«My Spelling is Wobbly» – Causes and consequences of word-level disfluencies in written

Abstract

The aim of this PhD-project was to explore word-level processes involved in writing, and in particular word-level disfluencies. I have investigate what predicts word-level processes and disfluencies, and how word-level disfluencies can influence aspects of the final text. Two broad questions were addressed; What are the causes of wordlevel disfluency in written production?, and What, if any, are the consequences of word-level disfluency when the writer is composing full text? Article 1 investigates the writing process and the written product of a group of dyslexic students and a group of control students. Results from this article indicate that students diagnosed with dyslexia have a word-level focus when writing, and that this word-level focus is related to the writing process and not them struggling to read what they have written. Article 2 is an investigation of the spelling process and spelling accuracy in a group of 6th graders. Results indicate that the spelling process persists beyond typing onset. Moreover, word-split performance and non-word spelling accuracy predict spelling accuracy. Spelling response latency was predicted by non-word spelling response latency, and by key-finding speed. Keystroke intervals within words was predicted by word-split performance, non-word spelling RT and key finding speed. Article 3 investigates the relationship between spelling, motor execution processes involved in keyboarding, text composing processes and text quality measures. Results indicate that the transcription measures; copying, key finding and spelling, all influence word-level processes when producing text. Moreover, results indicate that word-level disfluencies have a negative impact on measures of text quality. Article 4 is a theoretical investigation of existing technical aids for writing support, and the general ideas underpinning these. A shift from having correction as the main element, to a writing aid having fluency as the main principle is suggested. My conclusion is that word-level disfluencies are related to spelling, and that wordlevel disfluencies can influence aspects of the final text.