Why Combine Conjoint Analysis with Qualitative Research?
Sometimes it’s valuable to have respondents complete conjoint analysis surveys in qualitative research. Having respondent’s utilities available in real time allows the researcher to probe into the whys behind the preferences.
On occasion clients have wanted this for quantitative surveys, but more often I’ve had them ask for it as part of one-on-one qualitative research.
Guess what, we have a method for that.
Introducing Best-Worst Conjoint
The method was originally called “Best Worst Conjoint” when Jordan Louviere invented it (Louviere 1993, 1994) but has more recently come to be called Best Worst Case 2 Scaling (Louviere, Flynn and Marley 2015).
I first used it in a focus group setting after seeing Jordan present it at a conference in 1993.
A typical Choice-Based Conjoint (CBC) study would have questions that look like this:
Notice that there are three product profiles defined in terms of 5 attributes with 2–3 levels each, and we have the respondent choose among them.
How Best-Worst Conjoint Works
In a Best-Worst Conjoint we show just one product and the respondent’s task is to choose:
- The best (most motivating) level
- The worst (least motivating) level
The Best-Worst Conjoint questions are easy to program in our MaxDiff software (all the levels become items in the MaxDiff, and we add constraints so that a single question canʼt contain two items that are levels of a single attribute).
We can then make use of the “on the fly” estimation capability in our Lighthouse Studio software to generate MaxDiff utilities and display them on the screen, as soon as the respondent completes the Best-Worst Conjoint questions.
In repeated empirical tests we’ve found Best-Worst Conjoint to produce utilities that are highly correlated with those from a forced-choice CBC (i.e., one without an opt-out alternative).
Why This Is Valuable for Qualitative Research
As you can imagine, this could be valuable for one-on-one interviews and for qualitative research in general.
I’ve programmed this kind of experiment for use in qualitative research. Having these insights available during a one-on-one interview gives the researcher insight into a respondent’s preferences and also provides the groundwork for a discussion with the respondent about those preferences.
I’ve also programmed Best-Worst Conjoint experiments for physicians to give to their patients so that they can understand the individual patient’s therapy preferences—given attributes like the varying benefits the patient could gain from the therapy along with potential risks like side effects, drug interactions, and so on.
This information then allows a deeper discussion of options, informed by the patient’s values and preferences.
References
- Louviere, J. J. (1993), “The Best-Worst or Maximum Difference Measurement Model: Applications to Behavioral Research in Marketing,” The American Marketing Association's 1993 Behavioral Research Conference, Phoenix, Arizona.
- Louviere, J. J. (1994). Conjoint Analysis. In R. P. Bagozzi (Ed.), Advanced Methods of Marketing Research (pp. 223–259). Basil Blackwell, Cambridge, MA.
- Louviere, J.J., T.N. Flynn and A.A.J. Marley (2015). Best Worst Scaling: Theory, Methods and Applications. Cambridge: Cambridge University Press.