I think it's well known -- and supported by evidence -- that personality, chemistry and rapport with the therapist matter more than the specific kind of therapy

I think it's well known -- and supported by evidence -- that personality, chemistry and rapport with the therapist matter more than the specific kind of therapy. (With some exceptions; for example, cognitive-behavioural therapy is specifically better suited for OCD than other modalities such as psychodynamic therapy are, while CBT alone probably isn't a good fit for severe trauma. AFAIK these are backed by evidence.) So the specific theoretical frameworks aren't really scientific, at least for the most part. But you can do research on stuff such as whether outcomes are on the average better than without therapy, or whether modality or individual therapists matter more. Of course the evidence is largely still from the questionnaires, with all of their problems, but the questionnaires aren't necessarily administered by the people giving the therapy so it may not be as biased as you might think.