For academic journals, I think that what you propose is totally reasonable. I'll try to give a small taste for those not familiar: For a new HIV test, I want t

For academic journals, I think that what you propose is totally reasonable. I'll try to give a small taste for those not familiar: For a new HIV test, I want to know the likelihood ratios, not other parameters which tend to be driven by the background HIV prevalence in the researcher's specific catchment area. My catchment area's HIV prevalence will differ, so other aspects of the test that are dominated by prevalence will not be the same in my population. However, the likelihood ratio will be. The point that I didn't finish making before getting kicked out of the coffee shop was that, at some point, people will want to compute a posterior, so they'll need to figure out a prior. Your suggestion above is helpful. Edit - Also, you're right in that I saw just the 1st paragraph when I initially replied.