The data below represents 6 retail categories with the probability of a
respondent shopping for each category. e.g. Person 1 has a 0%
probability of shopping for category 1 and a 100% probability of
shopping for category 3 at this store.
I'm looking for a multivariate method that would indicate which
categories are the "most important".
Methods like regression require a dependent variable which I don't have.
Is it OK to derive a dep var --> depvar=mean(cat1 to cat6).
Any suggestions ?
| People |
cat1 |
cat2 |
cat3 |
cat4 |
cat5 |
cat6 |
| 1 |
0 |
30 |
100 |
50 |
60 |
30 |
| 2 |
20 |
100 |
100 |
0 |
0 |
0 |
| 3 |
30 |
50 |
50 |
50 |
100 |
100 |
| 4 |
40 |
20 |
20 |
40 |
100 |
50 |
| 5 |
50 |
100 |
100 |
50 |
66 |
30 |
| 6 |
60 |
0 |
0 |
0 |
0 |
0 |
| 7 |
100 |
50 |
50 |
50 |
60 |
20 |
| 8 |
50 |
100 |
0 |
100 |
0 |
100 |
| 9 |
40 |
40 |
50 |
50 |
50 |
0 |
| 10 |
20 |
50 |
40 |
100 |
30 |
100 |
--
Mark Webb
Line +27 (21) 786 4379
Cell +27 (72) 199 1000
Fax to email +27 (86) 5513075
Skype webbmark
Email [hidden email]
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
