Have an idea?

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

"Most likelies" instead of BYO in ACBC

Hello everybody,

I want to integrate an ACBC excercise in my study. However I would like to avoid displaying prices in the BYO (not only price levels, but prices at all). I found a Sawtooth Research Paper in which a "Most likely" question is used instead of the BYO. However I couldnĀ“t find information about how t he question is impemented.

This is the link to the paper:

Does anybody has experience with including a "Most likely" question instead of a BYO in an ACBC and could provide instructions or an example?

Thank you in advance!
asked Jul 6 by JoGu (230 points)

1 Answer

+1 vote
That was early research that wasn't incorporated into the software.

Unless you use a BYO section, ACBC will not oversample the most preferred levels within attributes (included in the BYO section).

One possibility that the software offers is the ability to ask respondents to specify a full or a partial ranking of levels for attributes in a question outside the ACBC question.  Then, during ACBC/HB utility estimation, the utilities can be constrained (custom per person) based on responses to the "most likely" questions.  This is described in the paper you reference...though the results were not as good as we had hoped for with that data set.  Perhaps other data sets would show better results.

If interested, search the ACBC documentation for "Customized Constraints".
answered Jul 6 by Bryan Orme Platinum Sawtooth Software, Inc. (144,240 points)
Thank you very much for your answer! I checked the documentation about "Customized constraints" and think about using it.

However you mentioned yourself that without a BYO most preferred levels are not oversampled. Also the ranking of the attribute levels will prolong my already quite long survey.
Is there a possibility to use the BYO and somehow hide the prices, thus creating a "Most likely"-type question?
There are power tricks using CSS to "hide" the price per attribute (if you are using Summed Pricing in ACBC).  But, be very careful about doing this, because you are hiding the price variable in the BYO, but the coding of the BYO for price estimation still would think that respondents were trading off the attribute levels vs. price in the BYO.  Thus, if you include the BYO in the HB analysis, the utilities would be misinformed and thus biased.

So, here are three options:

1.  Use Summed Pricing option in ACBC and use the BYO question.  Use CSS and perhaps some Javascript to hide the price part in the BYO questions (this is a power trick and perhaps you would need to pay some money to achieve that customization--our technical support group could point you in the right direction).  Then, in utility estimation, you would need to omit the BYO questions, but at the same time I would think you should use the customized constraints option in the HB estimation to constrain the "most likely" level for each respondent to be preferred to the  other levels.

2.  Ask the Most Likelies question as a separate "Select" type question in Lighthouse Studio, and then skip the BYO questions.  This means the attribute levels will all be sampled equally for the remaining ACBC sections.  During HB estimation, you could constrain the utilities such that the most likely level is preferred to the other levels for each attribute.

3.  Avoid doing "Summed Pricing" per attribute and instead use price as a single separate attribute in the ACBC setup.  You could either: a) treat price as a separate attribute like any other attribute, b) use Summed Pricing, but only specify Price as a base price +/- 30% (or some variation of your choice).  This will oversample preferred levels per respondent in the design, but the prices will not tend to loosely track with the quality of the features for the product concepts.  So (like a regular CBC) respondents will sometimes see great product profiles at really cheap prices or bad profiles at really expensive prices.