Transfer

Transfer is where this literature is weakest and most often overstated. Near transfer is well supported; the claim that practice improves reasoning in genuinely unrelated domains is the one that keeps failing to replicate.