That's not a helpful answer. What's the scoring system? What are the visual cues used, and how is it used in a linear regression model? The article is making strong claims: it never tells us what the data are, and how they differ from "well that's just your opinion, man."

Trauma therapist-in-training here... There has been an increasing amount of research in recent years on the importance of the “therapeutic alliance” - Bruce Wampold has done a lot of work around this topic. There are also assessment tools that have been developed to help measure the strength of the therapeutic relationship, for example the “Working Alliance Inventory.” Measures like this may look at consensus among treatment goals; sense of confidence in the therapist; commitment to the work/relationship; mutual trust; actually liking of one another, etc. Not sure what type of measure or specific research was relied upon for the article (in fact I didn’t read it, was more interested in reading peoples’ comments actually), but hope that helps!

The reason that the article doesn't go into that sort of thing is it was written for the general public. I am not expert in this area. I do know that over the years I have read a great many articles for the general public that refer to scientific study of infant-mother interaction. If you are seriously interested, do some research and read up about it.

Let me add that, I do understand that people in the therapy field often make claims that are unsupported by scientific research, and so I can understand that you are suspicious in this case. But this field is different, there is lots of sound research.