Decentring as a moderator of the associations of anticipatory and post-event processing with social anxiety

ABSTRACT

Recent studies have shown that decentring protects against social anxiety, but no research to date has explored the way it interacts with cognitive risk factors for social anxiety. The present study aimed to examine decentring as a moderator of the association of anticipatory and post-event processing with social anxiety. An unselected student sample (N = 444) completed questionnaires assessing anticipatory/post-event processing, decentring, and social anxiety. The data were analysed with structural equation modelling and the latent moderated structural equations (LMS) method. Results supported the moderating role of decentring in the relationship of anticipatory processing and social anxiety, but did not find evidence of moderation for the association of post-event processing and social anxiety, after accounting for the role of anticipatory processing. Limitations and clinical implications for the protective effects of decentring on social anxiety are discussed.

KEYWORDS:

• Decentring
• anticipatory processing
• post-event processing
• social anxiety

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Consistent with past work, we use decentring as an umbrella term that includes related constructs such as defusion and self-distancing (e.g., Bernstein et al., 2015).

2 We considered several relevant covariates, including age, gender, and race/ethnicity, which have been shown to vary based upon social anxiety and decentring (e.g., Caballo et al., 2008; Naragon-Gainey et al., 2020), as well as GAD symptoms (Generalized Anxiety Disorder-7 Scale; Spitzer et al., 2006) and depression symptoms (Centre for Epidemiological Studies – Depression; Radloff, 1977). These symptoms are known to be highly comorbid with social anxiety. However, the pattern, significance, and magnitude of findings was unchanged after including the above covariates, so we report and focus on the more parsimonious analyses that do not include covariates.

3 While we can’t test this definitively, it seems plausible that participants who couldn’t identify such an event would give ratings of zero (“not at all”) on all items. Only seven participants rated all items as “0” on this measure.