How do you avoid clearly better combinations? (Discrete Choice Conjoint Analysis)

Question

I'm trying to run a Discrete Choice Conjoint Analysis for a financial insurance product. When coming up with the cards for the experiment design how do you avoid combinations that are clearly "better"?

For instance if your attributes are

Benefit Amount - (\$1000, \$2000, \$3000)

Benefit Period - (10, 20, 30)

Price - (\$5, \$10, \$20)

You could potentially get a combination of

\$1000, 10, \$20

vs

\$3000, 10, \$5

which seems like a no brainer to select the second option. Should I be trying to avoid this, and if so, how?

Answer 1

It is actually not so easy to remove dominant tasks. If you use an orthogonal design (OD), then this type of tasks might be a necessary part of your design. You can decide to manually remove them from your OD. Cons: This will undermine the "good" properties of your design (i.e., orthogonality and level balance) - However it might not be a big issue because this type of tasks would generate no information about underlying preferences for the factors (assuming that ALL participants behave as expected). Pros: This might help keeping participants engaged in the choice experiment (Some participants could interpret dominant tasks as a "mistake" and then start question the credibility/quality of the experiment). Practical solution: Put the dominant tasks at the end (i.e. last task faced by the participants).

If you use an efficient design (ED) and have prior info about the preferences (e.g. cost increase expected to have a negative effect on choices), then it becomes possible to compute for each task a measure of utility balance (i.e., relatively desirability of the different options within the task) and to remove tasks with huge imbalance. Remark: The NGENE software does that pretty easily.

Theoretically speaking there is no such thing as dominant task. In your example A={$1000;10;$20} vs. B={$3000;10;$5}, option A seems to dominate B, but what if I don't care about the 1st and 3rd attributes? In that case you ask me to choose between A={10} and B={10}, which option would dominate the other now?

Answer 2

You should simply write your experimental design procedure in such a way that it removes them. You can use Conjoint.ly for your experimental design procedure. You can download the experimental design from there.