I don't see the surprise however.
> Prior research has documented a strong and positive correlation between completed education and adults’ mental health.
...
> Results from our analyses suggest that the observed association between education and mental health is attributable to confounding on unobserved variables.