Towards dimensional classification of psychopathology: a latent profile analysis of personality traits



Journal Title

Journal ISSN

Volume Title


Tartu Ülikool


Personality-based profiling can help elucidate associations between pathological symptomatology and address shortcomings of current diagnostic systems. Furthering these aims, this thesis investigated personality-based profiles within a female sample that included both patients (n = 313) and healthy controls (n = 114). It was hypothesised based on eating disorder (ED) and transdiagnostic studies that 3–5 classes resembling high functioning, undercontrolled and overcontrolled profiles would emerge. Anxiety, stress susceptibility, mistrust, detachment, irritability and embitterment were expected to, in addition to impulsivity and perfectionism, distinguish between profiles. Using latent profile analysis, a 5-class solution proved best-fitting and the extracted profiles included a high-functioning, a well-adapted, a moderately impulsive and socially dysregulated, an anxious-perfectionistic, and an emotionally and behaviourally dysregulated class. Statistically significant differences in depression, state anxiety and disturbed eating occurred, and diagnostic distribution across classes was also meaningful with a large number of bulimia nervosa patients falling in the impulsive and socially dysregulated class and the anxious-perfectionistic class displaying depression and generalised anxiety. The emotionally and behaviourally dysregulated class exhibited the most comorbidity and severe psychopathology. This hints at profiles’ clinical relevance in predicting current symptomatology. Due to over-representation of EDs in the sample, future studies are warranted to replicate the found profiles. Additionally, longitudinal studies are needed to assess treatment outcome and classification stability.



personality-based profiling, LPA, LPA, latent profile analysis, impulsivity, clinical sample, isiksusepõhine profileerimine, latentse profiili analüüs, kliiniline valim