Question Regarding Ordered Probit Model Results

[Zahra Vakili]

Dear Professor Bierlaire,

I hope this message finds you well.

I have estimated an ordered probit model, but I encountered an issue: the intercepts of some indicators are not statistically significant, while the sociodemographic variables show significant effects. I am unsure how to proceed in this situation.

Given the small sample size (n = 92), should I ignore the insignificance of the intercepts and accept the model as it is, or should I consider removing those indicators?

I have attached the model results as image files for your reference. I would greatly appreciate your advice on how to address this issue.

Best regards,

Zahra Vakili

[Michel Bierlaire]

> On 13 Apr 2025, at 10:12, Zahra Vakili <zahrava...@gmail.com> wrote:
>
> Dear Professor Bierlaire,
> I hope this message finds you well.
> I have estimated an ordered probit model, but I encountered an issue: the intercepts of some indicators are not statistically significant,

Saying that something is “statistically significant” does not mean much on its own. What you probably mean is that the estimates are not very precise. This is not surprising if you don’t have much data. The solution is simple: you can collect more data, or accept that the standard errors of your estimates will stay high.

> while the sociodemographic variables show significant effects. I am unsure how to proceed in this situation.
> Given the small sample size (n = 92), should I ignore the insignificance of the intercepts and accept the model as it is, or should I consider removing those indicators?
> I have attached the model results as image files for your reference. I would greatly appreciate your advice on how to address this issue.
> Best regards,
> Zahra Vakili
>

> --
> You received this message because you are subscribed to the Google Groups "Biogeme" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to biogeme+u...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/biogeme/1e16daea-d0f6-4182-b757-55e144843bf3n%40googlegroups.com.
> <888.png>

Michel Bierlaire
Transport and Mobility Laboratory
School of Architecture, Civil and Environmental Engineering
EPFL - Ecole Polytechnique Fédérale de Lausanne
http://transp-or.epfl.ch
http://people.epfl.ch/michel.bierlaire