The R sound is famous for a reason. It is one of the last sounds children master, it is hard to see, and it can resist months of practice. This guide explains how SLPs teach /r/ — using techniques that have real published support. Every claim links to its source, so you can check us.
A note on who this is for: parents can use the home-practice ideas, but eliciting a new /r/ is skilled work. If your child is past the typical age and /r/ has not arrived, the right first step is a speech-language pathologist. Our free sound development checker shows where your child’s age falls.
Why is the R sound so hard?
Three reasons, all well documented:
R is late. In US data, 9 out of 10 children master /r/ only during the 5-year-old range (Crowe & McLeod, 2020). A shaky /r/ at age 4 is normal.
R is invisible. The tongue does its work far back in the mouth. A child cannot watch your lips and copy, the way they can with /p/ or /f/.
There are two correct tongue shapes. MRI research shows American English speakers produce /r/ with either a “retroflex” shape (tongue tip raised) or a “bunched” shape (tip down, back of tongue raised) — and both sound essentially the same (Zhou et al., 2008). A child who fails with one shape may succeed with the other.
Techniques with published support
A 2020 tutorial in Language, Speech, and Hearing Services in Schools pulled together the motor-based strategies with actual literature behind them (Preston et al., 2020). The short version:
Technique | How it works | Source |
|---|---|---|
Shaping from /l/ | Start from /l/ (tongue tip up), then slide the tongue back into /r/. The oldest published elicitation path. | Shriberg (1975), via Preston et al., 2020 |
Shaping from “ee” | The vowel /i/ uses the same side-of-tongue bracing as /r/. Hold “ee”, then pull the tongue back. | |
Lateral bracing cue | Teach the sides of the tongue to press against the upper back teeth — the one placement cue with a clear rationale. | |
Trying both tongue shapes | If retroflex fails, try bunched (and the reverse). Neither shape is “better”; the right one is individual. | |
Facilitative contexts | Some syllables make /r/ easier (for many children, “kr/gr” contexts or a strong “er”). Find the context that works, then expand from it. |
Notice what these have in common: they all start from something the child can already do, and move in small steps toward /r/. Expect the first correct sound in a syllable, not a sentence — and expect it to take repetition to become stable.
Popular tricks without evidence
Honesty matters more than a longer list, so here is what we could not find support for:
The “growl” trick. Everywhere on blogs, nowhere in the peer-reviewed literature. It may help some children by accident of context — but it is folklore, not a method.
Nonspeech mouth exercises. Blowing, tongue push-ups and chewing tools do not improve speech sounds. The evidence here is actually negative (Preston et al., 2020, citing Forrest & Iuzzini, 2008). Practice speech, not mouth gymnastics.
Pressing under the chin. Also unsupported, and possibly counterproductive.
R is really a family of sounds
Part of what makes /r/ therapy confusing is that “the R sound” is not one target. There is /r/ before a vowel (rabbit, run), /r/ inside consonant clusters (train, bridge), and vocalic R — /r/ fused with a vowel, as in bird, car, corn, chair. A child can be perfect in one context and lost in another, so therapy starts by probing them all and treating the ones that break.
This is why word lists organized by context matter. We keep free, pronunciation-checked lists for every context: R by word position, R blends like BR and TR, and each vocalic R variant separately — ER, AR, OR, AIR, EAR and IRE.
A realistic teaching sequence
Whatever technique elicits the first good /r/, the road after it looks the same. Each step earns the next:
Isolation or syllable. One good “er” or “ra”, produced on purpose, repeatedly. This step can take a while — that is normal.
Words. The new sound goes into short words, starting with the easiest context found during probing.
Phrases and sentences. The word survives next to other words: “a red rabbit”, then “I see a red rabbit.” Our verified R phrases and R sentences are built for exactly this step.
Conversation. The sound holds up in real talk — the slowest step, and the one home practice supports best.
Two rules of thumb hold across the research: practice at the level where the child is mostly successful (roughly 8 in 10 correct), and count trials, not minutes — many short correct repetitions beat long frustrated sessions.
Older kids and stubborn R: biofeedback
When /r/ errors persist into ages 9 and beyond (“residual errors”), a newer family of tools shows the child their own speech in real time:
Ultrasound biofeedback shows the tongue’s shape live. Two systematic reviews (2019 and 2026) report positive but variable results (Sugden et al., 2019).
Visual-acoustic biofeedback shows the sound’s acoustics as a picture the student learns to match. It has some of the strongest trial evidence for residual /r/, including a randomized trial with 108 children (McAllister Byun et al., 2016).
These are clinic tools, not home apps — but they are worth asking about if regular therapy has stalled on /r/ for a long time.
Practicing R at home (the safe part)
Once an SLP has elicited a correct /r/ and set the level, home practice is where the repetitions happen. Keep it short and frequent — a few minutes a day beats one long weekend session. You will need words matched to the target: our free, pronunciation-checked lists cover initial R, medial R, final R, plus R blends and vocalic R — or generate a printable R worksheet with the word position, age and theme your child needs, phonetically checked before you print.
Frequently asked questions
Why can’t my child say the R sound?
Usually because /r/ is genuinely late and hard — 9 out of 10 children only master it during the 5-year-old range (Crowe & McLeod, 2020). If your child is past that range, or frustrated by not being understood, see an SLP. Check the age math with our sound development checker.
Bunched or retroflex R — which should be taught?
Neither is superior. Both produce a normal-sounding /r/ (Zhou et al., 2008), and treatment research suggests choosing per child: if one shape stalls, switch (McAllister Byun et al., 2014).
Can the R sound still be fixed in teens or adults?
Yes. Residual /r/ errors are the most-studied target in biofeedback research, with gains reported well into the teen years (McAllister Byun et al., 2016). Later is harder, not hopeless.
How long does R therapy take?
There is no honest single number — it depends on the error type, age, practice frequency and the child. Beware of anyone promising “R in X weeks.” What speeds things up in every study: more correct practice trials, at the right level.
Written from the published literature (all sources linked above; see also ASHA’s Practice Portal on speech sound disorders). This article is educational and does not replace an evaluation by a speech-language pathologist.