Writing Difficulties: What Actually Works — and What the Evidence Does Not Yet Support

If your child can tell you a good story out loud and then hands in four flat sentences, you have probably already been offered a remedy: a handwriting programme, a pencil grip, an app, a tracing workbook. This page is about which of those the research actually supports, with the numbers, and — just as important — where we looked for evidence and did not find it.

Quick answers

1. Three different problems look identical on the page

A short, messy, error-filled piece of writing can come from three very different bottlenecks, and the remedy for one does nothing for the others:

Most children who struggle have some mixture, but one usually dominates. Finding out which one is the single most useful thing you can do, and it costs nothing: ask the child to dictate the same task to you while you write it down. If what you transcribe is markedly better than what they produce themselves, the obstacle is in getting it onto the page. If the dictated version is also short and shapeless, the obstacle is in the planning.

2. What has the strongest evidence: teaching the process, explicitly

The best-evidenced approach for struggling writers is an explicit, staged method of teaching children how writers actually work — known in the literature as Self-Regulated Strategy Development. It is not a worksheet. It runs through six stages, and the point of each is that nothing is left for the child to infer:

1. Build background knowledge — what good writing in this genre looks like, and the memory aid for the strategy.
2. Discuss it — when and why you would use this strategy, and the child's own goal.
3. Model it — the adult writes in front of the child, thinking out loud, including the part where it goes wrong and gets fixed.
4. Memorise it — the steps are practised until recalling them costs no effort.
5. Support it — guided practice with feedback, supports gradually withdrawn. Usually the longest stage.
6. Independent performance — writing without the organiser, still setting goals and self-monitoring.

Why the third stage matters more than it looks: modelling out loud is the stage most often skipped at home, because watching an adult struggle visibly feels like a bad lesson. It is the opposite. A child who has only ever seen finished writing concludes that competent people produce it in one pass, and that their own crossing-out is evidence of being bad at it.

3. The numbers — and what they do when someone else runs the study

We read two independent documents, on 12 October 2026, that report on this method. Both report large effects. They do not report the same number, and the gap is instructive rather than embarrassing.

Read the table downwards. The meta-analytic figures of around 1.15 come substantially from smaller studies run close to the people who developed the method. The independent trial, delivered by ordinary classroom teachers across eleven schools, got 0.87 on writing quality. That is still a large effect by any standard in education — but it is about three-quarters of the headline figure, and that shrinkage is the normal fate of every education method that leaves its inventors' hands. Anyone quoting you 1.17 as what your child will get is quoting the best case.

The largest effect in the trial, 1.87, was for adhering to the prompt — writing about what was actually asked. That is worth naming because it is the complaint teachers make most often and parents hear least: the child wrote something, but not the thing that was set.

The authors' own cautions, published here as they stated them: the number of lesson-fidelity observations for some teachers was low, so what was observed may not represent the teaching across the whole period; the intended amount of teaching was not verified for every teacher; no measures were taken of whether teachers or children found it workable; and some children in the control group reported having practised the final writing prompt in advance, which would have pushed the control scores up and the measured effect down.

4. How much teaching, and for whom

• Dosage: typically 9 to 12 sessions of 20 to 25 minutes. The source adds that more intensive or longer sessions may be required. This is a useful benchmark for judging an offer: a programme promising transformation in two sessions, or one that runs for a year without a measured check, is not describing what was studied.
• Best-researched group: children with identified learning disabilities in upper primary and middle school. This is where the evidence is strongest.
• Younger children: there is evidence of effectiveness with children as young as Grade 2, and the source itself calls it preliminary. We are repeating that word on purpose.
• Children with emotional difficulties: the method has been used with this group to a much lesser extent, so the evidence base is thinner.
• Settings: it has been used in whole classes, small groups and one-to-one.

5. What we could not find evidence for — stated plainly

This is the section that is usually missing, and it matters more than the rest.

Neither of the two documents we read reported any effect size for handwriting instruction, spelling instruction, or transcription. Both are about composition strategy. So we are not in a position to tell you how well a handwriting programme or a spelling programme works, and we are not going to imply an answer in either direction. It does not follow that handwriting teaching is useless — it follows that we did not verify it here, and that a confident claim about it from anyone, including us, needs a source you can check.

We also cannot tell you that this improves exam results. The practice alert states directly that little is known about the method's effects on standardised test performance for students with learning disabilities, and that more research is needed on whether the self-regulation skills persist over long periods or transfer to other kinds of writing. The studies measured the quality of pieces of writing. That is not the same thing as a grade on a paper written eighteen months later, and we will not pretend the second follows from the first.

And one disclosure. We sell tutoring. The honest reading of the evidence above is that a good part of what helps a struggling writer is a teaching method that a school or a parent can apply without buying anything — explicit modelling, a memorised plan, goals the child sets, supports withdrawn slowly. We would rather say that than sell you a package you did not need.

6. What to do over the next two weeks

1. Run the dictation comparison from section 1 twice, on different tasks. Keep both versions.
2. Keep a two-week written note: which subject, how long the task took compared with what was set, and whether the trouble showed up in the hand, the spelling or the content. This note is the most useful thing you can carry to a teacher or a specialist.
3. Talk to the teacher and compare their observation with yours. Where they differ is itself information.
4. Rule out the dull explanations first: vision, seating and desk height, whether the pen suits the size of the hand, and whether the sheer volume of written homework is the real problem.
5. Stop correcting the handwriting of a child whose bottleneck is planning. It adds friction to the task they already avoid, and it does not address the thing that is going wrong.
6. If you do start structured support, set a review point — eight to ten sessions, with a before-and-after writing sample you keep. If nothing has moved by then, the plan is wrong, not the child.

7. When you do not need to do anything yet

• Your child is in their first year of learning to write. Inconsistent letters and spacing are expected at that stage.
• Handwriting is untidy but speed, spelling and structure are fine. Untidiness on its own is not evidence of a difficulty.
• The aversion is to the quantity of homework, not to writing. That is a conversation with the school, not a programme.
• The change was sudden. Writing that was fine and then dropped away quickly has other explanations that deserve looking at first; a learning difficulty does not usually begin that way.
• It shows up in one subject with one teacher only. Look at the context before looking at the child.
• You have not yet found out which of the three bottlenecks it is. Buying a remedy before this point is a coin flip with your money and your child's patience.

8. Our role and its limits

We do not diagnose and we do not treat. Assessment is the work of qualified specialists, and this page is background reading, not a substitute for an individual assessment. Our role comes after you know what you are dealing with: structured one-to-one online teaching aimed at the bottleneck that was identified, with a review point and a kept writing sample so that progress is something you can see rather than something we assert. No promised outcome and no promised timeline — the range between children is genuinely wide.

Full learning-support guide for parents · Learning-support sessions

📱 WhatsApp +971 56 660 8336 — bring your two-week note and we will go through it with you

FAQ

What is the single most effective thing for a child who writes badly?

On the evidence we read, explicit instruction in how to plan and compose — modelled out loud by an adult, with a memorised plan, child-set goals and supports withdrawn gradually. Meta-analyses report mean effect sizes of 1.14 and 1.17 for writing quality, and an independent trial of 418 children reported 0.87. All three are large.

Will handwriting practice fix it?

We cannot tell you, and we are not going to guess. Neither of the two sources we read reported any effect size for handwriting, spelling or transcription instruction — they are about composition. What we can say is that handwriting practice does not address a planning bottleneck, so identifying the bottleneck comes first.

How many sessions should it take?

The method as studied typically runs 9 to 12 sessions of 20 to 25 minutes, with the source noting that more intensive or longer sessions may be needed. Treat that as a benchmark: a two-session transformation and an open-ended year-long plan are both outside what was tested.

Will it improve my child's exam grades?

Unknown. The practice alert states that little is known about the method's effects on standardised test performance for students with learning disabilities. The studies measured the quality of pieces of writing, which is not the same as a grade on a later exam.

Why do the effect sizes differ between sources?

Because the smaller, earlier studies that feed the meta-analyses were often run close to the people who developed the method, while the independent trial was delivered by ordinary teachers across eleven schools and obtained 0.87 rather than about 1.15. Methods usually shrink when they scale. The honest figure to plan around is the smaller one.

My child speaks far better than they write. Is that a sign?

It is the most informative single observation available to you, and it is worth testing deliberately: have the child dictate a task to you while you write. If your transcription is much better than what they produce alone, the obstacle is in getting words onto the page. If the dictated version is also short and shapeless, the obstacle is in planning.

Does any of this apply to writing in Arabic?

The composition part — planning, structure, goal-setting — is about organising ideas and is not specific to a script. The transcription part is not transferable in the same way: Arabic has its own obstacles, including letter shapes that change by position in the word and dots that carry the whole distinction between similar letters. We cover those in the Arabic version of this guide.

Is it too early to act if my child is in Grade 2?

There is evidence of effectiveness with children as young as Grade 2, and the source calls that evidence preliminary. Meanwhile the steps that cost nothing — identifying the bottleneck, ruling out vision and seating, keeping a two-week note, reducing copying volume — are reasonable at any age.

Elmadrasah.com team — 12 October 2026. Written from two independent documents read on that date: a professional special-education council's practice alert on strategy instruction in writing, and a 2026 report of a cluster-randomised trial that also cites earlier meta-analyses. Where the two give different figures we have published both. Where we found no evidence — handwriting and spelling instruction, and effects on standardised test scores — we have said so rather than filling the gap. This page is background reading and does not replace an individual assessment by a qualified specialist.