Does chronic disease influence susceptibility to the effects of air pollution on depressive symptoms in China?

Table 4 Association of air pollution emission per capita/per unit area and depressive symptoms and the role of chronic disease: Robustness test

Variables	Influence of SO₂ emission intensity on depressive symptoms				Influence of TSP emission intensity on depressive symptoms
	Model 1		Model 2		Model 3		Model 4
	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.
Association of air pollution emission per unit area (tons per square kilometer) and depressive symptoms and the role of chronic disease among respondents without mental-related chronic disease: Tobit model (N = 10,508)
Log of SO₂ emission per unit area	0.191	0.163	0.156	0.153	–	–	–	–
Log of TSP emission per unit area	–	–	–	–	0.490***	0.132	0.419***	0.161
Chronic disease	–	–	0.856***	0.107	–	–	0.861***	0.094
Log of SO₂ emission per unit area × chronic disease	–	–	0.161*	0.093	–	–	–	–
Log of TSP emission per unit area × chronic disease	–	–	–	–	–	–	0.284**	0.124
Sigma(u)	2.088***	0.063	2.040***	0.079	2.092***	0.057	2.043***	0.072
Sigma(e)	3.453***	0.042	3.456***	0.040	3.449***	0.037	3.451***	0.040
Log likelihood	− 29,375		− 29,329		− 29,368		− 29,320

	Model 5		Model 6		Model 7		Model 8
	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.	Marginal effects	Boot. Std.
Association of air pollution emission per capita (tons per 10 thousand people) and depressive symptoms and the role of chronic disease among respondents without mental-related chronic disease: Tobit model (N = 10,508)
Log of SO₂ emission per capita	0.359***	0.122	0.332**	0.137	–	–	–	–
Log of TSP emission per capita	–	–	–	–	0.615***	0.121	0.564***	0.122
Chronic disease	–	–	0.860***	0.114	–	–	0.865***	0.088
Log of SO₂ emission per capita × chronic disease	–	–	0.145	0.105	–	–	–	–
Log of TSP emission per capita × chronic disease	–	–	–	–	–	–	0.211*	0.118
Sigma(u)	2.089***	0.072	2.041***	0.057	2.093***	0.081	2.046***	0.076
Sigma(e)	3.452***	0.041	3.455***	0.034	3.447***	0.048	3.449***	0.037
Log likelihood	− 29,374		− 29,329		− 29,365		− 29,319

Models 1–8 are estimated using the xi:xttobit-command in Stata 14. Decentralization was calculated using the center-command, and Marginal effects was calculated using the margins-command
Control variables included individual’s demographic, socioeconomic status, health behaviors and city-level climate variables in living areas. City dummy variables were also controlled
* P < 0.10; ** P < 0.05; *** P < 0.01

ISSN: 1752-4458