Skip to content

Pre-reading Questioning

Principle · All principles
Evidence · 16 claims (5 for, 9 mixed, 2 against) · 25 studies (8 causal, 8 review, 7 quant-synthesis, 1 associational, 1 theoretical), q2–q4 · 9 of 25 report an effect size · 7 claims rest on one study

Conditional relationship

For a reader about to meet an expository text (or a lecture or video segment) whose central ideas they do not yet know, attempting a written answer or best guess to a few questions that target those central ideas, then reading a text that answers them, then checking the attempts against it, is expected to improve later memory of the questioned content (immediate-performance or delayed-retention on items about the same ideas) compared with reading the same text for the same time without questions. The attempt does not have to succeed: in the recorded experiments almost every pre-question was answered wrongly or left blank, and the benefit came on those items.

The relationship is conditional on five things. Alignment: the gain is on what the questions ask about; factual pre-questions can reduce learning of material they do not ask about, and higher-order questions spread their benefit further. Attempting: reading the questions without answering them helped less than answering them. Resolution: the gain depends on the reading (or feedback) that follows supplying the answer; a pre-question with no answer to follow is a guess left standing. The learner's response to failing: in one online course, pre-tests reduced how many learners kept going, so a benefit to those who stay can sit beside a loss of those who leave. What the question activates: a question that asks for topic knowledge did not improve primary pupils' comprehension in one experiment, while one that activated knowledge about how to read did.

One claim tests a narrow form of this relationship; none tests the broad form. The tested form is undergraduates in a laboratory attempting pre-questions on an expository essay, with retention of the questioned concepts as the outcome. No claim in the wiki tests pre-reading questions in a classroom, with children or second-language readers, on conceptual-understanding or inference, against other pre-reading activities (a preview, a vocabulary pre-teach, a KWL chart), or as a way of "setting a purpose" in general. The default design below keeps the earlier page's guidance as labelled proposals.

How the converted siblings sit around this page: Activation holds the general model of making prior knowledge visible before instruction, of which a pre-question is one form; this page narrows it to questions asked before a text and the attempt–read–check cycle. Retrieval Practice holds retrieval of material already studied; a pre-question is an attempt on material not yet studied, where failure followed by the answer is the expected case, so the two should not be assumed to share moderators. Cognitive Activation holds the model for tasks that require generating an explanation or a prediction, which is what a higher-order pre-question asks for. The pre-reading questioning element and the strategies prequestions, pretesting, anticipation guide, KWL and generating questions before reading are concrete forms; none is in this format yet.

Default design, while the relationship is untested

A concrete default a designer can act on and revise. Steps 1–3 and 5 rest on the laboratory evidence; the rest are proposals from the earlier page, labelled as such.

  1. Write two to four questions per text, aimed at the ideas the later check will test. Pick the text's central concepts or claims, not its most quotable details, and ask about them in words the learner can read. Gains are on the questioned content (Pretesting enhances learning [+M]); factual pre-questions can cost learning of content they do not ask about (Question prompts improve learning [~M]). The number is an untested proposal (the pretesting claim page's own discussion suggests "two or three" central questions, with no study behind the count).
  2. Make at least one question ask why, how or what follows, not only what. Higher-order adjunct questions generalised to related and unrelated test material more than factual ones in the review on the question-prompts claim (Question prompts improve learning [~M]; abstract only, and it pools questions placed before and after text).
  3. Ask every learner to write an answer or a best guess before reading, privately, in about a minute a question (three to five minutes in all). Attempting beat reading the questions without answering them (Pretesting enhances learning [+M], Experiment 5). The timing is an untested proposal.
  4. Say plainly that the questions are not scored and that most guesses will be wrong. In one online course, pre-tests reduced persistence (Pre-tests reduced adult learners' persistence in a MOOC, though they improved post-test scores among those who completed it [-M]); whether framing prevents that is untested. Never grade the attempts.
  5. Read with the questions in view, then check. Learners mark where the text answers each question, then compare their guess with the text, correct it in writing and note which guess changed. The answer must actually arrive: the recorded benefit came only where reading or feedback supplied it (Retrieval Fails Without Encoding [~M]). Allow five to ten minutes for the check (untested proposal).
  6. Ask how sure each guess was, and spend the check on confident wrong ones. A confident wrong prediction that the text contradicts is the case to discuss explicitly: name the idea, say what in the text refutes it, and ask for a fresh prediction on a new case (Instruction that confronts science misconceptions … produces conceptual change more often than standard exposition [~M]; High-confidence errors are more likely than low-confidence errors to be corrected on a later test after feedback. [~M], a review and one experiment on corrective feedback, not on reading). Untested as a pre-reading routine.
  7. Protect what the questions leave out. Tell learners the questions sample the text, and end with one task on the whole text (a three-sentence summary, or "what else did the text say that the questions missed?"). Untested proposal against the narrowing cost on Question prompts improve learning [~M].
  8. Move the writing of questions to the learner across sessions. A default arc, all thresholds untested proposals:
  9. Early sessions (first two or three texts): the teacher writes all questions; learners attempt, read and check as above.
  10. Middle sessions: the teacher writes two questions and learners write one more from the title, headings and first paragraph, using question stems ("Why does…?", "What would happen if…?", "How is … different from …?"); the class compares which questions the text answered.
  11. Late sessions: learners write two or three questions of their own before reading and check them after; the teacher adds a question only for a text with a known trap.
  12. Moving on: a learner works without teacher questions once, on two consecutive texts, at least two of their own questions are answered by the text's central ideas rather than its details. Training learners to generate their own questions improved comprehension across intervention studies (Self-questioning improves comprehension [+M]), and so did reciprocal teaching, where questioning is one of four rotated strategies (Reciprocal Teaching Improves Reading Comprehension [+M]); those were trained over many sessions and mostly asked questions during reading, so the arc carries them to a pre-reading routine without a test.
  13. Check later, on both kinds of content. At a delay of a week or more, include items on questioned and unquestioned content, so both the gain and any narrowing cost show up. Untested proposal; the recorded experiments do not report their delays on the claim page.

Dose, as a proposal: three to five minutes of questions before, five to ten minutes of checking after, per text, two or three texts a week in a reading-heavy course. When a condition is not met: if the text does not answer the questions, rewrite the questions or choose another text; if learners already answer most questions correctly before reading, the routine is a review, so switch to one higher-order question or to their own questions.

Fitting the design to a situation

The three facts that most change the decision:

  • What the later check will ask about. If it tests the whole text, factual pre-questions on part of it may cost the rest; if it tests a few central ideas, pre-questions on those are the best-evidenced case. If unstated, ask: after reading, what must the learner be able to say or do, and about which parts of the text?
  • Whether the text resolves the questions, and whether someone checks the attempts against it. The benefit on record depends on the answer arriving after the attempt. If unstated, ask: does this text answer each question, and when will learners compare their guesses with it?
  • How the learner will take being wrong. Most pre-question attempts fail; a learner who reads that as proof of inability may stop. If unstated, ask: who are the readers, how confident are they, is anything scored, and can they leave (an optional course, a self-paced module)?
If the brief says… Then change… Basis
Readers are novice on the topic and will know almost none of the answers Keep the attempt-before-reading step; tell them most guesses will be wrong; keep to two or three questions Pretesting enhances learning [+M] (undergraduates, items answered at floor, laboratory)
Readers are advanced on the topic, or answer most questions before reading Replace teacher questions with one higher-order question, or with the reader's own questions about what they still want to know untested; the earlier page's caution against over-directing applies
Readers are primary-age children (child) Ask two questions aloud and record answers for them; add one question about how they will check they understood ("What will you do if a part doesn't make sense?"), not only questions about the topic Advance organizers aided learning …, but activating prior topic knowledge did not improve primary pupils' text comprehension in one experiment [~M] (88 primary pupils: metacognitive activation helped, topic activation did not)
Readers are second-language learners or have low literacy Write the questions in simpler language than the text; pre-teach the key words the questions use; accept guesses in the first language or as a drawing; move to learner-written questions only within a longer strategy programme Reading Strategy Instruction Improves Comprehension [+M] (English L2 meta-analysis g = 0.72; whole-classroom interventions d = .186 on standardized tests), carried from multi-strategy programmes to this routine; the wording moves are untested
The goal is understanding or inference (conceptual-understanding, near-transfer), not memory of facts Make most questions why/how/predict questions; in the check, ask for the reasoning in the text, not just the answer Question prompts improve learning [~M] (higher-order adjunct questions generalised further; abstract only); no claim tests pre-questions on inference
The later check covers the whole text Use fewer factual pre-questions, add a whole-text summary after reading, or move some questions to after each section instead of before Question prompts improve learning [~M] (factual pre-questions depressed learning of unquestioned content)
online-self-paced course or module, no teacher present Keep pre-questions to two, framed as a puzzle with no score; show the answer straight after the relevant section; track drop-off at the pre-question screen and remove it if completion falls Pre-tests reduced adult learners' persistence in a MOOC, though they improved post-test scores among those who completed it [-M] (399 adults in a MOOC: pre-tests reduced persistence; post-test gains only among those who finished)
Mixed modes: some learners in the room, some remote, or some on phones only Put the questions in one place everyone answers privately (a poll or a shared form) before the text opens; on phones, one question per screen with a short typed or multiple-choice answer; run the check in the same tool so remote learners' guesses are seen too untested proposal
One teacher with a large class Collect every learner's guesses at once (cards, mini-whiteboards, a poll), scan for the commonest confident wrong prediction, and spend the after-reading check on that one, with the text's refutation untested proposal; the confront-the-misconception move is from Instruction that confronts science misconceptions … produces conceptual change more often than standard exposition [~M] (science and reading syntheses, abstracts only)
One session only Two questions, attempt, read, check, one whole-text summary; do not claim a lasting effect untested beyond the laboratory sessions on Pretesting enhances learning [~M]
Learners have low confidence, or a history of failing at reading Keep answers private; offer forced-choice guesses instead of blank lines; open the check with what the class predicted well; never display individual guesses untested; the low-confidence risk is the motivation claim's own speculation, not a finding (Pre-tests reduced adult learners' persistence in a MOOC, though they improved post-test scores among those who completed it [~M])
Adults in work settings reading a procedure, policy or case Write questions as decisions they will face ("A patient refuses X: what does the policy let you do?"); check against the exact clause; ask what they do now that the text changes untested proposal
Learners are reading sources for their own inquiry or for a product an audience will see Let the learner's project question be the pre-reading question; ask what the source must show to be usable, and after reading, what it does and does not support untested proposal; links to Inquiry-based Learning
The pre-questions are scored, or the course is high stakes Remove the score from pre-questions; score only the later check Pre-tests reduced adult learners' persistence in a MOOC, though they improved post-test scores among those who completed it [-M] (persistence fell even unscored); scoring itself untested
Open brainstorming ("What do you know about…?") is planned instead of questions Replace it with two specific questions the text answers Prior-knowledge activation instructions before study had little effect on retention in Experiment 2 [-M] (1.5 minutes of talk produced about one relevant idea per topic)

Eleven of the fifteen rows cite a claim, but only the novice and one-session rows carry the tested form (laboratory pre-questions on an essay, undergraduates); the others carry a neighbouring finding (adjunct questions, strategy programmes, a MOOC, conceptual-change syntheses, topic activation) to this setting and say so. The advanced-reader, mixed-mode, adult-workplace and inquiry rows are untested proposals.

Observation, state and explanation

What a designer can observe: the written guess to each question, how sure the learner said they were, how long they took, whether they marked where the text answers, what they changed at the check, and how they did later on questioned and unquestioned items. "Has activated relevant knowledge", "holds a misconception" and "is reading with purpose" are inferred states. A blank answer before reading is the expected case and does not show disengagement; a correct answer after reading may come from copying the marked sentence.

Response Plausible explanations Observation that could change the interpretation
Blank or "don't know" on every pre-question Nothing to guess from (expected for new content); reluctance to be wrong in front of others; the question's wording is the barrier Offer a private forced-choice version; ask the learner to read the question aloud and say what it asks.
A confident wrong prediction that survives the reading A misconception the text did not confront; the learner did not find the answering passage; the text is ambiguous on the point Ask the learner to point to the passage that answers the question; give a refutation of the specific idea; re-ask on a changed case a week later.
Good on pre-questioned items later, weak on everything else in the text Attention narrowed to the questions; the questions sampled too little; the rest of the text was harder Compare questioned and unquestioned items of similar difficulty; add a whole-text task and see whether the gap closes.
Learner-written questions are all literal or about details The learner cannot yet tell central from peripheral ideas; decoding takes the effort; they think easy questions are wanted Ask them to rank their own questions by importance; check reading fluency on a short passage; show one central question from the teacher and ask for one more like it.
Fewer learners continue after the pre-question step (online) Discouragement at failing; time cost; the step looks like a test Compare completion with and without the step, or with a "no score, guesses expected" version; ask a confidence rating before and after.
Better later recall of pre-questioned items than of others The attempt made the answer stand out when it arrived; the question flagged importance; extra time on those items Compare items attempted with items whose questions were only read, at matched time, as the pretesting experiments did.

These rows are proposals, untested with learners as diagnostics. A pre-question is itself an intervention: asking it changes what the learner attends to, so record every question and probe given.

Evidence and qualifications

  • Pretesting enhances learning [+M]: across five laboratory experiments, undergraduates asked questions about an expository essay on human vision before reading it (nearly all answered wrongly or blank) did better on a later test of the same material than those given equivalent extra reading time; in Experiment 1, 75% against 56% correct (d = 1.1), counting only items failed at pretest. Emphasising the same concepts in italics or bold did not account for it, and reading the questions without attempting them helped less. A second set of experiments (trivia and word pairs, undergraduates) found failed attempts followed by the answer beat reading question and answer together (41% against 31%, d = 0.58). This is the closest test of this page's relationship. Read another way ([~M]), the same two experiment sets show that the failed attempt helps the encoding of an answer that follows it, not learning on its own: every trial ended with the answer shown or the passage read, which is why the default design requires a text that answers the questions and a check. It does not show effects on unquestioned content, comprehension or inference, classroom use, children or second-language readers, and the delay is not recorded on the claim page. Not yet checked against its sources.
  • Question prompts improve learning [~M]: a review of adjunct (inserted) questions in prose finds that questions of all levels strongly help on the same questions later and weakly on related ones; higher-order questions generalise to related and unrelated higher-order items; and factual pre-questions can hurt learning of unquestioned material. Effects varied with text length, question density and format, not with learner age or delay. Abstract only; no pooled effect. This is the limiting claim: it sets the narrowing cost against the gain.
  • Pre-tests reduced adult learners' persistence in a MOOC, though they improved post-test scores among those who completed it [-M]: 399 adults in a museum's climate-change MOOC, randomly assigned to pre-tests (with no, basic or elaborate feedback) or no pre-test: pre-tests reduced persistence; across all students they did not change post-test learning; among completers they improved it. Abstract only; no effect size. The claim's title says motivation, but the measured outcome is persistence. It qualifies the default design: an unscored, quick pre-question is still a cost where learners can leave.
  • Advance organizers aided learning in a 1980 meta-analysis, but activating prior topic knowledge did not improve primary pupils' text comprehension in one experiment [~M]: in a randomised 2 × 2 experiment, 88 primary-school students' text comprehension improved when metacognitive knowledge was activated, not when topic knowledge was, once prior knowledge was controlled; a meta-analysis of 135 studies found advance organizers helped learning and retention. It qualifies the earlier page's main rationale (activating background knowledge) for young readers' comprehension.
  • Self-questioning improves comprehension [+M]: a review of intervention studies teaching students to generate their own questions about text reports a median effect of 0.36 on standardized comprehension tests and 0.86 on experimenter-developed tests; a review of 35 experimental studies with K–12 students with and without disabilities calls self-questioning effective. Both abstract only. It supports handing question-writing to learners over time; it tests trained generation, mostly during reading, not a pre-reading warm-up.

Learners, settings, materials and outcomes differ across these claims (laboratory undergraduates, MOOC adults, primary pupils, intervention programmes), so do not rank them by effect size.

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 claims found near it. None has been re-read against the model above as core evidence; each says what it is and how far it bears.

Objective and learner-valued goal

The designer's objective is usually that the learner can recall or use the text's central ideas on a later check, with the check's items, criterion and delay stated. The pre-question attempts and the corrected answers are means, not the outcome. Ask separately what the learner wants from the text, and record agreement or divergence as an observation. A learner who wants to finish quickly may write the shortest guess; one who fears being wrong may leave every line blank.

For example, a care assistant reading a revised infection-control policy may want to know what changes on her next shift, while the trainer's objective is recall of the policy's main requirements on a quiz. Pre-questions written as shift decisions ("A visitor arrives with a cough: what does the policy say you do?") can serve both; a quiz alone cannot show whether her practice on the ward changed, so record what she says she will do differently as well.

What would revise this model?

The expectation should weaken if classroom studies with comparable readers and texts find no advantage of pre-questions over reading for the same time, or find that the gain on questioned content is matched by an equal loss on unquestioned content, so that whole-text learning is unchanged. It should be narrowed to retention of questioned facts if pre-questions do not help inference or comprehension. If unscored, framed pre-questions still reduce persistence in courses learners can leave, the online default should drop them. If topic pre-questions fail for young readers while questions about how to read succeed, the page should recommend the second for children. Do not protect the model by calling every null "badly written questions" after the fact.

The present evidence establishes no number of questions, no length of attempt, no delay over which the gain holds and no effect on comprehension of a whole text.

Claims

Examples

  • Prediction and purpose prompts: Learners answer a small set of questions before reading and revisit them afterward.
  • Question-generation warm-up: Learners create their own questions from a title, heading set, or opening image.
  • KWL-style framing: Learners identify what they know, want to know, and later what they learned.
  • Source anticipation: Before reading a document, learners predict what perspective, evidence, or problem the text might contain.
  • Concrete forms in the wiki: prequestions, pretesting, anticipation guide, KWL, generating questions before reading, and the pre-reading questioning element.

Key Sources

  • Ajideh, P. (2003). Schema theory-based pre-reading tasks: A neglected essential in the ESL reading class. The Reading Matrix, 3(1).
  • Clinton, V., & van den Broek, P. (2012). Interest, inferences, and learning from texts. Learning and Individual Differences, 22(6), 650-663. https://doi.org/10.1016/j.lindif.2012.07.004
  • Rasheed, H. (2014). Examining the effectiveness of pre-reading strategies on Saudi EFL college students' reading comprehension. English Language Teaching, 7(11), 79-91. https://doi.org/10.5539/elt.v7n11p79
Page metadata
Field Value
type principle
id pre-reading-questioning
title Pre-reading Questioning
description For a reader about to meet an expository text whose central ideas they do not yet know, attempting a few aligned questions before reading and checking the attempts against the text afterwards is expected to improve later memory of the questioned content over reading alone, possibly at some cost to unquestioned content; one laboratory claim tests this with undergraduates, and no claim tests its classroom, comprehension or purpose-setting forms.
status review
canonical True
generated by claude/unspecified on 2026-10-05
trust unverified — no one has confirmed this page's content
sources 2 — clinton-2012, rasheed-2014

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