A Milestone in Research on Conjoint Analysis


There has been a lot of research on conjoint analysis in the last decades. From validity to hierarchical bayes and from self explicated approaches to hybrid conjoint, the streams and topics of research are overwhelming.

But still, there are some pearls in this never ending stream of new papers. One of them is the paper on incentive-aligned conjoint by Ding, Grewal, and Liechty (2005) published in the Journal of Marketing Research.

The authors show a promising way of how to improve validity of conjoint analysis. Generally, conjoint analysis was performed in a so-called hypothetical setting, meaning that the respondents did not have an incentive to really tell their truth while answering the conjoint questionnaire. This led to rather a low predictive value and validity of the results. The approach of Ding, Grewal, and Liechty, on the other hand, is aimed at solving these issues by giving the respondents an incentive to answer truthfully. The incentive is, that there exists a possibility that the respondent would have to buy the product in question for a price that is built based on the respondents stated preferences. So, if the respondent misstates her preferences, she might end up being forced to pay too much for the product that she intended to pay.

Thus, using the incentive-aligend approach might help market researchers to improve the validity of their conjoint results. It will be interesting to see if the incentive-aligned approach gets accepted and applied by the market researchers and if it the process will be understood by the subjects of future studies.

Bookmark this page...

Author: Reto Hofstetter

3 Responses to “A Milestone in Research on Conjoint Analysis”

  1. Patrick Says:

    What exactly is ment by the formulation “never lasting stream of new papers”? Is the research concerning conjoint analysis so much advanced, that further research wouldn’t add new information to the subject, or is the method itself considered becoming obsolete in the near future, so that noone whises to produce another paper?

    Ok, joke. It seems obvious, that the alignment of incentives with “telling the truth” should create better results. On the other side, the method also has its limitation. It won’t be easy to find participants (and still have a representative sample), when the price of the good one possibly has to buy is too high, or if the potential participant simply doesn’t need the product in question (if someone already has a computer, a car, a sofa, a television set…., why then participate in a survey that maybe forces him to buy another one?).

    Another problem may occur, when the price of the product in question is too low. When it creates a utility for the participant to cheat (e.g. beeing socially accepted or simply fun), and the price is to low to cover this utility, then the incentive is not strong enough to reveail the truth.

    See ya, Patrick

  2. Reto Says:

    Hey Patrick,

    I wanted to write “never ending stream…”, never mind😉

    You brought up interesting points here. First, it might be true that people are concerned about the binding character of such a survey or that they simply do not understand how the incentive-alignment works and if it is any good for them. I think there even exists literature on the BDM mechanism that brings up this topic. These authors found that people tend to report values that are too low when measuring WTP with the incentive-aligned BDM approach.
    However, considering the fact that when asking hypothetically you do not have a great incentive to answer truthfully, incentive-aligned measurement might still produce more valid results.

    Thanks for starting the discussion, Reto

  3. Patrick Says:

    Hello Reto

    Tought, you wanted to write “ever lasting stream…”🙂

    Of course I agree, that one gets better results, if one can attach an incentive to reveal the truth. I only suggest, that one has to be careful with the interpretation of the results. With an incentive one basically changes the situation, but for many goods (especially durable goods) you won’t be able to create a situation which is exactly similar to the situation of a person who really want’s to buy a certain good. So there remains a distortion. That is ok, if the user of such a method is aware of this. Only, sometimes I feel, that members of the scientific community try to imply there is a 100 %-solution for a specific problem, and for many methods and models, this is not the case, because they only work under certain (more or less restrictive) assumptions.

    Unfortunately I do not have the time to go deeper into the matter, at least at the moment. But I hope your blog develops well.

    Till then, Patrick

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: