Skip to content

Explicit Instruction: Online Reading Strategies

Principle · All principles
Evidence · 25 claims (15 for, 10 mixed) · 40 studies (15 causal, 10 quant-synthesis, 8 review, 3 associational, 2 qualitative, 2 theoretical), q1–q4 · 9 of 40 report an effect size · 15 claims rest on one study

Conditional relationship

This page owns one capability: reading and making sense of linked online texts once they have been found — deciding which links serve the reading goal, keeping track of meaning across pages and tabs, resisting drift, and putting what several pages say together. For a learner who reads comparable print text adequately (decoding is reasonably automatic) but is a novice online reader, and whose observed starting response on an open online reading task is to click whatever looks interesting, lose track of the question after two or three pages, end with notes copied from one page, or be unable to say where a piece of information came from, explicit instruction in a small set of online reading moves (set a goal and questions before opening any link; choose links against the goal; keep a running record of pages and how they connect; notice drift and return to the goal; combine pages into an answer), modelled by thinking aloud on real web pages, practised on authentic linked texts with prompts and feedback, and faded across sessions, is expected to produce better comprehension and integration of information from several linked pages, and more goal-directed navigation, on new online reading tasks (immediate-performance, near-transfer) than reading online without such instruction. The target is a metacognitive-strategy used in service of comprehension of concept and principle content; whether the gain lasts (delayed-retention) or transfers to unlike tasks (far-transfer) is not settled by anything here.

No claim in the wiki tests this relationship. No claim here compares taught online reading strategies with untaught online reading, or compares any version of such teaching with another. The nearest evidence comes from four directions, each carried to online reading from somewhere else and labelled where it is used:

The relationship is conditional on the learner's print reading (strategy instruction does not make up for effortful word recognition), prior knowledge of the topic and self-regulation (both named as moderators of learning from hypermedia), the kind of online reading task (skimming to orient, reading for evidence and synthesising across pages call for different moves), how open the link environment is (a curated set of pages against the open web), and whether anyone or anything sees the learner's path while it happens.

How this sits beside the converted siblings. Explicit Instruction: Internet Search owns finding the pages (framing a query and revising it); this page starts where the learner has results or a start page in front of them. Evaluating Sources owns whether to trust a page, including the move off the page; this page hands over to it whenever a reading decision turns on credibility and does not repeat its model. Pre-reading Questioning owns questions attempted before reading a single text; the goal-setting step here borrows that move for a set of pages. Note-taking, Annotating and Graphic Organizers own how notes, marks and maps are made; the running record here is one use of them. Self-Regulated Learning holds the general plan–monitor–evaluate relationship of which this is an application to linked texts, and Guided Practice and Scaffolding and Fading hold the practice phase and how support is withdrawn. Clear Structure and Cognitive Load Management hold the designer's side: building digital material that needs fewer navigation decisions in the first place.

Default design, while the relationship is untested

The earlier page's guidance (define a goal and questions before reading, map the information while reading, limit the links explored, model skilled choices, explain why one clicked, skipped or combined), kept as a concrete default a designer can act on and revise. Each step says what stands behind it; where a claim is cited, it is carried from print reading, computer-based learning or general learning research, not tested on online reading strategy instruction.

  1. Check print reading and the online starting response first (10 minutes, once). Give one short print passage with two comprehension questions and one short online task (a start page with four or five links, one question). Watch: does the learner read the print text adequately, and online do they state a purpose, pick links, lose the question, or copy? If print comprehension or word reading is weak, work on that first or reduce the reading load (shorter pages, text read aloud); online strategy teaching will not make up for it. Basis: the discussion on Reading Strategy Instruction Improves Comprehension [~S] says strategy instruction cannot compensate for weak word recognition (a qualification stated in the claim's discussion, not in its two entries); the probe itself is an untested proposal.
  2. Set the goal and two or three questions before opening any link (2–3 minutes per task). The learner writes what they need to find out and what they will do with it, as process steps ("find two explanations of X and say how they differ"), not only a product ("finish the worksheet"). Basis: Process goals lead to better skill acquisition for novices than outcome goals. [+M] (84 high-school girls learning a writing-revision skill; carried); in one MetaTutor study, following the prompt to set subgoals went with larger gains (suggest-subgoal compliance [+M], associational, 28 college students). The questions-before-reading move is modelled on Pre-reading Questioning.
  3. Model the moves by thinking aloud on real pages (8–10 minutes, first two sessions). On a projected or shared screen, read a small set of linked pages against a stated goal and say each decision aloud: what I expect behind this link and whether it serves my question; why I am not opening that one; "I have drifted — back to my question"; how this page connects to the last one. Name each move after modelling it. Basis: explicit, teacher-modelled strategy instruction is what the print evidence tests (Reading Strategy Instruction Improves Comprehension [+S], print texts; carried); the think-aloud for online pages is an untested proposal. The routine is spelled out in the strategy Think-Alouds for Online Reading Strategies.
  4. Limit the links, then widen them. Early sessions use a curated set of three to five linked pages and a budget ("open at most four links beyond the start page for this question"); middle sessions a larger set with some irrelevant and distracting links; late sessions the open web with the learner setting their own budget. Basis: disorientation, distraction and overload are the usability problems a hypermedia review names (Usability problems … hinder … self-controlled hypermedia learning [~M], review, no effect sizes); working memory holds few new units (Chunking reduces working memory load by grouping information into fewer, more meaningful units. [+S], carried from memory research). The budget numbers are untested proposals. Handing learners control by itself is not the active ingredient: across 18 studies, learner control in educational technology had an effect near zero (Letting learners pace instruction helped on complex material in one experiment, while learner control in general showed almost no effect across 18 studies [~S]).
  5. Keep a running record while reading. One line per page opened: where it is, what it says for my question, and how it links to a page already read; in later sessions, a node-and-link map built by the learner. Before answering, the learner reviews the record and revises their summary. Basis: learners who construct concept maps from what they study learn more than in most comparison conditions, though not more than time-matched retrieval practice (Concept mapping improves learning over most comparison conditions by a moderate average amount, but not over time-matched retrieval practice [+S], carried from map construction on study texts); in the MetaTutor study, following prompts to review notes and revise a summary went with larger gains (review-notes compliance [+M]; revise-summary compliance [+M]; both associational, one study), while following the prompt to open the diagram did not (no effect for four explicit prompts [~M]).
  6. Ask for the reason at two or three decision points per task. The learner says or writes why they clicked, skipped, left a page, or joined two pages ("I left because it repeated page 2"; "this explains the cause the first page only named"). Basis: Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S] (carried from explaining worked examples and expository text, including a print text read by eighth graders; immediate tests). Explaining navigation choices is an untested proposal.
  7. Prompt and respond during practice, then fade. While learners practise, the teacher or the software prompts at visible drift or stalls: a plan prompt ("what is your question?"), a monitor prompt ("does this page answer it?"), an evaluate prompt ("what do you still need?"), with feedback on what the learner does next. Give more help to those who need it and less after a success. Basis: Metacognitive prompts improve learning [+S] (computer-based environments; effects vary with whether prompts carry feedback, how task-specific they are and whether they adapt to the learner; in one reviewed study, lower-prior-knowledge learners benefited only once trained to use the prompts); Contingent scaffolding improves learning more than fixed or absent support. [+M] (mostly one-to-one tutoring; carried).
  8. End each task with a synthesis that uses two or more pages (5 minutes). The learner answers the goal questions in a few sentences, names the pages used, and says how they relate (agree, add, conflict). Untested proposal; from the earlier page's objective of integrating information across pages. If the answer turns on whether a page is trustworthy, hand over to Evaluating Sources.
  9. Plan the arc across sessions, not one lesson (a default of six sessions of 40–50 minutes over two to four weeks). Sessions 1–2: check (session 1 only), goal-setting 3 minutes, teacher think-aloud 8–10 minutes, guided practice on a curated set of three to five pages with a printed prompt card 20–25 minutes, synthesis 5 minutes, debrief on one common drift 5 minutes. Sessions 3–4: no full think-aloud (a 3-minute reminder of one move), a larger set with distracting links, learners explaining choices to a partner, prompts only on request or at visible drift. Sessions 5–6: open-web tasks with the learner's own goal and link budget, prompt card withdrawn, the running record kept unprompted; end with one unaided task on a new topic. Move a learner on (to open-web tasks, or out of the sequence) when on two consecutive unaided tasks their synthesis answers the goal questions from at least two relevant pages and their record shows a return to the goal after any drift; keep a learner on the curated set while they lose the question or copy from one page. The arc, the minutes and the threshold are untested proposals. Learners who already read online this way should skip the modelling and do one check task: guidance that helps novices can become redundant as expertise grows (Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M]).
  10. Judge progress on an unaided online task like the target, not on assisted practice or a strategy quiz. Whole-class print strategy programmes showed a very small effect on standardised comprehension tests but larger ones on researcher-made tests and on strategy measures (Reading Strategy Instruction Improves Comprehension [~S]), so "uses the strategies" is not the same result as "comprehends better". Untested as a rule for online reading.

When a condition is not met: if no one can watch learners' paths (a large class, self-paced online work), build the goal entry, the page record and two or three prompts into the environment and review records afterwards; if learners already read online purposefully, skip to open-web synthesis tasks.

Fitting the design to a situation

The three facts that most change the decision:

  • The learner's print reading and their observed online behaviour. A learner who struggles to read the words needs reading support before online strategies; one who already keeps to a goal and joins pages needs little modelling. If unstated, ask: can they read a page of this difficulty in print and answer two questions on it, and what do they do with a start page and five links when given one question?
  • How open the link environment is, and who sees the path. A curated set of pages with a teacher watching, a self-paced course page with links, and the open web with nobody watching need different supports. If unstated, ask: which pages will learners read, can they leave them, and who (or what) will see the pages they open and the notes they keep, and how soon?
  • The reading task and what counts as success. Orienting on a topic, finding evidence for one point, and combining several sources into an answer use different moves, and each needs a different check. If unstated, ask: what must the learner produce from their reading, for whom, judged how, and when?
If the brief says… Then change… Basis
Learners are advanced online readers, or keep to a goal and join pages on the check task Drop the think-aloud and the curated sets; give one open-web synthesis task with a short debrief on their choices Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M] (cognitive-load studies; carried; gives no crossover point)
Learners have weak word reading or low literacy, or low confidence as readers Work on print reading first, or reduce the reading: short pages, text read aloud by a screen reader, two links per page, a picture-based record; praise and record each return to the goal Discussion on Reading Strategy Instruction Improves Comprehension [~S] (strategy instruction does not make up for weak word recognition; print); the adaptations are untested proposals
Learners are child (primary age) Curated sites only; one question; two or three links; an adult-made record template with boxes to tick and one line to write; the teacher checks paths one-to-one or in small groups untested for online reading; contingent help with children is carried from one-to-one long-division tutoring (Contingent scaffolding improves learning more than fixed or absent support. [+M])
Learners are adolescent or adult second-language readers Keep the strategy routine and name each move in plain language (the learners' first language allowed for the record); choose pages slightly below their reading level for the first sessions Reading Strategy Instruction Improves Comprehension [+S] (meta-analysis of strategy instruction for English L2 learners, print texts; carried to online)
adult learners in work settings Use the pages they actually read at work (an intranet, a policy portal, a supplier's documentation) and a real work question; the synthesis is the answer they would give a colleague untested proposal
Setting is online-self-paced, no teacher present Build the supports into the course: a goal box that must be filled before links open, a notes panel beside the reading, a prompt at each page change ("does this page answer your question?"), and a "review your notes and revise your summary" step before the answer is submitted Metacognitive prompts improve learning [+S] (computer-based environments); review-notes compliance [+M] (one study, associational)
classroom with one teacher and a large group Projected think-aloud; everyone keeps the page record on paper or in one shared document the teacher can scan; two checkpoints per task where the teacher reads records and gives whole-class feedback on the commonest drift, then visits those who lost the question untested proposal; whole-class strategy programmes showed much smaller effects than researcher-led ones (Reading Strategy Instruction Improves Comprehension [~S]) and contingent-support evidence is mostly one-to-one (Contingent scaffolding … [~M])
Mixed modes: some learners remote, some in the room, or phone-only Share the think-aloud screen with remote learners and record it for replay; one shared record document all can write in; phone-only learners work in one tab with a lower link budget (two or three) and keep the record on paper or in a single notes app untested proposal
One teacher or tutor with little time (volunteers, a busy shift, brief check-ins) Give learners a prompt card with the four moves and a model synthesis to check their own against; the tutor's one check-in reads the record and asks one question ("where did you leave your question?") untested proposal; the prompt card stands in for live prompts (Metacognitive prompts improve learning [+S], computer-based prompts; carried)
Time is one session only 45 minutes: goal and questions 3, think-aloud 8, guided practice on a curated set of four pages 22, synthesis 6, debrief on one drift 5, and one unaided question 1–2; record what it shows and claim no lasting change untested proposal
A fixed short time budget (for example 20 minutes inside another lesson) Goal 2 minutes, a 4-minute think-aloud of one move only (choosing links against the goal), practice on three pages 10 minutes, a two-sentence synthesis 4 minutes untested proposal
The brief names its own routine for reading or research (a school's research log, a KWL chart, an inquiry template) Keep it; add only what the evidence here points to: a goal written before any link opens, one "why this link?" line per page, and a review of notes before answering untested proposal; the additions rest on process goals [+M] and self-explanation [+S], both carried
The brief gives its own categories, criteria or rubric for the product Use them as the headings of the running record and the synthesis; do not substitute this page's moves as the criteria; if the rubric rewards tool use or link use, keep it and add one unaided question beside it untested proposal
The reading is private or sensitive (a health condition, personal finances, immigration status) The learner keeps the record and decides who sees it; teach and check the moves on a neutral practice topic, and let the learner apply them privately untested proposal
The brief wants a whole-task performance kept whole (a research report, a decision brief, a presentation) Do not split it into separate strategy exercises; teach each move inside the task at the point it is needed and check the moves in the record the task already produces untested proposal
The product goes to a real audience who should see the answer, not the reasoning The synthesis becomes the product; the page record and the reasons stay with the learner and the teacher untested proposal
The pages are full of advertising, pop-ups, autoplay video or notifications Before reading: close other tabs, silence notifications, use a reader view if available; pick practice pages without decorative clutter for early sessions Interesting but irrelevant details can impair comprehension and transfer, with effects that vary with the material and the learner [~S] (instructional materials, not web pages; carried); college students reporting goal setting and environment structuring reported less digital distraction (SRL strategies associated with lower digital distraction [+M], associational survey)
The brief proposes free exploration of a rich hypermedia site as the method Keep exploration, but add the goal, the link budget and the record; do not count on the freedom itself Letting learners pace instruction helped on complex material in one experiment, while learner control in general showed almost no effect across 18 studies [~S]; Learner characteristics … moderate hypermedia learning effectiveness [~M] (review)
High stakes, or the assessment is unaided Judge readiness only on an unaided online task like the target, with the record kept; never on prompted practice or a strategy-use questionnaire untested proposal; strategy measures moved more than standardised comprehension in whole-class print programmes (Reading Strategy Instruction Improves Comprehension [~S])
The real difficulty is whether to trust a page Hand over to the lateral-reading model on Evaluating Sources rather than adding credibility checks here untested proposal

Ten of the nineteen rows cite a claim (five of them only as a limit or a carried finding beside an untested proposal), and none of those claims tests online reading strategy instruction: each carries a finding from print strategy instruction, computer-based prompts, hypermedia reviews, instructional materials or general learning research to online reading and says so. The remaining rows are untested proposals.

Observation, state and explanation

What can be observed is a reading path under stated conditions: the goal written (if any), the pages opened and in what order, the time on each, the record kept, where the learner left the question, the prompts given and what the learner did next, and the synthesis produced. "Disoriented", "distracted", "overloaded" and "lacks the strategy" are inferred states, and the same path fits more than one.

Response Plausible explanations Observation that could change the interpretation
Opens many pages quickly, synthesis is thin No goal held; the pages are too hard to read so the learner skims for something easier; the learner thinks the task is to collect, not to answer Ask the learner to state the question mid-task; give one of the pages in print and ask two questions on it; show a model synthesis and ask what the task wants.
Stays on the first page, copies from it Does not know other pages could help; reading the first page used all their effort; believes the first result is the answer Point to one link and ask what they expect behind it; offer a shorter page on the same point; ask whether another page might disagree.
Follows a chain of links far from the question Interest pulled them (the material is engaging but irrelevant); lost track of the question in working memory; the question was vague Look for the goal they wrote and how specific it was; ask at the drift point what they were looking for; give the same task with the question on screen at all times.
Record is full, synthesis does not join pages Notes taken page by page with no linking; the learner cannot see how pages relate; the task did not ask for a relation Ask "how does page 3 relate to page 1?" orally; give a two-column record (what it says / how it connects) and compare.
Uses the moves when prompted, not without prompts The routine is not yet the learner's; prompts are doing the monitoring; the unaided task is harder or on a less familiar topic Give a matched unaided task on a familiar topic, then one with only the goal box; compare records.
Strategy questionnaire improves, comprehension does not The learner can name the moves but does not use them; the comprehension measure is not aligned with online reading; print reading limits comprehension Compare the record from an unaided task with the questionnaire; check the print probe again.

These rows are proposals, untested with learners as diagnostics. A probe is itself a prompt and changes the path it observes, so record each one, and keep the explanations that remain open rather than choosing the one whose name fits.

Evidence and qualifications

No claim in the wiki compares explicit instruction in online reading strategies with online reading without it, or compares one version of such instruction with another. The claims below test neighbouring parts of the relationship.

Learners, tasks and outcomes differ across these claims (second-language and school readers on print, college students in a prompting environment, pupils with animations, high-school writers), so do not rank them by effect size. None measures online reading comprehension after strategy instruction, and none measures it at a delay.

The evidence by question

  • Does explicitly teaching reading strategies (modelled, practised and named) improve comprehension beyond reading practice alone, and how far does that hold in ordinary classrooms? Reading Strategy Instruction Improves Comprehension [+S] (print texts; the page carries it to online reading)

Further evidence, not yet read against this model

Claims read for this rewrite that bear on the page less directly. Every claim the earlier page cited is used in the model above.

Objective and learner-valued goal

The designer's objective is usually comprehension and integration of information from several linked pages on an unaided task like the one the learner will meet, judged on a product the learner makes from their reading. Elicit separately what the learner wants from reading online: to get an assignment done fast, to settle a practical question (a benefit entitlement, a medical instruction, a product choice), to follow an interest, or to avoid looking slow in front of others. Agreement and divergence are both observations. A learner who values speed may skip the goal and the record; one following an interest may value the drift this page treats as a problem.

For example, an adult in a community literacy class may want to find out from a council website whether they qualify for a reduction in a bill, while the designer's objective is synthesising information across three linked pages. One guided task on the council site serves both; a curated set of pages about a science topic serves only the designer. Record both, so that a good synthesis on practice pages is not read as the capability the learner came for.

What would revise this model?

The expectation should weaken if comparisons of taught online reading strategies against untaught online reading, with comparable learners and an unaided online comprehension task, show no advantage, or an advantage on strategy measures only (as whole-class print programmes partly did). If learners whose print comprehension is controlled show no extra difficulty online, the case for online-specific instruction, as opposed to ordinary strategy instruction, weakens. If goal-setting and a page record alone, without modelling, produce the same gains, the think-aloud can be cut. If link budgets or curated sets make no difference to novices, step 4 should go. If a recall check after reading serves integration as well as a map does, as time-matched retrieval practice beat map construction in one experiment, replace the map. Do not protect the model by calling every failure "too few sessions" or "not enough modelling" after the fact.

Navigation behaviour, strategy use, comprehension of the pages read, integration across pages, transfer to new online tasks and use in the setting the learner values are separate outcomes. The present evidence establishes no number of sessions, no link budget and no hand-over criterion for online reading.

Examples

Key Sources

  • Afflerbach, P. A., & Cho, B. Y. (2010). Determining and describing reading strategies: Internet and traditional forms of reading. In Metacognition, strategy use, and instruction.
  • Azevedo, R. (2005). Using hypermedia as a metacognitive tool for enhancing student learning? Educational Psychologist, 40(4), 199-209. doi:10.1207/s15326985ep4004_2
  • DeStefano, D., & LeFevre, J. A. (2007). Cognitive load in hypertext reading: A review. Computers in Human Behavior, 23(3), 1616-1641. doi:10.1016/j.chb.2005.08.012
  • Coiro, J. (2003). Reading comprehension on the Internet. The Reading Teacher, 56(5), 458-464.
Page metadata
Field Value
type principle
id explicit-instruction-online-reading-strategies
title Explicit Instruction: Online Reading Strategies
description For a learner who reads print adequately but navigates linked online texts without a goal, explicit modelled instruction in a few online reading moves (goal and questions before opening links, choosing links against the goal, a running record of pages, returning from drift, combining pages), practised on curated then open sets of pages with prompts that fade, is expected to improve comprehension and integration across pages; no claim tests this, and the nearest evidence is print strategy instruction and computer-based metacognitive prompts.
status review
canonical True
generated by claude/unspecified on 2026-10-15
trust unverified — no one has confirmed this page's content
sources 2 — azevedo-2005, destefano-2007

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