One-Minute Silent Video Clips: A Database of Valence and Arousal

Area of Science:

• Behavioral and social sciences
• Psychophysiology
• Affective computing

Background:

• Linear discriminant analysis (LDA) is used for classification and variable identification in behavioral and social sciences.
• Longitudinal data and repeated measures analysis are crucial for capturing temporal dynamics and complex correlations.
• Psychometric data often violate multivariate normality assumptions, necessitating robust analytical methods.

Purpose of the Study:

• Introduce a comprehensive dataset of affective video clips with normative valence and arousal ratings.
• Facilitate research in affective computing, psychophysiology, and cross-cultural studies.
• Provide a standardized resource for selecting video subsets tailored to specific research questions.

Main Methods:

• A dataset of 160 one-minute affective video clips was curated.
• Each video was evaluated by 30 subjects, with each subject rating at least 20 videos.
• Normative values for valence and arousal were collected for each clip.

Main Results:

• The dataset includes videos across negative, neutral, and positive valence categories.
• A quadratic relationship between valence and arousal was observed, with high arousal for extreme valence clips.
• A negativity bias was indicated, with some negative videos showing higher arousal than positive ones.

Conclusions:

• The presented dataset offers a valuable, versatile resource for researchers in affective science.
• Its standardized duration and cultural neutrality enhance its utility in various psychophysiological and cross-cultural research designs.
• The dataset supports the selection of specific video subsets, advancing nuanced investigations into emotion.