Serial mediation effects
Three models were tested (Fig. 1). In the first model, the effects of the impact of COVID-19 on depression (c path) (B: 0.123, SE: 0.011; 95% confidence interval [CI]: 0.100, 0.145; t: 10.823, p = 0.001) and emotional dysregulation (B: 0.499, SE: 0.031; 95% CI: 0.438, 0.560; t: 16.038, p = 0.001) were positive and significant. On the other hand, the effects of the impact of COVID-19 on perceived social support (B: −0.242, SE: 0.040; 95% CI: −0.321, −0.163; t: −6.033, p = 0.001) and attachment anxiety (B: −0.008, SE: 0.002; 95% CI: −0.012, −0.003; t: −3.497, p = 0.001) were negative and significant. The first mediator, perceived social support, had a negative and significant effect on emotion dysregulation (B: −0.183, SE: 0.033; 95% CI: −0.248, −0.118; t: −5.542, p = 0.001), a positive and significant effect on anxious attachment (B: 0.012, SE: 0.002; 95% CI: 0.009, 0.016; t: 6.364, p = 0.001) and a negative and significant effect on depression (B: −0.040, SE: 0.010; 95% CI: −0.061, −0.020; t: −3.858, p = 0.001). These effects were significant because the p values were 0.001, and the CIs did not include zero (Fig. 1a).
The second mediator, emotion dysregulation, had a positive and significant effect on depression (B: 0.110, SE: 0.014; 95% CI: 0.083, 0.137; t: 7.918, p = 0.001) and a negative and significant effect on anxious attachment (B: −0.022, SE: 0.003; 95% CI: −0.027, −0.017; t: −8.839, p = 0.001). Finally, the third mediator, attachment anxiety, had a negative and significant effect on depression (B: −0.349, SE: 0.228; 95% CI: −0.797, 0.100; t: −1.528, p = 0.127). When the impact of COVID-19 and all the other mediating variables were added into the model at the same time, the effect of the impact of COVID-19 on depression remained significant (c′ path) (B: 0.200, SE: 0.010; 95% CI: 0.181, 0.220; t: 20.249, p = 0.001) (Fig. 1a).
The total effect, indirect effect, and total indirect effect values were 0.123, 0.200, and 0.077, respectively. In terms of indirect effects, the effect of the impact of COVID-19 (X) on depression (Y) through perceived social support (M1; bootstrap = 0.010, 95% CI = 0.004, 0.017), emotion dysregulation (M2; bootstrap = 0.005, 95% CI = 0.003, 0.009), and attachment anxiety (M3; bootstrap = 0,001, 95% CI = 0.001, 0.003) was significant. The effect of the impact of COVID-19 on depression (Y) through perceived social support (M1) and emotion dysregulation (M2) (bootstrap = 0.001, 95% CI = −0.000, 0.001), and through perceived social support (M1) and attachment anxiety (M3) (bootstrap = 0.055, 95% CI = 0.038, 0.073) was also significant. However, the effects of the impact of COVID-19 on depression (Y) through emotion dysregulation (M2) and attachment anxiety (M3) (bootstrap = 0.004, 95% CI = −0.001, 0.010), and through perceived social support (M1), emotion dysregulation (M2) and attachment anxiety (M3) all together (bootstrap = 0.003, 95% CI = −0.001, 0.008) were not significant. Thus, perceived social support, emotion dysregulation, and attachment anxiety did not have serial mediation effects in the relationship between the impact of COVID-19 and depression (Fig. 1a).
In the second model, the effects of the impact of COVID-19 on stress (c path) (B: 0.147, SE: 0.012; 95% CI: 0.122, 0.171; t: 11.949, p = 0.001) and emotion dysregulation (B: 0.499, SE: 0.031; 95% CI: 0.438, 0.560; t: 16.038, p = 0.001) were positive and significant. On the other hand, the effects of the impact of COVID-19 on perceived support (B: −0.242, SE: 0.040; 95% CI: −0.321, −0.163; t: −6.033, p = 0.001) and attachment anxiety (B: −0.008, SE: 0.002; 95% CI: −0.012, −0.003; t: −3.497, p = 0.001) were negative and significant. The first mediator, perceived social support, had a negative and significant effect on the second mediator, emotion dysregulation (B: −0.183, SE: 0.033; 95% CI: −0.248, −0.118; t: −5.542, p = 0.001), a positive and significant effect on the third mediator, attachment anxiety (B: 0.012, SE: 0.002; 95% CI: 0.009, 0.016; t: 6.364, p = 0.001), and a negative and significant effect on stress (B: −0.025, SE: 0.011; 95% CI: −0.047, −0.003; t: −2.221, p = 0.027). These effects were significant because the p values were <0.05, and the CIs did not include zero (Fig. 1b).
The second mediator, emotion dysregulation, had a positive and significant effect on stress (B: 0.129, SE: 0.015; 95% CI: 0.099, 0.158; t: 8.559, p = 0.001), and a negative and significant effect on attachment anxiety (B: −0.022, SE: 0.003; 95% CI: −0.027, −0.017; t: −8.839, p = 0.001). The third mediator, attachment anxiety, did not have a significant effect on stress (B: −0.354, SE: 0.247; 95% CI: −0.839, 0.131; t: −1.435, p = 0.152). When the impact of COVID-19 was added into the model together with all the other mediating variables at the same time, the direct effect of the impact of COVID-19 on stress remained significant (c′ path) (B: 0.230, SE: 0.011; 95% CI: 0.210, 0.251; t: 21.690, p = 0.001) (Fig. 1b).
The total effect, indirect effect, and total indirect effect values were 0.147, 0.230, and 0.084, respectively. In terms of indirect effects, the effect of the impact of COVID-19 (X) on stress (Y) through perceived social support (M1) (bootstrap = 0.006, 95% CI = 0.001, 0.013) and emotion dysregulation (M2) (bootstrap = 0.006, 95% CI = 0.003, 0.010) was significant. On the other hand, the effect of the impact of COVID-19 (X) on stress (Y) through attachment anxiety (M3) (bootstrap = 0.001, 95% CI = −0.001, 0.003) was not significant. The indirect effects of the impact of COVID-19 (X) on stress (Y) through perceived social support (M1) and emotion dysregulation (M2) (bootstrap = 0.001, 95% CI = −0.000, 0.001), and through perceived social support (M1) and attachment anxiety (M3) (bootstrap = 0.064, 95% CI = 0.045, 0.086) were significant. Contrary to these findings, the effects of the impact of COVID-19 (X) on stress (Y) through emotion dysregulation (M2) and attachment anxiety (M3) (bootstrap = 0.004, 95% CI = −0.003, 0.011), and through perceived social support (M1), emotion dysregulation (M2), and attachment anxiety (M3) (bootstrap = 0.003, 95% CI = −0.001, 0.009) were not significant. Thus, perceived social support, emotion dysregulation and attachment anxiety did not have serial mediation effects in the relationship between the impact of COVID-19 and stress (Fig. 1b).
In the third model, the effects of the impact of COVID-19 on anxiety (c path) (B: 0.140, SE: 0.012; 95% CI: 0.116, 0.163; t: 11.689, p = 0.001) and emotion dysregulation (B: 0.499, SE: 0.031; 95% CI: 0.438, 0.560; t: 16.038, p = 0.001) were positive and significant. On the other hand, the effects of the impact of COVID-19 on perceived social support (B: −0.242, SE: 0.040; 95% CI: −0.321, −0.163; t: −6.033, p = 0.001) and attachment anxiety (B: −0.008, SE: 0.002; 95% CI: −0.012, −0.003; t: −3.497, p = 0.001) were negative and significant. The first mediator, perceived social support, had a negative and significant effect on the second mediator, emotion dysregulation (B: −0.183, SE: 0.033; 95% CI: −0.248, −0.118; t: −5.542, p = 0.001), and had a positive and significant effect on the third mediator, attachment anxiety (B: 0.012, SE: 0.002; 95% CI: 0.009, 0.016; t: 6.364, p = 0.001). On the other hand, the effect of perceived social support on anxiety was not significant (B: −0.015, SE: 0.011; 95% CI: −0.037, 0.006; t: −1.378, p = 0.169). These effects were significant because the p values were <0.05, and the CIs did not include zero. Furthermore, the reason for the non-significance was because the p values were >0.05, and the CIs included zero (Fig. 1c).
The second mediator, emotion dysregulation, had a positive and significant effect on anxiety (B: 0.083, SE: 0.015; 95% CI: 0.054, 0.112; t: 5.669, p = 0.001), and a negative and significant effect on attachment anxiety (B: −0.022, SE: 0.003; 95% CI: −0.027, −0.017; t: −8.839, p = 0.001). The third mediator, attachment anxiety, had a negative and significant effect on anxiety (B: −0.618, SE: 0.240; 95% CI: −1.091, −0.146; t: −2.572, p = 0.010). When the effect of the impact of COVID-19 and all the other mediating variables were added into the equation at the same time, the effect of the impact of COVID-19 on stress remained significant (c′ path) (B: 0.202, SE: 0.010; 95% CI: 0.183, 0.222; t: 20.389, p = 0.001) (Fig. 1c).
The total effect, indirect effect, and total indirect effect values were 0.140, 0.202, and 0.062, respectively. In terms of indirect effects, the effect of the impact of COVID-19 (x) on anxiety (Y) through perceived social support (M1; bootstrap = 0.004, 95% CI = −0.001, 0.010) was not significant. On the other hand, the effects of the impact of COVID-19 (X) on anxiety (Y) through emotion dysregulation (M2; bootstrap = 0.004, 95% CI = 0.002, 0.007), through attachment anxiety (M3; bootstrap = 0.002, 95% CI = 0.001, 0.004), through perceived social support (M1) and emotion dysregulation (M2) (bootstrap = 0.001, 95% CI = −0.000, 0.001), and through perceived social support (M1) and anxious attachment (M3) (bootstrap = 0.041, 95% CI = 0.025, 0.062) were significant. The effects of the impact of COVID-19 (X) on anxiety (Y) through emotion dysregulation (M2) and attachment anxiety (M3) (bootstrap = 0.007, 95% CI = −0.001, 0.015), and through perceived social support, emotion dysregulation, and attachment anxiety (bootstrap = 0.005, 95% CI = 0.001, 0.013) were also significant. Thus, there were serial mediating effects of perceived social support, emotion dysregulation, and attachment anxiety in the relationship between the impact of COVID-19 and anxiety (Fig. 1c).