Game Mechanics Systematic Review Dataset

This dataset supports a PRISMA 2020 systematic review of empirical studies that relate an identifiable element of a digital game, analysed through player interaction data, to learning measured independently of the game. It contains the search, screening, eligibility, data extraction, coding, quality appraisal and synthesis data for the 36 included studies (91 analysed game elements). Studies were identified through Scopus, Web of Science and ERIC (journal articles, 2017–2026, English or Spanish, open access) and one iteration of backward and forward citation searching via OpenAlex. Eligibility required (E1) a digital game with an identifiable element, (E2) player interaction data, (E3) a learning measure independent of the game, and (E4) an analysis linking the element to that measure. The workbook includes the following sheets: systreview: bibliographic data (APA 7) and verbatim extraction for each included study: term and definition of the element, analysed element, interaction indicator, learning measure, declared limitations and link result with its statistic, with page references. fields: definition of every field in every sheet and the list of codes. syntax: search strings, dates and citation-searching procedure. screening: title/abstract decisions for 3,603 database records and 410 citation-search records. fulltext: full-text eligibility decisions with exclusion codes and evidence. Descriptives: year, journal, source, educational level, domain, design and quality. PI1–PI4: data and summary tables for each research question: conceptualisation of the element; nature of the element in the game system (Järvinen, 2008) and relation to content; inference logic, indicators and learning measures; direction of the element–learning relation. MMAT: methodological quality appraisal (Mixed Methods Appraisal Tool 2018) by criterion, with justifications. coding: element-level codes, consensus and justifications. SWiM: synthesis without meta-analysis (vote counting based on direction of effect, Wilson confidence intervals, sign test, subgroup and sensitivity analyses). Study identifiers correspond to record IDs from the deduplicated database search (R0001–R3603) or the citation search (S001–S410) and are consistent across all sheets. Abstracts are not included for copyright reasons; each record is linked through its DOI.