Important: Read this before posting to this forum

  1. This forum is for questions related to the use of Apollo. We will answer some general choice modelling questions too, where appropriate, and time permitting. We cannot answer questions about how to estimate choice models with other software packages.
  2. There is a very detailed manual for Apollo available at http://www.ApolloChoiceModelling.com/manual.html. This contains detailed descriptions of the various Apollo functions, and numerous examples are available at http://www.ApolloChoiceModelling.com/examples.html. In addition, help files are available for all functions, using e.g. ?apollo_mnl
  3. Before asking a question on the forum, users are kindly requested to follow these steps:
    1. Check that the same issue has not already been addressed in the forum - there is a search tool.
    2. Ensure that the correct syntax has been used. For any function, detailed instructions are available directly in Apollo, e.g. by using ?apollo_mnl for apollo_mnl
    3. Check the frequently asked questions section on the Apollo website, which discusses some common issues/failures. Please see http://www.apollochoicemodelling.com/faq.html
    4. Make sure that R is using the latest official release of Apollo.
  4. If the above steps do not resolve the issue, then users should follow these steps when posting a question:
    1. provide full details on the issue, including the entire code and output, including any error messages
    2. posts will not immediately appear on the forum, but will be checked by a moderator first. We check the forum at least twice a week. It may thus take a couple of days for your post to appear and before we reply. There is no need to submit the post multiple times.

Estimating quadratic preferences

Ask questions about how to estimate models and how to change your settings for estimation.
Post Reply
manes
Posts: 24
Joined: 15 Jun 2023, 13:13
Location: Israel
Contact:

Estimating quadratic preferences

Post by manes »

Hi all,

I'm estimating an MNL model for a DCE with two alternatives representing different levels of monitoring oil spills from drilling rigs and an opt-out option.

One of the attributes in the model is the number of species types saved as a result of monitoring, with the levels being:
1 = Fish,
2 = Fish + Invertebrates,
3 = Fish + Invertebrates + Mammals.

When this attribute was included as a categorical ordinal variable, maintaining its natural ordering, its coefficient was non-significant.

However, when I used three dummy variables and kept the coefficient for 'fish' (b_fish) fixed, the coefficients for 'fish + invertebrates' and 'fish + invertebrates + mammals' were positive and negative (respectively) and significant.

Given these results, if there is evidence of quadratic preferences, would it make sense to incorporate both a linear and a quadratic parameter for the attribute while treating it as a categorical ordinal variable in the utility specification?

Anat
Anat Tchetchik
stephanehess
Site Admin
Posts: 1235
Joined: 24 Apr 2020, 16:29

Re: Estimating quadratic preferences

Post by stephanehess »

Hi Anat

when you say "categorical ordinal", I think you are in fact still treating the attribute as linear in your first specification, given that you are estimating a single coefficient.

It is only when you dummy code it that you are in fact treating it as categorical, and are able to capture the non-monotonic trend.

With only three levels, there is no benefit at all to use linear + quadratic; you are already using the most flexible possible treatment

Stephane
--------------------------------
Stephane Hess
www.stephanehess.me.uk
manes
Posts: 24
Joined: 15 Jun 2023, 13:13
Location: Israel
Contact:

Re: Estimating quadratic preferences

Post by manes »

Thank you Stephane
This attribute was indeed dummy coded. I also understand why, with three levels, it is not possible to estimate quadratic preferences.

My question is: is there any way to isolate and elicit preferences for 'not harming sea invertebrates' and 'not harming sea mammals', independently?
(See the levels below):

Fish
Fish + Invertebrates
Fish + Invertebrates + Mammals
I was considering the following approach (for a model with two generic alternatives + opt-out):

Code: Select all

V[["alt1"]] = asc_alt1 
             + b_fish * alt1_fish  
             + b_invertebrates * (alt1_fish_invertebrates - alt1_fish)  
             + b_mammals * (alt1_fish_invertebrates_mammal - alt1_fish_invertebrates) 
             + .......
The model runs, but I am unsure if this formulation makes conceptual sense.

Anat.
Anat Tchetchik
stephanehess
Site Admin
Posts: 1235
Joined: 24 Apr 2020, 16:29

Re: Estimating quadratic preferences

Post by stephanehess »

Anat

doing this inside the model is not going to help as it can only lead to the same results as what you get.

Basically, if you fix fish (beta1) to 0 and estimate the other two parameters, then beta2 will tell you the added utility for Invertebrates compared to fish (which you indicated was positive). The fact that beta3 is negative means that Fish + Invertebrates + Mammals is worse than fish. That could be because mammals is so negative that it exceeds the benefit of invertebrates, but it could also be due to a negative interaction between them

Stephane
--------------------------------
Stephane Hess
www.stephanehess.me.uk
manes
Posts: 24
Joined: 15 Jun 2023, 13:13
Location: Israel
Contact:

Re: Estimating quadratic preferences

Post by manes »

Thanks a million Stephane!
Anat Tchetchik
Post Reply