OBJECTIVE We replicated research by ten Brinke and Porter (2012), who reported that a combination of four behavioral cues (word count, tentative words, upper face surprise, lower face happiness) could accurately discriminate deceptive murderers from genuinely distressed individuals, pleading for the return of a missing relative. HYPOTHESES We hypothesized that each of the four behavioral cues identified in the original study would be similarly related (i.e., size, direction, significance) to veracity in a novel set of pleaders. With these cues as predictors, we also hypothesized that logistic regression models-separately testing the original and replication samples-would produce similar accuracy rates exceeding chance in discriminating genuine from deceptive pleaders. METHOD We gathered a new sample of public appeals, including 82 genuine and 14 deceptive pleaders. After establishing ground truth, we transcribed video-recorded pleas and coded them for the presence of upper face surprise and lower face happiness. We used Linguistic Inquiry and Word Count to determine word count and the proportion of tentative words in each appeal. RESULTS We found support for several hypotheses. Tentative words were used significantly more by deceptive (vs. genuine) pleaders in both the original and replication samples. Deceptive pleaders used significantly fewer words in both samples, although this relationship was significant only in the original sample. Liars in both samples smiled more than truth-tellers, although this relationship was statistically significant only in the replication sample. However, predictive accuracy was poor and did not differ from chance in the replication sample. CONCLUSIONS Findings do not provide a tidy picture of the reliability of behavioral cues to deception. Although some behavioral cues did replicate across samples, others did not. More research will be necessary to understand the factors that produce variable findings across samples, despite using the same methods of investigation. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

中文翻译：

---

情绪化、高风险欺骗的行为检测：在注册报告中复制。


目的 我们复制了 ten Brinke 和 Porter （2012） 的研究，他们报告说，四种行为线索（字数、试探性词、上脸惊讶、下脸快乐）的组合可以准确地区分欺骗性杀人犯和真正痛苦的人，恳求失踪亲人回来。假设 我们假设原始研究中确定的四种行为线索中的每一种都与一组新颖的辩护人的真实性类似相关（即大小、方向、重要性）。以这些线索作为预测因子，我们还假设 logistic 回归模型 - 分别测试原始样本和复制样本 - 在区分真实和欺骗性辩护时会产生类似的准确率，超过机会。方法 我们收集了一个新的公众呼吁样本，包括 82 个真实的 Sently 和 14 个欺骗性的辩护。在确定基本事实后，我们转录了视频录制的请求，并针对上脸惊喜和下脸幸福感进行了编码。我们使用 Linguistic Inquiry 和 Word Count 来确定每个申诉中的字数和试探性单词的比例。结果 我们发现了对几个假设的支持。在原始样本和复制样本中，欺骗性（相对于真实）辩护词明显更多地使用试探性词语。欺骗性辩护在两个样本中使用了明显较少的词语，尽管这种关系仅在原始样本中显著。两个样本中的说谎者都比说真话的人笑得更多，尽管这种关系仅在重复样本中具有统计学意义。然而，预测准确性很差，并且在复制样本中与偶然性没有差异。 结论 研究结果并没有提供欺骗行为线索可靠性的整齐画面。尽管一些行为线索确实在样本中复制，但其他行为线索则没有。尽管使用相同的调查方法，但仍需要更多的研究来了解在样本中产生不同结果的因素。（PsycInfo 数据库记录 （c） 2025 APA，保留所有权利）。