Digital game-based learning in Brazilian graduate theses: educational levels, disciplines and game mechanics in the national repository

Introduction: Digital game-based learning is widely studied, but the profile of Brazilian graduate research on it, by educational level, discipline and game mechanic, has not been described from the national repository. Methods: Descriptive study of open metadata from the Brazilian Digital Library of Theses and Dissertations. Records deposited from 2013 to 2025 were retrieved with a digital game query and screened with explicit rules. Educational level, discipline and game mechanic were coded with a bilingual dictionary applied to titles, keywords and abstracts. Growth was modelled with Poisson regression with total deposits as offset, and groups were compared with Holm-adjusted tests. Results: Of 2,787 records retrieved, 1,156 theses were eligible (907 master's and 249 doctoral). Eligible theses rose from 0.53 to 2.42 per 1,000 deposits, a rate ratio of 1.08 per year (95% CI 1.06-1.11). Basic education appeared in 35.1% of theses, and the level was not identifiable in 44.7%. Portuguese language and literacy (20.2%), health sciences (18.2%) and computing (16.9%) led the disciplines. A game mechanic was identifiable in 46.2% of theses, most often points and rewards (14.0%) and quizzes (13.5%). Simulation, health sciences and engineering were more frequent in higher education theses, and mathematics in basic education theses. Conclusions: Brazilian theses on digital game-based learning grew faster than total deposits, concentrated in basic education and in language, health and computing, and described game mechanics only partly.