I'm sure it's hard to compare but if one child says "I can has cheeseburger?" and a different child says "May I have a cheeseburger?" we'd generally judge the second child to be more skilled in the language than the first. So, it's arguably not impossible to come up with some measure of when a child is speaking to some level of correctness and compare those across languages.
If you want a blind test you'd just ask lots of kids at different ages to do something "ask for a cheeseburger", transcribe the results, have native speakers, who don't know where the answers came from, judge which answers are not making grammar/conjugation mistakes. Some languages children will stop making mistakes much earlier than other languages which would at least suggest that some languages are harder than others (I'm sure there are confounding factors)
Which linguists claim this and how do they measure fluency?