DSC 410/510 - Multivariate Statistical Methods
Assignment 17 (deadline: 4pm on Wednesday, December 1)
Practical matters
Chocolates example:
Use SAS to reproduce the discrete choice analysis described in
the class notes for the chocolates dataset (available in the
"chocs.xls" spreadsheet on the data page).
Send an e-mail to ipardoe at
lcbmail.uoregon.edu with answers to the following questions:
As noted in slide 9, the dark, hard chocolate with nuts is the
most popular, with choice probability 0.504.
Write a description of the second most popular chocolate.
What is this chocolate's choice probability?
Chocolate preferences could also be investigated using a
traditional conjoint analysis. For example, the number of times each
chocolate was chosen leads to the following ranks for chocolates 1
through 8: 6.5, 2.5, 6.5, 6.5, 2.5, 1, 4, 6.5. Running a conjoint
analysis on these ranks leads to the following aggregate results:
| Attribute |
Relative importance |
Level |
Part-worth |
| dark |
36.4 |
no |
-1.000 |
|   |
  |
yes |
1.000 |
| soft |
50.0 |
no |
1.375 |
|   |
  |
yes |
-1.375 |
| nuts |
13.6 |
no |
-0.375 |
|   |
  |
yes |
0.375 |
Say whether the following four statements are True or False:
The discrete choice analysis and the conjoint analysis both show
that the most important attribute is soft/hard followed by dark/milk
followed by nuts
The discrete choice analysis and the conjoint analysis both show
that dark is preferred to milk, hard is preferred to soft, and nuts
are preferred to no nuts
In general, discrete choice analysis can be used to determine
individual preferences within a sample as well as aggregate
preferences, whereas conjoint analysis is restricted to finding
aggregate preferences across a sample
Discrete choice analysis is perhaps more realistic than
traditional conjoint analysis because individuals choose just one item
from those on offer rather than providing ratings/rankings for all
items on offer
This assignment is worth 15 points rather than the usual
10.
Reading
Discrete choice analysis is not covered in the text-book, so for some
background reading I recommend A Brief
Introduction to Discrete Choice Analysis in Marketing Research.
© 2004, Iain Pardoe, Lundquist College of Business,
University of Oregon
Last updated November 29, 2004