7.1.2 Research Design

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 this context, design is the carefully planned ‘arrangement of conditions for

analysis and collection of data in a manner that aims to combine relevance to

the research purpose with economy of procedure’ (Selltiz et al., 1981).

In any situation, do you want to know whether your male interviewees

construe the topic differently to the females? Then, whatever else you do when

you categorise the constructs, you’re going to need to keep the males’

constructs separate from the females’.

Or, suppose you’re working with a company’s marketing and sales

department, and you suspect that your sales staff – who have direct contact

with the customers – think about the products and the discounts they can give

on sales differently from the staff back in the office. Then you’ll have to

aggregate the constructs separately for the two groups (sales force and nonsales

force) and see what differences there might be.

You’re essentially grouping your respondents according to one or more

variables of interest, and then seeing how each separate group construed the

topic in question. This means that you have to label, code, or otherwise

identify each interviewee’s constructs as you elicit them, so that you can

identify which subgroup they belong to during the content analysis. It also

means that you need to think through how you intend to analyse the

information in the grids as you put the individual grids together, and this

needs doing in advance – before you start on the interviews, and not after!

We’ll look at how to do a generic content analysis in Section 7.2, and then see

how subgroups can be handled.

Finally, the systematic use of a supplied, ‘overall assessment’ construct with

the full sample of respondents lies at the heart of the procedure presented in

Section 7.3. It is this matching of ratings on this supplied construct with all the

elicited constructs which makes it possible to make use of a substantial

proportion of the personal information in each of the grids being contentanalysed.