I have been approached by a colleague with what appears to me to be a tricky question. I'll try to be as succinct as possible: They are measuring heart rates of mice and they wish to compare two incubation treatments so they have a data set, i believe which contains three measurements taken from each individual in treatment 1 (to correct for measurement error; these I presume will contribute to a single average) on day 1. They then have values of measurements of heartbeats taken from individuals in treatment 2 on day 2. I believe to standardize across days (instead of just blocking... which would have been nice). They took measurements on a sort of standard (I'm assuming it was say 10 individs not associated with either treatment sampled on the different days).

Moral of the story is I'm not sure what the best way to go about this without wasting data would be. Should I subtract all treatment values first by the mean of the standard for that day then do a two-sample t-test? Since they're not independent I can't really toss them in as controls. I don't believe there is any way to do an ANCOVA because the correlated measurements (to standardize for temp) don't match specifically to any individual treatment measurements. Perhaps I should weight each treatment by the mean of the standard? I know little about weighting and it seems somewhat black box to me...

Perhaps there is a standardized way of approaching this problem that I've just been missing but it seems to me to be the classic case of why we set up experimental designs with blocks.

Cheers!