Trust me, if I had more time, I would readily discredit the mental health profession. I worked in that field for sometime and have little respect for their largely unscientific approach to research, in that they, failing controlled studies and failing sufficient information, preemptively jump to conclusions. Also, while there is a huge difference between psychology and psychiatry, both fields unfortunately commit similarly flawed studies.
That little line you quoted bothers me on a multitude of levels. There is no indication as to whether these previously diagnosed children were receiving treatment, whether medical or counsel, and thus had no need to self-medicate; there is no indication as to how many children fell into this little "sample;" there is nothing to indicate these children had more or less accessibility to cannabis.
These sorts of studies are based on behavior which, by default, are not controllable. I.e., you cannot readily perform controlled studies on behavior of humans, at least not without breaking a plethora of laws and being labeled as a human rights violator. No, behavioral studies are simply not reliable, and the samplings of these studies are so grossly limited as to make these sorts of studies a mere feeding for doctoral dissertations on psychology/sociology, and not otherwise reliable, nor viable except as dinner table banter.
Far more reliable studies can be performed with long-term, blind research on the influences of drugs on the brain and body, which can more readily be performed, in a controlled environment, on animals, using utilities such as MRIs, chemical samplings, and biopsies. Such studies are not, however, what you are discussing. No, what you're discussing are little more than masturbative studies on causal contrivance.