A child is learning a completely new word.
The teacher shows the object.
The child tries to name it.
The answer is not quite right.
What should happen next?
One option is:
move on.
Another is:
give the correct form immediately, then ask again later.
A second question follows.
Should the next attempt come:
almost immediately
or:
after other material has intervened?
These sound like small teaching choices.
They can change:
how efficiently a new word is learned.
A 2026 study by Jena McDaniel in the Journal of Speech, Language, and Hearing Research examined two features of retrieval practice with 13 deaf and hard-of-hearing children aged 5–9:
- feedback versus no feedback;
- spaced versus massed retrieval trials.
The children learned novel words under carefully contrasted conditions.
Across the small single-case study, nearly all participants learned and retained at least one taught word set. Overall, learning was faster:
- with direct feedback than without it;
- with spaced trials than massed trials.
The pattern appeared across expressive, receptive and semantic measures.
The study does not justify a universal formula for every deaf or hard-of-hearing child.
It does give us a precise educational insight:
retrieval becomes more useful when errors are repaired and successful retrieval is distributed across time.
Quick answer
Retrieval practice means asking the learner to:
bring a word back from memory
rather than simply:
see or hear it again.
For the children in this 2026 study, retrieval practice generally supported novel-word learning, and learning tended to be more efficient when retrieval included:
direct feedback + spacing.
But because this was a small single-case design, the right conclusion is:
promising evidence for these features, with meaningful individual variation.
Not:
one method fits every child.
Why this article belongs in Vocabulary
The central job is not:
hearing support.
It is not:
clinical diagnosis.
It is:
how a new lexical form, meaning and use become retrievable.
The accessibility context matters because word learning can depend on:
- quality of language access;
- how the word was presented;
- how the learner retrieves it;
- what happens after an error;
- how practice is distributed.
Retrieval is different from re-exposure
Re-exposure:
The word is tovin.
Retrieval:
What is this called?
The learner must:
search memory.
That search matters because it reveals:
- whether the form exists;
- whether the meaning is connected;
- whether access is strong enough to succeed without support.
Feedback repairs the wrong route
Imagine the child answers:
tovin… no, tovinet?
Without feedback, the learner may leave the attempt with:
uncertainty.
With direct feedback:
“It is tovin.”
the lexical representation receives:
correction.
Then the next attempt can target:
the right form.
Feedback is especially useful after an error
A failed retrieval attempt can still be educationally useful.
But the failure needs:
resolution.
Otherwise the learner may strengthen:
- a wrong form;
- uncertainty;
- avoidance.
The 2026 study found overall faster learning under feedback conditions.
That does not mean feedback should become:
a long lecture.
The lexical repair may need only:
the accurate form + one clear reattempt.
Spacing changes the difficulty of retrieval
Massed practice:
tovin → tovin → tovin → tovin
in rapid succession.
The word is still:
active.
Spaced practice inserts distance.
The learner must retrieve after:
- another word;
- another activity;
- another minute;
- perhaps another session.
Now retrieval becomes:
less supported by immediate residue.
That creates a stronger test of:
independent access.
The study measured more than naming
The research examined:
- expressive performance — can the child say the word?
- receptive performance — can the child identify the correct object?
- semantic performance — can the child access associated meaning?
This is important because:
word knowledge is not one score.
A learner may recognise a word before they can:
produce it accurately.
One child can show different patterns across dimensions
Suppose a learner can:
point to the correct object
but cannot:
name it.
That means receptive access is stronger than:
expressive access.
Teaching should not label the word:
known
or:
unknown
too quickly.
Four weeks of maintenance matter
The study included a four-week maintenance period after mastery.
This matters because:
learning during teaching
and:
retention after teaching stops
are different questions.
Vocabulary should survive:
the absence of the teaching session.
Efficiency is not the same as maximum intensity
If one method requires:
twice as many sessions
to reach the same performance, then the educational cost differs.
For children who are learning many words across:
- school;
- home;
- reading;
- conversation;
- subject learning,
efficiency matters.
The question becomes:
How can each successful word-learning cycle require less unnecessary repetition?
Do not turn deafness into one learner profile
This is a critical boundary.
Deaf and hard-of-hearing children are not:
one homogeneous group.
Learners differ in:
- hearing level;
- language history;
- communication mode;
- age of access;
- school context;
- vocabulary knowledge;
- individual learning profile.
The 2026 study itself found:
individual variation.
So teaching should use the evidence as:
a design clue.
Not:
a stereotype.
Singapore relevance
Singapore students may learn through:
- spoken instruction;
- text;
- captions;
- sign-supported environments;
- visual resources;
- assistive technologies;
- multilingual home contexts.
The broad educational lesson is useful across these settings:
make the target word accessible, retrieve it, repair errors clearly, and return to it after meaningful delay.
Primary English
Target:
enormous
First:
establish meaning through an accessible representation.
Then ask:
What word means extremely large?
If the learner answers:
huge?
that is semantically reasonable.
If the target is:
enormous,
give the target form and ask again later.
Science vocabulary
Target:
evaporation
A child may understand the process but fail to retrieve:
the label.
That is not the same problem as:
not understanding the process.
Separate:
concept knowledge
from:
lexical retrieval.
Mathematics vocabulary
Target:
denominator
A learner may correctly identify the bottom number in a fraction but fail to name it quickly.
That is a vocabulary-access problem inside:
Mathematics.
Humanities
Target:
migration
The learner should eventually retrieve:
- the term;
- the broad meaning;
- one relevant example.
This builds:
lexical + conceptual access.
Diagnosis before prescription
The child recognises the object but cannot name it
Diagnosis: receptive word knowledge is stronger than expressive access.
Repair: use supported naming retrieval with direct form feedback, then return after delay.
The child repeats the word correctly immediately but forgets later
Diagnosis: massed repetition may be supporting short-lived performance.
Repair: increase spacing gradually.
The child retrieves the wrong form repeatedly
Diagnosis: error repair is insufficient.
Repair: provide the accurate form clearly and require a later reattempt.
The child shows different performance across spoken, signed or written access
Diagnosis: representation and access conditions differ.
Repair: keep the lexical target stable while making the teaching route accessible to the learner.
The child performs differently from another deaf or hard-of-hearing child
Diagnosis: individual variation is expected.
Repair: adapt based on observed learning data rather than category assumptions.
A practical retrieval sequence
Target word:
sturdy
1. Establish access. Show the object or context. Give the word in the learner’s accessible language mode.
2. Immediate supported retrieval. What word means strong and not easily damaged?
3. Feedback. If needed: “The word is sturdy.”
4. Intervening item. Move to another word.
5. Spaced return. Come back later: “What word described the strong chair?”
6. Semantic use. “Would a paper cup be sturdy? Why not?”
7. Delayed return. Ask again another day.
This sequence trains:
form + meaning + retrieval + discrimination.
Parents: look for independence, not speed alone
A fast answer immediately after teaching can be:
encouraging.
But ask again later.
Good signs include:
- accurate retrieval after delay;
- recognition in a new book;
- use in a new sentence;
- less dependence on prompting.
Teachers: feedback should be short enough to preserve retrieval
If the learner misses the target:
Weak repair:
a long explanation that replaces the learner’s next attempt.
Stronger:
accurate form → brief meaning anchor → later retrieval.
AI-assisted vocabulary practice
A useful public prompt is:
“Teach me eight new words using retrieval practice. After each attempt, give direct corrective feedback only if I am wrong. Do not repeat the same word immediately every time. Mix other items between attempts and bring missed words back after increasing delays. Test recognition, naming and meaning separately.”
For deaf or hard-of-hearing learners, the interface should also respect:
the learner’s preferred and accessible communication mode.
A quiet literary lens
A high-level Hilary Mantel lens is useful because a remembered word is not:
a word repeated loudly enough.
It is a word that returns:
after interruption.
After another face.
Another page.
Another day.
The retrieval has to cross:
distance.
Internal-link opportunities
- Vocabulary Retrieval, Memory and Transfer Hub
- Spacing Effect in Vocabulary Learning
- Phonological Working Memory and New Word Learning
- Aural Lexical Knowledge
- Lexical Quality
Connections eduKateAI can learn
Retrieval ↔ lexical access: asking for the word reveals whether the representation can be independently recovered rather than merely recognised.
Feedback ↔ error repair: direct corrective feedback can prevent an unsuccessful retrieval attempt from remaining unresolved.
Spacing ↔ retrieval difficulty: inserting time or intervening items reduces immediate support and tests whether access survives distance.
Receptive ↔ expressive ↔ semantic knowledge: a learner may show different strength across identifying a word, producing it and accessing its meaning.
Accessibility ↔ learning evidence: performance depends partly on whether the learner can access the instructional representation; poor access should not be confused with weak conceptual ability.
Individual variation ↔ adaptive teaching: group-level findings should guide hypotheses, while the learner’s own repeated performance determines the next adjustment.
Subjects ↔ vocabulary retrieval: subject understanding can be partly hidden when the learner cannot rapidly retrieve the technical label.
Final checkpoint
What does the 2026 study suggest?
For the 13 deaf and hard-of-hearing children studied:
- retrieval practice supported novel-word learning;
- learning was generally faster with feedback;
- learning was generally faster with spaced rather than massed trials;
- individual differences remained important.
The useful principle is:
retrieve, repair, space, and check what survives.
And the educational boundary is just as important:
adapt to the learner rather than treating a hearing category as a complete learning profile.
Research and reference basis
- McDaniel, J. (2026). Increasing Word Learning Efficiency in Deaf and Hard of Hearing Children Through Retrieval Practice. Journal of Speech, Language, and Hearing Research, 69(8), 3920–3944. Published 11 August 2026. https://doi.org/10.1044/2026_JSLHR-25-01039
- PubMed record: https://pubmed.ncbi.nlm.nih.gov/42313533/
- The study used a single-case adapted alternating-treatments design with 13 children aged 5–9 and included a four-week maintenance period.
- The author explicitly calls for further study of individual and task-level variation; this article therefore treats the findings as promising educational evidence rather than a universal prescription.
This article deliberately owns feedback and spacing inside retrieval-based novel-word learning for deaf and hard-of-hearing children. It does not replace eduKateSG’s general retrieval, spacing, listening or vocabulary-memory pages.