The Litkey Corpus (in english)

Investigating Spelling and Writing Development Based on Authentic Texts Produced by German Primary School Children

Ronja Laamann-Quante (Uni Bochum)

In this talk, I will present the Litkey Corpus, a longitudinal collection of descriptions of picture stories written by German primary school children. In total, the corpus comprises 1,922 texts produced by 251 children between second and fourth grade. All texts were manually transcribed and annotated with a target hypothesis that corrects orthographic errors. Furthermore, the corpus was semi-automatically enriched with various annotations (especially on the phonological, morphological and orthographical level) and comes in different formats, which makes it a treasure chest for studying spelling and writing development from various perspectives. I will showcase some of the research that the corpus has been used for so far, including the prediction of spelling errors using machine learning, and discuss possible future extensions of the corpus.