Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Complete ignore ommited levels of constructed lists in ACBC part worth estimation?

I´m focussing an issue relating to a recent study. We conducted an ACBC task, the sample is splitted in "active customers" and "unactive customers" without an annual fee-contract.

This information was used to create a different price level, the 1st group got something like "+ 10$ per month on top" and the 2nd group "110$ per month" - with 100 = annual fee and 10/XX/YY $/month the acual levels of price increase.

We set up a parent list with 5 "active customers" levels and 5 "unactive" levels - brought into ACBC with constructed list.

As you can see, there is no need of estimating hypothetic part worths for items which are comepletely irrelevant for the segment. But in HB estimation the price attribute is handled as 10-level attribute, with estimating part worths for all levels.

No I´m considering:
1. Is it possible to seperate the ommited levels before starting the utility runs - rescaling the price level to the 5-level attribute as it was intended? Even with "unavailable" selection in HB settings i´m getting some hypothetic part-worths, making every remaining price point "look" completely different as any other attribute´s zero centered diffs.

2. If 1. not: Is there a way to rescale the part-worths back to the "normal" look an feel of zero centered diffs?

3. If both is not possible: I´d do the workaround to not present price part-worths (the usefullness is questionable anyway) and only report price as sensitivity curve out of SMRT, asuming that my relative Information in part-worths of chosen levels aren´t polluted due to this issue?

thanks in advance,

asked Apr 27, 2012 by anonymous
edited Sep 13, 2012 by Walter Williams

Your answer

Please only use this to answer the original question. Otherwise please use comments.
Your name to display (optional):
Privacy: Your email address will only be used for sending these notifications.
Anti-spam verification:

To avoid this verification in future, please log in or register.