CHAPTER 7 ANALYSING MORE THAN ONE 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 

In the previous two chapters, we addressed the issues involved in analysing

single grids, and provided opportunities to practise the techniques involved.

In reflecting on the activity, you’ll remember that, in Section 5.2, I emphasised

that all knowledge is socially defined, and stressed the importance of being

able to communicate your understanding of the interviewee’s construing to

other people. This becomes particularly important when you’re dealing with

several grids at a time.

If the grids consisted entirely of supplied constructs, and the constructs were

identical across all the grids with just the ratings being different (see Section

4.2.7), you’d have no problem in analysing them. You’d treat each grid as a set

of rating scales which just happen to have been printed close together; and

you’d analyse the whole set as you would any set of rating scales. You’d work

out the average rating in each cell of the grid across all the grids, or you might

prefer to calculate the sum of the ranks and rank these sums – whatever you’re

accustomed to do with rating scales. But, to the extent that different constructs

appear in the different grids (and, to my mind, that’s much more interesting

and, indeed, is the whole point of all this work with repertory grids!), the

results would be meaningless.

7.1 The Nature of the Problem . . . . . . . . . . . . . . . . . . . . . . . 146

7.2 Generic Approaches to Content Analysis . . . . . . . . . . . . . 148

7.3 Honey’s Content Analysis . . . . . . . . . . . . . . . . . . . . . . . . 169

7.4 In Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

Things to Do . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

Things to Read. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

So how can you aggregate the different meanings presented? The constructs

are all different, having been elicited rather than supplied! Somehow, you

have to summarise the various meanings present in all the grids for the sample

as a whole; but, at the same time, you need to preserve as many of the different

interviewees’ personal meanings as possible.

We assume, first of all, for the sake of simplicity, that the elements are identical

for all of the respondents, or have been elicited by using identical elicitation

categories (see Section 3.2.2 and Table 3.1). That leaves us free to deal with the

differing constructs, classifying the different types of constructs used by all the

interviewees.

Content analysis is the technique in question. A variety of different

approaches is available to you, and the exact approach you take depends on

two distinct factors: how you intend to put the information together, which is a

matter of design, and the number of grids you have to deal with, which is a

matter of sample size.