Detecting the dark side of personality using social media status updates

Organizations use personality assessments to inform recruitment decisions as they are predictive of work-related outcomes. While accurate, these assessments are time-consuming and expensive. Using digital records of behavior to assess personality may offer an alternative solution to overcome these limitations. In this study, we explore whether the “dark side” of personality (non-clinical dysfunctional dispositions) can be inferred through the language used in Facebook Status Updates. Using the Hogan Development Survey (HDS), machine learning methods, and the Linguistic Inquiry and Word Count, language use was found to hold a relationship with HDS scores. The Excitable, Dutiful and Bold scales held the strongest relationship with language (R = 0.27, 0.25 & 0.22, respectively), while the Cautious, Colorful and Leisurely scales held the weakest relationship (R = 0.06, 0.07, & 0.08, respectively). This study extends previous research by demonstrating that the full spectrum of dysfunctional dispositions can be measured using online language. Implications for theory and practice are discussed.