Skip to content

Explicit Instruction: Vocabulary

Principle · All principles
Evidence · 32 claims (25 for, 6 mixed, 1 against) · 38 studies (13 causal, 11 quant-synthesis, 10 review, 2 theoretical, 1 associational, 1 qualitative), q1–q4 · 6 of 38 report an effect size · 22 claims rest on one study

Conditional relationship

For a learner whose understanding of a text, an explanation or a task is limited by particular words they do not yet know (everyday words for a young or second-language learner, general academic words, or the technical terms of a subject or job), explicitly teaching a small, selected set of those words (meaning in plain language, examples and non-examples, word parts where they help) and then arranging repeated encounters and use of them across reading, listening, speaking and writing is expected to produce better knowledge of those words, and better comprehension of texts that use them, than leaving the words to be picked up from exposure alone. The capability is word knowledge (verbal-association for form and meaning, concept where the word names a disciplinary idea) and its use in comprehension and production. The expected change is on measures that use the taught words, at an immediate horizon; comprehension of untaught texts, standardised comprehension and delayed-retention are separate claims, and the evidence for them is much weaker.

Claims test the core of this relationship, not its dose. Two syntheses of school-age vocabulary interventions found that teaching word meanings improves comprehension of passages containing the taught words, and does little for standardised or generalised comprehension (Vocabulary Instruction Improves Comprehension [+S]). A 1986 meta-analysis, read from its abstract, concluded that the most effective methods combined definitions with context, gave more than one or two exposures and asked for deeper processing (Pairing Contextual Encounters With Explicit Instruction Produces Stronger Vocabulary Outcomes Than Either Alone [+M]). Second-language syntheses show that exposure alone teaches only a small share of the words met (Incidental vocabulary exposure produces reliable learning, but learners pick up only a small share of the unknown words they meet [+M]). For English learners from pre-K to middle school, a What Works Clearinghouse practice guide rates teaching a set of academic words intensively across several days on six studies meeting its standards, five of them randomised trials (A WWC practice guide rests its recommendation to teach academic vocabulary intensively across several days on six studies meeting WWC standards (five RCTs, one quasi-experiment) [+M]; the claim page records the guide's rating, not the studies' effects). How the explicit teaching is done also matters in one small study: six middle-school students with learning disabilities scored higher on weekly content-vocabulary tests under a concept model than under definitions and sentence writing (The concept model of vocabulary instruction produced larger gains… [+M]). No claim here sets how many words to teach at once, how many encounters a word needs, how to space them, or when a word counts as learned; those parts of the design below are proposals.

The relationship is conditional on four things: which words matter for the learner's next texts and tasks (selection is the designer's first decision, and no claim settles it); the learner's current vocabulary and language proficiency (in second-language reading, a learner who knows too few of the surrounding words cannot use context to infer the rest); whether the learner can decode the written word at all; and the outcome and horizon that count. The population is a scope qualifier: most evidence below is from school-age first-language readers or from second-language learners at school or university, and carrying it to adults in work settings or to young children's oral vocabulary is an extrapolation.

How the converted siblings sit beside this page: Direct Instruction holds the general relationship between explicit explanation and modelling and less guided discovery; vocabulary is one domain of it, with its own evidence. Accessible Vocabulary & Syntax holds the decision about which wording to remove from a task and which essential terms to keep and teach; this page holds how to teach the words that are kept. Phonics holds decoding, which a learner needs before written vocabulary work can help. Scaffolding and Fading holds how glosses, word banks and sentence frames are withdrawn, and Spaced Learning and Retrieval Practice hold the scheduling and recall-practice models this page borrows for review.

Default design

Each step names its evidence status. The direction (teach selected words explicitly, then give many varied encounters and uses) rests on claims; the quantities are proposals to set locally and record.

  1. Select a small set of high-value words before the unit. Take the words the learner will meet in the next texts and tasks and must understand or use: words that recur, that carry the key ideas, or that the learner will need to say or write. Leave aside rare words a gloss can handle. A default dose of 5–8 new words per week of instruction, 3–4 per session, is an untested proposal (the earlier page asked for "a manageable set"); the reason to keep the set small is that the evidence for gains is on taught words, not on general comprehension (Vocabulary Instruction Improves Comprehension [+S]).
  2. Introduce each word explicitly, in about 2–3 minutes per word. Say and show it, give a learner-friendly meaning, two or three examples in contexts like the coming text, one non-example, and, where the word has transparent parts (prefix, root, suffix), point them out. Teaching word meanings directly is supported for taught-word comprehension (above); teaching word parts has a small effect on school-age children's vocabulary (Morphological Instruction Improves Vocabulary [+M]). For a word that names a concept, teach it as a concept (its defining features, examples and non-examples) rather than as a definition to copy into a sentence: in a multiple-baseline study of six seventh-graders with learning disabilities, mean weekly test scores rose from 63.75% under a definition/sentence-writing model to 90.67% under a concept model (The concept model of vocabulary instruction produced larger gains… [+M]; no effect size printed). The minutes are an untested proposal.
  3. Put the words straight into a text or task that uses them. Read, hear or watch the text in the same or the next session, with the taught words recurring. Combining definitions with contextual encounters is the authors' conclusion in one meta-analysis (Pairing Contextual Encounters… [+M]); exposure without teaching yields a small share of words (Incidental vocabulary exposure produces reliable learning, but learners pick up only a small share of the unknown words they meet [+M]).
  4. Ask learners to do something with each word, not only recognise it. Explain why an example fits or does not, answer a question that needs the word, use it in a sentence about their own experience or the content, sort words by meaning. The Wright and Cervetti review found that instruction involving active processing typically did more than giving a definition or using a dictionary (Vocabulary Instruction Improves Comprehension [+S]). One second-hand report links productive use in a task that needs the word to retention (Task-essentialness and productive use of new words… [+W]), and another reports that, among 125 Chinese ESL freshmen, reading tasks with more search and evaluation (glosses plus dictionary use) gave better word retention (Tasks with higher involvement load… [+W]).
  5. Plan repeated encounters across at least two to three weeks. A default of at least six encounters per word in varied contexts (two in reading, two in talk, two in writing or a task) is an untested proposal. The number of encounters goes with incidental word learning in second-language studies, a correlation rather than a tested dose (Vocabulary Knowledge Grows Incrementally With Repeated Encounters In Varied Contexts [+M]). The variety of contexts has one laboratory test: pseudowords trained in high-variability paragraphs were identified 26 ms faster than ones trained in uniform contexts (Words encountered in diverse contexts are identified faster… [+M]), a recognition-speed measure, not meaning or use.
  6. Review with spaced recall, 5 minutes at the start of later sessions. Bring each word back after about 1, 3, 7 and 14 days, asking learners to recall or use it before showing the meaning (an untested schedule). Supplemental computer-based spaced repetition gave much higher end-of-semester credit on taught items than class and homework alone, in one double-blind study of EFL cadets (Supplemental computer-based spaced repetition… [+M]).
  7. Respond to misuse as it appears. When a learner uses a word in the wrong sense, collocation or register, give a short contrast (the word in the right use beside theirs) and ask for a second attempt, rather than re-teaching the whole set. Contingent support is drawn from one-to-one tutoring in other domains (Contingent scaffolding improves learning more than fixed or absent support. [+M]); for words it is untested. Where learners will use the words in talk, also teach the moves for asking about and checking a word's meaning with a partner ("what do you mean by…?"); that rests on a theoretical argument with no data (Vocabulary instruction should focus on procedures of meaning negotiation… [+W]).
  8. Shift the source of meaning from teacher to learner across the unit. Early sessions: teacher gives the meaning and examples, glosses are on the page. Middle sessions: learners work out a meaning from context or word parts first, then check against the gloss or teacher. Late sessions: no glosses for the taught words; learners meet them in a new text and use them unprompted. Moving from explicit explanation to independent inference over time is from the earlier page; how to withdraw supports is modelled in Scaffolding and Fading, whose fading evidence comes from worked examples and a scaffolding review (Fading support promotes the transfer of responsibility from instructor to learner. [~S], carried by analogy). Untested for vocabulary.
  9. Move a word from "being taught" to "maintained" on use, not on a definition. A default criterion: the learner uses the word correctly, without prompting, in two different tasks (for example a spoken explanation and a written answer) in two different sessions. Then it joins a smaller maintenance review. Untested proposal; a definition test alone shows recognition, not use.
  10. Judge the design on the texts and tasks that matter. Check comprehension of a new text that uses the taught words, and use of the words in the learner's own speaking or writing, at the horizon the objective sets. Do not expect a general reading score to move from a small word set (Vocabulary Instruction Improves Comprehension [~S]).

When a condition is not met: if the learner cannot decode the written word, teach the word orally and through pictures or demonstration first, and pair it with decoding work (Phonics); if the learner knows too few of the surrounding words for context to help, gloss more words in the text rather than expect inference (Correct word inference in reading requires knowing roughly 95–98% of surrounding words… [~W]); if the learner already knows the words, skip to use in harder texts.

Fitting the design to a situation

The three facts that most change the decision:

  • Which words the learner must understand or use, in which texts and tasks, and by when. A designer who knows the coming texts can pre-select and recycle; one who does not will teach lists that never recur. If unstated, ask: what will the learner read, hear, say or write in the next few weeks, and which words in it would stop them?
  • The learner's current vocabulary, language background and decoding. A first-language child, a second-language beginner, a fluent adult entering a new field and a learner who cannot yet read the word need different selection, glossing and modes. If unstated, ask: in which language does the learner read best, how many of the words in a typical coming text do they already know, and can they read the target words aloud?
  • Whether the outcome is recognising words, comprehending texts that use them, or using them productively. Recognition can be reached with a few explicit encounters; productive use needs speaking and writing tasks; general comprehension is barely moved by a small set. If unstated, ask: what should the learner be able to do with these words, and how will you check it?
If the brief says… Then change… Basis
Learners are child (pre-K to primary), or are weaker readers Keep the set small (3–4 words per session), teach orally with pictures and actions, read texts aloud with the words in them; expect taught-word gains, not general reading gains Vocabulary Instruction Improves Comprehension [+S] (pre-K–12; students with reading difficulties gained most on custom measures); for students with learning disabilities, teach content words as concepts rather than definitions: The concept model of vocabulary instruction produced larger gains… [+M] (six middle-school students); doses untested
Learners are second-language learners, novice in the language Gloss unfamiliar words in the text (first-language or picture glosses), teach the most frequent words first, and do not rely on inference from context until most surrounding words are known Glossing yields medium-to-large positive effects on L2 vocabulary learning… [+W] (second-hand report of meta-analyses); Correct word inference in reading requires knowing roughly 95–98% of surrounding words… [~W] (second-hand); Proficiency moderates gloss effects… [~M] (one EFL study: benefits at middle proficiency levels only); first-language glosses: All three L1 gloss conditions significantly improve immediate vocabulary recall… [+M] and, from the same study of 135 Taiwanese undergraduates, only the gloss-with-L2-example and in-text gloss forms held on all three recall tests (Only L1-gloss-L2-ex and L1-in-text-gloss conditions… [~M]), so prefer those to marginal glosses where retention matters
Learners are fluent adults or professionals entering a new field (technical or workplace terms) Take the words from their real documents and tasks; teach word parts that recur in the field; check use in the job task (a handover, a report, a conversation with a client), not in a quiz untested; carried from school-age evidence on taught words and word parts (Morphological Instruction Improves Vocabulary [~M], school-age children only)
The goal is productive use (speaking, writing) rather than reading Add at least two speaking or writing tasks per word in which the word is needed to complete the task; give a sentence frame at first, then withdraw it Task-essentialness and productive use of new words… [+W] (second-hand, five learners); otherwise untested
The goal is reading comprehension in general, not of particular texts Say plainly that a small taught set will not move a general comprehension score; spend more time on wide supported reading with glosses, and teach words that recur across many texts Vocabulary Instruction Improves Comprehension [~S] (d = 0.10 on standardised measures against 0.50 on researcher-built ones)
Setting is online-self-paced, phone-only, or no teacher present Build each word as a short card (meaning, example, picture, audio) with recall prompts scheduled by the app over two to three weeks; put the words into short readings or audio inside the course; ask for one typed or spoken use per word Supplemental computer-based spaced repetition… [+M] (EFL cadets, alongside classroom teaching, not instead of it)
Some learners are remote and some in the room Put the word set, examples and the shared text where both groups see them (shared screen or document); take uses in the chat or a shared document as well as aloud, and pair a remote learner with an in-room one for the speaking task untested proposal
One teacher with a large group, or scarce facilitator time (a volunteer, one short check-in a week) Teach the set to everyone at once in 10 minutes; let spaced review run on cards or an app; in the check-in, look at one thing only: a written or spoken use of two or three of the week's words untested proposal; the review part rests on Supplemental computer-based spaced repetition… [+M]
A single session only, or a fixed short time budget (for example 20 minutes) Teach 3 words, not more: 6–8 minutes introducing them, 8–10 minutes reading or doing a task that uses them, 3–4 minutes of use in the learner's own sentence; give a take-away card for later review; do not claim retention untested proposal
Learners have low confidence, low literacy or anxiety about speaking Let learners show they know a word by pointing, sorting pictures or acting before saying or writing it; use pairs before whole-group use; never correct misuse in front of the group untested proposal
The brief names its own routine (a word wall, a Frayer-style chart, a vocabulary log, a glossary its course uses) Keep the routine; change only what the evidence says matters: put each word into a text or task soon after it is entered, bring it back several times across weeks, and ask for one use, not only a definition Pairing Contextual Encounters… [+M]; Vocabulary Knowledge Grows Incrementally… [+M]
The brief gives its own word list, categories or rubric (a course list, a standards list, a job's terminology) Teach from that list in its own categories; select within it by what the next texts and tasks need, rather than replacing it untested
Learners keep a personal word journal or notebook, or the material is private The learner decides what goes in it and who sees it; check use through a separate task, not by reading the journal untested proposal
The output is a product or performance for a real audience (a presentation, a client conversation, a published text), and the brief wants it kept whole Teach the words the performance needs beforehand and rehearse them in parts of the real task, then let the performance run whole; do not turn it into a vocabulary test untested proposal
Stakes are high (an exam with a set word list, a safety or medication term at work) Raise the criterion for "known": correct unprompted use in three tasks, plus recognition in the exact form the exam or job uses; keep spaced review until the event untested proposal

Six of the fifteen rows rest on a claim for their main move (child learners, second-language novices, productive use, comprehension goals, self-paced settings, a named routine), two more cite one for a part or a limit (adults, scarce facilitator time), and seven are untested proposals. The adult row carries school-age evidence to adults and says so.

Observation, state and explanation

What can be observed is a response to a word under stated conditions: the item or text it appeared in, whether a gloss or picture was available, what the learner said, wrote or chose, and how long after teaching. "Knows the word", "has partial knowledge" and "cannot infer from context" are inferred states. Recognising a meaning in a multiple-choice item, recalling it, understanding it in a new text and using it correctly in one's own sentence are different capabilities.

Response Plausible explanations Observation that could change the interpretation
Gives the definition, cannot use the word in a sentence or a new text Recognition without productive knowledge; the new text's other words or syntax are the barrier; the concept behind the word is not held Ask for a use in a short spoken explanation with a frame, then present the word in a new passage with the surrounding words glossed; see which, if either, changes the response.
Understands the word in the taught passage, not in a new text Knowledge tied to one context; a different sense of the word in the new text; too many unknown surrounding words Give a second new text with familiar surrounding words and the same sense; then one with a different sense.
Uses the word in the wrong collocation or register ("make a research") The meaning is held but its use is not; first-language transfer; the examples given showed only one pattern Ask the learner to choose between two uses, and to say which sounds right in their first language; give two more examples of the right pattern.
Knew the words at the end of the unit, much weaker two weeks later Too few encounters after teaching; no recall practice; the words did not recur in later texts Re-test with a new item at the delay and count the encounters and reviews each word actually had.
Cannot read the word aloud, though the meaning was taught A decoding gap; an irregular spelling; the word was taught only orally Present the word spoken and written separately, and ask for the meaning each time.
Avoids the taught words in speech or writing, using simpler ones Uncertainty about use; fear of error in front of others; the task did not need the word Give a task where the word is needed, privately or in pairs, and compare.

These rows are proposals, untested with learners as diagnostics. A probe is itself another encounter with the word, so record each one, and keep the explanations that remain open rather than choosing the one whose name fits.

Evidence and qualifications

The limiting claims.

Learners, settings and outcomes differ across these claims (school-age first-language readers, school-age English learners, students with learning disabilities, second-language university students, military cadets, laboratory participants), so do not rank them by effect label. None tests a number of words per session, a number of encounters, or a criterion for a word being learned, and none follows adults in work settings.

The evidence by question

Further evidence, not yet read against this model

Claims this page cited before it was rewritten as a conditional model, and others found while rewriting it. Read against the model on 2026-10-15 and left here because none tests vocabulary teaching; each says how far it bears.

Objective and learner-valued goal

The designer's objective names the words and the capability: which terms the learner must understand in which texts, or use in which tasks, judged how and when. Elicit separately what the learner wants from the words: to follow a lecture or a briefing, to sound competent in a meeting, to pass an exam, to read to their child, to stop feeling lost in a new job. Agreement and divergence are both observations. A learner who values passing a vocabulary quiz may memorise definitions and never use the words; one who values speaking at work may want the everyday phrases around a term more than the term.

For example, an adult second-language learner in a work-readiness programme may want to describe their experience confidently in a job interview, while the designer's objective is the general academic words of a reading syllabus. Teaching the interview words (and the phrases they sit in), and checking them in a practice interview, serves the learner's goal; a reading quiz on academic words serves the designer's. Record both, so that a quiz score is not read as the capability the learner came for.

What would revise this model?

The expectation should weaken if well-controlled comparisons of explicit teaching plus contextual encounters against contextual encounters alone, with comparable time and learners, show no advantage on taught-word comprehension or use, or show that the advantage disappears at a delay while incidental learning holds. If studies with adults in work settings, or with young children's oral vocabulary, find different results from the school-age and second-language evidence, the scope should narrow. If explicit teaching of a small set keeps failing to move comprehension of untaught texts, keep that limit rather than claiming general reading gains. If a study finds a number of encounters, a spacing schedule or a set size that changes outcomes, replace the proposals above with it. Do not protect the model by calling every failure "too few encounters" after the fact.

Recognising a word, recalling its meaning, comprehending a text that uses it, using it in one's own speech or writing, and keeping it at a delay are separate claims. The present evidence establishes the first three for taught words over the short term in school-age and second-language learners; it establishes no dose, no schedule and no criterion for a word being learned.

Claims

Examples

Key Sources

  • Bruce Taylor, D., Mraz, M., Nichols, W. D., Rickelman, R. J., & Wood, K. D. (2009). Using explicit instruction to promote vocabulary learning for struggling readers. Reading & Writing Quarterly, 25(2-3), 205-220. doi:10.1080/10573560802683663
  • Madrigal-Hopes, D. L., Villavicencio, E., Foote, M. M., & Green, C. (2014). Transforming English language learners' work readiness. Adult Learning, 25(2), 47-56. doi:10.1177/1045159514522432
  • Riccomini, P. J., Smith, G. W., Hughes, E. M., & Fries, K. M. (2015). The Language of Mathematics: The Importance of Teaching and Learning Mathematical Vocabulary. Reading & Writing Quarterly, 31(3), 235-252. doi:10.1080/10573569.2015.1030995
  • Snow, C. E. (2010). Academic language and the challenge of reading for learning about science. Science, 328(5977), 450-452. doi:10.1126/science.1182597
  • Mohamed Farrag Ahmed Badawi. (2019). The Effect of Explicit English Morphology Instruction on EFL Secondary School Students’ Morphological Awareness and Reading Comprehension. English Language Teaching, 12(4). https://doi.org/10.5539/elt.v12n4p166
Page metadata
Field Value
type principle
id explicit-instruction-vocabulary
aliases explicit-morphology-in-integral-efl-curriculum
title Explicit Instruction: Vocabulary
description For a learner whose comprehension or use of language is limited by particular unknown words, explicitly teaching a small selected set and then arranging repeated encounters and use across contexts is expected to improve knowledge of those words and comprehension of texts that use them more than exposure alone, which claims test for school-age and second-language learners, but not for general comprehension, dose or schedule.
status review
canonical True
generated by claude/unspecified on 2026-10-15
trust unverified — no one has confirmed this page's content
sources 5 — bruce-taylor-2009, madrigal-hopes-2014, riccomini-2015, snow-2010, mohamed-farrag-ahmed-badawi-2019

Frontmatter is the page's machine-readable record — see the Schema & Guide.