Fig 1.
Path diagram mediation models.
In the initial mediations (a and b) there was no covariate. Follow-up mediations (a and b) included a covariate, and in those models the paths for both the total effect (IV-DV) and IV-Mediator are interpreted with the effects of the covariate removed, the direct effect (IV-DV) with the covariate and Mediator-DV path removed, and the Mediator-DV path with the covariate and IV-DV path removed. IV = independent variable, DV = dependent variable, Att Anx = attachment anxiety, Att Avoi = attachment avoidance, Cog Reapp = cognitive reappraisal, Supp = suppression, Soc Anx = social anxiety.
Fig 2.
The relative sizes of the direct and indirect effects as indicated by the standardized beta weights (β) when presented as absolute values.
Two betas were negative (-ive), as noted to the left of the bars. The figure shows four mediations, all had social anxiety as the dependent variable and two mediators, cognitive reappraisal (Reapp) and suppression (Supp). The independent variable (IV) was either attachment anxiety (Att Anx) or attachment avoidance (Att Avoi), as indicated, and the covariates, when included, were either attachment anxiety or attachment avoidance. Significance levels: * p < .05, ** p < .01, n.s. = not significant.
Table 1.
Correlations, means, standard deviations, minimums and maximums of the measured variables (N = 253).
Table 2.
Multiple regressions of attachment anxiety (Att Anx) and attachment avoidance (Att Avoi) predicting, in three separate regressions, social anxiety, suppression, and cognitive reappraisal (Cog Reapp) (N = 253).
Table 3.
The unstandardized (b) and standardized (β) beta weights of the four mediations.