Does Background Music Matter in Data Videos? A Study of Music's Impact on Persuasion, Engagement, and Recall

Paper Title

Does Background Music Matter in Data Videos? A Study of Music's Impact on Persuasion, Engagement, and Recall

Publication Info

• Topic area: The role of background music in data videos and its impact on viewer experience.
• Keywords: Background music, data videos, persuasion, engagement, information recall, multimedia learning, Bayesian modeling, thematic analysis, narrative alignment, visual storytelling.

Background and Problem

• Problem / challenge: While background music is widely believed to enhance emotional and persuasive effects in multimedia, its specific impact in data videos remains unexplored. Prior research on music's influence in other domains has yielded mixed results, with some studies showing positive effects and others finding no impact or even detrimental outcomes.
• Significance: Understanding the role of background music in data videos could inform design practices for effective data-driven storytelling, which is increasingly used to communicate scientific and social issues to the public.
• Motivation and related work: Previous studies have examined design elements like animations and narrative structures in data videos but have largely overlooked background music. Research in other domains suggests music can influence decision-making, engagement, and emotional responses, but its effects on persuasion, engagement, and recall in the context of data videos remain untested.

Solution

• Proposed approach: A preregistered, between-subjects online experiment comparing six widely viewed data videos with and without background music to evaluate its impact on persuasion, engagement, and information recall.
• Novelty:
  1. Empirical evidence showing negligible measurable effects of background music on persuasion, engagement, or recall across six real-world data videos.
  2. Qualitative insights revealing nuanced perceptions of music, including its potential to enhance enjoyment, focus, or emotional resonance when well aligned with the video’s tone.
  3. Design implications emphasizing the context-dependent role of background music, with recommendations for tailored, incidental, or adaptive music in data storytelling.
• Procedure and key techniques:
  • Selection of six data videos from a corpus of 347 using a multi-stage filtering process.
  • Random assignment of 477 participants to watch videos with or without background music.
  • Measurement of persuasion, engagement, and recall using Bayesian modeling and thematic analysis.
  • Collection of qualitative feedback to explore subjective experiences of music.

Results

• Concrete findings:
  • Persuasion: Effect size of 0.0036 (95% CI [-0.023, 0.031]), indicating negligible impact of music on attitudinal shifts.
  • Engagement: Effect size of 0.0033 (95% CI [-0.022, 0.030]), showing no significant influence of music on viewer involvement.
  • Recall: Effect size of 0.01 (95% CI [-0.030, 0.049]), suggesting minimal impact on information retention.
• Advantage over baselines: No measurable advantage of background music over no-music conditions across all metrics.
• Experiments / evaluation:
  • Participants watched one of six videos in either music or no-music conditions.
  • Metrics included persuasion (attitudinal shift), engagement (11-item questionnaire), and recall (quiz-based assessment).
  • Bayesian models were used to estimate overall and per-video effects.
• Limitations and future work:
  • Limited generalizability due to the small number of videos and reliance on pre-existing music tracks.
  • Immediate post-viewing measurements may overlook long-term or subtle effects.
  • Future research could explore bespoke or adaptive music, longitudinal impacts, and broader cultural contexts.

Summary

This study investigated the impact of background music on persuasion, engagement, and recall in data videos through a preregistered experiment with 477 participants. Quantitative results showed negligible measurable effects of music across all metrics, while qualitative feedback highlighted subjective and context-dependent influences, such as enhanced emotional resonance or distraction. Findings suggest that music’s role in data videos is contingent on alignment with narrative and visual elements, rather than universally beneficial. Future research should explore tailored or adaptive music designs and employ broader evaluation methods to capture subtle or long-term effects. These insights provide a foundation for rethinking the use of background music in data-driven storytelling.