Abstract
As an investigation into the feasibility of recording personality status from questionnaire data in younger people we used a three phase Delphi survey to assess items from the Christchurch Health and Development Study, administered at ages 12 to 16 years. Twelve experts took part in Phase I, and 22 in Phases II and III, 16 of whom were experts in adult personality disorder (PD), and 6 were experts who work with children. In total, 189 questions (55% of the total (238) in the questionnaires) were identified as possibly being related to personality abnormality in one or more clusters with high consensus. Experts who work with children were less likely to label features as related to personality than experts in PD (p < 0.001), and the four personality factors (equivalent to Mulder and Joyce's antisocial, asocial, asthenic, and anankastic) chosen for assessment showed variable agreement. Confirmatory factor analysis showed the best fitting model of the data was a 3 factor solution involving asocial/asthenic, antisocial, and anankastic factors. This represents the first attempt to use existing recorded data to code personality status and the results of this Delphi survey give some grounds for optimism that this approach has potential in the early identification of personality features.