Feasibility of Big Data Analytics to Assess Personality Based on Voice Analysis

Area of Science:

• Psychology
• Speech Analysis
• Human-Computer Interaction

Background:

• Early 20th-century research suggested non-verbal cues like speech and posture reveal personality.
• Previous studies on linguistic cues yielded inconsistent results.
• Advancements in speech analysis technology and big data enable new personality prediction research.

Purpose of the Study:

• To explore the feasibility of an automatic personality assessment system.
• To investigate the predictive power of voice features on personality dimensions.
• To assess the utility of voice analysis in personnel selection.

Main Methods:

• 100 participants' voices were recorded during individual interviews.
• Participants completed the NEO-FFI personality inventory.
• Personality was also assessed by significant others and expert raters viewing interview recordings.

Main Results:

• Specific voice features predicted personality dimensions with moderate accuracy (0.3–0.4).
• Voice features also predicted personality as estimated by significant others and experts.
• The specific voice features varied slightly between self-report, peer, and expert assessments.

Conclusions:

• Voice analysis using ordinary devices shows potential for personality assessment.
• Controlled, non-restricted speech situations are suitable for voice-based personality evaluation.
• This method could be a valuable tool for personnel selection and other applications.