. 2018 Dec 11;13(12):e0208162. doi: 10.1371/journal.pone.0208162

Table 2. Regression results for the mediation model.

Model
Model 1: Total effect model
R	R²	F	df₁	df₂	p	B	SE	t	p
0.33	0.11	35.21	4	1080	< 0.001
Constant						3.56	0.40	9.01^***	< 0.001
Gender						-0.03	0.03	-0.84	>0.05
Age						-0.05	0.02	-2.27^*	<0.05
Grade						0.06	0.03	1.03^*	<0.05
SNS addiction						0.25	0.02	11.59^***	< 0.001
Model 2: Mediator variable model
R	R²	F	df₁	df₂	p	B	SE	t	p
0.45	0.20	56.64	4	1080	< 0.001
Constant						0.43	0.78	0.55	>0.05
Gender						0.05	0.07	0.77	>0.05
Age						0.07	0.04	1.61	>0.05
Grade						-0.10	0.06	-1.60	>0.05
SNS addiction						0.68	0.05	14.54^***	< 0.001
Model 3: Dependent variable model
R	R²	F	df₁	df₂	p	B	SE	T	p
0.39	0.15	38.50	5	1079	< 0.001
Constant						3.51	0.39	8.97^***	< 0.001
Gender						-0.03	0.03	-1.03	>0.05
Age						-0.06	0.02	-2.65^**	< 0.01
Grade						0.07	0.03	2.44^*	< 0.05
SNS addiction						0.17	0.02	7.83^***	< 0.001
SNS fatigue						0.11	0.01	7.61^***	< 0.001
Specific effect analysis
		B		SE	LLCI	ULCI	Ratio of direct effect or indirect effect to total effect
Total effect		0.25		0.02	0.20	0.29	Ratio of direct effect or indirect effect to total effect
Direct effect		0.17		0.02	0.13	0.21	68%
Indirect effect		0.08		0.01	0.06	0.10	32%

Note. N = 1085. Unstandardized regression coefficients are reported. Bootstrap sample size = 5000. LL = low limit, CI = confidence interval, UL = upper limit.

*p< 0.05.

**p< 0.01.

***p< 0.001.