Do you know, this is one area where you could present me with a cast iron, gold sealed, peer reviewed piece of research and I would still ignore it ...
Interesting thing this research does look reasonably reliable, peer reviewed etc but you have to look at the question its asking.
The study is looking at TA input as a measure of how much time each child spends with a TA, when all other factors such as SEN and previous educational attainment are controlled for. The kids who spent more time with the TAs made less relative progress.
Now all that actually says is that if a child spends relatively more time being taught by a person who is (probably) less qualified to teach, they will make less educational progress than with a qualified teacher
There is no measurement there of what happens if you have a class TA whose role is to enable the teacher to teach, and to divide extra time fairly equally among the kids under the teachers direction. Does that improve the outcomes of a whole class compared to a whole class who don't get that enabling influence?
Nor does it tell you what happens if TAs deliver carefully planned interventions to specific groups of kids under the overall guidance of the class teacher. Which might not be a lot in terms of 'time spent with TA' but might well show an impact in the overall outcomes of the class.
Yes, probably from this research it is not an effective use of money to use poorly trained and managed TAs to 'babysit' the most disruptive or underachieving kids (of whom I have one, by the way). But no way does it say that the TA role itself is useless!
Before you jump to conclusions based on the answer, have a damn good look at the question 