Ce travail explore l’intérêt de l’utilisation d’une technique électrophysiologique (la réponse électrodermale [RED]) pour évaluer les effets émotionnels en lien aux souvenirs des attentats de janvier 2015 à Paris. L’analyse des résultats obtenus permet de pointer les limites et l’impact de l’interprétation des paramètres psychophysiologiques appliqués en psychologie expérimentale. Ce travail conclut sur l’importance de stimuler l’esprit critique des futurs psychologues, face à la fascination qu’exercent certains outils psychophysiologiques.

This work focuses on the importance of critical thinking by future psychologists, faced with psychophysiological tools. Galvanic skin response is related to the functioning of sweat glands innervated by the sympathetic nervous system chain (Lykken, 1959). Dermal conductance is a good indicator of emotional processing of stimuli (Bechara et al., 2002), and there is a large body of clinical work studying these unconscious reactions such as the evaluation of pain (Naveteur, 2008), anxiety disorders (Stankovic et al., 2014), in marketing (Droulers et al., 2013), and some police investigations (Breska et al., 2014; Honts et al., 2015). In the laboratory, these recordings are made in a standardized and rigorous manner (Fowles et al., 1981; Boucsein, 2011). However, this is inconsistent with “on the ground” observations. Some electronic equipment manufacturers offer mobile sensors, but this work calls into question these devices by showing EDA interpretation errors (Picard et al., 2015). To measure the autonomic nervous index, the unconscious emotions to words associated with the series of attacks in January 2015 were analyzed.

A group of 68 second-year psychology students, aged 19–23years old, participated in this study. Electrodes were attached to the palms of both hands and connected to an operational amplifier. The subject was asked to relax, and then the experimenter randomly pronounced a word from a list. After an interval of 1 to 3seconds, the amplitude of EDA was displayed. Only 6 words were retained in this experience: “violence”, “attack”, “freedom”, “humor”, “kill” and “religion”. It provided a simplified analysis of the results using a ordinal scale, ranking the EDA “zero” (0), the EDA “weak” (1), EDA “average” (2) and EDA “strong” (3). Each individual used the same scale (0–3) and noted their attitude regarding the attack on 7 January 2015. This survey classified the group into three samples: those who were upset (level 3, category “Charlie”: 21 subjects), those affected moderately (level 2: 31 subjects), and those who remained indifferent to this event (scales 0 to 1, “non-Charlie”: 17 subjects).

The analysis focused on the relationship between the intensity of emotion felt during the attacks in January and galvanic skin response for each word. Comparing the average EDA by topic, ANOVA indicated a difference between the three groups (F(1–67)=5.34, P=0.01). A clear difference was observed between “Charlie” and “non-Charlie” (t(36)=3.20, P<0.01), the reaction being greater in “Charlie”. A graph showing the average EDA showed a notable difference between the two categories for the word “freedom” (t(36)=3.14, P<0.01). There was also a difference in the word “humor” (t(36)=2.26, P<0.06). Finally, a correlation analysis was performed by test words and a difference was observed according to the categories.

Despite the simplification of hardware from galvanic recording, unconscious reactions were obtained in this population of students. But, in general, results using EDA are especially valuable in large samples (Ben-Shakhar, 2012). This experimental study simply checked statistical variations related to the effects of current events in the study population. Care must be taken with these analyzes because this measure remains delicate. It is common to find in some articles illusory interpretations of the psyche (Poirel, 1994). In this world of “big data”, new devices are attractive but sometimes they give an impression of omnipotence among researchers who use powerful tests, but are offset by the object of study. Pedagogically, students (and future researchers) must master the psychophysiological parameters and key statistical concepts, but should also be aware of the flaws in these measures.