Dissociation in focus: replicating and extending network analytic insights based on the dissociative experiences scale.

Background

Factor-analytic studies of the Dissociative Experiences Scale (DES) have reported between one- and seven-factor solutions, leaving its dimensional architecture unresolved. This inconsistency compromises the interpretability of scores and limits the scale's utility for screening and treatment planning. To ensure diagnostic relevance and capture dissociation across mental disorders, we focused on a clinical cohort. Emerging evidence from network-analytic approaches suggests a stable three-factor structure, though replication is limited.

Objective

We employed Exploratory Graph Analysis (EGA), a network-psychometric method, to replicate and refine the DES structure in a large UK clinical cohort.Method: Participants were 1,260 treatment-seeking adults drawn from personality disorder and NHS IAPT services (73.8% female; M = 32.3 years). Data were collected between 2013 and December 2024. A Gaussian graphical model based on Spearman correlations was estimated using EGA with 1,000 parametric bootstraps. Item- and dimension-stability indices guided iterative pruning. Dimensional solutions were compared to prior models using the Adjusted Rand Index (ARI).

Results

Seven items with stability indices below .75 were excluded, resulting in a 21-item, three-factor solution reflecting amnesia, derealisation/depersonalisation, and absorption. Each factor emerged in at least 93% of bootstrap samples, with the full configuration replicated in 96.1%, indicating high structural robustness. Cluster correspondence was perfect with the original DES model (ARI = 1.00) and moderate to substantial with two recent network-based studies (ARI = 0.51; 0.80).

Conclusions

EGA recovered the original three-factor configuration of the DES across diverse diagnostic groups, suggesting that amnesia, derealisation/depersonalisation, and absorption represent stable core dimensions of dissociation. Further longitudinal and multi-instrument studies are needed to evaluate temporal invariance and expand construct coverage beyond the DES.