictating the number of elements you have, then use nine. Then you can write a 3 x 3 matrix –
– and use this to order your triads. So you’ll have 123, then 456, then 789, then 147, 258, 369, and if you need more you can go diagonally 159, 267, 348, and so on. This makes it much easier to control the cards, because you’re not searching for the next triad, and it also has the advantage of giving you every element in the company of every other element in the shortest time. So, if you think you need it, give yourself a practice with nine elements.
By this time you probably won’t feel completely competent in managing the cards and the recording, but you ought to feel as if it’s within your grasp. The test is whether you find that it’s becoming easier to listen to the actual constructs, rather than just being a recorder. So it’s time to try some laddering.
The purpose of laddering is to learn more about what the constructs mean to the interviewee. Personal construct theory refers to people having a ‘hierarchy’ of constructs, with a few ‘core constructs’ which represent their own core values at the top, and peripheral constructs at the bottom. (This is much oversimplified. The most important thing to remember is that core constructs must be handled with care).
Start by Laddering Up. There are two ways of doing this. Suppose that you’re doing a counseling interview in which the elements are key people in the interviewee’s childhood and teenage years, and you’ve been given the construct took an active interest in my education – paid no attention to my education. One way of laddering up is to present that construct and ask which pole the interviewee prefers, in terms of the purpose: so your question would be something like ‘Which kind of people did you prefer – those who took an active interest in your education or those who paid it no attention?’ You might get an answer like ‘Those who took an interest.’ Then you go on to ask ‘Why is that important to you?’ Suppose the answer is ‘Because there were subjects I would have liked to have studied and I would have been good at.’ Then you ask ‘And why is that important to you?’ and you might get an answer like ‘Because I’ve always felt at a disadvantage compared with my brothers.’
Feel how the repeated ‘Why?’ questions are going deeper, getting closer to the heart of the interviewee. Before we discuss laddering up in more depth, let’s look at the other way of asking the question. The other way is to present the construct and say ‘You drew a distinction between people who took an active interest in your education and those who paid no attention to it. Is that an important distinction between people in your childhood?’ and if the answer’s Yes, you ask why, and go on asking why just as in the previous example.
Note: because core constructs are, by definition, very important to the interviewee, don’t treat them lightly. We recommend that for most purposes – certainly for ‘extractive’ Grid interviews – you shouldn’t go through more than three levels of asking Why, and you shouldn’t ladder up any more than you need to know. You can usually tell when you’re getting close to someone’s core constructs because they become iterative – they’ll use a phrase which indicates that that’s the way it is for them and they can’t explain further. The body language may give them away also – people tend to use ‘going round in a circle’ type gestures, or – if you’ve been insensitive– gestures inviting you to stop.
Further note, on recording and process management: the reason for using a large card for each construct and writing the construct about two-thirds of the way up is that you can write the answers to the laddering up questions above the construct – you can draw a small arrow pointing upwards to show that’s what you have done. Note that the answers to the ‘Why’ questions won’t usually be phrased as a bipolar construct, just a comment; for nearly all purposes this is enough. Also, if you get an answer like the one cited above – because there were subjects I would have liked to have studied and I would have been good at – you can ask what the subjects were; learning to follow up clues like this is one of the signs that you’re becoming a skilled interviewer. And you don’t have to start the laddering process with the first construct produced – look for one which you think is likely to be easy and interesting for the interviewee.
Practice both ways of laddering up. The choice between which method to use should be dictated by the need to ask a sensible question; for example, if you get a great many constructs where there is obviously a positive pole and a negative pole you could feel stupid asking a series of questions about ‘which do you prefer as companions, people who have appalling table manners or people who are well-mannered …. people who have body odor or people who wash …. people who kick your dog or people who don’t?’ In this case, it’s much better to ask whether that’s an important distinction and why.
Also, note that when you’re laddering (up or down) you’re laddering the construct, not talking about the specific elements which generated that construct. Observe that the examples given above quote ‘people who are X or people who are Y’, and you may need to clarify this by saying ‘In general …….’
Finally, if time is an issue or there are other good reasons, you could give all the construct cards back to the interviewee and ask them to sort them into high, medium, and low priority, and then just ladder the high priority ones. And once the interviewee’s seen the point of laddering and given you several answers, you can show the remaining cards and ask ‘Are we going to get anything new out of this one?’
Now try laddering down. The objective of laddering down is to ‘unpick’ the constructs into their component parts, to get information about how the interviewee defines them in practice. The standard laddering down question is ‘Can you tell me more about how elements that are (one pole) differ from elements that are (other pole)?’ So, using our example of took an active interest in my education – paid no attention to my education, you would ask ‘Can you tell me more about what the people who were actively interested in your education actually did, in practice, which made them different from the ones who paid it no attention?’ A question like this will generate one or more constructs which should describe behavior: for example you might get Went to the parents’ evenings – didn’t go to the parents’ evenings, helped me with my homework – never helped me with homework, encouraged me to use the library – said that the library was a waste of time, and so on. What these new constructs do is give observable, behavioral examples of the original construct in practice. Constructs obtained from laddering down should have both poles defined. Again, observe that the laddering down question refers to the constructs, not to the elements which gave it. Record the new constructs underneath the primary constructs – use a little down-pointing arrow to show that this is what you have done.
Note: when you ask the laddering down question, some people will go ‘across’ – that is, they’ll give you both poles of the construct – and some will go ‘down’, giving you lots of discriptors of one pole. If this happens, go with the flow and then take each pole and ask how they would describe the other pole by contrast
Time to Take Stock
Time to pause and take stock. You should have given yourself enough practice that you feel comfortable managing the cards, you’re able to listen to the content, you can find a new way of phrasing the questions if you get stuck. You should also have learned to discipline yourself not to suggest any of the content, and not to be afraid of the silence because almost always it means that the interviewee’s thinking. Try two or three different people to practice with, and different subjects, so you get some idea of how interviewees differ, and ask your practice interviewees for feedback.
If you can arrange it, try to get someone to accompany you while you practice interviewing. In the best of all worlds, you would go out with an experienced Grid interviewer who conducted the interview while you ‘shadowed’ them – that is, wrote your own record cards so that you could compare, and then you could swap places.
Frequently Asked Questions
Before we move onto the next stage of Grid, how about some Frequently Asked Questions:
Q1. How long does it usually take to get to this stage?
Answer: it depends on the topic, and on whether it’s an extractive or reflective Grid. If you use element creation questions to get your element set, that will take some time. If it’s an element set which the interviewee will find easy (for example, the team s/he works with) it’ll be quick, but if it’s an unrehearsed element set (for example, times in their life when they’ve tried to be assertive, which are likely to be harder to retrieve) then it will take longer but the time will have been well spent. For eliciting the constructs and laddering, in an ‘extractive’ Grid you should allow about an hour as a rule of thumb. If interviewing a sample of managers in a study of corporate culture, you will probably use colleagues as elements and you can ask each manager for a ninety-minute appointment. You may not get a complete ‘download’ from each manager, but aggregating the results over the sample will compensate. For a reflective Grid, where you’re there to perform a service such as counseling, the answer is ‘it takes as long as it takes’ – but you may in any case want to build in time for reflection.
Q2. What order do I do things in – when do I start laddering, for example?
Answer: it depends on the flow of the interview. Your primary aim should be to make the interviewee comfortable with the process as quickly as possible – some people slip into the two-against-one comparison easily, others take more time. You’ll probably get some propositional constructs first; write them down before you start to emphasize the qualifier. (By the way, propositional constructs can be very interesting when laddered, so don’t discard them). Remember also that the corollary of the Grid’s being free from observer bias is that it’s impossible to fake one, and if your interviewee isn’t comfortable they’ll send up distress signals. So, at the start, go with the flow where it’s flowing. Moving on to laddering can be a good way of helping the insights flow again if the interviewee’s dried up. You can move between construct elicitation to laddering, back to some more constructs, do a spot more laddering. If you’ve succeeded in creating an atmosphere of the interview being a joint exploration through structured conversation, the interviewee may give you some signals which you can follow.
Q3. How many constructs ought I to expect?
Answer: it depends on the subject, on your purpose, and your analysis method. A very rough guide is that when someone’s talking about a topic that’s familiar but not heavy with emotional investment or technical complexity – for example, managers giving constructs about their colleagues as quoted above – One might expect somewhere between twenty and forty primary constructs. The number of constructs you should be content with – a slightly different question – depends on whether your purpose will permit you to use the 80/20 rule: ‘let’s see how many we get in a couple of sweeps over the territory’. Where it really matters is if your analysis is going to include a frequency count of the number of constructs, and/or a content analysis which includes a frequency count. The basis of an analysis of frequency count is the reasonable assumption that people won’t have many constructs about topics where they have little experience: for example, the average person can probably muster no more than four constructs about the sea, a sailor probably has a very sophisticated construct system. If the sailor were to try to teach me to ‘read’ the sea as s/he does, then a simple robust evaluation of the success of this teaching would be the increase in my constructs. To take another example, if in the corporate culture survey interviewing managers about their colleagues 60% of their constructs were about conflict and there were none about innovation, you will need to be pretty sure of the material [have a good number of constructs] before you confront the management committee with this interesting result. To summarize this point: if you are going to draw significant conclusions from the number of constructs, you must give yourself a good sample and/or the opportunity for each interviewee to dig really deep.
Q4: Do you have any guidance on sample size?
Answer: it depends on the purpose. If you are sampling a homogenous group of people about a subject with which they are familiar but which doesn’t carry a great deal of emotional investment, then generally you stop getting new information between 15 and 20 interviews. So if you’re doing a research project in a company with 32 managers, political prudence says you interview them all so as not to make anyone feel excluded. Otherwise, you may have to work with what you are given if the numbers are smaller, or think about turning your constructs into a questionnaire or survey if the numbers are much greater.
Q5: Does this guide start every answer with ‘It depends?’
Answer: Yes. Because once you’ve mastered Grid and learned to do a ‘pure’ interview – that is, a conversation in which you provide the structure and do the listening, and the interviewee provides the content – then you’ll be able to work out the answer to almost every question yourself. People often ask ‘Can the interviewer suggest an element? …..a construct?’ Answer Yes if you know that’s what you’ve done and why. Learning to do a ‘pure’ interview will train you to recognize when you’ve intervened, and that’s what’s important. Sometimes it can be useful for the interviewer to suggest that we add another element, especially what’s called an ‘ideal’ element – MYSELF AS MY FATHER WANTS ME TO BE, or MY FATHER AS I WOULD HAVE LIKED HIM TO BE. Or if you’ve interviewed a Frequent Flyer on his or her experiences with airline service, you will probably want a construct like made me decide never to travel with them again – not as important as that, and if the interviewee hasn’t offered it then you can. Almost anything is permissible, as long as you know what you’ve done and have a good reason for it.
The flow is broken at this point for two reasons. One is that after you’ve collected (and probably laddered) a number of constructs, you might want to engage the interviewee in some feedback. What you’ve done by this point is establish most of the dimensions on which the interviewee construes their world, and it could be useful just to talk about these dimensions before going on to rate each element on each dimension.
However, if your purpose is to reveal the interviewee’s constructs as they are used, you need the next stage, which is to turn each construct into a scale, and rate each element on each construct. This will give you a matrix which you can analyze and perhaps develop further.
Explain to the interviewee that the next stage is to see how each of the elements is rated on each of the constructs – for example, ‘You’ve told me about the people who were important in your childhood, and you’ve given me these constructs – wouldn’t it be useful to see how each person rates on each construct?’
You may be using a computer program which presents each element for rating on each construct, which simplifies things mightily. Or you may have to draw up a Grid with one column for each element and the poles of the constructs at either end of each row, with one pole labeled 1 and the other pole labeled with your chosen number. In either case, the question to the interviewee is relatively straightforward; but there are some administrative issues which could arise, thus:
• How many points on the scale? Most people seem content with five.
• Positional Response Bias. This is an issue you need to stay alert to – even experienced facilitators can get caught in it. It’s very important to realize that constructs in general don’t necessarily have a ‘good’ and ‘bad’ pole; they may simply denote differences, such as when someone distinguishes their acquaintances as likes to talk things through – likes to work things out alone but doesn’t have a preference. So when you start rating the elements on the constructs, the extreme left will be 1 and the extreme right will be 5, and if there is a preferred pole for a given construct it’s as likely to be on the right as on the left. However, people can slip into mistakenly always giving the preferred pole a 5 (or a 1 – depends on their experience of rating scales!) and so you need constant vigilance to make sure that the interviewee is working from the actual pole descriptions.
• Range of Convenience of Constructs. Not every element in the world can be rated on every construct in the world: try rating FALSE TEETH on the construct religious-atheist. So there will be times when you – taking your cue from the interviewee – want to drop a construct out, or rewrite it. For example, in a Grid about careers you might have been given the construct working with women – working with men, which fitted that triad, but when you try applying it more widely it works better as two constructs: working with women – working with both sexes, and working with men – working with both sexes.
Lets take a look at what you might have collected. Let us suppose that the domain that you chose to explore was family and your qualification is relationships. You have selected 1) the child [the subject of this exploration], 2) the mother, 3) the father 4) the teacher, 5) the clinician [you], 6) the child as the mother would want me to be, and 7) the child as my father would like me to be as the elements listed on the chart. You have placed each pole of the constructs solicited on the chart and you ask the child to rate each element by each construct. You might get a rating something like this.
Rating #1 1 2 3 4 5 6 7 Rating #7
listens to me talks all the time
thinks I am okay thinks I am stupid
What might this tell you? Even with only three sample constructs we know that the child generally thinks of people as either nice or mean. Nice people listen to the child and mean people don’t. We would have to ‘ladder’ up to discover what the child means by okay. Teacher and dad are defined negatively which may have to do with each demanding a performance. Mother is nice and listens – but what is her relationship with father. Hopefully, you get the idea. While a statistical analysis might be nice, a minimum of fifteen constructs across seven to nine elements should give you a picture of your client at least within the selected domain.
A Few Final Points
That is about all there is to say about getting the actual Grid done. This section has deliberately concentrated on how to learn the skills, manage the administration, and prepare yourself for the first few real interviews, and avoided talking about content. So, just a few final points:
• You must practice in a safe place, otherwise you’ll get into a horrid muddle, but the learning curve is steep, especially if you have someone to give you feedback.
• Learn to do a ‘pure’ interview – that is, one where you know that you have suggested none of the content – because then you’ll be able to make sound decisions about adapting the process.
• Let the works show – the sooner your interviewee sees the point of it all, the easier it will become for both of you.
• Don’t forget to ‘sign off’ by asking if the interviewee has had any further thoughts. Sometimes the experience can cause the interviewee to restructure their insights and you’ll get some little gems.
• Most important of all – don’t forget that Grid is a structured conversation. It is much more than a matrix of figures to be analyzed.
This section makes some assumptions – that you have practiced your interviewing technique and can be sure that you are not imposing your own framework on the interviewee; that you have absorbed the twin messages of the importance of planning your analysis from the start, and that the Grid is a conversation; and that you don’t need advice on the interpersonal skills needed to be a good listener and counselor. In that case, there are three golden rules for feedback:
• give it;
• always relate it to the purpose;
• ask for it.
In good Grid feedback the principle is for the interviewer to act, as much as possible, as a skilled mirror: that is, to ask questions and give information which will encourage the interviewees to see things for themselves, rather than the interviewer offering an interpretation or judgment. This is why it’s important to be aware of when, and how, you offer your own thoughts. There may be a time when you need to; and there will certainly be a point where you have to make the connection between what the Grid tells you and the broad purpose which the Grid interview(s) is meant to address. But it’s best if you can guide the interviewee to the insight, rather than do it yourself.
It’s also important to be clear about whom you will be giving feedback to. This is where the difference between ‘extractive’ and ‘reflective’ Grids becomes important. If you’re doing a research project for a client which means that you will have to do lots of interviews but probably with a limited time for each (therefore relying somewhat on the 80/20 rule) your main feedback will be a report to your client, with appropriate suggestions for actions, and the people you have interviewed may get a summary report. But if you’re doing a counseling interview, and therefore in ‘reflective’ mode, it’s your counseling client who needs the feedback and action planning.
Discussing how to give feedback is best illustrated by a series of examples, from which you can draw your own ideas – after all, no two Grid interviews are the same. The constant which should run through all interviews is encouragement, especially at the beginning – if you follow the advice to ‘let the works show,’ you’ll be making a pile of cards with constructs written on them, or a series of entries into a computer program, and once the interviewee has understood the two-against-one principle you can refer to the buildup of constructs and make a remark like ‘Now you see how it works, the more of these you can give me (add a phrase relevant to the purpose if appropriate) the better.’
The purpose of giving feedback, besides ordinary politeness, is that it will often facilitate the interviewee to give more information, or reframe the issues. Unless you have decided to take the 80/20 rule – that is, to interview a number of people on the same topic and rely on the sample size to give you what you need – you can’t expect the interviewee to give you everything you both need in a nice neat orderly fashion, moving smoothly from elements to constructs to laddering to rating to matrix analysis and on to action planning. Often the most important insights have to be sweated for, because the Grid interview may be the first time that your interviewee has done some systematic introspection. This is why it was suggested earlier that you should be ready to move around within the process itself, for instance by moving on to laddering for a while before going back to look for more constructs.
The question which you’re likely to need most, for the interviewee and for yourself, is: ‘Can you see any patterns in here?’ It’s a good idea if you can see some, otherwise it’s a risky question, and as a general rule it’s best if the client can answer rather than your supplying it. Patterns can be obvious surprisingly early, but you must use your judgement about when to raise them.
If you have used element creation questions to get your elements – suppose that you’re doing a wide scale training needs analysis and you’ve asked for critical incidents as elements – then provided that you have stuck to the same order of questions for all your interviewees, there may be a pattern which you can detect in the answers. For example, if you’ve asked for two really tough incidents, two which the interviewee thought would be tough but weren’t, two which the interviewee thought would be easy but weren’t, a routine enjoyable event, a routine disliked event, and one more ad lib – you may have something very interesting for your client, if there are clear patterns. (HINT: it’s always a good idea to ask the client to help you analyze the data. It helps you retain the ‘no interviewer bias’ standpoint, and it builds in ownership by the client. Also they might be able to read things which you can’t, because they work or live there.
Never stop an interview simply to discuss patterns in the elements, because you could get stuck there. But it is often appropriate to stop and discuss patterns in the constructs, especially when the pattern relates to your purpose. Suppose for example that you were counseling someone who knew that they ought to get fit and take more exercise: you’d probably have an element set of ‘methods of getting fit’, like aerobics and swimming and tennis (This element set would probably have been derived by your asking the client to name as many methods of getting fit as she could think of). Suppose furthermore that you could see a theme running through the constructs to do with not wanting to make a fool of herself in public, and another theme about not wanting to let the rest of the side down. You could ask her if she could spot any major themes running through her constructs. Or, if you were using cards or could print off her list of constructs, you could ask her to arrange them into themes; or you could ask her to sort them into high, medium, and low priority. If by that point she hasn’t grasped what’s obvious to you, you could try laddering up the high priority constructs and see if these themes emerged as you got closer to core constructs. If by that point she still hasn’t seen the theme, you have two choices: to go back to your noninterventionist role as a Grid interviewer, or to say ‘Well, I can see a couple of themes – would you like me to show you?’ and sort the cards yourself and pray for the insight to occur naturally, or you could come right out with it yourself. In making this decision, your guiding skills must be your skills as a counselor – your reading of body language and tone of voice, and the other ways you have learned of knowing when to speak and when to stay silent.
One point, though – when first doing counseling Grids, you might feel frightened by the speed with which the problem seems to become obvious to you, though not necessarily for the client. Being anxious not to fall into the trap of construing other people’s construing, you might ask advice from more experienced practitioners. Their answers could be summed up as: ‘You’re probably right; this is one consequence of the lack of redundancy in the Grid process, because the interviewee can’t waffle on; but hold yourself back in order to give the interviewee time to see things, and be prepared to be wrong.’
Many ‘extractive’ uses of Grid use only construct elicitation (and laddering, of course) if they have a large sample to interview. This is often the best available research design, because while the technology to share and compare actual Grids exists, it usually imposes restrictions on the research design which are difficult to manage. Construct elicitation is often used to measure corporate or family culture, usually as part of a change program and/or to develop [child] management competencies. Standard procedure is to ask people to name colleagues [family members] as elements (keeping them anonymous) and then elicit constructs ‘in terms of how they behave’. The analysis is a simple content analysis into the categories which suggest themselves from the constructs, and it is really helpful to enlist some people from the client organization to help with the analysis (ownership and all that). Feedback is then to the person who commissioned the work. Present the construct groupings, starting with the largest group first; and the question is ‘If these are your hopes for the business [family], and these are the terms in which your managers [family members] judge effectiveness, will this view of effectiveness support your achievement of the business [family management] plan?’ If so, fine; if not, we work on how it will have to change.
Meaning lies in Function
Back to ‘reflective’ uses of Grid, where you have chosen to use a computer program to analyze your data. Here the possibilities and strategies for feedback are so many that we can only give a few examples, and leave it to your own practice to learn what seems to be appropriate. The principle to bear in mind is that Meaning lies in Function – in other words, you only know what a concept means to your interviewees if you see how they use them. You can base a good deal of feedback around this principle: for example, the best place to start can be looking at elements (or constructs) which are closely correlated, and asking if this represents the truth as they see it. So if the interview is about close relationships, you might ask ‘You’ve described yourself as very similar to your mother, but very different from your sister – would that be true?’ If your counseling skills tell you that you’ve probably hit a hot spot, then you’d probably probe this or ask if the interviewee wants to leave it for later. If the response is ‘No, I’m more different from my mother than this suggests,’ you can ask for a new construct which rates Mother at one end and Self at the other, and recalculate the Grid. Probing similarities is often very useful when examining the constructs, because it gives you a sight of the interviewee’s stereotypes: if you get two constructs which are semantically different but are closely correlated – for example religious-atheist and bully-not a bully – the presumption is that the interviewee associates religious people with bullying behavior, and atheists are seen as much less likely to be bullies. Obviously this is an area you would want to probe with your counseling skills.
The Ideal Element
Finally, another technique you can use when giving feedback is to invent an ‘ideal’ element, or offer a construct. So if we go back to the lady who wants to get fit but has problems with feeling incompetent and letting the side down, you can use the constructs – in the Grid, or just on their own – to develop an element called MY IDEAL WAY OF GETTING FIT. Rating it on the constructs will give you the criteria, which you can put in priority order; the interviewee then has a shopping list of questions to ask, or you may be able to make a suggestion yourself. Similarly, you could offer a construct if it is appropriate to the purpose and hasn’t appeared naturally.
Feedback is Essential
To summarize: feedback is an essential part of any Grid project, but as far as possible in the early stages you should try not to interpose your own interpretation; better to do it by open questions. At some point you will come to the action planning stage, which is where your own experience and wisdom will be in demand. Most important is to be able to know, yourself, when you have stepped out of the ‘I provide the structure, you provide the content’ role and started to share what you see.
And Never Ever Forget That the Grid Is A Conversation!!!!
Little has been said about the response phase, because this is much more a part of you, then of the grid. You will either need to develop a report which interprets your analysis of the data of an extractive interview, or you will need to create and implement interventions to help the person with problems in living based upon your analysis of the person’s cognitive map and the distress it may cause. Just because little can be said here about the response does not mean that it is not important. It is true that in a reflective interview the client may gain insight into his/her own way of seeing the world and may be able to make changes based on their own analysis. However, if they have come to you to get help or to design a plan of change to enable help to be helpful, this requires your professional perspective. Normalize, don’t traumatize! You may intuit constructs which cause concern, but the feedback from your client is vital – does s/he have concern, distress, etc. Based on the nonbiased integrity of the Grid – the design of the interventions should be based on the goals of the client – not on the needs defined by the professional.
Reminders, Tips, and Wrinkles
This final part is intended to help you if you get stuck, or need a spot of inspiration. There’s no continuous theme; just a collection of Handy Household Hints.
- Don’t forget that Grid is a conversation. Yes, that has been said before, but no apologies for saying it again. The same goes for:
- Pilot your session and Plan Your Analysis as part of the design.
- Don’t assume that somewhere out there is a generic Grid analysis program which you can call up at will. There are many kinds of analysis – including a fair number which don’t use a computer – and you need to choose the appropriate analysis in advance.
- No analysis program will spare you the task of interpreting the results in terms of your purpose.
- Consider where in the interview you will learn the most – which could be anywhere from the answers to element creation questions to the relationships in the analyzed matrix, and plan your process and timing accordingly.
- Set out your contract with your interviewee: why you’re doing the Grid, what will happen to the results, any issues about confidentiality, etc.
- If your interviewee seems to get stuck early on, remember that people can’t do a Grid about something they know nothing about. If that’s not the problem, remember that you will hardly ever go wrong by making your elements more concrete: things, people, time-bound events or activities. It’s a common mistake to make a feature of an element into an element.
- You’ll probably get propositional constructs early on. Don’t worry, because the first task is to establish the two-against-one principle with the interviewee (and you can use propositional constructs in laddering). But if you keep getting propositional constructs, try being really explicit about including the qualifier (the ‘…. in terms of’ phrase) in your question. If that still doesn’t break the logjam, you may have a problem with your interviewee’s comfort level: because Grid can’t be faked, an uncomfortable interviewee only has two choices: to go silent or to give propositional constructs. Go back and check your contract and whether the interviewee is comfortable with it.
- The Repertory Grid gives you lots of different possibilities. You can do a few constructs, a spot of laddering, a few more constructs, some rating …. the important thing is to keep a good flow going and make the interviewee feel comfortable and able to see where you’re going
- Try to get constructs in the form X – Y, rather than X – not X; both poles should carry equal weight. Creative – Has no Imagination is better than Creative – Not Creative.
- It is not the case that one pole of the construct should be ‘good’ and the other ‘bad’.
- When referring to construct poles, talk about the contrast pole rather than the opposite pole. Learn to ask the question ‘How would you describe the other one, by contrast?’
- It’s OK to rewrite a construct when you start using it to rate all the elements.
- Not all the elements in the world can be rated on all the constructs in the world. If you find that it’s difficult to use a construct to rate more than a few of the elements, ask the interviewee if they’d like to drop it or rephrase it.
- An in-depth interview in ‘reflective’ mode may need to be spread over several sessions while the interviewee processes what’s happened so far.
- Don’t construe other people’s construing – it’s not up to you to decide what’s important, or high priority, or meaningful. Ask the interviewee. It’s a joint exploration.
- Practice being noninterventionist, and not offering your own contributions, so that you can be sure of the point when you stopped being a recorder and moved into the ‘helpful interpreter’ mode.
- Feedback, especially in a ‘reflective’ Grid, is part of the process, not the end-point.
- Leave time and space at the end for you to check the interviewee’s comfort level, ask for feedback, repeat what will happen to the data, ask if there are any last-minute thoughts.
- The Repertory Grid is a technique which can serve you all your life. It can change the way you listen, think, ask questions, and be aware of yourself. It’s worth learning properly. The good news is that the learning curve is steep: if you are reflective and/or lucky enough to have someone observe you, most new facilitators will have mastered the basics with four or five practice interviews. You then spend the rest of your life learning more.
The repertory Grid Technique is an excellent addition to the mentor’s tool kit, offering sophisticated means of revealing the unique ways that clients perceive the world around them. This technique is extremely powerful as it does not assume a pre-defined framework and then try to ‘fit’ everyone into a limited set of ‘boxes’. In fact, participants’ grids may be wildly different from one another causing analyses of the data generated to require some thoughtfulness on the part of the facilitator. The technique was designed to bypass cognitive defenses and give access to the person’s underlying construction system. Kelly described an individual’s constructs as transparent templates created with the intention to make them fit over the realities of world in order to anticipate coming events – then revealing these predictions to be either correct or misleading. This anticipation may be passive, as in a prediction, or active, as in action.