Compared with nonusers, noncurrent and monthly users had greater odds of poor academic performance (adjusted odds ratio [aOR], 1.30–2.20), poor impulsivity and self-regulation (aOR, 1.26–2.19), and adverse emotional states (aOR, 1.1–1.42). Adjusted odds of all adverse psychosocial categories, excluding low social engagement, showed a consistent dose-response trend. Effect sizes were small for poor academic performance (d = 0.39–0.44), small to medium for poor impulsivity and self-regulation (d = 0.43–0.55), small for adverse emotional state (d = 0.33–0.40), and none to small (d = 0.03–0.18) for low social engagement.

This study doesn't seem to be actually determining if cannabis consumption is causal. Simply that there is a correlation. I'd argue that in the case of teens that both want the escapism of weed and have a high level of access, there are probably other external factors contributing to poor emotional health and academic performance.

I think that a longer term study looking at health and social outcomes would be more useful than simply comparing how often a kid smokes pot vs their grades.