Using Artificial Intelligence in Civics Assignment: Rethinking Assessment, Learning Engagement, and Trust in Digital Technology

This dataset contains the primary data collected from 28 undergraduate students enrolled in a Civic Education course. The study employed a mixed-methods pre-test and post-test design to evaluate the impact of AI tools (specifically ChatGPT) on task perception, learning engagement, and digital trust.

1. Quantitative Data (Sheet: Quantitative_Data)

This sheet contains the raw Likert-scale scores (1–5) from surveys distributed before and after the AI-assisted assignment intervention.

Core Variables: Measures perceived task difficulty, grade expectations, and self-perceived essay quality.

Trust Metrics: Utilizes the validated Multi-Dimensional Measure of Trust (MDMT) V2 scale, specifically focusing on the "Ethicality" and "Benevolence" dimensions of the AI agent.

Responsibility: Tracks the shift in student comfort levels regarding taking legal or academic responsibility for machine-generated content.

2. Participant Demographics (Sheet: Demographics)

Provides context regarding the participant pool to ensure data transparency and replicability.

Profile: Includes age (ranging from 19 to 22), gender, and faculty affiliation (Education, Social Sciences, and Law).

AI Background: Notes the students' prior experience with artificial intelligence, categorized as None, Basic, or Intermediate.

3. Qualitative Responses (Sheet: Qualitative_Responses)

Contains the verbatim transcripts of open-ended responses, providing narrative depth to the quantitative findings.

Scope: Students reflect on the technical challenges, the shift in their role from "writer" to "editor," and the limitations of AI in understanding local Indonesian contexts (Pancasila and national identity).

Thematic Coding: Each response is paired with a "Key Theme" (e.g., Co-Creation, Human-AI Gap, Critical Thinking) derived from the thematic analysis process described in the manuscript.

4. Statistical Output Summary (Sheet: Statistical_Output)

Provides a summary of the inferential statistical analysis that supports the claims made in the article.

Paired Sample T-Test: Documents the significant increase in perceived difficulty (p=0.05) and the significant growth in trust regarding AI ethics (p<0.01).

Reliability Statistics: Reports Cronbach’s Alpha values (all >0.80), confirming the high internal consistency of the research instruments.