BookR: School of Health Core Facilities Booking

The problem: People use social media platforms to chat, search, and share information, express their opinions, and connect with others. But these platforms also facilitate the posting of divisive, harmful, and hateful messages, targeting groups and individuals, based on their race, religion, gender, sexual orientation, or political views. Hate content is not only a problem on the Internet, but also on traditional media, especially in places where the Internet is not widely available or in rural areas. Despite growing awareness of the harms that exposure to hate can cause, especially to victims, there is no clear consensus in the literature on what specific impacts this exposure, as bystanders, produces on individuals, groups, and the population at large. Most of the existing research has focused on analyzing the content and the extent of the problem. More research in this area is needed to develop better intervention programs that are adapted to the current reality of hate.

Objective: The objective of this review is to synthesize the empirical evidence on how media exposure to hate affects or is associated with various outcomes for individuals and groups.

Search methods: Searches covered the period up to December 2021 to assess the impact of exposure to hate. The searches were performed using search terms across 20 databases, 51 related websites, the Google search engine, as well as other systematic reviews and related papers.

Selection criteria: This review included any correlational, experimental, and quasi-experimental study that establishes an impact relationship and/or association between exposure to hate in online and traditional media and the resulting consequences on individuals or groups.

Data collection and analysis: Fifty-five studies analyzing 101 effect sizes, classified into 43 different outcomes, were identified after the screening process. Initially, effect sizes were calculated based on the type of design and the statistics used in the studies, and then transformed into standardized mean differences. Each outcome was classified following an exhaustive review of the operational constructs present in the studies. These outcomes were grouped into five major dimensions: attitudinal changes, intergroup dynamics, interpersonal behaviors, political beliefs, and psychological effects. When two or more outcomes from the studies addressed the same construct, they were synthesized together. A separate meta-analysis was conducted for each identified outcome from different samples. Additionally, experimental and quasi-experimental studies were synthesized separately from correlational studies. Twenty-four meta-analyses were performed using a random effects model, and meta-regressions and moderator analyses were conducted to explore factors influencing effect size estimates.

Results: The 55 studies included in this systematic review were published between 1996 and 2021, with most of them published since 2015. They include 25 correlational studies, and 22 randomized and 8 non-randomized experimental studies. Most of these studies provide data extracted from individuals (e.g., self-report); however, this review includes 6 studies that are based on quantitative analysis of comments or posts, or their relationship to specific geographic areas. Correlational studies encompass sample sizes ranging from 101 to 6829 participants, while experimental and quasi-experimental studies involve participant numbers between 69 and 1112. In most cases, the exposure to hate content occurred online or within social media contexts (37 studies), while only 8 studies reported such exposure in traditional media platforms. In the remaining studies, the exposure to hate content was delivered through political propaganda, primarily associated with extreme right-wing groups. No studies were removed from the systematic review due to quality assessment. In th