“Comparing apples to oranges” is quite a common phrase being invoked when one attempted to compare two incomparable items, i.e. an invalid comparison.
In my O&G experience, when I was asked to compare benefits of two different technologies or methods being implemented on wells, my head first started spinning and eyes were rolling over database to find the two “apples”, i.e. the twin wells that are identical in many ways. These twin-well comparisons are far and few in between and yet we have used these methods on hundreds of wells. I practically ignored majority of data and focused on apple vs apple comparison. It is not to say the comparison is invalid, but that might not be statistically adequate to conclude one method is better than the other.
Back to the question of comparing apples and oranges, I have been pondering how scientists can publish “vaccine A improves change of survival by 10%”. Surely they did not compare the same person with and without the vaccine, there will be one group with vaccine injection and one without. But what about apples and oranges? Different age, genes, lifestyles, genders. They must have some ways to have a valid apple vs orange comparison.