9.1.1 Identifying Personal Change: The Simple Change Grid

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 

Working with identical elements and constructs is straightforward. This

situation will arise if you have supplied the interviewee with their original

grid, with the ratings removed, asking them to re-rate the elements using the

same constructs (as part of a study of relatively short-term changes, perhaps).

It might also occur if you elicit the second grid from scratch, using the same

elements, soon after eliciting the first grid, and your interviewee finds no

reason to use different constructs.

In both instances, the obvious thing to do is to focus directly on the differences

between the two sets of ratings, as follows.

(1) Elicit two grids in succession from the same interviewee as outlined in

Section 3.1.2. After the first grid has been completed, make a photocopy, Tipp-

Exing out the ratings, and use the resulting form for the second grid.

Alternatively, simply elicit the two grids in succession from scratch. In this

case, note no. 2 below.

(2) Make sure that the elements and constructs are in the same position in

both grids, before you start the analysis. In other words, shuffle the columns

and rows of the second grid around (together with their ratings, of course!) so

that the elements are in the same columns in both grids, as you look at the

columns from left to right; and that the constructs are in the same rows as you

read from top to bottom.

(3) Cell by cell, record the difference between the ratings in corresponding

positions in the two grids. The easiest way to do this is to work with a second

photocopy of the original grid, with the original ratings Tipp-Exed out. Just fill

in each of the cells with the difference between the rating in the first grid and the

rating in the second grid. (Literally: cell by cell, subtract the rating in that cell in

the first grid from the rating in that cell in the second grid, putting the answer

into that cell in the third grid.) At this point, you have a choice:

. Working out the absolute differences (ignoring minus signs) will provide

you with an overall impression of change. Totalling the sum of differences in

the change grid for each element can be informative in this regard!

. Working out the arithmetic differences (using minus signs) will make it

easier to discuss the changes with your interviewee, and address the issue of

why the various changes have occurred.

 (4) At the bottom of each column in the change grid, sum the differences you

recorded in that column, to provide a rough-and-ready indication of the

overall extent to which ratings of each element have changed. (If you recorded

arithmetic differences for purposes of discussion, ignore any minus signs just

when you compute this total.) Table 9.1 shows an example.

(5) Consider the analysis procedures described in Chapter 5. Process

analysis (Section 5.3.1) shouldn’t be forgotten, even though your present

procedure is focused on the numeric changes in the ratings. The way in which

the second grid was completed (especially if it has been elicited rather than

filled in using a photocopy), the interaction between you both, and the whole

elicitation process may, on reflection, help you to interpret the numeric

changes in an informative way. What was happening when the different

ratings were being provided?

Eyeball analysis (Section 5.3.2) is the most relevant procedure of those

outlined in Chapter 5, since it focuses attention on the ratings themselves, and

the ways in which they are being used. Discuss the changes with your interviewee,

and set them in context: why have the changes been made? Particularly,

are there any changes in the supplied constructs, and why might this be?

Characterising constructs (Section 5.3.3), how are the constructs being used?

Do the differences in the ratings point to constructs being used in a less

propositional way? Less pre-emptively?

As you can see, all of these procedures make sense of the change by putting

the purely numeric changes in context. The remainder understand the change

in purely numeric terms.

(6) Consider any of the analysis procedures described in Chapter 6. The

values you’re working with have rather low variance (they’re mainly 1s and

2s), and you may feel that the information available from the following

procedures doesn’t add a lot to a careful and detailed use of the analyses at

step 4. But, for the record: simple relationships between elements (Section

6.1.1) focuses attention on which elements have changed the least, and which

the most. Is there a pattern? Is there anything in common between constructs

on which the ratings have changed a lot? Avoid step 7 of that procedure.

Simple relationships between constructs (Section 6.1.2) does the same for

constructs. Again, there’s no need for step 7 of that procedure.

In both cases, it’s important to remember that your basic data concern change

in construing rather than construing itself. You need to be careful not to

overinterpret the differences you’re working with, and the best way of

avoiding overinterpretation is to discuss the changes with the interviewee.

Cluster analysis (Section 6.2.2) may be useful, summarising the overall

changes in a way which makes them perceptible at a glance. Principal

components analysis (Section 6.3.2) is, in my view, best avoided. The power of

Table 9.1 Change grid analysis before and after a course on ‘post-war prime ministers’

Attlee Churchill Macmillan Wilson Thatcher Blair

Before the course

Principled 2 1 2 4 4 4 Opportunist

Experienced 1 3 2 3 4 3 Inexperienced

Populist appeal 3 1 5 2 1 2 Distant from the public

Ensured a succession 3 2 1 3 4 4 Left in a hurry or likely to

Politically successful 3 3 3 2 4 4 Political failure

More effective 3 1 3 2 3 4 Less genuinely effective

After the course

Principled 1 3 4 5 1 5 Opportunist

Experienced 2 1 2 3 4 5 Inexperienced

Populist appeal 3 2 3 2 1 2 Distant from the public

Ensured a succession 4 1 1 2 5 3 Left in a hurry or likely to

Politically successful 1 2 4 1 2 3 Political failure

More genuinely effective 2 3 1 2 2 4 Less genuinely effective

Difference grid

Principled 1 2 2 1 3 1 Opportunist

Experienced 1 2 0 0 0 2 Inexperienced

Populist appeal 0 1 2 0 0 0 Distant from the public

Ensured a succession 1 1 0 1 1 1 Left in a hurry or likely to

Politically successful 2 1 1 1 2 1 Political failure

More effective 1 2 2 0 1 0 Less genuinely effective

Sum of differences


6 9 7 3 7 5

this kind of analysis comes from identifying underlying ‘components’, the

derivation of which may be problematic (since you’re dealing with data with

little variance), and the meaning of which will be difficult to establish through

numeric procedures alone. (Far better simply to discuss the nature of the

change as under ‘simple relationships between constructs’ above.)

(7) Discuss all of these changes with your interviewee. In case you need

reminding! In all cases, you’re setting out to understand what the change

means to your interviewee, why it might have occurred, and what it might

lead to so far as s/he is concerned.

Table 9.1 exemplifies a common application, where the impact of a teaching

programme, training course, or other intervention is being examined. At this

stage, you might like to use this table to address the questions posed in

Exercise 9.1.

Try Exercise 9.1 now.