Individuals have a large amount of personal information on social media (SM), which provides companies with new opportunities for talent selection. However, researchers’ understanding of the effectiveness of assessments based on SM is relatively ambiguous, and the conclusions of empirical studies remain controversial. The Realistic Accuracy Model provides theoretical and methodological support for the application of SM information in zero-acquaintance contexts. Accordingly, we collected and matched 160 sets of Chinese SM assessment (other-assessment) and employee self-assessment data. Through a two-step data analysis, we conducted a consistency check and verification of behavioural predictions. The results suggested that in terms of general suitability, as well as knowledge, skills, abilities, and other characteristics (KSAOs), the other-assessments and self-assessments were consistent. Furthermore, the general suitability and KSAOs of the other-assessments were predictive of behavioural intention (i.e., openness to change). This study empirically tested the accuracy of SM talent assessments, and finally, the research limitations and future trends were discussed.

个人在社交媒体（SM）上拥有大量个人信息，这为公司提供了人才选择的新机会。然而，研究人员对基于SM的评估有效性的理解尚不明确，实证研究的结论仍存在争议。现实准确性模型为零相识环境中SM信息的应用提供了理论和方法论支持。因此，我们收集并匹配了160套中文SM评估（其他评估）和员工自我评估数据。通过两步数据分析，我们进行了一致性检查和行为预测验证。结果表明，就总体适用性以及知识，技能，能力和其他特征（KSAO）而言，其他评估和自我评估是一致的。此外，其他评估的总体适用性和KSAO可以预测行为意图（即，改变的开放性）。本研究通过经验检验了SM人才评估的准确性，最后讨论了研究的局限性和未来趋势。