Have an idea?

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

CBC attribute independence

Dear all,

I am wondering whether I have a problem regarding attribute independence:

I am studying preferences towards electric vehicles and one attribute is driving range in kilometers. Another attribute is charging time in minutes per 100 kilometers.

Now, in the setup the waiting time for getting a full charge changes with driving range and keeping charging time equal. e.g. when a concept shows 300 km range & 10 min charge time for 100 km, that would make 30 min waiting time. But 600 km range & 10 min charging time makes 1 hour waiting time. So the waiting time changes even though the charging time is the same, because of the attribute range.

Is there a problem with that?

Thank you!
asked Jul 23 by Ida27 (220 points)

1 Answer

+1 vote

You could accommodate this kind of dependency by building what's called an alternative-specific design.  Our Lighthouse Studio software allows you to build alternative-specific designs in two ways, which are called Conditional Display and Alternative-Specific.  Either of these methods should work for you to allow you to build in this dependency while preserving the statistical properties of your design.  Conditional display might be the easier of the two to use in your situation.
answered Jul 23 by Keith Chrzan Platinum Sawtooth Software, Inc. (93,025 points)
Thank you Keith. Would it be wrong to do it without conditional display or alternative-specific design?
Well, without using one of those approaches, the only way you'll be able to get the choices to display as you want them (with these dependencies) if by using prohibitions, which tend to detract much more substantially from the quality of your designs.  I guess the other option is to just ignore these dependencies and treat charge time and miles as if they were independent, which can also work if the respondents don't find the combinations unbelievable.
Thank you very much! So if there are just few people who are very deep into the topic and think it is unbelievable, it might still work, correct? They can simply choose a very attractive offer or simply not choose a way too bad offer. So if I keep it as it is it would be still correct regarding the CBC method?