Student's Conceptions of Artificial Intelligence and Machine Learning: Dataset of a Systematic Review

This dataset contains the supporting research data from a systematic literature review on K-12 students' conceptions of artificial intelligence and machine learning. The review synthesizes evidence from peer‑reviewed journal and conference articles published within the last 20 years, focusing on learners in ISCED levels 1–3. The aim is to identify common pre‑instructional conceptions that students hold about AI and ML, as well as to document the methods and age groups addressed in existing research. Five major academic databases were systematically searched using a predefined keyword strategy. After removing duplicates, studies were independently screened by multiple reviewers at title/abstract/keywords and full‑text levels. Metadata of the included studies were extracted using standardized forms, while qualitative data on student conceptions were analyzed through in‑vivo coding and categorized using the AI4K12.org Five Big Ideas of AI (Grade Band Progression Charts 2020-2022 version) as a framework. The resulting dataset provides a structured overview of the current evidence base and supports further research and development of age‑appropriate AI education. The data is structured as follows: screening-data.xlsx (Literature Screening Data): A spreadsheet with five worksheets that describe the individual steps of data aggregation. screening-data_screening.csv: All records exported from the databases, with duplicates removed, along with the inclusion/exclusion decisions made by reviewer 1 and reviewer 2 for each record, as well as the reason for exclusion, if applicable. screening-data_fulltext-screening.csv: All records included in the screening step, for which the full text was retrieved and screened. For each record, the inclusion/exclusion decisions made by reviewer 1, reviewer 2, and reviewer 3, as well as the reasons for exclusion and the final inclusion/exclusion decision for the review, were added. screening-data_snowballing-screening.csv: All records identified by snowballing of the retrieved records, with duplicates removed, along with the inclusion/exclusion decisions made by reviewer 1 and reviewer 2 for each record, as well as the reason for exclusion, if applicable. screening-data_snowballing-fulltext-screening.csv: All records included in the snowballing screening step, for which the full text was retrieved and screened. For each record, the inclusion/exclusion decisions made by reviewer 1, reviewer 2, and reviewer 3, as well as the reasons for exclusion and the final inclusion/exclusion decision for the review, were added. screening-data_included-in-review.csv: All records that were evaluated as eligible during the initial full-text screening and snowballing full-text screening and were analyzed during the review. codes.xlsx (Qualitative Analysis Data): A spreadsheet with three worksheets that contains all the in-vivo codes from the data analysis, including their category assignments, as well as the conceptions identified as "highly prevalent" and the codebook for assigning the corresponding concepts. codes_conceptions.csv: Contains all student conceptions extracted directly (in vivo) from the data, the corresponding publications from which they were taken, and their inductive categorization on two levels. codes_authors.csv: Contains the mapping between the short titles of the publications (filenames) and the full references. codes_codebook-concepts.csv: Contains the codebook for assigning concepts from the Grade Band Progression Charts (2020–2022 version) of AI4K12.org’s Five Big Ideas of Artificial Intelligence to the extracted conceptions. To date, this assignment has been limited to the highly prevalent conceptions identified using the audit procedure described by Akkerman et al. (2008). codes_highly-prevalent-conceptions.csv: Contains the conceptions identified as “highly prevalent,” which were published in the associated article (see below). In addition to a unique ID (C01–C29), these entries include information on their overall frequency and the number of publications in which they appeared, as well as an explanation of how they are assigned to the concepts. documents-basic-data.xlsx / documents-basic-data.csv: A spreadsheet summarizing the study data from the included publications. This includes age/grade level, school type, country of origin, sample size, gender distribution, date of data collection, survey method(s), evaluation method(s), information on whether they address the AI4K12.org Five Big Ideas of AI or conceptual change, and information on data availability. This data publication is part of the article accepted for publication in Computer Science Education (https://www.tandfonline.com/journals/ncse20): Kreinsen, M., Mercan, M. and Schulz, S. (2026). Student’s Conceptions of Artificial Intelligence and Machine Learning in K–12 Education: A Systematic Review. Computer Science Education. http://dx.doi.org/10.1080/08993408.2026.2738165