Study habits and artificial intelligence use among university students: A proportionally stratified survey dataset

This repository contains a survey dataset on study habits and artificial intelligence (AI) use among undergraduate students at the Manizales campus of a Colombian public university. The data were collected during the first academic term of 2025 using a proportional allocation across 14 academic programs.
Within each program, data collection followed a quota-based approach: responses were collected until the target number of students for that program was reached. The final dataset includes 357 observations and preserves the intended distribution of the sample across programs.
The repository includes both the original and processed versions of the data. The file “survey_raw_spanish_anonymized.csv” contains the original questionnaire responses in Spanish, with identifying timestamp detail removed. The file “survey_cleaned_english.csv” provides a cleaned analytical version with English variable names, harmonized categorical responses, a cleaned GPA variable, a standardized AI tool field, and a derived count of reported AI tools.
Additional files support interpretation and reuse. The “sampling_frame_by_program.csv” file documents the population and proportional allocation by academic program. The “codebook.csv” provides variable definitions, response formats, and missing-value conventions. The “questionnaire_bilingual.csv” contains the survey instrument in Spanish and English, organized by thematic sections and linked to coded items. The “analysis_script.R” file reproduces the data cleaning, transformation, and descriptive analysis steps.
The dataset includes variables on sociodemographic characteristics, study habits, academic self-perception, AI tool use, perceived usefulness, academic integrity, perceived dependence, creativity, and prompt engineering knowledge. Most variables are ordinal categorical, making the dataset suitable for descriptive and exploratory analysis.
This dataset is intended for secondary use in educational research, including studies of study habits, AI adoption in higher education, academic integrity, and cross-program comparisons. It is based on self-reported responses from a single campus and follows a cross-sectional design; therefore, it should not be used to draw causal conclusions.

本数据集仓库收录了哥伦比亚某公立大学马尼萨莱斯校区本科生学习习惯与人工智能（Artificial Intelligence）使用情况的调研数据集。本次调研数据采集于2025年第一学期，采用按14个本科专业进行比例分配的抽样方法。
在每个专业内部，数据采集采用配额抽样法：持续收集问卷回复直至达到该专业预设的样本量目标。最终数据集共包含357条有效观测值，且保留了各专业预设的样本分布比例。
本仓库同时提供数据集的原始版本与处理后版本。文件"survey_raw_spanish_anonymized.csv"包含西班牙语版的原始问卷回复，已移除可识别的时间戳信息。文件"survey_cleaned_english.csv"则提供经过清洗的分析版本，采用英文变量名、统一化的分类回复、清洗后的平均学分绩点（Grade Point Average）变量、标准化的人工智能工具字段，以及基于报告的人工智能工具使用数量衍生出的计数变量。
此外，仓库还包含辅助数据解读与复用的额外文件："sampling_frame_by_program.csv"记录了各专业的总体规模与比例分配方案；"codebook.csv"提供了变量定义、回复格式与缺失值处理规范；"questionnaire_bilingual.csv"包含西班牙语与英语双版调研问卷，按主题板块组织并与编码条目关联；"analysis_script.R"可复现数据清洗、转换与描述性分析的全流程。
本数据集涵盖社会人口学特征、学习习惯、学业自我认知、人工智能工具使用、感知有用性、学术诚信、感知依赖度、创造力以及提示工程（Prompt Engineering）知识相关变量。多数变量为有序分类变量，因此本数据集适用于描述性与探索性分析。
本数据集旨在用于教育领域的二次研究，包括学习习惯研究、高等教育领域人工智能工具采用情况研究、学术诚信研究以及跨专业比较研究。本数据集基于单一校区的自我报告回复，采用横断面研究设计，因此不可用于推导因果性结论。