Conjoint Analysis - Collecting Trade-off Data
Conjoint Analysis - Collecting Trade-off Data
Respondents can reveal their trade-off judgments by either considering two attributes at a time, or by making an overall judgment of a full profile of attributes.
Full-Profile Approach
In a full-profile approach respondents are given cards that describe complete product or service configurations. For example, two possible full-profile descriptions of package tour holidays are shown in Figure 19-2. Not all possible combinations of attribute levels have to be presented in order to estimate the utilities. For example, in Figure 19-2, even with six attributes
each described at three levels, there are 18 profiles to compare.5 Respondents can be asked either to rank order the profiles in order of preference, or assign each of the 18 cards to a category of a rating scale measuring overall preference or intentions to buy. The advantage of the rating scale is that it can be administered by mail, whereas a ranking task usually entails a personal interview.
Trade-Off Approach (Two Attributes Simultaneously)
Respondents in a trade-off approach are asked to rank each combination of levels of two attributes from most preferred to least preferred. The matrix shown in Figure 19-3 illustrates this approach, with the numbers in the cells representing one respondent's rankings. In this matrix, there are nine possible alternatives to be ranked. The best and the worst alternatives are obvious. The interesting evidence is that the respondent doing the ranking in this example is willing to walk five minutes to find comfortably warm water. However, there are six attributes in all, so potentially there could be
n(n - 1) H- 2 = 6(5) - 2 = 15
such matrices for each respondent to fill in. Fortunately, it is not necessary to present all pairs in order to extract statistically the utilities without confusing the contributions of the various attributes.
Respondents can reveal their trade-off judgments by either considering two attributes at a time, or by making an overall judgment of a full profile of attributes.
Full-Profile Approach
In a full-profile approach respondents are given cards that describe complete product or service configurations. For example, two possible full-profile descriptions of package tour holidays are shown in Figure 19-2. Not all possible combinations of attribute levels have to be presented in order to estimate the utilities. For example, in Figure 19-2, even with six attributes
each described at three levels, there are 18 profiles to compare.5 Respondents can be asked either to rank order the profiles in order of preference, or assign each of the 18 cards to a category of a rating scale measuring overall preference or intentions to buy. The advantage of the rating scale is that it can be administered by mail, whereas a ranking task usually entails a personal interview.
Trade-Off Approach (Two Attributes Simultaneously)
Respondents in a trade-off approach are asked to rank each combination of levels of two attributes from most preferred to least preferred. The matrix shown in Figure 19-3 illustrates this approach, with the numbers in the cells representing one respondent's rankings. In this matrix, there are nine possible alternatives to be ranked. The best and the worst alternatives are obvious. The interesting evidence is that the respondent doing the ranking in this example is willing to walk five minutes to find comfortably warm water. However, there are six attributes in all, so potentially there could be
n(n - 1) H- 2 = 6(5) - 2 = 15
such matrices for each respondent to fill in. Fortunately, it is not necessary to present all pairs in order to extract statistically the utilities without confusing the contributions of the various attributes.
Comments
Post a Comment