I'm planning a discrete choice experiment (DCE) and intend to analyze the data using latent class analysis (LCA). My design has 4 attributes, 10 choice sets, and 3 alternatives per choice set (2 substantive alternatives + a no-change/opt-out option).
I ran a pilot survey with 20 participants and estimated the required sample size following de Bekker-Grob et al. (2015), which is based on a mixed logit model.
My question: how do I adapt or extend this kind of power/sample size analysis for a latent class model, given that the number of latent classes is not known in advance? Specifically, rule-of-thumb approaches like de Bekker-Grob et al. assume a single (or mixed) population-level model; LCA instead estimates class-specific parameters for an a priori unknown number of classes with unknown relative sizes. Is there an established way to translate a mixed-logit-based sample size estimate into a minimum-N recommendation for LCA?
Any guidance, relevant references, or rules of thumb would be very helpful.
References
de Bekker-Grob EW, Donkers B, Jonker MF, Stolk EA. Sample Size Requirements for Discrete-Choice Experiments in Healthcare: a Practical Guide. Patient. 2015 Oct;8(5):373-84. doi: 10.1007/s40271-015-0118-z. PMID: 25726010; PMCID: PMC4575371.
Important: Read this before posting to this forum
- 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.
- 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
- Before asking a question on the forum, users are kindly requested to follow these steps:
- Check that the same issue has not already been addressed in the forum - there is a search tool.
- 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
- Check the frequently asked questions section on the Apollo website, which discusses some common issues/failures. Please see http://www.apollochoicemodelling.com/faq.html
- Make sure that R is using the latest official release of Apollo.
- Users can check which version they are running by entering packageVersion("apollo").
- Then check what is the latest full release (not development version) at http://www.ApolloChoiceModelling.com/code.html.
- To update to the latest official version, just enter install.packages("apollo"). To update to a development version, download the appropriate binary file from http://www.ApolloChoiceModelling.com/code.html, and install the package from file
- If the above steps do not resolve the issue, then users should follow these steps when posting a question:
- provide full details on the issue, including the entire code and output, including any error messages
- 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.
Sample size / power analysis for latent class analysis based on a DCE — unknown number of classes
-
dce.farmers
- Posts: 10
- Joined: 03 Sep 2024, 16:27
Re: Sample size / power analysis for latent class analysis based on a DCE — unknown number of classes
Hi,
I am not aware of techniques for efficient experimental design for discrete choice latent class models. That said, Ngene is the leading software for generating discrete choice experimental designs, so you may want to look into it. A quick glance through their manual did not return a match for latent class designs, but I recommend you search further.
In principle, if you know the asymptotic covariance matrix of your model (be it latent class or any other), you could write a software routine to search for an experimental design with a low D-error. Not sure how easy it would be, though, and you would still need priors for the class allocation parameters, as well as a good guess at how many latent classes you would get. If you are interested in this path, Alcorta & Mariel (2026) already calculated the asymptotic covariance matrix of a latent class choice model with linear utility functions.
Best wishes,
David
I am not aware of techniques for efficient experimental design for discrete choice latent class models. That said, Ngene is the leading software for generating discrete choice experimental designs, so you may want to look into it. A quick glance through their manual did not return a match for latent class designs, but I recommend you search further.
In principle, if you know the asymptotic covariance matrix of your model (be it latent class or any other), you could write a software routine to search for an experimental design with a low D-error. Not sure how easy it would be, though, and you would still need priors for the class allocation parameters, as well as a good guess at how many latent classes you would get. If you are interested in this path, Alcorta & Mariel (2026) already calculated the asymptotic covariance matrix of a latent class choice model with linear utility functions.
Best wishes,
David