How looks, voice, movement and scent draw us to future lovers and friends

Attraction is often influenced by facial features like symmetry and averageness, but it also involves other senses.

A person’s voice, scent, and body movement can shape how appealing they seem, often without conscious awareness.

These multisensory cues play a vital role in forming first impressions and guiding social or romantic interest.

Schirmer, A., Franz, M., Krismann, L., Nöring, V., Große, M., Mahmut, M., & Croy, I. Attraction in every sense: How looks, voice, movement and scent draw us to future lovers and friends. British Journal of Psychology. https://doi.org/10.1111/bjop.12787

Key Points

• Focus: The study examines how facial appearance, voice, body motion, and scent each contribute to perceptions of interpersonal attractiveness.
• Aims: The research aimed to investigate the relative attractiveness of different non-verbal modalities, the extent of redundancy across these channels, and the roles of common versus personal preferences in both same- and opposite-sex interactions.
• Findings: Multimodal (audio–video) stimuli were rated as most attractive, while body odour was rated least attractive. Attractiveness ratings across modalities were partially redundant, and personal preferences played a stronger role in same-sex evaluations.
• Implications: Attractiveness is shaped by multiple sensory channels, and personal preferences can influence non-romantic social bonding. This expands the concept of attraction beyond mate selection to broader social contexts.

Rationale

Traditional attraction research has focused on facial features and opposite-sex evaluations in mate selection contexts.

However, humans use multiple senses when forming impressions, and social relationships—including friendships—are influenced by attraction.

This study expands on existing findings by including olfactory, auditory, and motion-related cues, and considers both common and personal preferences in same- and opposite-sex interactions.

Method

This study used a multimodal design involving 61 “agents” who provided stimuli and 71 “perceivers” who rated the agents’ attractiveness across five modalities: voice, facial photos, videos (muted and unmuted), and body odour.

Procedure

• Perceivers rated 8 agents (4 males, 4 females) in 5 modalities:
  • Body odour samples collected via underarm pads after exercise.
  • Photographs taken with neutral expressions.
  • Videos showing a spoken greeting and a sustained “Aaah.”
  • Audio-only extracted from the videos.
  • Muted video stimuli separated from audio.
  • Audio–video (unmuted video).
• Order of presentation: odour → photo → audio → muted video → audio–video.
• Ratings included attractiveness on a 0–6 scale.
• Only the attractiveness rating was used in analysis.

Sample

• Agents: 61 participants (34 women, 27 men), average age ≈ 21.9 years.
• Perceivers: 71 participants (37 women, 34 men), average age ≈ 22.1 years.
• Participants had C1-level German proficiency and no reported psychological or major physical health conditions.
• Sexual orientation was self-reported on a 1–5 scale; distribution was similar between male and female participants.

Measures

• Attractiveness rating: 7-point Likert scale (0 = not at all attractive; 6 = very much).
• Kinetic index: Calculated by regressing video ratings on photo ratings to isolate motion-specific effects.
• Modality-specific stimuli: Audio, video, photos, odour, and combined audio–video presentations.

Statistical Measures

• Cumulative link mixed effects models: Used for ordinal data.
• Variance partitioning (ICC): Distinguished variance due to agent (common preferences) and perceiver (personal preferences).
• Wald tests & AIC improvement: Assessed significance and effect sizes.
• Bootstrapped ICCs: Compared contributions of agents and perceivers by sex and agent sex (same vs. opposite).

Results

• Multimodal stimuli were rated most attractive, especially audio–video.
• Body odour was rated least attractive across all perceiver-agent pairings.
• Redundancy existed among modalities: voice and video were most strongly correlated; body odour showed weaker, but some, associations with movement.
• Same-sex attraction showed greater influence of personal preferences (perceiver variance > agent variance).
• Opposite-sex attraction showed balanced contributions of personal and common preferences.

Insight

This study confirms that attractiveness is not merely visual but multisensory, with voices, movement, and odour influencing perceptions.

Multimodal integration enhances attractiveness, perhaps due to sensory convergence.

The findings also highlight the importance of personal preferences, particularly in same-sex social evaluations, indicating that attraction is shaped by individual compatibility and context.

The study refines our understanding of interpersonal attraction and expands it beyond mating to broader social bonding.

Clinical Implications

• Highlights the complexity of first impressions in social and therapeutic contexts.
• Emphasizes the need to consider non-verbal and multisensory cues in social skills training, therapy, or communication-based interventions.
• Suggests personal biases in impression formation may affect assessments in clinical, hiring, or educational settings.

Strengths

This study had several methodological strengths, including:

• Multisensory design: Included visual, auditory, movement, and olfactory stimuli.
• Balanced gender representation: Considered both same- and opposite-sex interactions.
• Novel kinetic index: Isolated body movement effects.
• Advanced modelling: Used ordinal regression with random effects for rigorous analysis.
• Bootstrapped ICCs: Offered robust insights into variance sources.

Limitations

This study also had several limitations, including:

• Limited ecological validity: Stimuli were brief, static or scripted; real-world interactions are more dynamic.
• Olfactory samples: Artificial collection and presentation may not fully capture natural scent cues.
• Sample homogeneity: All participants were young adults from similar cultural backgrounds.
• Sexual preference: Self-reported orientation was not tightly controlled across conditions.
• Motion index approximation: Derived from residuals; may not fully isolate body movement from other variables.

Socratic Questions

1. What are the benefits and drawbacks of using isolated sensory modalities in studying attraction?
2. How might cultural norms or individual differences shape personal preferences in attractiveness?
3. Could the methods used to collect olfactory or movement data affect the findings?
4. How might the findings differ in older, more diverse populations?
5. To what extent can the findings about same-sex attraction inform our understanding of platonic relationships?
6. Do the results suggest that attraction is more about compatibility than about universal standards?
7. What implications do these findings have for digital dating platforms that emphasize visual features?
8. How could future research improve the realism or ecological validity of multimodal attractiveness studies?