Items included in the factors.

Personality traits consistently relate to and allow predicting body mass index (BMI), but these associations may not be adequately captured with existing inventories’ domains or facets. Here, we aimed to test the limits of how accurately BMI can be predicted from and described with personality traits. We used three large datasets (combined N ≈ 100,000) with nearly 700 personality assessment items to (a) empirically identify clusters of personality traits linked to BMI and (b) identify relatively small sets of items that predict BMI as accurately as possible. Factor analysis revealed 14 trait clusters showing well-established personality trait–BMI associations (disorganization, anger) and lesser-known or novel ones (altruism, obedience). Most of items’ predictive accuracy (up to r = .24 here but plausibly much higher) was captured by relatively few items. Brief scales that predict BMI have potential clinical applications—for instance, screening for risk of excessive weight gain or related complications.

人格特质（personality traits）与体重指数（body mass index, BMI）始终存在关联，且可用于预测后者，但现有人格量表的维度或侧面未能充分捕捉这类关联。本研究旨在探究利用人格特质预测与描述体重指数的极限精度。本研究使用3个大型数据集（合并后总样本量约10万），包含近700条人格测评条目，以完成两项目标：(a) 实证识别与体重指数相关的人格特质集群；(b) 筛选出可尽可能精准预测体重指数的精简条目集。因素分析结果显示，共提取得到14个人格特质集群，其中既包含已被广泛证实的人格特质与体重指数的关联（如混乱、愤怒），也涵盖鲜为人知或全新的关联（如利他、服从）。多数条目的预测精度（此处相关系数r最高可达0.24，但实际潜力或更高）仅需极少量条目即可实现。可精准预测体重指数的精简人格量表具备潜在临床应用价值——例如用于筛查体重过度增长或相关并发症的风险。