To look for a bias between 2 different machines measuring the same variable.
an individual performs 3 or 4 spirometric blows (fev1, FTC and peak flow are measured with each blow). Two of the blows have to be reproducible (within 150ml for fev1 and Fvc and say 40l/s peak flow). The best results are picked from the combination of blows. This is repeated during the same day on another machine. I could just do a repeated measures t-test of the overall result on each machine and a bland & Altman to compare the level of agreement between machines But this doesn’t take into account of the individual variability as each individual performs at least 3 blows on each machine. Any ideas of what statistical approach to use to compare the agreement between machines?