Abstract
Objective
This study adopted a person-centred approach to identify subgroups of Dark Triad (DT) and Light Triad (LT) personality traits, examine differences in depression and anxiety levels across these subgroups and explore differences in projected network effects of simulated symptom node manipulations.
Design
A cross-sectional survey design with convenience sampling was used. A total of 2228 university students were assessed for personality traits and mental health using the Light Triad Scale, Dirty Dozen, Patient Health Questionnaire and Generalized Anxiety Disorder Scale.
Methods
Latent profile analysis was employed to identify personality subgroups. Ising network models and the NodeIdentifyR algorithm were integrated to estimate the projected network effects of simulated node manipulations within depression and anxiety symptom networks across personality subgroups.
Results
Three personality subgroups were identified: high DT (9.25%), high LT (52.51%) and medium traits (38.24%). The high DT subgroup exhibited significantly higher depression and anxiety scores compared to the other subgroups (p < .001), while the high LT subgroup showed the lowest scores. Simulated node-manipulation analyses indicated that the high LT subgroup showed the largest projected network changes, with a decrease rate of up to 45% (node D4) and an increase rate of 67% (node A4). The high DT and medium traits subgroups had decrease rates of 30% and 30% and increase rates of 20% and 28% respectively.
Conclusions
Individuals with high LT demonstrate better mental health but show larger projected network changes following simulated node manipulations, whereas those with high DT exhibit poorer mental health and smaller projected symptom changes. This study confirms variations in mental health outcomes and projected effects of simulated node manipulations across personality subgroups, highlighting the role of personality traits in shaping projected network effects of node manipulations.