Department of Psychology, University of Turin, Turin, Italy.
Department of Psychology, The University of Toledo, Toledo, Ohio, USA.
J Pers. 2023 Dec;91(6):1410-1424. doi: 10.1111/jopy.12817. Epub 2023 Feb 22.
Since the first study linking recorded smartphone variables to self-reported personality in 2011, many additional studies have been published investigating this association. In the present meta-analyses, we aimed to understand how strongly personality can be predicted via smartphone data.
Meta-analytical calculations were used to assess the association between smartphone data and Big Five traits. Because of the lack of independence of many included studies, analyses were performed using a multilevel approach.
Based on data collected from 21 distinct studies, extraversion showed the largest association with the digital footprints derived from smartphone data (r = .35), while remaining traits showed smaller associations (ranging from 0.23 to 0.25). For all traits except neuroticism, moderator analyses showed that prediction performance was improved when multiple features were combined together in a single predictive model. Additionally, the strength of the prediction of extraversion was improved when call and text log data were used to perform the prediction, as opposed to other types of smartphone data CONCLUSIONS: Our synthesis reveals small-to-moderate associations between smartphone activity data and Big Five traits. The opportunities, but also dangers of the digital phenotyping of personality traits based on traces of users' activity on a smartphone data are discussed.
自 2011 年首次研究将智能手机记录的变量与自我报告的个性联系起来以来，已经发表了许多其他研究来调查这种关联。在本荟萃分析中，我们旨在了解通过智能手机数据可以多准确地预测个性。
使用元分析计算来评估智能手机数据与大五人格特质之间的关联。由于许多纳入的研究缺乏独立性，因此使用多层次方法进行了分析。
基于从 21 项不同研究中收集的数据，外向性与智能手机数据衍生的数字足迹显示出最强的关联（r=.35），而其他特征则显示出较小的关联（范围从 0.23 到 0.25）。除神经质外，对于所有特征，调节分析表明，当将多个特征组合到单个预测模型中时，预测性能会得到提高。此外，当使用通话和短信记录数据进行预测时，而不是使用其他类型的智能手机数据时，预测外向性的准确性会提高。
我们的综合研究揭示了智能手机活动数据与大五人格特质之间的小到中等关联。讨论了基于智能手机数据用户活动痕迹对人格特质进行数字表型的机会，但也存在危险。
