A serious problem confronts the psychologist whose patient threatens, within the privacy of a therapy session, to inflict violent harm upon some third person. Therapists in Texas face a risk of unjust legal liability because of a lack of widely accepted, clearly and fully articulated standards. A questionnaire was submitted to Texas psychologists and Texas judges of mental illness courts. It involved a hypothetical case of a patient who threatened to kill his girlfriend. The hypothesis that no consensus exists at present among psychologists or judges appears to be supported by the data. Comparisons are made of the attitudes of psychologists and judges. Correlations between psychologist attitudes and certain demographic and practice variables are reported. The need for new legislation in Texas concerning legal liability of therapists for the violent behavior of patients is discussed. Proposed legislation for Texas is set out. Among its important features are (1) recognition that continued therapy is itself a protective strategy and (2) establishment of good faith as the standard by which the behavior of the therapist is to be judged.

Les réseaux sociaux et les mediums de télécommunication ont donné naissance à une abondance de propos menaçants par lesquels sont véhiculés des messages de haine qui sont possiblement annonciateurs d’une violence future (Laforest & al., 2017). À ce jour, les études ne permettent pas de décrire le contexte dans lequel les menaces proférées par un medium technologique s’insèrent. Elles ne permettent pas non plus de comprendre les facteurs qui influencent les sentences de tels comportements. À partir d’une approche méthodologique mixte et d’un échantillon de 93 décisions judiciaires canadiennes, l’objectif principal de ce mémoire consiste, dans un premier temps, à décrire les facteurs circonstanciels entourant les menaces de mort et de lésions corporelles à l’ère numérique. En général, les résultats montrent que les menaces qui ont donné lieu à une décision écrite présentent un risque considérable de se concrétiser. De plus, nous avons découvert que même si dans certaines situations ce sont des témoins qui ont dénoncé les menaces aux autorités, ce sont principalement les victimes elles-mêmes qui dénoncent le plus souvent leur assaillant, craignant pour leur sécurité et celle de leurs proches. Dans un deuxième temps, nous cherchions à comprendre la réponse juridique canadienne pour de telles infractions. Plus précisément, un test de régression logistique a été élaboré afin de comprendre les facteurs contextuels qui influencent le recours à une sentence d’emprisonnement. Les résultats de cette analyse montrent que parmi les facteurs susceptibles d’influencer de telles sentences, quatre se sont avérés significatifs : (1) un lien conjugal avec la victime, (2) l’utilisation des médias sociaux, (3) la consommation d’alcool ou de drogues et (4) la gravité de la menace. Finalement, à partir de nos analyses, nous pouvons conclure que de nombreux enjeux subsistent dans le processus de détermination de la peine pour les menaces proférées par un medium technologique. En effet, plusieurs éléments peuvent interférer dans l’analyse du juge, notamment la crédibilité de la victime et l’intention criminelle de l’accusé. / Social networks and telecommunications tools have risen an abundance of threatening words through which violent messages are conveyed. They are possibly predictors of future violence (Laforest & al., 2017). To date, studies do not allow us to describe the context in which the threats uttered by a technological medium occur. They also do not allow us to understand the factors that influence the sentences of such behaviors. Based on a mixed methodological approach and a sample of 93 Canadian court decisions, the main objective of this study consists, firstly, in describing the circumstances factors surrounding threats of death and bodily harm in the digital age. Overall, the results show that threats that resulted in a written decision have a significant risk of materializing. Furthermore, we discovered that although in some situations it was witnesses who reported the threats to the authorities, it was mainly the victims themselves who most often reported their assailant, fearing for their safety and that of their loved ones. Second, we sought to understand the Canadian legal responses for such offences. More specifically, a logistic regression test was developed to understand the contextual factors that influence the use of a prison sentence. The results of this analysis show that among the factors likely to influence such sentences, four were found to be significant: (1) a marital relationship with the victim, (2) the use of social media to make the threat, (3) the use of alcohol or drugs and finally (4) the seriousness of the threat. Finally, based on our analyses, we can conclude that that many issues remain in the sentencing process for threats made by a technological medium. Indeed, several elements can interfere in the judge’s analysis, in particular the credibility of the victim and the criminal intentions of the accused.