<![CDATA[Bibliometric profiles based on relevance-ranked searches may conflate patterns of research publication with patterns of record retrieval. This study examined the descriptive characteristics and interpretive limits of a bounded Crossref corpus on artificial intelligence in education. On 17 September 2026, the first 100 records returned for the query artificial intelligence education, restricted to journal articles published during 2016–2025, were obtained. Records were assessed in retrieval order, and the first 30 with an author field and the complete phrase artificial intelligence in their titles, excluding editorial or corrective items, formed the analytical corpus. Publication year, journal, authorship, Crossref citation counts, and seven non-exclusive title dictionaries were analysed descriptively. Twenty-six records (86.7%) were published in Computers and Education: Artificial Intelligence and four (13.3%) in the International Journal of Artificial Intelligence in Education; 25 appeared during 2022–2025. School-related terms occurred in eight titles, teacher-related terms in five, and ethics-related terms in two. Median authorship was two, and the median recorded citation count was 100. These distributions reflect a fixed search and stopping rule, not population-wide publication frequencies. The study demonstrates how transparent metadata analysis can identify observable patterns while exposing the inferential constraints introduced by search formulation, ranked retrieval, and restrictive eligibility criteria. ]]>
Copyrights © 2026
