3.2.2 Choosing Elements

К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 
119 120 121 122 123 124 125 126 127 128 129 130 131 

Next, your job is to choose terms which sample or represent that topic. The

best set of elements is one that covers the whole field of the topic evenly.

Imagine you’re drawing a map of (a part of) your interviewee’s mind: a mental

cartographer indeed! Cartographers work by surveying the physical terrain

from a series of triangulation points spread evenly over the landscape, and

you should usually try to do the same.

Anything can be an element: people, places, institutions, job responsibilities,

teaching skills, policies, business strategies – the list is endless, provided it

doesn’t include constructs. Try to avoid words and phrases which have clear

opposites, or words which represent qualities rather than actions or things.

Elements which are nouns are easier to handle than those which are verbs.

Concrete nouns are easier than abstract ones. Names, people, textbook titles,

job roles, occupations, products, brand images – all of these are nouns or can

be expressed as such.

Where you’re using verbs, try to express them as activities, each ending in

‘-ing’, since this is easier to handle when you present each triad. Thus,

‘deciding’, ‘delegating’, and ‘chairing’ are a more user-friendly trio to offer

than ‘to make a decision’, ‘the delegation of job responsibilities’, and

‘chairmanship skills’.

A usable set of elements has an obvious ‘neatness’ about it in representing the

topic. The set should be ‘all of a kind’. If you can, try not to mix abstract nouns

with concrete nouns, activities, and complicated verbal forms. How would

you construe ‘the Olympic ideal’, ‘mountain bikes’, and ‘Saturday afternoons

spent acting as a sports coach’? It’s possible, but rather messy to handle for

you and the interviewee.

The set should certainly consist of mutually exclusive items – in other words,

one element mustn’t include another. It would prove impossible to work with

an element set which included ‘cats’; ‘dogs’; ‘Siamese’; ‘budgerigars’; and

‘Airedale terriers’.

A good set of elements will evoke a feeling of ownership on the part of the

interviewee, and here you have four feasible alternatives in deciding how far

to involve the interviewee in the choice of elements.

Elements Chosen by Investigator

You choose the elements based on your background knowledge and reason for

conducting the grid interview. And, particularly if you decided on the topic,

you should know how best to represent it by means of elements. By definition,

though, you don’t know how the topic appears to the interviewee, and this

runs the risk of omitting elements which are important to him or her.

Elements Chosen by Interviewee

You let the interviewee choose the elements. This will ensure that the topic is

represented from his or her point of view. But it may omit elements, and hence

issues, that you’re interested in.

Elements Chosen by Negotiation Between Investigator and Interviewee

You share the reasons for your investigation of the topic with your

interviewee, and identify a set of elements jointly, by discussion and

negotiation. One particular variant is for you to agree to add one or two key

elements to the set which the interviewee has proposed: elements which

encourage the interviewee to focus on himself or herself, as one of the

elements – ‘self elements’, in fact. You’ll find a description of the rationale for

this in Section 4.2.8.

Elicited Elements

The 10-step procedure described in Section 3.1.2 elicits constructs, as you

know. You can also elicit elements, as a first step before beginning to compare

elements in order to elicit constructs. This is done by providing general

categories, which cover the range of the topic, that the interviewee responds to


For example, if your topic deals with ‘people I know’, you might ask your

interviewee to name ‘someone who I get on well with’, ‘someone I respect’,

‘someone I dislike’, ‘someone good at their job’, and so on, covering a range of

possibilities, each one of which becomes a particular named element, known

Table 3.1 Examples of elicited element categories

If the topic was . . . . . . categories for eliciting elements might be

‘My degree’ A course that you liked

A course that you disliked

The most difficult course you studied

The course most relevant to getting you a job

after graduation


‘My friends’ My best friend

My oldest friend

My closest friend, emotionally

Someone I like who avoids me


‘Employee effectiveness’ A reliable employee

A well-trained employee

An employee with promotion potential

An ineffective employee


‘A job analysis for a telephone

receptionist’s job’

An important task

A time-consuming task

A difficult task

A task requiring special training


‘Potential sources of counselling


A clergyman

A counsellor

A parent

A physician

A lecturer

A psychiatrist

A psychologist

A departmental secretary

Another student

This procedure is useful when you are carrying out a set of grid interviews with several

interviewees. It results in elements which are unique to each individual interviewee, using

categories which are shared by all of them. This proves useful when you come to analyse the grids

as a complete set (see the introduction to Chapter 7).

to the interviewee, to be used in the grid. Table 3.1 provides you with further


This has the advantage of allowing you to offer the same categories of

elements to a sample of interviewees, knowing that, though they will each be

thinking of different named instances, the existence of common categories can

help you when you come to analysing the grids as a set. By choosing

appropriate categories, you can test a hunch, or hypothesis, you may have

about the way in which the whole sample construes the topic. There is more

on this in Section 7.2.1.

It does, however, carry the possible disadvantage of suggesting the kinds of

constructs you are looking for, in a way which may constrain the sorts of

constructs you will be offered. Thus, the elements for the job-analysis grid

shown in Table 3.1 may cause the interviewee to get stuck on the more obvious

constructs which reflect management priorities to do with the importance,

duration, difficulty, and training requirements of tasks that comprise the job,

while omitting to mention more complex and revealing personal constructs

such as:

I like it – I dislike it

When you do the task is up to you – The task has to be done regularly

You can get away with slipshod work – Mistakes are easily spotted

Only our receptionists do this job – This task is done by all receptionists

Done this one all of my life – This job was new to me

Generally speaking, if you choose to provide elements by elicitation, it’s useful

if you have a strong rationale, perhaps based on some previous work on the

issue, for choosing the particular categories you use.

And if you are using grids as part of a research study, your literature review, or the

research question which you are addressing, may suggest ‘obvious’ categories to

you. I once did a piece of research on counsellor outreach programmes (Kaczmarek

& Jankowicz,1991) inwhich the research questionwaswhether students at the end of

a Masters course in counselling and guidance used more sophisticated constructs

about the role of a ‘helper’ than first-year psychology students. (Yes, they did, when

‘sophisticated’ was defined as ‘the way the professional counsellor thinks’.) The

obvious categories to be used in this instance were the ones shown as the last

example in Table 3.1 ^ since this set represents roles found to be important in the

literature on counsellor approachability.

The elicited-element form is a generalised variant of Kelly’s original Role Construct

RepertoryTest (Kelly,1955/1991: 221), in which 24 standard categories that he found

useful in his clinicalwork were used.

Once you’ve decided on a suitable set of elements, you might consider writing

each one of them on a separate card, and offering the appropriate cards three

at a time as you elicit each construct from the interviewee. Many people find

that it helps them to think about and clarify their constructs if they have

something physical to move around on the table.