How to handle price in ACBC with one very expensive level?

Question

Hi,

I'm designing an ACBC study that is a challenge in terms of how to handle price. Most of the features in the design don't make much sense for summed pricing, but there is one binary feature where if it is included, the total price would e about $175k more expensive in the real world.

The total range of prices applicable is about $150k to $450k, but I'd like to prohibit that one feature from ever being shown with a price below a certain point (maybe $300k-ish). All other prices would be set up in increments of $25k and there would be no constraints about features/levels tied to price values. It's just the one feature I can't quite work out what the best approach would be.

Any suggestions or considerations I should know? Would just having a hard rule where that one feature cannot be shown below a certain price point work out okay?

Thanks in advance!

Answer 1

I would think a "corner" prohibition (prohibiting one level of your binary non-price attribute) with one or two levels of the price attribute could work, as long as the Price attribute had 4 or more levels.  ACBC's "test design" capability could help you see how much precision you would be losing on the two attributes if you did such a corner prohibition.  I'd generate at least 300 robotic respondents for that test design report; and it will probably take around 10 to 20 minutes to generate.

I'm a bit nervous about your statement that the prices would be set in $25K increments in the range for $150K to $450K.  Generally, I don't like to use more than about 7 levels of price in conjoint experiments when you are using price in the usual manner.