Investigating the predictors of academic language competences in primary school children: A machine learning approach
Preschool and Primary Education · 01.03.2026 · 14 (1) · 144-162
DOI: 10.12681/ppej.40307 ↗Kurz gefasst
Fachartikel in der Zeitschrift Preschool and Primary Education, online erschienen am 01.03.2026. Verfasst von Astrid Rank und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).
Themen: Künstliche Intelligenz in der Pflege, Kinderkrankenpflege
Abstract (Originalsprache)
Academic language plays a critical role in students’ ability to succeed in school. Since multilingual and socioeconomically disadvantaged students often face challenges in acquiring the necessary language register, academic language proficiency is widely recognized as a key factor in promoting educational equity. To investigate how academic language skills develop – and what factors influence this process over time, the Eva-Prim study (Rank et al., 2021) analysed longitudinal data from 570 German primary school students. Academic language comprehension and production in mathematical contexts were assessed. Drawing on over 1,000 student-, family-, and school-related variables, a machine learning approach (Random Forests) (Breiman, 2001) was employed to identify the most relevant predictors. The findings show that at the beginning of primary school, distal factors such as intelligence, parental education, and socioeconomic status were associated with academic language comprehension. Over time, however, the influence of these background characteristics declined, while trainable skills – particularly vocabulary, reading fluency, and mathematical competence, became increasingly important. Academic language production in grade four was predicted mainly by these proximal, trainable skills. Demographic factors – including migration background, played only a minor role in predicting academic language performance. These results suggest that academic language development is shaped less by static background variables and more by dynamic, educationally influenceable skills. Thus, supporting vocabulary and reading fluency within subject teaching may be key to fostering academic success for all students, regardless of their background.
Zitieren
APA
Rank, A., & Kraus, E. (2026). Investigating the predictors of academic language competences in primary school children: A machine learning approach. Preschool and Primary Education, 14(1), 144-162. https://doi.org/10.12681/ppej.40307
Vancouver
Rank A, Kraus E. Investigating the predictors of academic language competences in primary school children: A machine learning approach. Preschool and Primary Education. 2026;14(1):144-162. doi:10.12681/ppej.40307
BibTeX
@article{Rank2026,
author = {Rank, Astrid and Kraus, Elisabeth},
title = {Investigating the predictors of academic language competences in primary school children: A machine learning approach},
journal = {Preschool and Primary Education},
year = {2026},
volume = {14},
number = {1},
pages = {144--162},
doi = {10.12681/ppej.40307}
}
EUVIMED