Exploring the Perspectives of Social VR-Aware Non-Parent Adults and Parents on Children's Use of Social Virtual Reality

TL;DR

This study investigates how parents and non-parent adults perceive children's use of social VR and the safeguarding measures needed. Using a mixed-methods online questionnaire (N=149; 79 parents, 70 non-parents), it reveals that 43% of under-13 children use social VR at least every two weeks, despite age restrictions, and that familiarity with VR and parental status shape perceptions of age-appropriateness and supervision. Qualitative findings expose pervasive concerns about harassment, immaturity, and lack of safeguards, while respondents propose a spectrum of interventions spanning pre-planned controls, real-time oversight, and post-event insight. The work highlights the need for evidence-based guidelines and age-appropriate safety tools to harmonize parent-child trust with platform safety, informing future design and policy in social VR environments.

Abstract

Social Virtual Reality (VR), where people meet in virtual spaces via 3D avatars, is used by children and adults alike. Children experience new forms of harassment in social VR where it is often inaccessible to parental oversight. To date, there is limited understanding of how parents and non-parent adults within the child social VR ecosystem perceive the appropriateness of social VR for different age groups and the measures in place to safeguard children. We present results of a mixed-methods questionnaire (N=149 adults, including 79 parents) focusing on encounters with children in social VR and perspectives towards children's use of social VR. We draw novel insights on the frequency of social VR use by children under 13 and current use of, and future aspirations for, child protection interventions. Compared to non-parent adults, parents familiar with social VR propose lower minimum ages and are more likely to allow social VR without supervision. Adult users experience immaturity from children in social VR, while children face abuse, encounter age-inappropriate behaviours and self-disclose to adults. We present directions to enhance the safety of social VR through pre-planned controls, real-time oversight, post-event insight and the need for evidence-based guidelines to support parents and platforms around age-appropriate interventions.

Figures (4)

• Figure 1: Mixed-methods questionnaire structure with information on question types and example items for each of the topics.
• Figure 2: Likert Scales showing percentages of children using (a) VR in general or (b) Social VR in our sample. Among children under 13 years, 24% use social VR daily or weekly and 24% use VR in general daily or weekly. 43% use social VR at least every two weeks.
• Figure 3: Violin and boxplots of inputted ages by parent and non-parent adult subsets corresponding to the minimum appropriate ages in social VR for three different supervision states (a, b, c). We conducted a three-way Aligned Rank Transform ANOVA (significant level 0.05) with age as the dependent variable and with parenthood groups (non-parent adult and parent), subsets (familiarity with VR and social VR) and questions (supervision states) as the factors involved. The main effect of familiarity with VR and social VR is statistically significant ($F$(1,54)=5.62, $p$= 0.022$<$0.05, $\eta_{p}^{2}$=.1, medium effect size). The main effect of parenthood is statistically significant ($F$(1,54)=5.60, $p$= 0.021$<$0.05, $\eta_{p}^{2}$=.1, medium effect size). The main effect of supervision states (Q1) without supervision, Q2) with supervision, Q3) only minors) is statistically significant ($F$(2,108)=13.49, $p$= 5.9e-06$<$0.01, $\eta_{p}^{2}$=.2, large effect size). The interaction between familiarity and supervision states is statistically significant ($F$(2,108)=6.14, $p$= 0.003$<$0.01, $\eta_{p}^{2}$=.1, medium effect size). The interaction between parenthood groups and supervision states is statistically significant ($F$(2,108)=5.03, $p$= 0.008$<$0.01, $\eta_{p}^{2}$=.1, medium effect size)
• Figure 4: Interventions presented to the participants and the selected age range. The violin boxplots summarise age ranges selected for four categories of interventions (pre-planned, knowledge, real-time actions and after-the-fact). Each intervention was selected by n participants. The mean of minimum ages given for all interventions is 10 years old and the mean of maximum ages is 15 years old.