[Limdep Nlogit List] Choice of data to use [Medard]

[Kassie, Girma Tesfahun (ICARDA-Morocco)]

Hi Medard -

You have an interesting discrete choice experiment (DCE). Let me share my take on where the problem lies and what you could possibly do.

It is important to remember that stated choice methods (SCMs) help us understand individual's preferences for different attributes of a given good or service, price included. This applies regardless of their current behavior. Therefore, those sample units who never used fertilizer should have been part of the DCE as they might be open to using organic solid waste fertilizer under certain conditions. The opt out option could not be assumed to reflect the stated preferences of the never users. Rather it is their revealed preference under their current circumstances. Including them in the DCE could have allowed you to explore if those people would change their preferences under different conditions. So, there is a design flaw that you would have to address.

Your approach of appending the never users in the analysis, as if they all opted out, is an innovative thing to do in my opinion. However, you would have to make a good justification as to why you are doing that given the problem discussed above. You can estimate the models with and without the never users and see how the marginal utility and wtp values behave.

The coefficient of price could be positive due to many reasons. If you believe that is has to have a negative marginal utility and unless your sample is too small, check whether there is preference heterogeneity that needs to be explained and whether you might revise your model specification [interactions or entering categorical vars as continuous or vice versa.]

Good luck -

Girma.

==========================================================

Girma T. Kassie (PhD)

Principal Ag. Market Economist

ICARDA

Rabat, Morocco



________________________________
From: Limdep on behalf of limdep-request at mailman.sydney.edu.au
Sent: Friday, March 14, 2025 4:00 AM
To: limdep at mailman.sydney.edu.au
Subject: Limdep Digest, Vol 200, Issue 1

Send Limdep mailing list submissions to
        limdep at mailman.sydney.edu.au

To subscribe or unsubscribe via the World Wide Web, visit
        https://mailman.sydney.edu.au/mailman/listinfo/limdep
or, via email, send a message with subject or body 'help' to
        limdep-request at mailman.sydney.edu.au

You can reach the person managing the list at
        limdep-owner at mailman.sydney.edu.au

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Limdep digest..."


Today's Topics:

   1. Choice of data to use (medard kakuru)


----------------------------------------------------------------------

Message: 1
Date: Thu, 13 Mar 2025 11:33:04 +0300
From: medard kakuru <medakseth at gmail.com>
To: Limdep and Nlogit Mailing List <limdep at mailman.sydney.edu.au>
Subject: [Limdep Nlogit List] Choice of data to use
Message-ID:
        <CALwnpeS6mPqar9cq1c0ry1QUVQapMNi0fLSxHREBbH_ZLimRGA at mail.gmail.com>
Content-Type: text/plain; charset=UTF-8

Dear Members,
Please advise on which data set to use.
My study is about estimating willingness to pay (wtp) for organic solid
waste fertilizer. I used a choice experiment (CE) with two alternatives and
an opt-out/no choice. During data collection, to ensure random sampling, I
interviewed farmers who had ever used any fertilizer and those who have
never. Those who said never were not given the choice sets to elicit wtp.

The first set of results is using the entire sample (both ever and never
used fertilizers). This implies that there is missing CE data for  those
who have never. I circumvent this by assuming that these farmers, if they
had been given the choice sets, would opt-out/ make no choice. I get good
results, even the price parameter is negative and significant.

The second set of results is using only a subset of only those given the
choice sets (ever used fertilizers). The price parameter is positive and
significant, contrary to economic theory.

Please, advise on which dataset I should consider for my analysis.

PS. I estimated mixed logit models in utility space and wtp space

Regards,

Medard


------------------------------

Subject: Digest Footer

_______________________________________________
Limdep mailing list
Limdep at mailman.sydney.edu.au
https://mailman.sydney.edu.au/mailman/listinfo/limdep


------------------------------

End of Limdep Digest, Vol 200, Issue 1
**************************************

Find out "What We Do" - watch the video!<https://url.au.m.mimecastprotect.com/s/VJ3MC3QNPBiRkzxL8HgfGUQQ-9O?domain=youtu.be>

[ICARDA]

This email and any files transmitted with it may be confidential and are intended solely for the use of the individual or entity to whom they are addressed. If you received this message in error, do not use or rely upon it. Instead, please inform the sender and then delete it. Thank you.