I've done a large number of tests (100). I also have access to a thesis where it is reported the same type of tests on the same type of material. The mean and COV values are reported, along with the histograms. I'd like to use this information, to join it with my data and from there do some model selection. Since the number of tests done by previous researchers (80) is similar to what I did I think Bayesian Theory can only be useful if one of the series of results is considered representative of the (unknown) population. Maybe joining all the data makes more sense. Any advise?