Writing content predicts benefit from written expressive disclosure: Evidence for repeated exposure and self-affirmation

Abstract

Expressive disclosure regarding a stressful event improves psychological and physical health, yet predictors of these effects are not well established. The current study assessed exposure, narrative structure, affect word use, self-affirmation and discovery of meaning as predictors of anxiety, depressive and physical symptoms following expressive writing. Participants (N = 50) wrote on four occasions about a stressful event and completed self-report measures before writing and three months later. Essays were coded for stressor exposure (level of detail and whether participants remained on topic), narrative structure, self-affirmation and discovery of meaning. Linguistic Inquiry and Word Count software was used to quantify positive and negative affect word use. Controlling for baseline anxiety, more self-affirmation and detail about the event predicted lower anxiety symptoms, and more negative affect words (very high use) and more discovery of meaning predicted higher anxiety symptoms three months after writing. Findings highlight the importance of self-affirmation and exposure as predictors of benefit from expressive writing.

Keywords:

• Expressive writing
• Anxiety
• Self-affirmation
• Exposure
• Stress and coping

Acknowledgements

We would like to thank Tania Reynolds, Bita Mesri and Annabelle Soars for their valuable contributions to coding the data included in the current manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Kappa for self-affirmation was .54 and for discovery of meaning was .68.

² The term objective is based on established terminology used in previous research (Hammen et al., 1985).

³ Despite high correlations between these variables, we chose not to create a composite due to theoretical distinctions between them.

⁴ We tested baseline depressive symptoms and negative affect separately as covariates, and the quadratic effect of negative word use on anxiety remained significant in the model.

⁵ Following a reviewer’s suggestion, we tested simplified models including all covariates and only one hypothesised predictor at a time. For anxiety, significant associations were found for detail (b = -.29, p = .012), self-affirmation (b = -.11, p = .033), discovery of meaning (marginal; b = .09, p = .058) and negative emotion word use (b = .33, p = .024). No other significant relationships were found between individual predictors and the outcomes (ps > .102).

Additional information

Funding

This project was funded by a grant [R01MH084116, Lieberman and Stanton] from the National Institutes of Health, Bethesda, MD. Andrea Niles received support from a training grant [MH15750 NIMH] in Biobehavioural Issues in Mental and Physical Health.