Logo: to the web site of Uppsala University

Machine learning of high-dimensional models have received attention for their ability to predict psychological variables, such as personality. However, it has been less examined to what degree such models are capable of generalizing across domains. Across two text domains (Reddit message and personal essays), compared to low-dimensional- and theoretical models, atheoretical high-dimensional models provided superior predictive accuracy within but poor/non-significant predictive accuracy across domains. Thus, complex models depended more on the specifics of the trained domain. Further, when examining predictors of models, few survived across domains. We argue that theory remains important when conducting prediction-focused studies and that research on both high- and low-dimensional models benefit from establishing conditions under which they generalize.

Objective: Although violent extremism is often attributed to clinical (dysfunctional) dispositions, it is also possible that violent Jihadists might be clinically "normal" but bear certain personality signatures. This alternative view has yet to be tested.

Method: In six studies, employing hard-to-reach Muslim samples, including one study of former Mujahideen, we investigated the relationship between basic personality traits and violent extremism. We further used a known group paradigm to validate the personality signatures of violent extremism, comparing a sample of former Mujahideen with another sample from Afghanistan.

Results: These studies and an internal meta-analysis revealed that Lower Openness to Experience, lower Emotionality, and lower Altruism were associated with more violent intentions to support Muslims. Higher Altruism was associated with higher levels of nonviolent intention to support Muslims. Supporting the validity of the nonviolent intention measure, similar associations were found in Study 3 with overt behavioral support of Muslims (donations). More important, compared to the nonMujahideen, the Mujahideen sample scored lower on average on, for instance, Openness, indicating that these results go beyond self-reported, findings.

Conclusion: We demonstrated that personality predicts violent and nonviolent defense of Muslims among four general populations of Muslims living in the West and in Asia (including the Middle East), and a sample of Mujahideen in Afghanistan.

Collective action is often equated with progressive politics, but are there aspects of group mobilisations that generalise across contexts? We examine general social and personality psychological factors behind endorsement of group-based violence across different types of violent group mobilisation. Specifically, we focus on the endorsement of group-based violence amongst supporters of the Black Lives Matter (BLM) movement (N = 394), an immigration-critical group (N = 252), and soccer supporters (N = 445). Across three preregistered studies, we tested an integrative model including personality and social psychological factors. Several effects were consistent across all three contexts, with group-based relative deprivation positively, and honesty-humility negatively, predicting support for violence. Further, amongst BLM supporters and the immigration-critical group, emotionality negatively predicted support for violence, violent intentions, and self-reported aggression/violence. Overall, our results suggest that individuals who endorse violence in different contexts have some psychological factors in common.

Harmful communication, such as toxic language and threats directed toward individuals or groups, is a common problem on most social media platforms and online spaces. While several approaches exist for detecting toxic language and threats in English, few attempts have detected such communication in Swedish. Thus, we used transfer learning and BERT to train two machine learning models: one that detects toxic language and one that detects threats in Swedish. We also examined the intersection between toxicity and threat. The models are trained on data from several different sources, with authentic social media posts and data translated from English. Our models perform well on test data with an F1-score above 0.94 for detecting toxic language and 0.86 for detecting threats. However, the models' performance decreases significantly when they are applied to new unseen social media data. Examining the intersection between toxic language and threats, we found that 20% of the threats on social media are not toxic, which means that they would not be detected using only methods for detecting toxic language. Our finding highlights the difficulties with harmful language and the need to use different methods to detect different kinds of harmful language.

Human language has many functions. Our communication on social media carries information about how we relate to ourselves and others, that is our identity, and we adjust our language to become more similar to our community - in the same way as we dress and style and act to show our commitment to the groups we belong to. Within a community, members adopt the community's language, and the common language becomes a unifying factor. In this paper, we explore the possibilities of identifying linguistic alignment - that individuals adjust their language to become more similar to their conversation partners in a community. We use machine learning to detect linguistic alignment to a number of different ideologies, communities, and subcultures. We use two different approaches: transfer learning with RoBERTa and traditional machine learning using Random forest and feature selection.

There is growing recognition that members of structurally advantaged groups experience group-based relative deprivation. We consider the idea that personality may explain these "entitlement-based" feelings of deprivation. Specifically, we predicted that modesty would be negatively associated with group-based relative deprivation among members of advantaged groups, but not amongst disadvantaged groups. Two studies focusing on White and Black Americans (N = 334), and Men and Women (N = 309) showed that modesty interacted with group membership. Modesty was negatively related to group-based relative deprivation amongst White Americans and men, but not amongst Black Americans and women. The findings help explain why some individuals espouse rhetoric that their group is being disfavored, even when group statistics and history suggest otherwise.

Suicide is a global health problem with more than 700,000 individuals dying by self-destruction each year, yet it is classified as a low base rate behavior that is difficult to prognosticate. Aiming to advance suicide prediction and prevention, we examined the potential use of machine learning and text analyses models to predict suicide risk based on written communications. Specifically, we used a dataset consisting of more than 27,000 general writings unrelated to suicide, 193 genuine suicide notes from individuals who committed suicide, and an additional 89 suicide posts shared on sub-Reddits for an in-the-wild test to examine the prediction accuracy of two machine learning models (SVM & RoBERTa) and a linguistic marker model. Our tests showed that the machine learning models performed better than the linguistic marker model when examined on the test data. However, the linguistic marker model achieved higher results in the wild, correctly classifying 88% of written communications as a "high risk of suicide" versus 56% and 70% of the machine learning models. The best in-the-wild performing model was adopted in an online suicide risk assessment tool called Edwin to honor Edwin Shneidman for his numerous contributions to the field of suicidology. Finally, discrepancies between training and real-world data, vocabulary variation across domains, and the limited number of benchmarks constitute limitations that need to be addressed in future research.

One of the most challenging threats towards the security of the society is attacks from violent lone offenders, individuals that act alone or with minimal help from others without any economic gains or direct orders from organizations. Over the past few years, several terror attacks have been accompanied by manifestos published on social media platforms that outline ideology, motivation, and in some cases tactical choices. The trend in publishing manifestos and other communication on social media sites before committing an attack has increased the need for threat assessment in digital environments. Most existing methods for threat assessment are developed to be used in offline settings where information about an individual is accessible and cases where the individual is present and can answer questions. In this paper, we present seven indicators that can be used to assess the potential threat of violence based on digital communication only. The seven indicators are designed to be used when analyzing texts and can be seen as a complement to other risk assessment protocols.

The HEXACO Personality Inventory-Revised (HEXACO-PI-R) has become one of the most heavily applied measurement tools for the assessment of basic personality traits. Correspondingly, the inventory has been translated to many languages for use in cross-cultural research. However, formal tests examining whether the different language versions of the HEXACO-PI-R provide equivalent measures of the 6 personality dimensions are missing. We provide a large-scale test of measurement invariance of the 100-item version of the HEXACO-PI-R across 16 languages spoken in European and Asian countries (N = 30,484). Multigroup exploratory structural equation modeling and confirmatory factor analyses revealed consistent support for configural and metric invariance, thus implying that the factor structure of the HEXACO dimensions as well as the meaning of the latent HEXACO factors is comparable across languages. However, analyses did not show overall support for scalar invariance; that is, equivalence of facet intercepts. A complementary alignment analysis supported this pattern, but also revealed substantial heterogeneity in the level of (non)invariance across facets and factors. Overall, results imply that the HEXACO-PI-R provides largely comparable measurement of the HEXACO dimensions, although the lack of scalar invariance highlights the necessity for future research clarifying the interpretation of mean-level trait differences across countries.

Objective: When dealing with the impact of malocclusion on self-esteem, the terms global and dental self-esteem are sometimes used. Although these terms are related to one another, they do not depict the same concept. The aims of this paper were to explore if the two forms of self-esteem are distinguishable, to find out if they represent different factors, and to investigate how they are related to malocclusion. Materials and methods: A sample consisting of 150 adolescents, aged 13 years, completed self-assessed measures of Dental and Global Self-Esteem. Orthodontic treatment need for each individual was assessed by the Dental Health Component of the Index of Orthodontic Treatment Need (IOTN-DHC). Data were analysed by factor analyses and a 5 (IOTN-DHC grades) by 2 (global vs. dental self-esteem) ANOVA, with the IOTN-DHC grades as the independent and self-esteem (repeated measure) as the dependent variables. Results: The factor analyses showed that the two forms of self-esteem, based on the measures, are distinguishable. More importantly, the results of the ANOVA revealed that Dental and Global Self-Esteem are differentially related to IOTN-DHC. Specifically, Dental Self-Esteem varied across IOTN-DHC scale while Global Self-Esteem did not. There was no effect of gender. Conclusions: Dental self-esteem is related to malocclusion while global self-esteem is not. These findings have implications in areas where the predictive power of dental self-esteem needs to be considered.