Choose Language Hide Translation Bar
Highlighted
Steffo
Level I

Intraclass correlation coefficient with 95% confidence intervall

Is it possible to calculate the intraclass correlation coefficient (ICC) with 95% confidence intervall in jmp?

1 ACCEPTED SOLUTION

Accepted Solutions
Highlighted

Re: Intraclass correlation coefficient with 95% confidence intervall

If you are using measurement system analysis (MSA), then it is available in the Measurement System Analysis platform. The point estimate of the ICC is provided. The confidence interval estimate of the ICC is not available.

See JMP Help menu for more details (Help, Books).

Learn it once, use it forever!

View solution in original post

2 REPLIES 2
Highlighted

Re: Intraclass correlation coefficient with 95% confidence intervall

If you are using measurement system analysis (MSA), then it is available in the Measurement System Analysis platform. The point estimate of the ICC is provided. The confidence interval estimate of the ICC is not available.

See JMP Help menu for more details (Help, Books).

Learn it once, use it forever!

View solution in original post

Highlighted
ron_horne
Super User

Re: Intraclass correlation coefficient with 95% confidence intervall

I know this is an old question by now, but i came across this issue myself the other day.

looking at the ICC definitions from this page: https://en.wikipedia.org/wiki/Intraclass_correlation#Calculation_in_software_packages

i have came up with the attached data table and calculating scripts. The ICC is the first Percent of total in each Random coefficients table.

The confidence interval can be derived from the 95% CI to the variance components. Any other confidence range (i.e. 90%) can be produced manually or by changing platform preferences.

I would appreciate it if someone can confirm the calculations and terminology are consistent with the ones presented on Wikipedia.

Perhaps someone can make a GUI  add in that will make a nice table with all the different ICC's and their confidence intervals.

 

All the best,

Ron

 

Article Labels

    There are no labels assigned to this post.