Sentence Repetition Speech Therapy

What Sentence Repetition Actually Is

Sentence repetition is a language assessment and intervention technique where a clinician or caregiver says a sentence and the client repeats it back. That's the basic mechanics of it. But what makes it useful clinically is the structure behind the sentences. They're designed with specific grammatical targets embedded in them — relative clauses, passive constructions, conjunctions, morphological markers like past tense -ed or plural -s. The point isn't just that the person repeats something. The point is that you can watch exactly where they break down. I've used this with kids who have developmental language disorder and adults recovering from stroke. The pattern tends to be the same: they'll nail short, simple sentences, then hit a wall when the syntax gets complex. A kid might repeat "The dog ran" perfectly but struggle with "The dog that the cat chased ran away." That second sentence has an embedded relative clause, and that's where the grammar processing falls apart. You can see it happen in real time.

Getting Started With Sentence Repetition Speech Therapy

Here's the practical part. You need sentences that are graded by length and grammatical complexity. There are standardized batteries like the Nonword Repetition Task or the Token Test, but those are for assessment. For therapy, you build or source your own list. Start with two-word subject-verb combinations, move to three-word sentences with a direct object, then add modifiers and conjunctions one piece at a time. Keep a tracking sheet. Write down each sentence, the target structure, and whether the client got it correct on the first try, needed a model, or couldn't produce it at all. You need to know what percentage they're getting right before you move up. The usual rule of thumb is 80% accuracy before advancing. Below that, they're not actually acquiring the structure — they're just matching sounds by rote. Delivery matters more than people realize. Say the sentence at a natural pace. Not slowly. Not exaggeratedly. Just normal conversational speed. Then ask them to repeat it. If they fail, give them the sentence again once. Don't break it into chunks unless they're completely lost — chunking trains them to rely on you rather than building their own syntactic processing. One full model, then try again. That's it.

You should be doing this in short bursts. Twenty to thirty minutes, three or four times a week. Longer sessions don't produce better outcomes and they do produce fatigue, which looks like regression on paper but is just exhaustion. I had a case once where a ten-year-old was doing an hour-long session five days a week and his scores were getting worse over three weeks. Cut the session to twenty minutes and he stabilized within two weeks. The problem wasn't the method. It was the volume.

Why This Works When It Works

Sentence repetition isn't magic, but it does tap into something real. When a person repeats a sentence, they're engaging multiple language systems at once — phonological short-term memory, syntactic parsing, morphological encoding, and articulatory planning. Damage to or delay in any one of those shows up as a specific error pattern. Missing a morpheme like -ed or -s tells you the morphological system is weak. Dropping a whole clause tells you working memory or syntactic capacity is the bottleneck. You get diagnostic information for free while you're doing therapy. There's a counter-intuitive thing about sentence repetition that beginners miss. Practice effects are real but narrow. If a child practices sentences with passive voice, they get better at passive voice. They don't necessarily get better at active voice or at understanding language in general. The improvement is structure-specific. So you have to vary your targets deliberately. Rotate through different grammatical constructions across sessions. Don't drill one structure for weeks and hope it transfers. Another thing: accuracy on repetition doesn't always predict spontaneous production. I've had clients who could repeat a complex sentence perfectly when prompted but never produced anything like it on their own. That's normal. Repetition taps into different neural pathways than spontaneous language generation. If your goal is functional communication, you need to bridge that gap. After they can repeat the sentence reliably, you ask them to use it in a different context — make it a question, change the subject, tell the story using that sentence structure. That's where the carryover happens.

Common Pitfalls

The biggest mistake I see is using sentences that are too long too early. People want to see progress and they think harder sentences equal faster progress. It doesn't work that way. If a client is below 50% on a given length, you're not helping them by moving up. You're just training frustration. Go back down two levels, rebuild accuracy, then climb again. Another issue is not controlling for vocabulary. A child might fail a sentence because of the grammar or because one of the words is unfamiliar. If you're working with a multilingual client or a child with a limited expressive vocabulary, pre-teach or confirm understanding of individual words before testing the full sentence. Otherwise you're measuring vocabulary, not grammar. Sentence repetition also breaks down as a standalone intervention for certain populations. Kids with severe apraxia of speech will struggle regardless of grammatical complexity because their issue is motor planning, not language processing. In those cases, you need a different primary approach — things like PROMPT or integral stimulation — and you can use sentence repetition only as a supplementary activity at a very slow pace. Same goes for clients with significant auditory processing deficits. If they can't hold the sounds in memory long enough to parse the grammar, you need to address the auditory component first.

And don't expect this to fix fluency disorders. Stuttering and sentence repetition don't have a reliable productive relationship. Some people who stutter actually fluency more on repeated utterances, but that's a separate phenomenon and shouldn't be confused with language therapy.

Building Your Own Materials

You don't need to buy a commercial kit. A well-structured list you make yourself works fine. Here's a basic progression I've used: Level 1: The boy runs. The dog barks. (subject + verb) Level 2: The boy runs fast. The dog sees the cat. (subject + verb + modifier/object)

Level 3: The big boy runs fast. The dog sees the big cat. (adjective + subject + verb + adverb/object) Level 4: The boy who is big runs fast. (embedded relative clause) Level 5: The boy ran fast because the cat ran. (coordinated clauses with conjunction)

Level 6: The cat that the dog chased ran fast. (embedded clause with object reversal) Each level introduces one new structural element. Don't stack them. The difference between Level 4 and Level 5 is the jump from embedding to coordination — that's a big cognitive leap for someone with language impairment. Keep them separate. I made the mistake once of combining adjectives and relative clauses in the same sentence during Level 4. A client who was doing okay on relative clauses alone collapsed completely. Turns out the adjective added enough lexical load that the working memory capacity was exceeded. Took me a while to figure out what happened. The fix was simple — strip it back to one new element per sentence and rebuild from there.

Tracking Progress

Record sessions when you can. Not for the recording itself — for going back and listening. You'll catch things in playback that you miss in the moment. A missing morpheme you didn't notice, a hesitation pattern, the exact point where the client starts to decompose the sentence. Keep a simple graph. X-axis is session number. Y-axis is percentage correct. You want to see a gradual upward trend. If it flatlines for three sessions, something is wrong — either the sentences are too hard, the material isn't varied enough, or the client needs a different approach entirely. Don't just push harder. That's the most common error. Pushing harder when the data says the current level isn't working just wastes time and erodes motivation. For standardized benchmarks, the Clinical Evaluation of Language Fundamentals and the Phonological Symmetry and Syllable Structures tests both include repetition components if you want norm-referenced data. But those are assessment tools, not therapy materials. Using them for therapy is possible but inefficient. You're better off building your own graded list and using standardized tests periodically to measure broader language ability.

The bottom line is that sentence repetition is a targeted tool, not a comprehensive therapy. It's good for grammatical structures and working memory. It's not good for vocabulary acquisition, pragmatic skills, or fluency. Know what it does, know what it doesn't do, and don't try to make it do things it can't handle.