CIs for categorical variables with more than two categories

[K_S]

Hi there, 

I have several categorical variables with multiple categories.

For example: has disease x, doesn't have disease x, unknown

I want to get the percentage (prevalence) of each with 95% CIs around the estimates. How do I go about doing this?

Thanks!

[FreelanceReinh]

Hi @K_S,

I remember that this topic was discussed here in 2016: 95% CI for Categorical Variables.

In 2017, Rick Wicklin published an even more comprehensive blog post about these confidence intervals: Simultaneous confidence intervals for multinomial proportions.

[Reeza]

Each is a proportion, just use the normal approximation to normal (assuming N is large enough) and get a CI that way? You may want a correction factor if you want to limit it to 0 to 100%.

https://documentation.sas.com/?docsetId=statug&docsetTarget=statug_freq_examples04.htm&docsetVersion...