Manuscript in preparation Multimodal Dataset

AVI6

A Trait-Activated Multimodal Dataset for Personality and Competency Assessment in Asynchronous Video Interviews

Tianyi Zhang1, Jinwenxi Shang1, Antonis Koutsoumpis2, Wei Shan1, Yuan Zong1,*, Reinout E. de Vries3, Wenming Zheng1

1 Key Laboratory of Child Development and Learning Science (Ministry of Education), School of Biological Sciences and Medical Engineering, Southeast University, Nanjing, China

2 School of Business and Economics, Vrije Universiteit Amsterdam, the Netherlands

3 Department of Experimental and Applied Psychology, Vrije Universiteit Amsterdam, the Netherlands

* Corresponding author: Yuan Zong (xhzongyuan@seu.edu.cn)

646 Participants
3,876 Interview Videos
~85h Video Footage
HEXACO Traits + Competencies

Psychometrically Grounded AVI Assessment

Existing multimodal personality datasets often rely on short, task-free social media videos and crowdsourced apparent-personality labels. AVI-Personality instead collects structured asynchronous video interviews in a simulated job application, with questions designed from Trait Activation Theory and annotations provided by trained psychologists and professional recruiters.

Overview of the AVI-Personality dataset: multimodal data, interview pipeline, and expert annotation.
Figure 1. Overview of AVI-Personality: multimodal interview data with HEXACO and competency annotations; a three-stage pipeline from question development to self-report collection; and expert annotation by trained personality psychologists and professional recruiters using BARS.
01

Trait-Activated Interview Protocol

Participants completed a simulated management traineeship application and answered two generic selection questions plus four personality-targeted questions designed to activate Honesty-Humility, Extraversion, Agreeableness, and Conscientiousness.

02

Expert-Validated Dual Labels

AVI-Personality provides HEXACO-100 self-reports, observer-rated personality from 12 trained psychologists, recruiter-rated job competencies from 5 professionals, interview performance, and ICAR cognitive ability. Observer ratings use Behaviorally Anchored Rating Scales (BARS).

03

Open Multimodal Dataset

The release includes interview videos with aligned audio and transcripts, psychometric labels, demographics, and an official subject-level split. The dataset is available to academic researchers after signing the user agreement.

The AVI-Personality Benchmark

AVI-Personality preserves a standardized recruitment context while providing multimodal behavioral signals and psychologically grounded labels for personality and competency modeling.

Collection Protocol

  • Recruitment: Native English-speaking adults in the United States via Prolific; 793 recruited, 646 retained after quality control and consent checks.
  • Task: Simulated management traineeship AVI. Each response was recorded once (no re-recording), with a target length of 1–2 minutes.
  • Questions: Two recruiter-informed generic questions and four TAT-based personality questions targeting H, X, A, and C.
  • Self-reports: HEXACO-100 personality inventory and 28-item ICAR cognitive ability test, plus demographics.
  • Annotation: At least 3 personality psychologists per participant and at least 2 professional recruiters, after 9-hour rater training.

Key Properties

Participants646 (309 men / 309 women / 28 non-binary)
Videos3,876 AVI responses (~85 hours)
Age36.69 ± 11.92 years
Work experience15.96 ± 11.36 years
Personality labelsHEXACO self-reports + observer ratings
Competency labels4 competencies + overall interview performance
Cognitive abilityICAR g-score (0–28, mean 14.51)
Official splitSubject-level 70 / 10 / 20 (452 / 64 / 130)
Prior datasets

Short in-the-wild clips | Crowdsourced first impressions | Coarse Big Five adjectives | Weak interview relevance

AVI-Personality

Structured AVIs | Trained psychologist & recruiter ratings | HEXACO + competencies | Trait-activated questions

Interview Questions

Question order is fixed. Generic questions elicit broad selection-relevant information; personality questions use a past-behavior format to activate specific HEXACO traits.

# Interview question Type Target trait
1 What would you consider among your greatest strengths and weaknesses as an employee? Generic
2 How would your best friend describe you? Generic
3 Think of situations when you made professional decisions that could affect your status or how much money you make. How do you usually behave in such situations? Why do you think that is? Personality Honesty-Humility
4 Think of situations when you joined a new team of people. How do you usually behave when you enter a new team? Why do you think that is? Personality Extraversion
5 Think of situations when someone annoyed you. How do you usually react in such situations? Why do you think that is? Personality Agreeableness
6 Think of situations when your work or workspace were not very organized. How typical is that of you? Why do you think that is? Personality Conscientiousness

From HEXACO Traits to Job Competencies

Each participant is annotated with six HEXACO domains (self-report and observer-report) and recruiter-rated workplace competencies. Observer personality ratings are provided separately for generic questions and trait-activating questions.

H

Honesty-Humility

  • Sincerity
  • Fairness
  • Greed Avoidance
  • Modesty
E

Emotionality

  • Fearfulness
  • Anxiety
  • Dependence
  • Sentimentality
X

Extraversion

  • Social Self-Esteem
  • Social Boldness
  • Sociability
  • Liveliness
A

Agreeableness

  • Forgiveness
  • Gentleness
  • Flexibility
  • Patience
C

Conscientiousness

  • Organization
  • Diligence
  • Perfectionism
  • Prudence
O

Openness

  • Aesthetic Appreciation
  • Inquisitiveness
  • Creativity
  • Unconventionality
Int

Integrity

  • Trustworthiness and ethical standards in workplace interaction
Col

Collegiality

  • Openness to others and willingness to support colleagues
Sv

Social versatility

  • Adapting behavior across social and organizational settings
Do

Development orientation

  • Motivation to broaden skills and grow professionally
Ip

Interview performance

  • Overall suitability for the management traineeship role

How to Obtain AVI-Personality

AVI-Personality is released for academic, non-commercial research. Please sign the user agreement and email it to Tianyi Zhang. Access is granted after the signed form is reviewed.

Request procedure

  1. 1Download the AVI-Personality User Agreement and print or export it as PDF.
  2. 2The form must be signed by a researcher with a permanent position (e.g., professor or assistant/associate professor). Students may be listed as participants.
  3. 3Email the signed PDF to Tianyi Zhang. Include your name, institution, position, and a brief description of the intended research use.

Contact: Tianyi Zhang
Email: t.zhang@seu.edu.cn
Subject: AVI-Personality Dataset Access Request

What is released

  • Asynchronous video interview recordings from 646 participants (6 questions each)
  • Aligned audio and textual transcripts
  • HEXACO self-reports and observer-rated personality scores
  • Recruiter-rated competencies, interview performance, and ICAR cognitive ability
  • Demographic information and the official subject-level data split

The dataset may be used only for academic, non-commercial, not-for-profit research. Redistribution is not permitted. Videos and still frames must not be used in any way that could cause the original participants embarrassment or mental anguish. See the User Agreement for the full terms.

How to Cite

If you use the AVI-Personality dataset in your research, please cite our paper. Venue, volume, and DOI will be added after submission.

@article{zhang2026AVI-Personality,
  author  = {Zhang, Tianyi and Shang, Jinwenxi and Koutsoumpis, Antonis and Shan, Wei and Zong, Yuan and de Vries, Reinout E. and Zheng, Wenming},
  title   = {AVI-Personality: A Trait-Activated Multimodal Dataset for Personality and Competency Assessment in Asynchronous Video Interviews},
  journal = {},
  year    = {2026},
  note    = {Manuscript in preparation}
}