turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- CBC: How to treat attribute level-dependent price ...

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

10-13-2016 04:33 AM - edited 10-13-2016 04:35 AM

Hi everyone,

let's say I am running a choice based conjoint analysis and a simple MNL model with proc phreg and the ties=breslow option.

Among other attributes (Color), I have an attribute level (Car) that is naturally higher priced that the another attribute level (Bike). Therefore, I connect the price attribute levels to Car and Bike, e.g.

If "Car" is in the product alternative in the CBC, a price of either $1000, $2000 or $3000 is shown.

If "Bike" is in the product alternative in the CBC, a price of either $400, $600 or $800 is shown.

I know these price don't make any sense but this is just an example.

In the regression, I have now coefficients for

- Red: 1,5

- Blue: 1,0

(White base product =0)

- Car: 1,8

(Bike base product =0)

- Price if Car:

$2000: -0,5

$3000: -0,75

($1000 base product =0)

- Price if Bike:

$600: -0,25

$800: -0,5

($400 base product =0)

Now let's say I want to calculate the relative importances of the attributes.

The question is **how do I treat the price attribute?**

Can I just take the relative importance for all the prices (then I could just substract -0,75 from 0 = -0,75) or do I have to treat the price effects of both Car and Bike seperately (then the importance for "Price if Car" would be more important (-0,75) than "Price if Bike" (-0,5).

Thank you very much for your help.

Best,

Alex