Using an app-based screening tool to predict deficits in written word spelling at school entry

Bibliographic Data

ID22166168
AuthorsMartin Schöfl (0000-0002-7254-2621, Johannes Kepler University of Linz, corresponding author), Gabriele Steinmair (0000-0001-6202-2705, University of Education Upper Austria), Sabine Zepnik (0000-0001-9167-1739, University of Education Upper Austria), Christoph Weber (0000-0002-6624-7909, Johannes Kepler University of Linz)
Year2024
Volume9
Publication date2024-11-04
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueFrontiers in Education (JOURNAL)
Journal identifiersISSN: 2504-284X • E-ISSN: 2504-284X
PublisherFrontiers Media SA (PUBLISHER • CH)
DOI10.3389/feduc.2024.1378493
OpenAlexW4404038665
LanguageEN
Citations received1
References cited69

Introduction The first year of schooling is crucial for the further development of spelling abilities in children, which makes early assessment and intervention essential. The aim of this study was to develop and validate an efficient and cost-free screening tool for identifying spelling problems in community school settings around the time of school entry. Methods A broad range of precursors of spelling (vocabulary, grammar, letter knowledge, phonological awareness, phonological working memory, rapid automatized naming) were assessed in 522 Austrian first graders (6–7 years of age) in the first weeks of schooling. At the end of first grade, spelling abilities were assessed by newly developed spelling tasks based on the trochaic foot. By applying logistic regression with the least absolute shrinkage and selection operator (LASSO), we aimed to select a set of important predictors of spelling problems at the end of grade 1 (i.e., scoring below the 16th percentile in the spelling test). Results Our analysis identified letter knowledge (i.e., an aspect of phonological information processing) and sentence repetition (i.e., a measure of grammatical knowledge) as important predictors of spelling problems. The screening tool has acceptable diagnostic accuracy [area under the curve (AUC) = 0.0.725 and DeLong 95% CI (0.666, 0.784)]. Further analyses indicated that the AUC differs neither between boys and girls nor between children with and without German as their first language. Discussion These results suggest that administering the screening tool during the first weeks of schooling is a valid approach to identifying spelling deficits, which in turn enables early targeted pedagogical interventions. Practical implications for spelling instructions are discussed

Human–computer interaction · Linguistics · Natural language processing · Speech recognition · Spelling · Text entry · Child Development and Digital Technology · Computer Science · Reading and Literacy Development · Writing and Handwriting Education · Artificial Intelligence

Tools

Open DOIOpen Access