That's not always how you find a study.

I map my ontology to wikipedia.

The closest relevant top is the Dodo BIrd verdict on wikipedia. From a quick scan I couldn't find what I was looking for, but I'm somewhat confident it'll have the right leads. Also, I remember another study suggesting new psychologists were more effective at CBT than experienced ones.

I would dig harder to find the study, but I reckon it's misleading in isolation anyway. The larger, wider body of evidence overwhelmingly suggests the biggest factor in therapeutic success in the 'therapeutic alliance', a relationship level factor, not the therapy itself, or the characteristics individually of the patient of clinician.