cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Discussions

Solve problems, and share tips and tricks with other JMP users.
Choose Language Hide Translation Bar
allstats
Level I

Interclass correlation in Measurement Systems Analysi

New JMP user here.  I'm trying to get an ICC for inter-rater reliability data.  I have two raters that each scored one measure from a bunch of subjects.  The dataset is organized as:

Subject   rater   score

100          1            43

100          2           41  

101           1   etc

Is it possible to get an ICC from this type of data?  Perhaps I don't understand what gets coded as what in the dialog boxes.

4 REPLIES 4
jthi
Super User

Re: Interclass correlation in Measurement Systems Analysi

How did you try to fill in the columns?

-Jarmo
allstats
Level I

Re: Interclass correlation in Measurement Systems Analysi

I tried a couple variations, but the one that seemed most parallel to the Gasket example in the help files was 

Y, response:  score

Part, Sample ID:  rater

all others were unfilled  Doing this doesn't return an ICC at all.

I just tried changing the rater values to a and b instead of 1 and 2.  It actually did give me a ICC of 0, which seems unlikely since the values from the two raters do have a significant Pearson r.  There are 185 pairs of scores in the dataset.

 

 

jthi
Super User

Re: Interclass correlation in Measurement Systems Analysi

I would guess Subject is Part, Sample, ID and rater would be X, Grouping

Quality and Process Methods > Measurement Systems Analysis > Example of Measurement Systems Analysis example from JMP Help.

-Jarmo
allstats
Level I

Re: Interclass correlation in Measurement Systems Analysi

Thanks jthi, I did try that allocation.  I get an error message:  "Not enough data to compute the process standard deviation.  Disabling options that require standard deviation."  This seems odd given my sample size of 185 pairs of scores.  The ICC returned is 1, which doesn't seem likely since there is variation between the paired scores.

Recommended Articles