Self-Explicated Conjoint Analysis: The Real Reason Why You Should Not Use It

The term self-explicated conjoint analysis confuses most people as being a form of conjoint analysis, which it’s not. It’s just a wanna-be copy technique that tries but is not successful in presenting results that are similar as conjoint. So this is an unapologetic warning; do not use self-explicated conjoint analysis technique thinking it’s the real deal that stands in place of a conjoint analysis study.

Similarities between conjoint analysis and Self-Explicated Conjoint Analysis

Both of these techniques assume that the final buyer decision is a result of the addition of choices for the several features, or attribute levels.

Differences between conjoint analysis and Self-Explicated Conjoint Analysis

The main difference between the two types of techniques is that in one, the normal conjoint analysis, respondents are presented with a set of purposefully designed questions about their preferred services or products. On the other hand, in self-explicated conjoint, respondents are presented with two sets of questions. In the first question, the attribute levels (which are grouped based on the attributes) are presented to the respondents and in each level the respondents are questioned on their preferences. There is then a second question where the levels are summarized in a question geared at assigning the relative attributes their importance scores, meaning respondents are expected to view the first question as indicators of their favorite attribute levels.

Why is Self-explicated conjoint analysis wrong?

This approach has been proven to be theoretically wrong and practically misleading. To begin with, the random utility theory does not support this approach. In real world cases, research studies that have been based on self-explicated conjoint have been proven to produce unreliable numbers, which makes one better off using Excel’s RAND function.

The reason why self-explicated conjoint analysis ends up with wrong results is because they fail to use the three main pillars that choice-based conjoint analysis excel is founded on. Firstly, Self-explicated conjoint analysis fails to incorporate the regression analysis model that in normal conjoint analysis is used to make simulations in the sound market share and also in the calculation of confidence intervals.

Secondly, in self-explicated conjoint respondents are not given an opportunity to make their independent decisions but are instead asked to do a self-assessment of the factors that are influencing their options. This type of assessment is not feasible in a real world scenario as most people are either ignorant or unaware of the reasons why they make particular purchase decisions, or those who are get too timid to let it be known.

Finally, while in self-explicated conjoint respondents are required to see attributes and levels as different entities, while in choice based conjoint analysis informants are encouraged to make trade-offs between products, this being viewed as ‘joint considerations.’

No reason whatsoever should compel you to use self-explicated conjoint if you have better alternatives. We recommend sound analytics techniques every time.

I like writing articles on various topics and hope readers would enjoy reading them.