If you know what you want to say, the process of getting your ideas down on paper – or, more commonly, on a computer screen – can be quick and easy. This will only be true, however, if you have the necessary linguistic, spelling, and handwriting (or keyboarding) skills to smoothly convert your ideas into words. If you find one or more of these processes difficult then it is likely that this will interfere with your thinking: Stopping in the middle of a sentence to struggle with how to spell a word, for example, can result in you, quite literally, forgetting what it was that you were going to say.

If you are a child in the early stages of school, or if you are learning to write in a second language, or if you have a specific learning deficit, then the demands of basic language processing can swamp attempts to express your ideas.

Lack of fluency when writing is, arguably, as much a problem as lack of accuracy. Spelling mistakes can be fixed retrospectively. Correcting failure to retrieve ideas is harder. A child who repeatedly hesitates mid-idea to struggle with a word but who finally gets the spelling right is probably at a greater disadvantage than a child who writes fluently but less accurately.

Disfluent writers are likely to become frustrated, to produce text with weak content, and simply to write less. Shorter texts provide less scope for teacher feedback, and therefore less opportunity for student learning.

“Spelling mistakes can be fixed retrospectively. Correcting failure to retrieve ideas is harder.”

Teachers therefore need to be alert to not just the accuracy and quality of a student’s completed texts, but to how fluently – how easily – the text was written. If the teacher also sees patterns in where their student hesitated then this would give direct insight into the specific aspects of text production that the student finds difficult.

Diagnosing underlying difficulties on the basis of fluency is something that happens easily and naturally when children are being taught to read: A teacher listening to a child read out loud can immediately identify where a child is struggling. Getting similar information from a child’s writing is much harder. Diagnosis of inaccuracy is straightforward: If a child is making errors in their spelling or sentence structure, then this is visible in their final text. But a final text does not provide any information about how fluently it was produced. This requires capture and analysis of the time course of writing as it develops, moment by moment, during production.

“Teachers therefore need to be alert to not just the accuracy and quality of a student’s completed texts, but to how fluently – how easily – the text was written.”

Collecting real-time writing data is relatively straightforward. If the child is writing by keyboard – as is increasingly common in some European schools – then, with appropriate software, each keystroke – the production of each letter – can be recorded with the time, to millisecond accuracy, of when it occurred. Capturing similar data from children writing by hand is trickier but is possible if the child writes on a digitising tablet and the resulting pen trace is marked up to identify letters and words.

A more difficult problem is how to take the resulting data and analyse it to enable robust inferences about underlying process. This is a non-trivial task, involving a combination of automated linguistic analysis and sophisticated statistical modelling. The EarlyWritePro Emerging Fields Group aims to contribute to these methods. Our long-term goal is to provide researchers and teachers with methods that will provide precise diagnosis of writing difficulties through automatic analysis of children’s writing disfluencies.


The EarlyWritePro Emerging Fields Group is a small, geographically diverse group (Belgium, Estonia, Germany, Netherlands, Norway, US, UK) with a unique combination of expertise in computational linguistics, statistical modelling (including machine learning) and educational and psycholinguistic processes in written production. Our goal is to develop methods for understanding early writing through analysis of process dysfluencies. We will report on progress and findings in future blog posts.

EARLI 2019

The European Association for Research on Learning and Instruction (EARLI) is an international networking organisation for junior and senior researchers in education. Representing over 2000 members in more than 60 countries, EARLI is the biggest educational association in Europe.

In times of constant change, the future is a moving target – difficult to predict and prepare for. Yet, education is doing just that. At the 18th Biennial EARLI Conference and the accompanying 23rd Conference of the Junior Researchers of EARLI, researchers in learning and instruction from all over the world come together to discuss current research findings. In order to think tomorrow’s education and education research, it is crucial to relate new findings to what we already know and elaborate how this will help foster sustainable learning processes and navigate what is yet to come. EARLI 2019 will take place in Aachen, Germany, from August 12th to 16th, 2019. More information at EARLI.org

Keep up to date with the BOLD newsletter