Here, we use data on 12 dichotomous neuroticism items (Supplementary Table 1; Eysenck Personality Questionnaire—Revised Short Form11), to investigate whether the items used to operationalize neuroticism are genetically homogeneous. Given its high correlation with depression (r = 0.58; genetic correlation rg = 0.60) and anxiety (r = 0.75; rg = 0.77)12,13, neuroticism is considered an important phenotype in psychiatric genetic research14. Indeed, neuroticism items often resemble items used to measure (anxious) depression (see Supplementary Table 1 for a substantive overview). For the present study, data was obtained from two samples of the UK Biobank cohort15 (sample 1: first UK Biobank release, available since 2015; sample 2: data added in the second UK Biobank release, available since July 2017; item-specific N range sample 1: 106,218–109,017; sample 2: 260,083–266,896; see Methods; Supplementary Tables 1, 2). We use sample 2 to replicate basic findings of sample 1, and then meta-analyze the results of both samples to compare item-specific genetic signals to the signal obtained in genetic analysis of the sum-score. Our study demonstrates that item-level analyses supplement genetic sum-score analysis, as the 12
