Can I merge 2 Choice Based Conjoint survey results?

Question

I have 2 surveys (Choice Based Conjoint or Discrete Choice) conducted in 2 different cities. The attributes and the related levels are the same but the surveys are conducted seperately. Can I merge these 2 city data and report the results (importance, simulation ect) under the name of the related State??

And if I had same attributes but different levels in these 2 different geo surveys, would I still be able to combine the results?

Thanks!

Answer 1

Hi,

I'm not an expert on these, but I wouldn't do it, since their utilities are calculated independently. I don't believe "5 utility" in one study is equal to "5" in the other. After all it's a zero-sum calculation method.

You could do a merged simulation however, build up two simulators and add their results. That is exactly what happens in real life anyway.

Comments

Hi, I think you may combine two results in simulators, and set up a filter to choose the city. Then it would be more realistic when you do city simulation.

Answer 2

For your first question, if you had the same conjoint exercise fielded twice, then there is no issue merging things together either for initial estimation of utilities or combining utilities for a single simulator (other than the manual work involved merging utilities if you wanted to use the SMRT software or our online simulator).  It would be akin to having two waves of people take the same online survey.

If you had the same attributes but different levels, it would probably be unwise to run utilities together, and you couldn't do it if you had a different number of levels to begin with.  Building a single simulator together could be done in something like Excel just to aggregate the results as Bahadir mentioned, but I would imagine that would be a little weird as you would have different levels to choose from for the different groups when configuring your products to simulate.