Table 1.

Assumed relationships between covariables of mental well-being when developing the DAG

Directed edges		Potential bidirectionality*	References
From	To
Sex	All variables except age and country of origin	No	[61, 66, 67]
Age	All variables except sex and country of origin	No
Country of origin	Education	No
	Social class	No
	Social support	No
Education	Employment status	Yes ( +)	[68, 69]
	Social class	Yes (+ +)
	BMI	No
	Smoking status	No
	Hours of sleep	No
	Mental disorder	Yes ( +)
Social class	Employment status	Yes (+ +)	[70–72]
	Social support	No
	BMI	No
	Smoking status	No
	Hours of sleep	No
	Physical disorder	No
	Mental disorder	No
Employment status	Economic difficulties	Yes (+ +)
	Social support	Yes ( +)
	BMI	Yes ( +)
	Smoking status	Yes ( +)
	Hours of sleep	Yes ( +)
	Mental disorder	Yes ( +)
Economic difficulties	Social support	No
	BMI	No
	Smoking status	No
	Hours of sleep	No
	Physical disorder	Yes ( +)
	Mental disorder	Yes ( +)
Social support	Mental disorder	Yes (+ +)	[14, 55, 73]
	Perceived health	Yes ( +)
Body mass index (BMI)	Physical disorder	Yes (+ + +)	[61, 66]
	Mental disorder	Yes (+ + +)
Smoking status	Physical disorder	Yes (+ +)
	Mental disorder	Yes ( +)
Physical disorder	Hours of sleep	Yes ( +)	[19, 60, 74, 75]
	Mental disorder	Yes (+ + +)
	Lack of autonomy	Yes ( +)
	Perceived health	Yes (+ +)
Mental disorder	Hours of sleep	Yes (+ +)
	Lack of autonomy	Yes ( +)
	Perceived health	Yes (+ +)
Lack of autonomy	Perceived health	Yes (+ +)

*Threatens to the face-validity (plausibility of the posited relationship) are assessed according to evidence, and “Yes” is given in the cases where reverse directionality is plausible. In such cases, the chance of reverse causality is qualitatively evaluated with (+ , +  + , +  + +), indicating less or more probability of bidirectionality, respectively