Why Unstructured Interviews Predict Almost Nothing

Lucas and Luna dig into the research showing that unstructured job interviews—the classic 'tell me about yourself' conversation—are barely better than random chance at predicting on-the-job performance. Lucas cites a 2024 meta-analysis of 117 studies finding a corrected correlation of just 0.20. They contrast this with structured interviews, which reach 0.57. The hosts walk through a concrete example: a mid-sized tech company that switched from free-form chats to a structured scorecard and saw new-hire performance ratings jump by 0.8 standard deviations in the first year. They also tackle the common pushback: won't structure kill rapport and chase away creative candidates? The answer, backed by data, is no. Episode 49 of Hiring & Firing with Fexingo. #UnstructuredInterviews #StructuredInterviews #HiringScience #InterviewValidity #MetaAnalysis #Prediction #CandidateExperience #BiasInHiring #PerformancePrediction #RecruitmentResearch #Business #WorkforceDecisions #FexingoBusiness #BusinessPodcast #HiringAndFiring #EvidenceBasedRecruiting #InterviewDesign #TalentAcquisition Keep every episode free: buymeacoffee.com/fexingo