Ethnophaulisms and Exclusion: The Behavioral Consequences of Cognitive Representation of Ethnic Immigrant Groups

Multimodal detection of hateful memes by applying a vision-language pre-training model

Detrimental to individuals and society, online hateful messages have recently become a major social issue. Among them, one new type of hateful message, “hateful meme”, has emerged and brought difficulties in traditional deep learning-based detection. Because hateful memes were formatted with both text captions and images to express users’ intents, they cannot be accurately identified by singularly analyzing embedded text captions or images. In order to effectively detect a hateful meme, the algorithm must possess strong vision and language fusion capability. In this study, we move closer to this goal by feeding a triplet by stacking the visual features, object tags, and text features of memes generated by the object detection model named Visual features in Vision-Language (VinVl) and the optical character recognition (OCR) technology into a Transformer-based Vision-Language Pre-Training Model (VL-PTM) OSCAR+ to perform the cross-modal learning of memes. After fine-tuning and connecting to a random forest (RF) classifier, our model (OSCAR+RF) achieved an average accuracy and AUROC of 0.684 and 0.768, respectively, on the hateful meme detection task in a public test set, which was higher than the other eleven (11) published baselines. In conclusion, this study has demonstrated that VL-PTMs with the addition of anchor points can improve the performance of deep learning-based detection of hateful memes by involving a more substantial alignment between the text caption and visual information.

Exposure to hate speech increases prejudice through desensitization

In three studies (two representative nationwide surveys, N = 1,007, N = 682; and one experimental, N = 76) we explored the effects of exposure to hate speech on outgroup prejudice. Following the General Aggression Model, we suggest that frequent and repetitive exposure to hate speech leads to desensitization to this form of verbal violence and subsequently to lower evaluations of the victims and greater distancing, thus increasing outgroup prejudice. In the first survey study, we found that lower sensitivity to hate speech was a positive mediator of the relationship between frequent exposure to hate speech and outgroup prejudice. In the second study, we obtained a crucial confirmation of these effects. After desensitization training individuals were less sensitive to hate speech and more prejudiced toward hate speech victims than their counterparts in the control condition. In the final study, we replicated several previous effects and additionally found that the effects of exposure to hate speech on prejudice were mediated by a lower sensitivity to hate speech, and not by lower sensitivity to social norms. Altogether, our studies are the first to elucidate the effects of exposure to hate speech on outgroup prejudice.

Sticks and Stones Can Break My Bones, But Ethnophaulisms Can Alter the Portrayal of Immigrants to Children

An archival study examined the portrayal of ethnic immigrants to children as a function of the prevailing cognitive representation of those ethnic immigrant groups in ethnophaulisms. The complexity in ethnophaulisms (and, to a lesser degree, the valence in ethnophaulisms) predicted the portrayal of ethnic immigrant groups. Overall, ethnic immigrant groups characterized in terms of ethnophaulisms of low complexity were less frequently present in children’s literature, children from these ethnic groups were described more in terms of physical appearance than in terms of personal traits, fictional child characters from those ethnic groups were portrayed with smaller heads and with lower verbal complexity, and the folksongs attributed to these groups had a more negative affective tone. The implications of these results for approaches to intergroup relations are considered.

Negative emotions versus target descriptions: Examining perceptions of racial slurs as expressive and descriptive

There is a debate about whether racial slurs operate primarily as descriptives (of the ethnicity of targets) or expressives (of negative emotions toward targets). In three studies (overall N = 471), we examined whether different racial slurs used in different situations led to slurs being perceived as descriptive versus expressive, and whether this distinction was related to the perceived offensiveness of the slurs. Our results showed the descriptive and expressive natures of racial slurs are directly related to their perceived offensiveness. Specifically, as the perceived offensiveness of slurs increase in intensity, the slurs are perceived as more negatively expressive, more descriptive, less positively expressive, and comparatively less descriptive and more expressive.

Labelling and discrimination: do homophobic epithets undermine fair distribution of resources?

This research investigated the behavioural consequences of homophobic epithets. After exposure to either a category or a homophobic label, heterosexual participants allocated fictitious resources to two different prevention programmes: one mainly relevant to heterosexuals (sterility prevention), the other to homosexuals (AIDS-HIV prevention). Responses on allocation matrices served to identify strategies that favoured the ingroup over the outgroup. Results indicated stronger ingroup-favouritism in the homophobic than in the category label condition. This study shows that discriminatory group labels have tangible effects on people's monetary behaviours in intergroup contexts, increasing their tendency to favour the ingroup when distributing resources.