. 2021 Feb 5;12:612007. doi: 10.3389/fpsyg.2021.612007

TABLE 3.

Testing the moderated mediation effect of problematic social media usage on anxiety.

Predictor	Model 1			Model 2
	Psychological capital			Anxiety
	b	se	t	b	se	t
Gender	0.11 (0.11)	0.04 (0.06)	2.66** (1.78)	0.03 (0.01)	0.04 (0.06)	0.81 (0.80)
Problematic social media usage	−0.30 (−0.33)	0.02 (0.03)	−14.03* (−11.34*)	0.26 (0.17)	0.02 (0.03)	12.60* (5.91*)
Academic burnout				0.17 (0.13)	0.03 (0.04)	6.58* (3.38*)
Problematic social media usage × Academic burnout				0.08 (0.04)	0.02 (0.02)	4.32*** (1.53)
Psychological capital				−0.25 (−0.30)	0.02 (0.04)	−10.16* (−7.87*)
Psychological capital × Academic burnout				−0.06 (−0.06)	0.02 (0.02)	−3.78*** (−2.57*)
R²	0.09 (0.11)			0.30 (0.26)
F	106.08* (68.11*)			146.99* (61.27*)

Gender was dummy coded such that 1 = male and 0 = female. Model 1 regressed psychological capital on problematic social media usage and gender; Model 2 regressed anxiety on problematic social media usage, academic burnout, problematic social media usage × academic burnout, psychological capital, psychological capital, academic burnout, and gender, respectively; the b values are standardized coefficients. The data outside the brackets were for university students whose academic performance had been affected by the pandemic (n₁ = 2056), and the data inside the brackets were for university students whose academic performance had been not affected by the pandemic (n₂ = 1067). *p < 0.05, **p < 0.01, ***p < 0.001.