Designing social media to foster user engagement in challenging misinformation: a cross-cultural comparison between the UK and Arab countries

The study addresses how to design social media features that encourage users to challenge acquaintances who post misinformation. Motivated by the persistent spread and acceptance of misinformation online and the limited willingness of users to publicly correct it, the authors propose leveraging persuasive system design (PSD) principles and communication theories to nudge prosocial correction behaviors. They highlight that prior work has been WEIRD-centric and that communication styles differ across cultures—particularly high-context (e.g., Arab) versus low-context (e.g., UK)—which may influence the effectiveness of persuasive techniques. The paper replicates and extends Gurgun et al. (2023) by comparing UK and Arab users to test whether perceived persuasiveness of seven design techniques differs across cultures and whether individual characteristics (gender, age, empathy, perspective-taking, Big Five traits) predict perceiving these techniques as more persuasive than a standard comment box. The research questions are: RQ1: Are there cross-cultural differences (UK vs. Arab) in the persuasiveness of social media design techniques to challenge misinformation? RQ2: Do gender, age, empathy, perspective-taking, and personality traits predict perceiving these techniques as more persuasive than a standard comment box within each culture?

The paper situates the problem within extensive research on the impacts of misinformation across political and health domains, cognitive biases such as repetition effects and anchoring, and the promise and limits of social and information interventions. Persuasive System Design (PSD) is introduced as a comprehensive framework (Primary Task Support, Dialogue Support, System Credibility Support, Social Support) with principles such as Reduction, Tunnelling, Self-monitoring, Liking, Suggestion, Verifiability, Normative Influence, and Recognition, which have been successfully applied in behavior change contexts including social media. Prior work (Gurgun et al., 2023) designed seven PSD-inspired techniques to motivate correction of misinformation, but tested only in the UK. Cross-cultural HCI literature and Hall’s high-/low-context communication theory indicate cultural variability in communication preferences: high-context cultures (e.g., Arab) rely more on nonverbal and contextual cues, whereas low-context cultures (e.g., UK) value explicit verbal communication. This suggests design techniques may vary in effectiveness across cultures. Individual differences literature shows personality (e.g., extraversion), empathy, perspective-taking, and age can shape participation in discussions and responsiveness to persuasive strategies. The authors hypothesize cultural context moderates these relationships, requiring culturally aware design.

Design: Online cross-sectional survey in English and Arabic using Qualtrics. Participants were shown eight high-fidelity prototypes modelled on Facebook: seven PSD-informed techniques—Private Commenting (Reduction), Predefined Question Stickers (Suggestion), Tone Detector (Self-monitoring), Fact Checker Badge (Recognition), Social Norm Messages (Normative Influence), Sentence Openers (Tunnelling), Thinking Face Reaction (Liking)—plus a Standard Comment Box (baseline). A scenario presented a widely shared false news article (CNN iReport asteroid story from 2014). The posting account was an acquaintance with a gender-neutral identity. Participants were told the post was false to focus on technique persuasiveness rather than detection. Participants: N=462 (UK=250; Arab=212). Inclusion: age ≥18, fluent in English (UK) or Arabic (Arab), active Facebook account with real identity, have encountered misinformation on Facebook, residing in UK or an Arab country. Recruitment via Prolific (both contexts) and Cint (additional Arab respondents). Ethical approvals obtained in UK and Qatar. Consent obtained. Quality controls: three attention checks (exclusion if ≥2 failed), removal for implausibly short completion time, contradictory answers, patterned responses, or gibberish in open responses. Compensation provided. Measures: Demographics (gender, age, education). Willingness to publicly challenge misinformation (7-point likelihood scale). Perceived impact of each technique on willingness to challenge (7-point scale from “Far too little” to “Far too much”). Personality via BFI-10 (two items per Big Five trait; 1–7 scale). Empathic Concern and Perspective-Taking via IRI subscales (7 items each; 1–5 scale), with internal reliability: Empathy α=0.82 (UK), 0.61 (Arab); Perspective-Taking α=0.79 (UK), 0.71 (Arab). Arabic translation via back-translation. Analysis: Descriptive statistics; Welch’s t-test for cross-cultural difference in baseline willingness to challenge (unequal variances/sizes). Mann–Whitney U tests for cross-cultural differences in perceived persuasiveness for each technique (non-normal distributions). Wilcoxon signed-rank tests to compare each technique vs Standard Comment Box (SCB) within each culture. Binomial logistic regressions predicting whether a given technique was rated more persuasive than SCB (binary difference score: 1=greater than SCB; 0=equal/less) from gender, age, empathy, perspective-taking, and Big Five traits. Linearity checked via Box–Tidwell; Hosmer–Lemeshow indicated good fit. Analyses conducted in JASP v17.

• Baseline willingness to challenge misinformation: Arabs higher than UK. Arab M=4.70, SD=1.80; UK M=3.30, SD=1.94; Welch’s t=8.01, p<0.001; Cohen’s d=0.75 (large).
• Cross-cultural differences in perceived persuasiveness (Mann–Whitney U): Arabs rated most techniques as more persuasive than UK, except Private Commenting (ns). Means (UK vs Arab) and p-values: • Predefined Question Stickers: 4.01 vs 5.10; U=17338; p<0.001; r=0.27. • Thinking Face Reaction: 4.49 vs 4.92; U=22763; p=0.008; r=0.12. • Private Commenting: 4.84 vs 5.02; U=23757; p=0.051; r=0.09 (ns). • Sentence Openers: 3.27 vs 4.61; U=15757; p<0.001; r=0.35. • Fact Checker Badge: 3.66 vs 5.18; U=14954.5; p<0.001; r=0.38. • Social Norm Message: 3.89 vs 4.99; U=17132.5; p<0.001; r=0.31. • Tone Detector: 3.86 vs 4.64; U=20162.5; p<0.001; r=0.21. • Standard Comment Box: 3.66 vs 4.33; U=20989.5; p<0.001; r=0.18.
• Technique vs SCB within-culture (Wilcoxon signed-rank): • UK: More persuasive than SCB: Predefined Question Stickers z=2.43, p=0.015; Thinking Face Reaction z=5.00, p<0.001; Private Commenting z=7.20, p<0.001. Less persuasive than SCB: Sentence Openers z=−2.70, p=0.007. Others ns. • Arab: More persuasive than SCB: Predefined Question Stickers z=−4.46, p<0.001; Thinking Face Reaction z=−3.37, p=0.001; Private Commenting z=−3.64, p<0.001; Fact Checker Badge z=−4.92, p<0.001; Social Norm Message z=−4.12, p<0.001. Tone Detector and Sentence Openers ns.
• Predictors of rating a technique more persuasive than SCB (binomial logistic regressions): • Predefined Question Stickers: Arab—Extraversion OR=1.25, p=0.027 (positive). UK—Age OR=1.04, p<0.01 (positive); Perspective-Taking OR=1.08, p<0.05 (positive); Openness OR=0.81, p<0.01 (negative). • Thinking Face Reaction: Arab—Agreeableness OR=0.79, p=0.019 (negative). UK—Openness OR=0.79, p<0.01 (negative). • Private Commenting: UK—Age OR=0.96, p<0.01 (negative). Arab—no significant predictors. • Social Norm Message (Arab only): Openness OR=1.25, p=0.024 (positive). • Fact Checker Badge (Arab only): no significant predictors. • Sentence Openers (UK only): no significant predictors. Overall: Arabs report higher willingness and greater perceived persuasiveness of most PSD-informed techniques; Private Commenting shows comparable impact across cultures.

Findings support both research questions. RQ1: Arabs exhibited higher willingness to challenge misinformation and perceived most PSD-inspired techniques as more persuasive than UK users, aligning with high-context communication preferences that favor nonverbal and contextual cues. Techniques like Thinking Face Reaction and Predefined Question Stickers, which rely on indirect signaling and structured prompts, resonated more with Arab participants. In contrast, Sentence Openers underperformed relative to SCB in the UK, suggesting that guidance requiring users to craft arguments may not reduce effort sufficiently in low-context cultures valuing explicit, self-composed messages. Private Commenting had similar influence across cultures, likely balancing directness with privacy, accommodating both communication styles. RQ2: Individual differences partly accounted for technique persuasiveness. Age and perspective-taking increased the appeal of Predefined Question Stickers in the UK, while openness reduced it; extraversion increased appeal in the Arab context. Thinking Face Reaction was less appealing for those higher in openness (UK) and agreeableness (Arab), indicating nuanced dispositional sensitivities to indirect expressions of doubt. Younger UK users were more persuaded by Private Commenting. These results underscore the interplay of cultural context and personal traits in designing effective social-correction features. Practically, platforms should culturally tailor persuasive elements—leveraging badges and normative cues in high-context settings and emphasizing clarity and ease for low-context users—while offering broadly acceptable options like Private Commenting.

The study demonstrates that persuasive design techniques can increase users’ willingness to challenge misinformation and that their effectiveness varies by cultural communication style and individual characteristics. Except for Private Commenting, all techniques had greater impact in the Arab context than in the UK. Compared to a standard comment box, Predefined Question Stickers, Thinking Face Reaction, and Private Commenting were beneficial in both cultures; Sentence Openers were less effective in the UK; Fact Checker Badge and Social Norm Messages were more effective only in the Arab context. Designers should adopt culturally aware strategies that align with high-/low-context communication preferences and consider user traits such as age and personality. Future work should move beyond self-reports to behavioral measures, explore additional cultures, and evaluate new interventions (e.g., gamification, nudges, co-creation) tailored to cultural communication styles.

The sample focused on active social media users willing to engage in public discourse, potentially excluding passive users and limiting generalizability. The study centered on challenging acquaintances on Facebook, not varying relational closeness. Outcomes relied on self-reported perceptions rather than observed behavior, subject to biases (e.g., social desirability, recall). Within-culture heterogeneity (regional, socioeconomic, demographic) was not examined. The context was social media only, and data were cross-sectional. Future research should employ behavioral measures in experimental/observational designs, assess relational closeness and within-culture differences, expand to other platforms and cultures, integrate nudges and gamification, and examine constructs like need for cognition and risk preferences.