Attacking the persons who ask for evidence is a standard trick which it may seem that you like in this specific case.
I’m not attacking you. I’m trying to clarify what it is you’re asking for.
I find it hard to conceive of a
meaningful study designed such that all variables other than passive/active are controlled for.
To illustrate, on one hand you could take a passive speaker and replace the crossover with an active crossover having identical frequency and phase response.
That would control for (most) other variables, but then you would have an active speaker lacking most of the inherent technical advantages active speakers offer; in other words, the study would be meaningless.
On the other hand, you could take a passive speaker and actively redesign the crossover and make other changes which do capitalise on the technical advantages of active speakers, but then you would end up with many other variables (eg amplitude response, phase response, polar response, nonlinear distortion, bass extension, etc) changed.
No doubt there would conceivably now be an audible difference, but this would be (at least arguably) attributable to the differences between
this particular active implementation vs
this particular passive implementation.
I wondered if you had a clearer idea than I do what you were asking for?