Your analogy is inapt; the whole reason dark matter is hypothesized is precisely because of the directly observable, inarguable effects it has; we can even see

Your analogy is inapt; the whole reason dark matter is hypothesized is precisely because of the directly observable, inarguable effects it has; we can even see that it is both not evenly distributed and not distributed congruently with matter, so we know enough of its nature to give a name to the phenomenon. Regardless, you've misunderstood my point, which is that if you expect religious phenomena to be undetectable (not "we haven't observed it" but "you can never observe it") and ineffectual (not "we haven't observed its effect" but "you can never observe any effect") then it simply isn't real in any useful sense: religious, scientific, or otherwise. And I don't think religious people believe that: I think they think their religion is real, in that it has effects (a volcano doesn't erupt because we've made appropriate sacrifices to it; my consciousness will live on in some way in the presence of other personhoods after my death, in another place; Florida will be hit by more frequent natural disasters because schoolteachers tell students that gay people exist) and is observable (my charismatic evangelical preacher speaks meaningful sentences in a foreign language when possessed by a spirit; unhappy dead emerge from the ground each year to be led by Hecate against the living). I'm pointing out that while, as a conciliation (essentially, because people dislike it so much and grow antagonistic and violent when challenged), we often say that religion and science are separate, or that religion, as something that is taken as faith, is not investigable by scientific methods; in fact if religion has any reality or meaning then it is investigable. Not necessarily as a practical matter (though to some extent, it is practicable), but there's no principle that precludes any religious hypothesis from being tested. To boil it down: if a religion were real, we could test it. An assertion that it's untestable is equivalent to saying that it's purely imaginary. That's different from saying that we just can't test it yet.