What Ian Stevenson collected were anecdotes, not evidence. Evidence would be statistically significant instances of characteristics, not transmittable through any other mechanism , verified against a control group. He never even proposed a mechanism for transmission that someone might actually investigate, or even prove or disprove, let alone actually conduct anything resembling a real test or experiment himself.
Yes I agree with your first statement.But then IMO one doesn'nt need to propose a mechanism for a phenomena in order to prove that the phenomena exists. Just statistically significant positive results balanced against a control group is enough.