Dataset for: AI-Driven Personalization of Gamification in Education: A Systematic Literature Review (2020–2025)
Description
This dataset supports the systematic literature review titled "AI-Driven Personalization of Gamification in Education: A Systematic Literature Review." It contains the complete data extraction, quality assessment scores, and coded analysis for all 49 included studies published between January 2020 and April 2026, retrieved from Scopus, IEEE Xplore, and Web of Science. The workbook includes: (1) search and screening records following PRISMA 2020 guidelines, (2) quality assessment scores across eight criteria (Q1–Q8) for each study, scored independently by two reviewers with substantial-to-excellent inter-rater agreement (Cohen's κ = 0.75, ICC(2,1) = 0.93), (3) full data extraction covering bibliographic information, AI techniques and functions, educational levels and disciplines, learning outcomes, gamification elements and AI–gamification integration patterns, methodological limitations, ethical risks, implementation challenges, and proposed frameworks, and (4) coded summaries for each of the five research questions (RQ1–RQ5). This dataset enables full reproducibility of the review findings and supports secondary analyses by other researchers.
Files
Steps to reproduce
1. Search and Screening: Searches were conducted across Scopus, IEEE Xplore, and Web of Science using Boolean combinations of AI, gamification, and education terms, limited to primary empirical studies (2020–2026). Records were exported and screened for deduplication based on exact DOI match and normalized title match, both treated as high-priority fields. From 7,836 initial records (Scopus: 4,438; IEEE Xplore: 1,949; Web of Science: 1,449), 2,148 duplicates were removed, 5,688 underwent title/abstract screening, 198 proceeded to full-text review, and 49 met all inclusion criteria: (a) empirical primary study reporting original research, (b) English-language publication, (c) explicit integration of AI techniques with gamification elements in an educational context, (d) gamification, rather than a complete game environment, as the primary intervention, (e) empirical evaluation with end users, and (f) publication between 2020 and 2026. 2. Quality Assessment and Data Extraction: Each study was scored against eight criteria (Q1–Q8: research question clarity, study design appropriateness, sample/context description, intervention reporting, data collection adequacy, validity/reliability, analysis appropriateness, and results/limitations reporting), on a 0/0.5/1 scale (max 8), independently by two reviewers, with substantial-to-excellent inter-rater agreement (raw agreement 92.9%, Cohen's κ = 0.75, ICC(2,1) = 0.93). A structured Excel form captured bibliographic data, AI techniques and functions, educational level and discipline, learning outcomes across four dimensions (academic performance, motivation, engagement, self-regulated learning), gamification elements and AI–gamification integration patterns, methodological limitations, ethical risks, implementation challenges, and proposed frameworks or design principles. 3. Analysis: Descriptive statistics were computed per research question. Bibliometric and thematic analyses (technique-family and integration-pattern coding, keyword co-occurrence clustering) were performed using Python (pandas, matplotlib, networkx). All figures were generated programmatically from the coded dataset. Tools: Microsoft Excel (extraction and quality assessment), Python 3.12 (analysis and visualization), Rayyan.