Lecture 3 (End of Chapter Questions) Flashcards
1
Q
What are the methods to generate attribute weights? Explain the differences between them.
A
- Swing method: does not require knowledge of attribute value functions
- Trade-off method: requires knowledge of attribute value functions, problematic if the attribute levels are discrete
- Direct-ratio method: widely used but unreliable (range effect)
2
Q
What is the drawback of the direct ratio method?
A
Tries to derive weights from general statement about the importance of attributes without considering their range -> no credible result
3
Q
Explain the details of the experiment in and the findings of Fischer (1995).
A
- 45 students got divided into two groups
- Each group have the attributes starting salary and vacation days
- The range for salary is the same and the range for vacation days is different in each group
- Students were asked to assign weights to the attributes using the direct method and trade-off method
Results:
- Students were unaware of connection between range and magnitude of weights
- Direct method: difference between both groups very small
- Trade-off method: differences much larger
- Conclusion: range sensitivity increases for the trade-off method
4
Q
Explain Splitting Bias with one example.
A
- People tend to overweight an objective that is broken down into subordinate objectives when compared to an objective that is not split up
- Example: Students were asked for importance of attributes with the objective good job, individual attributes get higher values then combinations of attributes