Proc mixed repeatability coefficient

[MrTh]

Hi all. this is a general question i'm afraid. I have data that repeat over time for each individuals and I want to compute a repeatability coefficient as r = variance within individual / (variance within individual + variance across individuals). Is there a way to get those parameters with SAS proc mixed? Thanks.

[SteveDenham · Accepted solution]

You might try:

proc mixed data=have;
class time subjid;
model response = time;
random time/subject=subjid <if you want to structure the relationship over time you would insert a type= here> solution;
ods output covparms=covparms;
run;

data covparms2;
set covparms;
totalvar=estimate+lag(estimate);
repeatability=lag(estimate)/(estimate + lag(estimate));
run;

it is actually quite a bit harder to do using a REPEATED statement, but this should get you started.

[MrTh]

Hi Steven

that is awesome! 

i was looking into that and you confirmed all. 

many thanks