Picturing It!: The Effect of Image Styles on User Perceptions of Personas

Title of the Paper

Picturing It!: The Effect of Image Styles on User Perceptions of Personas

Bibliographic Information

• Domain: Human-Computer Interaction (HCI)
• Keywords: persona design, images, realism, anthropomorphism, empirical HCI research, visual design, personality attribution
• Citation Format: Joni Salminen et al. (2021). Picturing It!: The Effect of Image Styles on User Perceptions of Personas. CHI Conference on Human Factors in Computing Systems (CHI '21), Yokohama, Japan, DOI: https://doi.org/10.1145/3411764.3445360

Research Background and Questions

• Problem Discovery and Challenges:

  1. In persona design, photos are often used to depict personas, but the impact of different image styles (e.g., photos, cartoons) on user perceptions has been underexplored.
  2. The choice of image style is often based on designers' personal preferences, lacking empirical data to support decision-making.
  3. Using photos may make personas appear more realistic but could also reinforce stereotypes and biases.

• Research Importance: Personas are widely used in design practices to simulate target user groups and help designers understand user behavior. Understanding the impact of image styles on user perceptions can optimize persona design, enhancing credibility, clarity, and empathy.

• Research Motivation and Related Work:

  1. Existing studies suggest that images are crucial for shaping the authenticity and emotional connection of personas, but these studies often lack statistical support or involve small sample sizes.
  2. This paper further investigates how different image styles influence user perceptions and personality attribution of personas.

• Research Questions:

  • RQ1: Does image realism enhance positive perceptions of personas?
  • RQ2: How does image realism affect users' judgments of personas' personality traits?
  • RQ3: Does the presence or absence of images significantly affect persona perception?
  • RQ4: Is there a "Uncanny Valley" effect in persona image styles?

Solution

• Research Methods:

  1. Design an experiment featuring different image styles, ranging from real photos to no images, across seven styles with progressively reduced visual information.
  2. Recruit 149 participants to evaluate three distinct personas, each presented with different image styles randomly assigned to participants.
  3. Use the Persona Perception Scale (PPS) and Ten Item Personality Measure (TIPI) to test changes in user perceptions and personality attribution.

• Key Experimental Design:

  • Seven Image Styles:
    • Real Photo (RP)
    • Realistic Cartoon (RC)
    • Human-like Cartoon (HC)
    • Symbolic Cartoon (SC)
    • Silhouette (SL)
    • Gender-neutral Silhouette (GN)
    • No Picture (NP)
  • Evaluation Dimensions:
    • Perception dimensions: Clarity, Completeness, Consistency, Credibility, Empathy.
    • Personality dimensions: Agreeableness, Conscientiousness, Emotional Stability, Extraversion, Openness.

• Innovative Contributions: Pioneering exploration of the "Uncanny Valley" effect in persona design, providing empirical data to guide decisions on image realism.

Research Findings

• Key Discoveries:

  1. Impact of Realism:
     • High image realism significantly improves persona perception scores, including clarity, completeness, credibility, and empathy.
     • Realistic images make personas more persuasive and enhance users' confidence in the persona content.
  2. Impact on Personality Attribution:
     • High realism images strengthen users' attribution of agreeableness, emotional stability, and openness (significant), but have no statistical impact on conscientiousness and extraversion.
     • Users' confidence in evaluating persona personality increases with higher image realism.
  3. Images vs. No Images:
     • Overall comparison showed no significant difference; however, real photos (RP) significantly improved persona perception scores compared to no images.
  4. Uncanny Valley Effect:
     • Realistic Cartoon (RC) images triggered the "Uncanny Valley" effect, lowering perception scores, suggesting that "too realistic but not fully authentic" styles may backfire.
     • Gender-neutral Silhouette (GN) scored even lower than the no-image option.

• Limitations and Future Directions:

  • Limitations:
    1. The sample primarily consisted of participants from Western societies, which may affect the generalizability of the results across cultures.
    2. The experimental tasks focused on personality evaluation, excluding other contexts (e.g., functional tasks) where image impact might differ.
  • Future Research Directions:
    1. Investigate the influence of cultural differences on image style selection.
    2. Examine the intensity of the "Uncanny Valley" effect in various design contexts, such as educational interactions and entertainment design.
    3. Use behavioral data (e.g., eye-tracking) to further validate how users process image information.

Summary

This study highlights the key advantages of high-realism images in persona design while uncovering the potential negative impact of the "Uncanny Valley" effect. It provides empirical support and design guidance for selecting image styles in persona development.