Bootstrapped Schemes

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 

After going to all the trouble involved in bootstrapping a category scheme,

one’s inclined to put it to further use! Now that it’s pre-existing, as it were, it’s

useful to put it to work in order to categorise new constructs should your

circumstances require a fresh set of interviews. This would seem entirely

appropriate if the topic remained the same from the first to the second

occasion, and you were satisfied that the new interviewees were, as a sample,

from the same population, so far as thinking about the topic is concerned.

Nevertheless, you’d probably want to review the literature relevant to your

topic area, to draw on some background theory to which you could appeal in

establishing this.

Many of the personnel-psychology applications of grid technique (e.g.,

Jankowicz, 1990) depend on an initial job-analysis grid in which the

different ways in which effective employees do their jobs are expressed in

the form of constructs, and then categorised. This scheme can then be used to

devise standard scales for use in the development of

. competency frameworks

. performance-appraisal questionnaires

. training needs analyses

. personal development need identification.

After a bootstrapping approach has been used to devise the original category

scheme using the constructs of a group of employees, the categories might

subsequently be applied to other groups of employees, if it is established that

the jobs done by the original, and the new, groups of employees are similar.

Section 7.3 provides further examples of applications in job analysis, and

development of employee-appraisal questionnaires.

Theory-Based Schemes

Perhaps the best-known standard category system is Landfield’s (Landfield,

1971). Twenty-two distinct categories, ranging over such themes as ‘social

interaction’, ‘organization’, ‘imagination’, ‘involvement’, and ‘humour’, are

offered, together with a set of detailed guidelines to their use; and careful

definitions of what is, and what isn’t, an example of each category are

provided. While this system has been used mainly in clinical work (Harter et

al., 2001, give a recent case example), it is sufficiently general to be helpful in a

variety of personal and interpersonal situations.

More recently, Feixas et al. (2002) have developed a 45-category system

divided into six overall themes, dealing with moral, emotional, relational,

personal, intellectual/operational, and value and interest-related constructs.

Superficial constructs (which I assume would include some types of

propositional construct: see Section 5.3.3), figurative constructs involving

comparisons with particular other individuals, and constructs applicable to

particular relationships have been excluded from the scheme. Great care has

been taken to ensure the reliability of the scheme, and the level achieved, for

what is a generic system, is comparable to the levels of reliability achievable

with the more specific category systems obtained by bootstrapping (as

outlined in the previous section). For those of a statistical inclination, a mean

percentage agreement figure of 87.3%, corresponding to a Cohen’s Kappa of

0.95 on the six themes and 0.89 on the 45 categories, and to 0.96 and 0.93,

respectively, using the Perrault–Leigh Index, are reported.

Viney has provided scales for assessing cognitive anxiety (broadly,

uncertainty-related anxiety) by drawing on Kelly’s personal construct theory

to revise earlier scales constructed in the psychoanalytic tradition (Viney &

Westbrook, 1976).

Duck (1973) has devised a category scheme for assessing interpersonal

relationships, based on a personality-role-interaction classification. Other

schemes for categorising constructs commonly used in social settings are

briefly described in Winter (1992: 31–32).

Combining Bootstrapping and Theory-Based Approaches

Finally, Hisrich & Jankowicz (1990) provide an example of a pilot study in

which bootstrapping and the use of a pre-existing scheme were combined. It

was known from previous theory and research that when they take an

investment decision, venture capitalists pay attention to three main

characteristics of the proposal: the managerial expertise in the company

seeking funds (MacMillan et al., 1987), the market opportunity, and the cashout

potential (Tyebjee & Bruno, 1984; MacMillan et al., 1987) for the whole

venture. Hisrich & Jankowicz (1990) followed the bootstrapping procedure to

arrive at nine categories classifying 45 constructs used by five venture

capitalists. They found that these reflected three superordinate headings:

. ‘management’ (which fits into what is known about managerial expertise)

. ‘unique opportunity’ (a match with what is known about the importance of

market opportunity in the literature)

. ‘appropriate return’, which, as its component constructs indicate, is

synonymous with ‘cash-out potential’.

This was a small-scale pilot study, with only 45 constructs in total. However,

as Table 7.7 shows, the value in doing a bootstrapping content analysis, rather

than relying entirely on the pre-existing analytic framework, is indicated by

the fact that nine of the constructs could not be accommodated within the preexisting

scheme. The nature of the proposal itself, the way in which it uses

technology, and the geographical location (local versus distributed) also play a

part, in ways which the pre-existing category scheme did not allow for.