Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Interpreting HB model results for an alternative-specific design

Hello, everyone,

I have some difficulty interpreting my HB model results with an alternative-specific design and would appreciate your help.

I have an attribute, let’s call it “brand”. And brand comes with the levels A, B and C. Furthermore, brand B always appears with one of the three eco-labels X, Y and Z. The other brands A and C don’t have a label.

Brand A has a sustainability score of 0%, B has 50% and C 100%. However, this information is hidden in the CBC tasks and is only accessible to the respondents via mouse-over.

Now, my HB model results look like this:

Brand A: -78
Brand B: 50
Brand C: 28

Eco-label X: -1
Eco-label Y: -6
Eco-label Z: 7

Now the question is whether the respondents were deceived by the eco-labels, i.e. the respondents did not recognize that brand C has the highest degree of sustainability despite the lack of a label?

It is obvious that even though Brand C has the highest sustainability score, brand B is preferred. My guess now is that I cannot separate the effects of brand B and eco-label and only the following conclusion is allowed: “If a respondent chooses brand B, it is of secondary importance which label is displayed.”
However, I cannot say that eco-labels – due to their low average importance - generally play a subordinate role.

Is that correct?

Thanks in advance for your advice.
asked Jun 9, 2020 by Nico Bronze (800 points)

1 Answer

0 votes
Nico,

You're right that you cannot separate Brand B from Eco-label - Brand B was never shown without an Eco-label, so we simply don't know what it's utility would be without any Eco-label, nor do we know what Eco-labels might be worth separately from Brand B.

Certainly the Eco-labels play some role in the choice of Brand B (because the levels have non-zero utilities) but it does look like they play less of a role than the other factors, whatever they are, that contribute to the differences in brand utilities.
answered Jun 9, 2020 by Keith Chrzan Platinum Sawtooth Software, Inc. (102,700 points)
Hi, Keith,

thanks for your quick response. A, B and C are not well-known brands, but fictitious brands, which in turn suggest some degree of sustainability:

Brand A (“Standard”): -78
Brand B (“Eco Basic”): 50
Brand C (“Eco Sustainable”): 28

Eco-label X: -1
Eco-label Y: -6
Eco-label Z: 7

The question now is whether the respondents prefer "Eco Basic + any Eco-label" over "Eco Sustainable + no label", although the latter is more sustainable. However, they only get this information - as already written - via a mouseover.

As A, B and C are not well-known brands or product names, I suspect that some of the utility currently attributed to B actually comes from the eco-labels. Otherwise, in my opinion, the preference for "Eco Basic" cannot be explained, unless the respondents do not like sustainable products.

Would you agree?
Yes, that explanation makes sense to me.  If brands A, B and C are not otherwise very informative, then other intangibles that might be associated with them probably are NOT the reason for differences in utilities.
...