Help with Intraclass Correlation Coefficient

I am working on a research project in which Doctors (n=50) and Lay people (n=50) were separated into two groups. Each evaluated the same variables by marking a visual analog scale. They were again retested with the same survey instrument at a later time. I would like to use the ICC test to determine intra-rater reliability. However, I am unsure as to which (random, mixed, fixed) ICC model to use. I am trying to see how reliable the different groups are at judging these variables over time.

Statistics is foreign to me and I am trying desperately to get a grasp on it!--any help is appreciated.