14 Jun How We Diagnose Dyslexia in a Second Language
Dyslexia is not a difficulty with any one language. It is a processing difference that travels with a child across every language they use, the same obstacle showing up in each. Getting this right matters most for children who speak English as a second, third, or even fourth language, where the ordinary signs of learning a new language are easily read as dyslexia, and real dyslexia is easily written off as just the language.
Can dyslexia be diagnosed in a second language?
Yes, and before a child’s English is strong. This cuts against the advice families in international schools hear most often, which is to wait. Wait until the English improves, the thinking goes, and then the real difficulty will show itself. The advice is well meant and usually wrong. The waiting does not sharpen the diagnosis, and it costs the child years of support in the meantime.
Why is dyslexia not tied to a particular language?
Because its source sits beneath language, in the cognitive systems that any language has to recruit. The core of dyslexia is a phonological processing difference, a difficulty in representing and manipulating the sounds of speech, together with slow retrieval of familiar names and reduced working memory. Those systems are used to read English, Spanish, Arabic, Thai or Mandarin alike.
This is why the difficulty transfers across languages rather than staying behind in the first one. The skills that go wrong in dyslexia are part of a child’s general linguistic machinery, and they carry over when a new language is learned.
A child whose phonological system struggles to map sounds to symbols in their first language meets exactly the same obstacle in their second, often a steeper version of it, because a new language adds unfamiliar sounds and patterns on top of a system that was already working hard.
The practical consequence is the single most reliable signal in the whole field. A genuine difficulty shows up in both languages. A language gap shows up only in the language the child is still learning.
An assessor who can see the difficulty in both places is looking at dyslexia. An assessor who sees a child reading and writing well in their home language while struggling only in English is most likely looking at language development, not a difficulty. Telling an ordinary English-language gap from a genuine learning difficulty is the question that sits underneath every referral in an international classroom, and dyslexia is the one it most often comes down to.
Does testing in a weaker language not make the diagnosis unreliable?
Not if it is built properly. It would, if the assessment relied on the language-loaded parts of a test, but it does not have to, and a good one does not. The skill is in separating the scores that a second language contaminates from the scores it leaves largely untouched.
Some measures depend heavily on acquired English. Vocabulary, verbal reasoning, and comprehension of complex English passages all reflect how much English a child has met, and a bilingual child can score low on them for reasons that have nothing to do with dyslexia. These are the scores that produce false alarms when read at face value.
Other measures depend very little on it. Naming speed, phonological awareness, working memory, processing speed, reasoning with shapes and patterns, and reading of nonsense words all tap the underlying systems directly, without requiring the child to know much English.
These are the scores that carry the diagnostic weight for a bilingual child. A valid second-language diagnosis is built by giving these measures the leading role and treating the language-loaded ones as context rather than verdict.
Which parts of a Psychoeducational assessment work regardless of language?
The language-fair measures, sometimes called language-light tasks. These are the components that have been shown to distinguish genuine difficulty from typical second-language development across very different language pairs, and they form the spine of any assessment of a bilingual child.
- Rapid automatised naming (RAN), assessed within the CTOPP-2. The child names a grid of familiar items, letters, numbers, or colours, as fast as they can. The items are simple and known, so the task measures retrieval speed rather than vocabulary, and slow naming is one of the most robust markers of dyslexia found across alphabetic and character-based scripts alike.
- Phonological awareness, also within the CTOPP-2. Blending, segmenting, and manipulating sounds reveal the phonological system directly. The difficulty here is not knowing English words but handling the sounds themselves.
- Nonword reading. Because nonsense words carry no meaning, a child cannot look them up in their vocabulary or guess them from context. This isolates the decoding mechanism, which is precisely where dyslexia lives, and it is one of the cleaner windows on a difficulty in a child whose English vocabulary is still growing.
- Working memory and processing speed, drawn from the WISC-V for children and young people or the WAIS for adults. Both sit at the heart of dyslexia and neither is principally a measure of how much English the child has acquired.
- Fluid and non-verbal reasoning indices within the same cognitive measures. These show the child’s reasoning ability with very little language demand, which gives the assessor a trustworthy estimate of underlying ability to read everything else against.
Attainment measures such as the WIAT are still administered, but for a bilingual child they are read as one input rather than as the answer, because attainment in English is the score most affected by how long the child has been learning it.
Does the assessor need to test in the child’s first language?
No. Formal testing in the first language helps where it is feasible, but it is not required, and its absence does not invalidate a diagnosis. What matters is that first-language evidence is gathered, by whatever reliable means are available.
In many international families, formal standardised testing in the home language is not practical. The student may have three languages, a home language with no standardised instrument, or a first language they speak well but have never been taught to read.
This does not leave the assessor without first-language evidence. A structured developmental and language history captures it i.e. whether the child reached early speech milestones on time, whether they read and write in the home language and to what standard, whether anyone in the family has a history of literacy difficulty, and how the child’s home-language literacy compares to their English. A child who reads fluently in their first language and struggles only in English is telling a different story from a child who struggles in both.
This is why first-language evidence carries weight even when no formal test in that language is possible. It is one of the threads the Psychologist weaves together with the test scores, and no single thread settles the question on its own.
Does dyslexia look the same in every language?
No, and this is where an assessor who understands a child’s language background separates a real difficulty from an artefact of the writing system. Dyslexia shows up differently depending on how regular the relationship is between letters and sounds in a given language, which is why the child’s orthographic history has to be read into the profile.
In transparent writing systems, where letters map onto sounds consistently, such as Spanish, Italian, German, or Finnish, dyslexia tends to surface less as reading errors and more as slow, effortful, dysfluent reading and slow naming speed. A dyslexic child reading a transparent language may read accurately but painfully slowly. The difficulty is real but quiet, and it can be missed by an adult watching only for mistakes.
In opaque writing systems, where letters and sounds line up unpredictably, English being one of the least regular of all, dyslexia shows up in accuracy as well as speed. The same underlying difficulty produces visible errors because the language gives the phonological system far less to rely on.
This has a direct consequence for international school students. A child who learned to read in a transparent first language may have looked like a competent reader there, accurate if slow, with the difficulty masked by the regularity of the script.
Moving into English exposes it sharply, and the English picture can look worse than the underlying difficulty because English is simply a harder orthography to read. An assessor who does not know the child came from a transparent-language background may misjudge the severity in either direction.
In character-based systems such as Chinese, reading recruits additional skills again, including morphological and visual-orthographic processing, though naming speed remains a dependable signal across scripts. Knowing where the child started is part of reading where they are now.
How do all the results map together into a diagnosis?
No single score decides anything. The diagnosis comes from reading the whole profile as a pattern, and the pattern answers one question: is the reading difficulty larger and more specific than the child’s English can explain?
The assessment starts by establishing what the child can do with language stripped out. The fluid reasoning and visual-spatial tasks of the WISC-V or WAIS, solving novel patterns and puzzles, give a clean estimate of underlying ability and set the expectation everything else is read against.
The English-loaded scores are then read for what they are. A bilingual child usually scores lower on the Verbal Comprehension Index, the vocabulary and verbal reasoning, than on fluid reasoning. That gap is expected. Reading it as low ability, or as a difficulty, is the most common way a capable child gets mislabelled.
The verdict turns on the layer beneath reading: working memory, processing speed, and the phonological awareness and rapid naming scores from the CTOPP-2. This is the cognitive signature of dyslexia, and it is language-fair. A child can carry a small vocabulary and an intact phonological system, or a large vocabulary and a weak one. These scores tell the two apart. The verbal scores cannot.
Attainment is read in that light. Word reading, spelling, fluency, and above all nonword decoding from the WIAT show literacy itself. The question is never whether English reading is weak, but whether it is weaker than exposure predicts and whether it tracks the cognitive signature.
Around all of this sits the context no test produces: the first-language and developmental history, the family history, the writing system the child first learned in, and how they have responded to good teaching.
The diagnosis lives in whether these agree. Two children make the point.
Both have strong fluid reasoning and weak English reading and spelling. On the literacy scores alone, both look dyslexic.
The first has normal phonological awareness and naming speed, intact working memory and processing speed, reads at age level in the home language, and is improving with teaching. Nothing beneath the English confirms a difficulty. This is language, not dyslexia, and a label would be a mistake.
The second has low phonological awareness, slow naming, reduced working memory and processing speed, particularly poor nonword decoding, the same difficulty in the first language, and a parent who recognises it from their own schooling. Everything beneath the English confirms it. This is dyslexia, and the child’s level of English changes nothing.
That is what the mapping does. It turns the second language from a barrier into a variable the assessor can account for. When the signature, the attainment, the first language, and the history point the same way, the diagnosis is sound however much English the child has. When they do not, a careful assessor says so, and a capable child is spared a label they do not need.
Should we wait until our child’s English is stronger before assessing?
Waiting does not make the diagnosis more accurate, because the measures that carry the diagnosis do not depend on English in the first place. A child’s phonological processing, naming speed, and working memory can be assessed accurately now, and they will not be measured any more accurately in two years.
What waiting does change is the cost to the child. Every term spent assuming the problem is only English is a term without the structured support a dyslexic child needs, and that support is most effective the earlier it starts. Waiting also lets the gap widen, because a child with an unaddressed difficulty falls further behind peers who do not have one, and the longer the difficulty is read as a language issue the more a child’s confidence and motivation erode under the assumption that they should be coping.
There is one sensible version of patience, and it is not the same as waiting. A school should put good, targeted English-language teaching in place and watch how the child responds to it, because the response is itself evidence. A language gap narrows with good teaching. A genuine difficulty barely moves. When a child has had sound provision and is still not progressing, that is the signal to assess, not a reason to wait longer.
How does Global Education Testing assess students who learn in English as a second, third, or fourth language?
This is the standard case in our work, not the exception. Global Education Testing assesses students across international schools worldwide where the language of instruction is English but the child speaks it as a second, third, or even fourth language, and the method is built around that reality rather than treating it as a complication.
The assessments are carried out remotely, over a secure video link, by HCPC-registered educational psychologists. Working this way means a family is assessed wherever they are, on whatever continent their work has taken them, without joining a waitlist or travelling to a clinic, which matters for families who move between countries on someone else’s timetable.
Our psychologists handle the multilingual picture through method rather than guesswork. Before interpreting a single score, the assessor builds the child’s full language history: every language spoken and by whom, the age the child met each one, where and how they learned to read in each, and any gaps or mid-system moves in their schooling. That history is what tells the psychologist how to read the scores that follow.
From there the assessment leads with the language-fair measures, the phonological and naming-speed tasks of the CTOPP-2, working memory and processing speed, fluid reasoning, and nonword decoding, and weights the English-loaded scores as a reading of exposure rather than ability. The orthography the child first learned in is read into the profile, so that a child who learned to read in a transparent language is not misjudged against an English yardstick.
Because our psychologists assess children from a wide range of language backgrounds rather than one population, they hold the comparison that makes the call possible: they know what ordinary second-language development looks like, and they know where it ends and a difficulty begins.
The result is a report written against DSM-5-TR and ICD-11 criteria, with the child’s language history read into the whole profile, that translates any finding into the classroom support and formal exam access arrangements recognised by the major international boards, including IB, Cambridge, Pearson Edexcel, and College Board.
Where the picture is a difficulty, the report says so and sets out what to do about it. Where it is language, it says that just as clearly. For a family weighing whether to assess now or wait for the English to catch up, the answer the assessment gives is already valid today.
Alexander Bentley-Sutherland is the CEO of Global Education Testing, the leading provider of Learning Development Testing tailored specifically for the International and Private School community worldwide.