[R] simultaneous confidence intervals for multinomial proportions: sample size

David Winsemius dw|n@em|u@ @end|ng |rom comc@@t@net
Sun Apr 9 15:25:06 CEST 2023

> On Apr 9, 2023, at 5:46 AM, paladini using trustindata.de wrote:
> Hello!
> I want to calculate simultaneous confidence intervals for a nominal variable with three categories: "yes", "no", "partially" and I expect that far more than 5 samples fall into each category.
> I have read that Glaz & Sison's method is only appropriate for variables with 7 or more categories. Therefore, the Goodman method seems like a good idea.
> I have found the R function "MultinomCI" {DescTools} to calculate these CIs.
> However, I have not found a way to calculate in advance the required sample size for the Goodman method at a fixed confidence level and length of the intervall.

Seems very unlikely that you can achieve both of those goals simultaneously. I would expect the CI width t vary with the size of the sample in each category in any case. So I'm wondering if your failure at searching derives from the impossibility of your goal. Why would you desire such a result?

> I would be very grateful for any suggestion.
> Many thanks in advance!
> Claudia
> 	[[alternative HTML version deleted]]
> ______________________________________________
> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list