Cold-blooded women can detect lies with greater accuracy than other women

ABSTRACT

Lies are notoriously difficult to detect. But it appears that some people are better at accomplishing this task than others even though the factors contributing to deception detection accuracy are not well understood. This study explored the influence of empathy on the detection of deception as a function of the detectors’ gender while dark personality traits were statistically controlled. Eighty men and 80 women were requested to judge whether individuals viewed in videos were giving their true opinion or not on current debatable issues (50% truthful and 50% deceptive narratives). Judges were divided into four groups according to their gender and their degree of empathy, as assessed using the Questionnaire Measure of Emotional Empathy. It was found that women with lower levels of empathy distinguished false from true opinions better than women with higher empathy, whereas no such difference was found in men. These results suggest that the degree of empathy in women influences their ability to detect deception and supports recent studies showing that emotional skills negatively affect deception detection ability. We suggest that less empathic women are less affected by emotional contagion and thus may be more able to focus on non-emotional cues that might reveal deception.

KEYWORDS:

• Empathy
• gender
• deception detection

Acknowledgement

This research benefited from a LabEx Cortex grant and a Université Lyon 2 doctoral grant. We would like to thank Cécile Reineri for checking the quality of the English in this article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Step 1: Individual proportions of correct responses were adjusted as a function of variables-to-control (degree of Psychopathy, Machiavellianism and Narcissism) by dividing each individual proportion by the covariates. Step 2: The individual number of trials was adjusted by dividing it by the covariates. Step 3: The proportions of adjusted correct responses were pooled across the whole sample. Step 4: The adjusted number of trials was pooled across the whole sample. Step 5: The pooled adjusted proportion of correct responses is multiplied by the unadjusted number of trials, then divided by the adjusted number of trials. Step 6: The adjusted polled proportions of correct responses and the unadjusted number of trials were used for the analyses.

2. In the case of signal detection, the control of the covariates was strictly identical to the procedure used for the proportion of correct responses except that, this time, the adjustment was carried out separately on the proportion of hits and false alarms before carrying out the signal detection analyses.

Additional information

Funding

This work was supported by the Université Lumière Lyon 2 [doctoral grant] (2015–45) awarded to the first author and by the LABEX CORTEX (ANR-11 -LABX-0042) of the Université de Lyon, part of the “Investissements d’Avenir” program (ANR-11-IDEX-0007) run by the French National Research Agency (ANR).