Analysis

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 
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 
136 137 138 139 140 141 142 143 144 145 146 

Once the construct-level scheme is complete, the concepts can be aggregated into higher

level (construct) categories. While some loss of information will be experienced, the ease

of interpretation is greatly increased. Table 6 lists a sample of raw statements, the

corresponding concept and construct-level categorization.

Validation of Concepts

Once the coding scheme has been developed the concepts should be validated to ensure

reliability of the scheme. The coding scheme approach (benchmarking or theory-driven)

will determine the most appropriate method of validation. When using the benchmarking

approach validation of the coding scheme with individuals who have domain expertise

but are not involved in a prior portion of the study is helpful. Validation can be

accomplished in multiple ways (e.g., card sort, electronic card sort). With a card sort, the

participants are provided the constructs and index cards with each concept. The

reliability between the participants is calculated by measuring the level of agreement on

the card sort. With the electronic card sort, participants are given an electronic spreadsheet

with all of the concepts listed on one sheet and the statements on another sheet.

The expert raters sort the concepts into the constructs. The results of each card sort are

compared to determine reliability of the coding scheme.

When using the theory-driven approach, the coding scheme should be validated against

the existing theoretical framework. When using the theory-driven approach validation

of the coding scheme can be accomplished using researchers who have knowledge of

the theoretical framework but are not involved in a prior portion of the study. The

researchers can validate the scheme by comparing the coding scheme with the theoretical

framework. Any discrepancies should be resolved through discussion.

Step 4: Recast “Raw” Maps into Revealed Causal Maps

Once the classification scheme is completed, the causal statements for each participant

are placed into the appropriate concept (and construct level) categories. The result is a

concept (and construct level) causal map for each participant. See Figures 4 and 5 for

concept and construct level maps respectively.

Figure 4. Concept level causal map

Method

Object

OO Development