gender-bias dataset

Gender bias dataset

 📝 Description

This dataset was created for the study “Detection of gender-based violence in digital environments using AI models.”  It contains comments from a teaching evaluation survey, each annotated as {0: non-gender bias; 1: gender bias}. Each row corresponds to a comment written by a student in a teaching survey.  The comments have been manually annotated according to the presence or absence of gender-based bias.  This dataset supports experiments in machine learning, NLP, and automatic bias detection in Spanish-language educational contexts.

Language: Spanish-Latam

Format: CSV

Structure of row: 

id_ev: Identifier

sesgo: Bias{0: non-gender bias; 1: gender bias}

sexo: Sex of who receives the bias {0: male, 1:female}
tamano (size)
comentario (comments)

4. Dataset components

The dataset is composed of the following files:

dataset960_gender_via.csv    - Total rows: 960

trainext_gender_bias.csv    - Training split    - Rows: 770   - Percentage: 80.2%

testext_gender_bias.csv   - Test split    - Rows: 190   - Percentage: 19.8%