Evanreads

Text-to-Speech for Dyslexia: Does It Actually Help?

Dyslexia · 6 min read · Updated 2026-08-10

Key points

• A meta-analysis found read-aloud tools improved comprehension for students with reading difficulties by an average effect of 0.35 — real, but small-to-moderate, with wide confidence intervals.
• It works because dyslexia's core difficulty is decoding, not language comprehension — hearing the words routes around the bottleneck.
• It helps least when listening comprehension is also weak, and it is an accommodation, not reading instruction.
• Follow the highlighted line with your eyes rather than listening passively, and pair it with recall.

Text-to-speech is one of the most commonly recommended accommodations for dyslexia, and one of the most commonly oversold. "It changed everything" and "it's a crutch" are both things you'll hear, usually from people with no evidence either way.

Here's the honest answer: on average it helps, the effect is real but moderate, and how much it helps you depends on something specific about your own reading. This guide covers what the research found, the mechanism, when it doesn't work, and how to set it up so you get the version that helps.

What does the research actually show?

The most direct evidence is a meta-analysis of text-to-speech and related read-aloud tools for students with reading difficulties. Pooling the studies, it found an average weighted effect size of 0.35 on reading comprehension, with a 95% confidence interval of 0.14 to 0.56 (Wood et al., 2018).

Worth unpacking that number rather than just quoting it:

• It's positive and statistically reliable. The whole interval sits above zero. This is not a placebo.
• It's small-to-moderate, not transformative on average. A 0.35 effect is a meaningful nudge in comprehension, not a different reader.
• The interval is wide, and study design explained some of the variation. That means individual results genuinely differ — some readers benefit considerably more than the average, some less.
• The authors call for more research, and they're right to. This is a real finding on a still-thin literature.

So: worth trying, likely to help somewhat, and not something anyone should promise you a specific result from.

Why it should help

The reason text-to-speech has a plausible mechanism — rather than just enthusiastic testimonials — comes from how reading comprehension decomposes.

Under the long-standing simple view of reading, comprehension is the product of two separable capacities: decoding (turning print into words) and language comprehension (understanding those words as language) (Gough & Tunmer, 1986). Dyslexia's core difficulty sits squarely in the first one — it is fundamentally a phonological processing difference, affecting the link between the sounds of language and the letters representing them (Shaywitz & Shaywitz, 2020).

That's the whole argument. If your decoding is effortful but your listening comprehension is intact — which for many dyslexic readers is exactly the profile — then hearing the words routes around the bottleneck and lets your language comprehension do what it was already capable of.

There's a second, related mechanism. Effortful decoding consumes working memory that comprehension also needs, and working memory can only hold a few things at once (Sweller, 1988). Every unit of effort spent sounding out a word is effort not spent on what the sentence means. Offloading some of that decoding frees capacity for understanding.

When it doesn't help much

Being straight about the limits is what makes the rest of this trustworthy.

If listening comprehension is also weak. The mechanism above depends on your language comprehension being the stronger of the two capacities. If following complex spoken material is hard for you too, text-to-speech has nothing to route around to, and the benefit shrinks.

If decoding isn't your bottleneck. Readers whose difficulty is attention, vocabulary, or background knowledge won't get the same effect — the tool is aimed at a specific problem.

If you listen passively. Audio you're not attending to is easy to let wash past, especially at speed. Comprehension requires engagement whichever channel the words arrive through.

And most importantly: it is not instruction. Text-to-speech makes today's reading accessible. It does not build reading skill. The strongest evidence for building that skill points to structured literacy — explicit, systematic instruction in phonological awareness, sound–symbol relationships, syllables, and morphology (Hall et al., 2023; International Dyslexia Association). The two do different jobs and neither substitutes for the other. That distinction is covered in more depth in reading with dyslexia.

Is it cheating?

No — and the accommodation-versus-instruction distinction above is exactly why the question doesn't land.

Nobody suggests glasses are cheating at seeing. An accommodation changes how the material reaches you so that the ability you already have can be applied to it. It doesn't answer the questions for you, and it doesn't make the ideas easier. If the point of the assignment is to understand a text, a tool that gets you to understanding is doing the assignment, not avoiding it.

The one caveat worth keeping honest: if the point of a specific exercise is decoding practice, then yes, listening skips the thing being practised. Know which kind of task you're doing.

How to set it up so it actually helps

The studies pool a lot of different setups. These are the choices that tend to separate a version that helps from a version that doesn't.

Read along — don't just listen. Follow the words with your eyes while you hear them. Both channels feeding the same content is the point; audio alone throws away the visual information and makes drifting easy.

Highlight the line being read. Losing your place is the fastest way for reading-while-listening to fall apart. A moving highlight keeps eyes and ears synchronized without effort.

Slow it down before you speed it up. People discover speed controls and immediately go to 1.5×. Start at a pace where you comprehend everything, and only raise it while comprehension holds.

Use a voice you can stand. Robotic voices are tiring, and fatigue ends sessions. If a more natural voice keeps you reading for twenty more minutes, that's not a luxury.

Keep the visual side tidy too. Increasing letter spacing improves reading speed and accuracy for dyslexic children with no training at all (Zorzi et al., 2012), and surrounding clutter measurably slows letter recognition (Pelli & Tillman, 2008). Reading one line at a time, well spaced, makes the text you're following easier to follow. Relatedly, dyslexic readers using short lines on an e-reader read faster and comprehended better than on paper (Schneps et al., 2013).

Work in short sections, and recall after each. Listening can feel like understanding in the same way re-reading does. Close the section and say what it covered — retrieval practice is what turns exposure into memory (Roediger & Karpicke, 2006).

The takeaway

Text-to-speech is a legitimate, evidence-supported accommodation with an average comprehension benefit of about 0.35 for students with reading difficulties — real, moderate, and variable between people. It works by bypassing decoding so your language comprehension can do its job, which means it helps most when decoding is genuinely your bottleneck. It is not instruction, and it is not cheating.

The practical test is simple: try it for a week on real assigned reading, follow the highlighted line rather than listening passively, and check your comprehension honestly afterwards. You'll know quickly.

Evanreads has this built in — text-to-speech that highlights the line as it reads, alongside adjustable letter and line spacing and one-line-at-a-time focus, so the audio and the visual reinforce each other instead of competing. Standard voices are included; more natural-sounding narration is available as a $5/month Premium Voice add-on for anyone who reads with it every day.

Try Evanreads free — no account needed →