Skip to content
Prev 169354 / 398502 Next
Original Raw

How to do ICC

pgseye 
Hi,

I'm essentially wanting to calculate intra- and inter-observer variabilities
for the first principal component of an optic disc shape measure of a sample
of individuals, so from what I can work out I need to work out an intraclass
correlation coefficient(s).

For the intra-data, I have 2 measurements taken on each individual by the
same observer. For the inter-data, 2 observers have taken one measurement
each on the sample.

I've found the 'icc' function in the {psy} package, but am not sure how I'd
perform the intra- calculation as the data argument is an 'n*p matrix or
dataframe, n subjects p raters'.

Would I just substitute the replica for the rater?

Hope this makes sense - still learning about it all.

Thank you.

Paul