I was thinking along similar lines. People who love fish might tend to fish more, being outside.
Or is the study controlled for that?
These statistical relationships always come down to some other causal relationship.
“Drinking 2 but not more than 5 glasses of red wine per week improves life expectancy by years.”
Who is it in society that drinks 2 but not more than 5 glasses of red wine per week? Rich people with good health care, not in prison, not alcoholic, etc.