Hello, I am developing a model where I try to estimate the amount of public concern about climate change in different countries, based on Likert-scale (i.e., ordinal) survey responses. Each survey question has a different number of answer options. So, the data might look like this: country----ques…

Sorry for the long delay. I’m only just getting back to the forums to try to mop up some unanswered questions. These long multi-part ones are the ones least likely to get answered because they’re so involved. [image] Clara_Vandeweerdt: Every question will have its own number of answer options…

Thanks for your response, Bob, and apologies for the delay on my side. I will work on implementing your advice, and will let you know what came out of it.

Bob, I finally got around to implementing your advice. I got stuck, however, at implementing the transformation of the step size vector into the threshold vector. This is what I would like to do in the transformed parameters block: real thres[H]; //question thresholds int<lower=1> thres_start[…

[image] Clara_Vandeweerdt: The problem I run into is that declaring integers in the transformed parameters block does not work. What’s the right way to approach this? If you need integer variables, you can declare them as local variables in the transformed parameter or model blocks. Just put…

Thanks a bunch. that solved the problem for the indexing variables. How should I deal with the thres_start variable in the block above? It’s supposed to be an array of indices, which keeps track for every question of where its series of threshold parameters start in the array thres. I am using that…

[image] Clara_Vandeweerdt: I suppose I could enter these arrays as data. Is that the best solution? Probably. Or you could compute them in transformed data. Either way, you can compute something once in the data block then reuse it for ever log density and gradient evaluation. If you want …

Ordered logit--three challenges

Modeling

Clara_Vandeweerdt November 3, 2017, 10:01pm 8

For posterity: a follow-up thread on this same model can be found here

Topic		Replies	Views
Hierarchical ordinal logistic regression Modeling specification	11	1408	March 21, 2019
Logistic regression with partially ordered parameters Modeling	1	327	February 6, 2024
Ordinal IRT model Modeling fitting-issues	9	2595	December 15, 2017
Ordinal probit response model (also translation from jags) Modeling	10	2807	March 3, 2022
Hierarchical Ordinal Probit model: priors for fixed thresholds and non-centered t-distribution Modeling specification , ordinal-response , hierarchical-model	2	937	September 22, 2021

Ordered logit--three challenges

Related topics