PaperlayerView exact revision on arXiv

Source-linked AI summary

Exploring spatial configuration and group familiarity when co-watching 360-degree stories in social virtual reality

Portia Wang, Monique Santoso, Yujie Tao, Jeremy Bailenson

arXiv:2610.30353v1cs.HC

TL;DR

How spatial arrangements and social relationships shape attention and experience during immersive co-watching remains underexplored, despite the social character of cinematic viewing. Three experiments examined boundary type, personal-space size, and group familiarity in social VR. Results show tradeoffs between social or narrative engagement and workload or discomfort, alongside associations between familiarity and reported experience.

Abstract

from arXiv · show

While traditional cinematic viewing is often characterized as a social experience, engaging with immersive stories is predominantly researched as an individual activity. This mismatch becomes increasingly problematic in social virtual reality (VR), where features designed for typical social activities (e.g., virtual locomotion, spatial and social context) can become meaningful levers for shaping the user experience. To address our limited understanding of these effects, we conducted three exploratory experiments (N$_1$=109, N$_2$=116, N$_3$=88) examining how spatial boundary type, size of allocated personal space, and group familiarity influence co-watching 360-degree stories in social VR. Results shows that small communal spatial boundaries heightened social presence towards others compared to individual boundaries, but also increased mental demand. Reducing the size of the personal spatial boundary increased social presence towards others but also increased disorientation and frustration. Finally, participants with higher group familiarity reported stronger attentional focus on the story, as well as higher spatial presence and social presence towards story characters, but also reported greater physical and perceptual demand. We discuss systematic tradeoffs between engagement and cognitive effort, and suggest practical guidelines for designing shared immersive narrative experiences.

I. INTRODUCTION

Immersive co-watching makes social and spatial context part of the viewing experience, yet how people negotiate space and attention—and how relationships shape that negotiation—remains underexplored. This paper examines those questions through three experiments and frames co-watchers as potential competing or reinforcing cues for attention and narrative engagement.

I. INTRODUCTION

  • How space and attention are negotiated during co-watching, and how social relationships shape that negotiation, remains largely unexplored.Virtual co-watchers may be perceptually salient, while movement and avatar presence can complicate attention to the story and to others.Source: §I · p. 1
  • Three experiments examine how spatial configuration and group familiarity shape psychological and behavioral outcomes during immersive co-watching of 360-degree stories.The first two experiments examine spatial configuration; the third explores group familiarity.Source: §I · p. 2

A. Theoretical Perspectives of Watching Immersive Stories

  • In immersive 360-degree stories, viewers can change their perspective by looking around, and attention often converges on shared centers of interest and salient cues.Prior research describes this convergence as attentional synchrony.Source: A. Theoretical Perspectives of… · p. 2
  • Co-watchers may divert attention from the narrative or guide it through gaze following and shared focus, while narrative engagement involves sustained effort and attention.The theoretical account also allows that co-watchers could guide attention to salient areas and lower cognitive demand.Source: A. Theoretical Perspectives of… · p. 2
  • Co-watching creates tension between narrative engagement and the social negotiation and maintenance required when watching with others.Prior work suggests realistic avatars can enhance social presence, but how social cues interplay with media consumption remains unclear.Source: A. Theoretical Perspectives of… · p. 2
  • The paper addresses practitioners’ limited design guidance with three experiments examining the psychological and behavioral effects of spatial and social contexts.The authors describe the topic as lacking empirical work.Source: A. Theoretical Perspectives of… · p. 2

B. Spatial Design of Co-watching Experiences

Spatial arrangements affect how people occupy shared environments, and immersive co-watching adds distinctive demands because viewers look around while virtual movement changes their arrangement relative to one another and the story. The paper motivates studies of spatial design and familiarity as complementary factors in this experience.

B. Spatial Design of Co-watching Experiences

  • In 360-degree co-watching, movement changes users’ positions relative to co-watchers and the story, not the story viewpoint, creating opportunities to avoid occlusion or align with camera movement.Avatars positioned inconsistently with the story world may also create perceptual conflict.Source: B. Spatial Design of Co-watching… · p. 3

C. Social Context Shaping Immersive Experiences

  • Prior relationships may shape immersive co-watching, but little research directly informs how they do so, leaving practitioners with limited design guidance across social contexts.The paper’s second research question addresses how familiarity relates to presence, workload, and interpersonal distance.Source: C. Social Context Shaping Immersive… · p. 3
  • Familiarity may reduce social regulation and free attention for the task, but familiarity bias may instead draw attention toward interaction with known co-watchers.The paper presents these as competing possibilities for immersive co-watching.Source: C. Social Context Shaping Immersive… · p. 3

D. Motivations for Prototyping Spatial Configurations

  • Communal spaces can leave users scattered or require active proxemic coordination, while personal boundaries can limit movement and make occlusion harder to resolve.These tradeoffs motivate comparing boundary configurations for immersive co-watching.Source: D. Motivations for Prototyping Spatial… · pp. 3–4
  • Spatial boundaries range from large communal spaces to personal boundaries; a small communal configuration offers an intermediate condition between them.The three boundary types motivate Experiment 1.Source: D. Motivations for Prototyping Spatial… · p. 4
  • Personal boundaries provide a standardized, scalable configuration that can reduce within-group variation in movement and interpersonal distance.They also make it easier to relate spatial characteristics to psychological outcomes across designs.Source: D. Motivations for Prototyping Spatial… · p. 4

III. METHODS

The study uses three between-subjects experiments with university-course participants to examine spatial boundaries, personal-space size, and group familiarity during social-VR co-watching. Shared procedures use ENGAGE XR, 360-degree stories, and custom virtual spatial configurations.

III. METHODS

  • Three pre-registered, between-subjects experiments varied the examined factors at two or three levels to study co-watching 360-degree stories in VR.The experiments examined spatial boundary type, personal-space size, and group familiarity.Source: §III · p. 4
  • Reserved seating configurations accommodated up to 16 participants in a four-by-four grid, with users moving inside outlined spatial constraints.The figure shows movement by teleportation or smooth translation.Source: Figure 2 · p. 5

A. Participants

  • The experiments recruited university-course participants, with samples of 109, 116, and 88, respectively.Participants were divided into groups that gathered in the same physical space for onboarding and co-watching.Source: A. Participants · p. 5

B. Study Setup and Equipment

  • Participants used Meta Quest 2 headsets and controllers in ENGAGE XR, where 360-degree stories were rendered as skyboxes and users could move physically or with controller-based translation and teleportation.Translation changes spatial arrangement, while head rotations change the story viewpoint.Source: B. Study Setup and Equipment · p. 5

C. 360-Degree Cinematic Stories

The study sampled five monoscopic 360-degree stories, distributing the first three across Experiments 1 and 2 and the final two across Experiment 3. A four-week procedure included onboarding and repeated group co-watching sessions under assigned spatial conditions.

C. 360-Degree Cinematic Stories

  • The study sampled five 360-degree cinematic story pieces of comparable length to increase generalizability; the first three were used in Experiments 1 and 2, and the final two in Experiment 3.Source: C. 360-Degree Cinematic Stories · p. 5, Table I · p. 6
  • The stories were selected to control duration, exclude single-user or horror content, and include virtual characters.Monoscopic 360-degree video was chosen for its prominence, sampling feasibility, and compatibility with tracking needs; stereoscopic content caused discomfort and motion sickness during pilot testing.Source: C. 360-Degree Cinematic Stories · p. 6
  • In Experiments 1 and 2, each group watched two of three stories, with each story equally sampled across conditions; Experiment 3 added two stories to prevent stimulus repeats after reshuffling.Source: C. 360-Degree Cinematic Stories · p. 6

D. Procedure

  • The four-week procedure included pre-study consent and demographics, week-one onboarding, and co-watching sessions in the following weeks.Onboarding included avatar creation and instruction on movement and looking around in VR.Source: D. Procedure · p. 7
  • In week two, participants moved to assigned virtual markers, familiarized themselves with the space, watched a story, and completed a post-VR questionnaire.Reserved-seating participants used individual 1×1 meter markers, while box and general-admission groups used a communal 20×20 meter marker.Source: D. Procedure · p. 7
  • The week-two protocol was repeated in weeks three and four, with spatial configurations set by group condition and some groups reshuffled in week four.Shuffled groups met in separate physical rooms for the co-watching exercises.Source: D. Procedure · p. 7

E. Self-Report Measures

The study combined post-experiment self-reports with VR-session behavioral measures to characterize presence, attention, workload, motion sickness, and group spatial dynamics.

E. Self-Report Measures

  • Social and spatial presence were measured with three-item Likert scales, with social presence assessed separately toward other users and story characters.The social-presence composites had α=.84 for users and α=.76 for characters; spatial presence had α=.78.Source: E. Self-Report Measures · p. 7
  • Story attention used an adapted three-item Narrative Engagement Scale, while low-reliability measures of attention to and from other users were analyzed item by item.The social-attention items had α=.22, so the authors caution that null results may reflect measurement noise.Source: E. Self-Report Measures · p. 7
  • Workload was measured with four NASA-TLX items, and motion sickness with the VR Sickness Questionnaire’s oculomotor and disorientation scores.Each workload item used a 7-point scale; motion-sickness scores were scaled from 0 to 100.Source: E. Self-Report Measures · pp. 7–8

F. Behavioral Measures

  • Behavioral measures were computed from story-playing periods and included interpersonal distance and movement per second.Movement aggregated physical head movement and virtual locomotion; interpersonal distance was averaged over the story duration.Source: F. Behavioral Measures · p. 8
  • The researchers measured both the number of co-watchers in a user’s field of view and the percentage of that view covered by avatars.FOV coverage accounts for avatar distance and estimates visual obstruction of story content; visible-avatar area was approximated from projected silhouettes.Source: F. Behavioral Measures · pp. 8–9

G. Data Analysis

The analyses used condition-focused linear models for self-reports and condition plus group size for behavioral outcomes; participants were young adults with varied prior XR experience.

G. Data Analysis

  • Self-report outcomes were modeled with condition as the predictor after mixed-effects models with group as a random effect rarely improved fit.Content and group size were also tested but did not predict most self-report outcomes.Source: G. Data Analysis · p. 9
  • Behavioral outcomes were modeled using condition and group size because group size predicted outcomes and mixed-effects models did not improve fit.Interpersonal distance and virtual movement were log-transformed to address violations of normality assumptions.Source: G. Data Analysis · p. 9
  • Experiment 1 treated spatial configuration categorically, while Experiments 2 and 3 modeled space size and familiarity continuously.Experiment 1 pairwise comparisons used Bonferroni correction; values beyond three standard deviations were winsorized, with α=.05 for significance.Source: G. Data Analysis · p. 9

A. Demographics

  • The Experiment 1 sample comprised 109 participants aged 18–29, with varied prior XR experience and multiple reported ethnic identities.The demographics passage reports 67 female and 42 male participants and prior XR experience ranging from none to at least six prior uses.Source: A. Demographics · pp. 9–10

B. Conditions

Experiment 1 compared three ecologically grounded spatial configurations; communal box seating was associated with higher social presence toward others, while reserved seating involved lower mental demand and less movement.

B. Conditions

  • The experiment compared general admission, communal box seating, and individually bounded reserved seating.General admission used a 400-by-400-meter square, box seating a communal 20-by-20-meter boundary, and reserved seating individual 4-by-4-meter areas arranged in a grid with 2-meter spacing.Source: B. Conditions · p. 10
  • Reserved seating changed both boundary type and spatial constraint size, so its contrast with the other conditions cannot isolate either dimension.The authors frame Experiment 1 as a comparison of ecologically grounded configurations; box versus general admission holds boundary type constant to examine constraint size.Source: B. Conditions · p. 10

1) Self-Report Differences Across Spatial Boundary Types:

  • Social presence toward others was higher in the communal box condition than in general admission and reserved seating, while box and general admission had higher mental demand than reserved seating.Boundary type did not significantly predict presence toward characters, spatial presence, attention, or motion-sickness measures; other workload measures also showed no significant differences.Source: 1) Self-Report Differences Across… · pp. 10–11

2) Behavioral Differences Across Spatial Boundary Types:

  • Reserved-seating users moved less than users in box or general-admission conditions, while box users had the highest interpersonal distance and user-avatar FOV coverage.General-admission users stood furthest from their group; the number of users in FOV did not differ by condition.Source: 2) Behavioral Differences Across… · p. 11

V. EXPERIMENT 2: SIZE OF ALLOCATED PERSONAL SPACE

Experiment 2 varied the size of individual reserved spaces and found that smaller spaces were associated with greater social presence toward others, disorientation, and frustration, while larger spaces increased interpersonal distance.

A. Demographics

  • The Experiment 2 sample comprised 116 participants aged 18–29, with varied prior XR experience and multiple reported ethnic identities.The demographics passage reports 69 female and 47 male participants.Source: A. Demographics · p. 11

B. Conditions

  • The between-subjects experiment randomly assigned groups to individual square spaces of 1x1, 4x4, or 8x8 meters, keeping grid spacing at 2 meters.Source: B. Conditions · p. 11

1) Self-Report Differences Across Personal Space Sizes:

  • Larger personal spaces predicted lower social presence toward others and lower disorientation, whereas smaller spaces were associated with greater frustration.Personal-space size did not significantly predict presence toward characters, spatial presence, story attention, oculomotor motion sickness, or the other listed workload measures.Source: 1) Self-Report Differences Across… · p. 11, Figure 8 · p. 12

2) Behavioral Differences Across Personal Space Sizes:

  • Interpersonal distance increased with personal-space size, while avatar coverage of users’ FOV decreased; virtual movement and the number of users in FOV did not vary significantly by space size.Group size predicted interpersonal distance and the number of users in FOV, but not FOV coverage or movement.Source: 2) Behavioral Differences Across… · p. 11, Figure 9 · p. 12

A. Demographics

Experiment 3 examined group familiarity using shuffled and non-shuffled groups, but the manipulation check failed, so analyses used self-reported familiarity. Higher familiarity was associated with greater story attention and presence toward story characters, while behavioral findings were mixed.

A. Demographics

  • The 88 participants were ages 18–26 (M=20.73, SD=1.36), and their prior XR experience ranged from none to at least six uses.The sample included 42 Asian or Asian-American participants and 18 Caucasian participants.Source: A. Demographics · p. 12

B. Conditions

  • The shuffled versus non-shuffled group manipulation did not predict familiarity (p=.72), so subsequent analyses used a self-reported 1-item, 5-point familiarity measure.One “Very well” response was winsorized to the next-highest ordinal value; no participant reported knowing group members “Extremely well.”Source: B. Conditions · p. 12

1) Self-Report Differences Across Group Familiarity:

  • Higher group familiarity predicted greater attention towards the story, spatial presence, and social presence towards story characters.It did not significantly predict social presence towards users, attention from or towards others, disorientation, oculomotor motion sickness, frustration, or mental demand; perceptual and physical demand were higher.Source: 1) Self-Report Differences Across… · p. 12, Figure 10 · p. 13

2) Behavioral Differences Across Group Familiarity:

VII. DISCUSSION

Across three experiments, greater engagement co-occurred with greater cognitive effort or workload. Spatial configurations and group familiarity were associated with distinct tradeoffs, while the exploratory findings leave some mechanisms and causal directions unresolved.

A. Systematic Tradeoffs During Immersive Co-watching

1) Spatial Boundary Type Shapes Social Presence and Mental Demand:

  • The configuration with the highest social presence toward co-watchers (box) was not the one with the lowest mental demand (reserved seating).Greater avatar coverage in users’ field of view tracked social presence, but simple counts of users in view did not capture experiential differences; overlapping avatars and interpersonal distance may affect perceptual salience.Source: 1) Spatial Boundary Type Shapes Social… · p. 13

2) Size of Allocated Personal Space as a Double-Edged Sword:

  • Smaller personal spaces and closer interpersonal distance were associated with higher social presence toward users, disorientation, and frustration.Greater avatar coverage may occlude the story, while misalignment between the story skybox and shared virtual space may add perceptual conflict; the two spatial characteristics in reserved seating were confounded, leaving their individual contributions unclear.Source: 2) Size of Allocated Personal Space as… · pp. 13–14

3) Group Familiarity is Associated with Story Engagement and Workload:

  • Greater self-reported group familiarity co-occurred with more story attention, stronger spatial and character social presence, and higher perceptual and physical demand.It did not negatively predict attention towards others; because familiarity was self-reported rather than manipulated, third variables could explain the associations.Source: 3) Group Familiarity is Associated… · p. 14

B. Practical Considerations and Outlooks for Co-watching of Immersive Stories

The discussion recommends treating spatial boundaries and social context as design variables for immersive co-watching. It also cautions that implications may vary across content formats, so the studied medium should guide application.

1) Crafting Spatial Configurations for Immersive Co-watching:

  • Use spatial boundaries to shape proxemics and balance social engagement with experiential demands; individual boundaries were associated with lower mental demand, while communal boundaries can elicit stronger social presence.Larger personal spaces did not increase movement in Experiment 2; moderately constrained communal space enhanced social presence, but the benefit diminished in the largely unconstrained configuration.Source: 1) Crafting Spatial Configurations for… · pp. 14–15

2) Designing for Individual Differences in Social Context:

  • Design for individual differences in familiarity, including within the same group, and consider customization of spatial arrangements or social-cue visibility.The discussion proposes adaptive cue salience, personalized tutorials, and gradual transitions between individual and communal viewing modes.Source: 2) Designing for Individual… · p. 15

3) Beyond Watching 360-Degree Stories:

  • Treat the implications as format-dependent: perceptual conflict, occlusion, and attention patterns may differ between monoscopic 360-degree, stereoscopic, volumetric, and 2D content.The authors position their implications as a starting point rather than a general account of co-watching.Source: 3) Beyond Watching 360-Degree Stories · p. 15

4) Co-Located vs. Remote Co-watching:

The study used co-located groups, so some findings may not transfer directly to remote co-watching. Its exploratory results also have limits in scope and support systematic tradeoffs rather than uniform experiential gains.

4) Co-Located vs. Remote Co-watching:

  • Co-location may shape social dynamics beyond avatar cues: participants could hear one another and remain aware of others, while remote groups would rely more on avatars, nonverbal behavior, and spatial audio.The authors expect perceptual and behavioral findings tied to virtual configuration to transfer more readily than social-presence findings.Source: 4) Co-Located vs. Remote Co-watching · pp. 15–16

VIII. LIMITATIONS AND FUTURE WORKS

  • The experiments did not vary group size or composition, and Experiment 3's failed manipulation check means familiarity findings are associations, not causal effects.Experiment 1 confounded boundary type with constraint size, while Experiment 2 could not isolate interpersonal distance from movement.Source: §VIII · p. 16
  • Familiarity findings reflect a narrow, low-familiarity university sample; whether they hold for close relationships or strangers remains open.Familiarity averaged 1.73 on a 5-point scale; no participant reported the highest level, and only one reported the second-highest.Source: §VIII · p. 16
  • The work is exploratory and analyzes session-level behavior; future research could test mechanisms, moment-to-moment behavior, additional outcomes, and validated co-watching measures.The authors also call for studying more interactive stories and populations beyond university students.Source: §VIII · p. 16

IX. CONCLUSIONS

  • Greater social presence often coincided with greater workload and motion sickness; smaller personal space also brought greater frustration and disorientation.Participants with greater group familiarity reported greater attentional focus, spatial presence and social presence towards story characters, alongside greater perceptual and physical demand.Source: §IX · p. 16

X. BIOGRAPHY SECTION

The biographies describe the authors’ education and research roles or interests.

X. BIOGRAPHY SECTION

  • Portia Wang is a Stanford communication Ph.D. candidate researching psychological and behavioral responses to immersive stories and tools for memory reliving.She earned degrees from Columbia University and Stanford University.Source: §X · p. 18
  • Monique Santoso is a Stanford communication Ph.D. candidate whose research concerns VR’s psychological, interpersonal, and behavioral implications.Her stated interests include climate adaptation, prosocial attitudes, and group interactions.Source: §X · p. 18
  • Yujie Tao is a Stanford computer science Ph.D. candidate studying computing systems that reconnect users to everyday interactions by augmenting the senses.Jeremy Bailenson is a Stanford professor and founding director of the Virtual Human Interaction Lab.Source: §X · p. 18
Loading 2610.30353v1…