Skip to content

Ask Experts

Principle · All principles
Evidence · 18 claims (6 for, 11 mixed, 1 against) · 30 studies (9 causal, 5 quant-synthesis, 5 associational, 4 review, 4 theoretical, 2 qualitative, 1 design), q1–q4 · 3 of 30 report an effect size · 11 claims rest on one study

Conditional relationship

This page owns the expert consultation: a learner brings a specific question, attempt or piece of work to someone with more expertise in the domain (an instructor, practitioner, mentor or more knowledgeable peer), and the expert responds to it. For a learner who is novice or intermediate on the task (they can attempt it, but their attempt stalls, misses what matters, or they cannot say what a good result would look like), a consultation in which the expert first has the learner explain their attempt, then responds contingently (a question, then a hint, then modelling the decision aloud, explanation last), and in which the learner applies what they heard to a revision or a new attempt soon afterwards, is expected to produce better unaided performance on that task (immediate-performance, near-transfer) than either no access to the expert or an expert who simply supplies the answer. Across a run of consultations the expert does less and the learner more. The knowledge type is usually a principle or complex-skill where expert judgement is not visible in the finished product (a diagnosis, a design critique, a revision decision, a debugging choice); the settings range from one-to-one tutoring to classroom office hours and workplace-clinical mentoring.

The relationship is conditional on what the expert does, not on their expertise alone: the nearest evidence suggests the learning comes from what the learner is prompted to construct, and one meta-regression found no tutor-background category (content experts included) that predicted learning in problem-based courses. It is also conditional on the learner's starting response (an advanced learner gains less from explanation and more from critique), on whether the advice is applied before it is forgotten, and on whether the learner can ask at all without social cost. No claim in the wiki tests access to an expert, as such, against no access, or compares ways of running a consultation. The nearest evidence is from one-to-one tutoring (contingent against fixed support; prompting against explaining), from prompted self-explanation, from fading support, and from process against outcome goals, all carried here to a setting they did not test.

How the converted siblings sit beside this: Scaffolding and Fading holds the general model of raising support after a failure and lowering it after a success, and its withdrawal; this page applies it to one kind of encounter and does not repeat it. The Cognitive Apprenticeship pattern is the full modelling–coaching–scaffolding–fading–articulation–exploration sequence, of which a consultation is usually the coaching or articulation step. Guided Practice holds the practice phase after a model. Communities of Practice holds sustained participation alongside experienced members over a term or more; a single consultation, or a short series of them, is the narrower object here.

Default design, while the relationship is untested

The earlier page's guidance, kept as a concrete default a designer can act on and revise. Each step is a design proposal; its evidence status is stated beside it.

  1. Bring real work, prepared. Before each consultation the learner writes a short question card: the task, what they tried, where it stalled, and what they think the answer might be. Unguided "ask me anything" sessions are replaced by consultations around an artefact, case or attempt. Untested proposal; from the earlier page's requirement of "a concrete learner question, artifact, case, or task".
  2. Learner explains first. The expert opens by asking the learner to talk through their attempt and their reasoning at the point where it stalled, before saying anything about it (about the first third of the time). Carried from prompted self-explanation [+S] (learners studying examples or solving problems, not consultations) and from tutoring where prompting alone matched explaining-and-feedback (tutoring claim [~M], eighth graders, 11 tutor–student pairs).
  3. Respond on a ladder, smallest help first. A question that points at the decision; then a hint; then the expert thinking aloud through the same decision on a parallel case, naming the cues they look at and the criteria they apply; a full explanation only if those fail. Evidence: contingent support [+M] beat fixed or no support in one-to-one tutoring of children on long division, and interactive tutoring (questions and hints) gave better transfer than explaining; untested for experts consulted by older learners.
  4. Point at process and criteria, not only the verdict. Name the strategy or criterion the learner's work needs ("what would you check before choosing this?"), not just whether it is right. Carried from process goals [+M] (high-school writing revision; once the skill was established, outcome goals helped), which is about goal-setting, not expert talk.
  5. Close with one next action. End each consultation with the learner stating, in their own words, the one or two changes they will make and why. Untested proposal.
  6. Apply within a week. The learner revises or attempts a comparable task within about a week (sooner for a single-session visit), and brings the revision to the next consultation or submits it for brief written comment. Untested proposal; from the earlier page ("expert advice that is not applied quickly is easily forgotten").
  7. Hand responsibility back across the series. A default arc of three to five consultations over a term, 15–20 minutes each per learner or small group (untested quantities): early, the expert models and explains on parts of the learner's work for up to half the time; middle, the learner presents attempt and diagnosis first and the expert asks and hints; late, the learner proposes the critique of their own work and the expert confirms, corrects or adds one point. Move a learner to the next stage when, on two consecutive consultations, they name the criterion the expert would have applied before the expert does, and their revision deals with the previous points without further help. The shape follows fading support [~S] (faded worked examples, near transfer); the thresholds are untested proposals.
  8. Make it safe to be uncertain. Ask for questions in writing or in small groups as well as aloud, accept partial and wrong ideas without comment on the person, and let low-confidence learners bring a question with a peer. Untested proposal; from the earlier page's requirement of psychological safety.

When a condition is not met: if the expert has no time to see each learner, collect question cards in advance and spend the session on the three or four commonest stalls, thinking aloud through each; if no expert is available live, record the expert thinking aloud through two or three representative pieces of work and pair it with a written response to each submitted question card.

Fitting the design to a situation

The three facts that most change the decision:

  • What the learner can already do and say about the task. A learner who cannot yet attempt the task needs a model before a consultation is useful; one who can already diagnose their own errors needs critique, not explanation. If unstated, ask: can the learner attempt one representative task now, and say where it went wrong?
  • How much of the expert's time each learner gets, and in what form. The evidence for contingent help is one-to-one; a guest visit, a drop-in office hour and a weekly mentor meeting allow very different designs. If unstated, ask: how many learners per expert, how many minutes each, how often, and live or written?
  • What the expert is expected to do. A practitioner with no teaching experience tends to tell; the design may need to brief them to ask first and to think aloud. If unstated, ask: who is the expert, have they taught or mentored before, and will someone brief them?
If the brief says… Then change… Basis
Learners are advanced on this task, or already self-diagnose most errors Drop the explanation and modelling steps; the learner presents work and their own critique, and the expert adds only what the learner missed; use peer critique between consultations Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M] (review of cognitive-load studies, not consultations; gives no crossover point)
Learners are child (primary age) Keep the consultation one-to-one or with two or three children, short (10 minutes), on one problem, with the expert moving up the question–hint–model ladder and back down after each success Contingent scaffolding improves learning more than fixed or absent support. [+M] (fourth and fifth graders, long division, 8 per condition, one-to-one)
Learners are adult professionals or trainees in a workplace-clinical setting Hold consultations on real cases from the learner's work; the mentor models the decision on one case, then observes the learner on the next and debriefs; plan when the mentor stops being present Mentor modelling and a school-wide behaviour program's strategies transferred to the mentee's classroom management practices [+W] (one preservice teacher, qualitative case study)
Goal is a concept, principle or transfer, not only a procedure Have the expert ask and prompt more and explain less; include one changed problem in each consultation for the learner to reason through aloud Contingent scaffolding improves learning more than fixed or absent support. [+M] (interactive tutoring gave better transfer with similar immediate scores) and Human tutoring has a medium effect over no tutoring, and in one small study students learned as well when tutors only prompted them as when tutors also explained and gave feedback [~M] (eighth graders, circulatory system)
The expert is a content specialist or practitioner with no teaching experience (guest, clinician, engineer) Brief them for ten minutes: ask the learner to explain first, give a hint before an answer, think aloud about what they notice; pair them with the course teacher, who manages questions and follow-up untested proposal; content expertise alone did not predict learning in one meta-regression of problem-based-learning tutors (In a re-analysis of PBL tutor-background data, neither a meta-regression nor subgroup comparisons found that tutor expertise predicts student learning [~M], PBL tutors, not guest experts)
The "expert" is a more knowledgeable peer Give the peer a rubric and annotated strong and weak examples, ask for comments tied to criteria, and keep the instructor's review for conceptual errors In two secondary-school writing studies, peer reviewers' own performance predicted how helpful their feedback was but not how accurately it identified problems [~M] (secondary and AP English writing: peers' expertise predicted helpfulness, not accuracy; with a well-specified rubric, accuracy did not differ by school performance)
Setting is online-self-paced, no expert live Use an asynchronous question queue: each question carries the learner's attempt; the expert answers the commonest ones with a recorded think-aloud and the rest in writing within a set time (say two working days); add an adaptive tutor or worked-feedback items for routine steps untested proposal; human and intelligent tutoring had similar effects over no tutoring in one review read from its abstract (tutoring claim [~M])
Some learners join remotely (video or phone) while others are in the room with the expert Keep remote learners in the live exchange rather than giving them a recorded or written reply afterwards: take their prepared questions first, pair each with an in-room partner who describes what the expert is showing and relays follow-ups, and ask the expert to say aloud what they see, check and rule out Untested proposal
classroom with one expert and a large group (office hours, drop-in) Collect question cards in advance, triage them, spend open time on the commonest stalls with a think-aloud, and book short individual slots for learners who have not come; do not count attendance as evidence of benefit untested proposal; drop-in tutoring visits were usually unrelated to course grades in one observational study (Tutoring center visits usually showed no significant difference in individual course performance, and sometimes tutored students performed worse [~M], college tutoring centre, self-selected visitors)
Time is one session only (a guest visit, a site visit) Collect learners' questions beforehand and send them to the expert; spend the session on the expert's reasoning about two or three of them; set a follow-up task within a week that applies one point, reviewed by the course teacher; do not claim lasting change untested proposal
Learners have low confidence, low status in the group, or are working in a second language Accept written or anonymous questions, consult in pairs or small groups, give question stems; avoid formats that put one person on the spot in front of everyone Passing individual microphones discouraged some students from asking questions during telelectures [~W] (one 1972 course, students' reports); otherwise untested
Stakes are high (licensure, patient or client safety) Judge readiness on observed unaided performance on representative tasks, not on how well the learner discusses them with the expert; withdraw the expert's presence in stages (present, then on call, then reviewing afterwards) untested proposal; the staging follows fading support [~S], which was tested on worked examples, not supervision
A long mentoring relationship (a year, a probation period) Plan from the start when the mentor will ask more and tell less, and check at the midpoint whether the mentee is proposing their own solutions untested proposal; one small comparison found fewer mentored than unmentored new teachers reached the most developed concern level and raised whether mentors held mentees back (Teachers without mentors reached the impact level more often than mentored teachers, raising the question that mentoring may inhibit development [~W], associational, 39 teachers)

Six of the twelve rows rest on a claim, and none of those claims tests consulting an expert as such: they carry tutoring, cognitive-load, peer-review and mentoring findings to this setting, and each row says so. Five rows are untested proposals that cite a claim only for its limit or as a neighbouring finding; one (a single session) is an untested proposal with no claim.

Observation, state and explanation

What can be observed in a consultation is what the learner brought, what they said when asked to explain it, which level of help was needed (question, hint, model, explanation), what they said they would change, and what the next revision or attempt shows. "Understands the expert's reasoning", "is dependent on the expert" and "is afraid to ask" are inferred states. Agreement during the conversation is not evidence of understanding; the next unaided attempt is.

Response Plausible explanations Observation that could change the interpretation
Brings no question, or only "is this right?" Cannot yet locate the difficulty; fears exposing ignorance; does not know what the expert could help with Ask them to talk through the work and stop them at the first choice; offer a written or private channel and see whether questions appear.
Nods and agrees, then the revision ignores the advice Did not understand and did not say so; understood but could not apply it; did not value the change Before leaving, ask for the change in their own words; at the next meeting, ask what they tried and what stopped them.
Applies the advice to this piece of work but repeats the error on the next Learned a fix, not the criterion behind it; the next task looks different enough to hide the same decision Give a changed task where the same criterion applies and ask which of the expert's points is relevant before they start.
Returns with the same kind of question every time Dependent on the expert's judgement; the expert has been answering rather than prompting; the task is too far beyond them Have them propose an answer first and rate their confidence; compare accuracy of proposals with and without the expert's confirmation.
Asks sharp questions but performs poorly alone Understands the criteria but lacks practice; the consultation has replaced practice Give a short unaided attempt between consultations and record its accuracy.

These rows are proposals, untested as diagnostics. A probe is itself part of the consultation and changes what it measures, 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 tests consulting an expert against not doing so, or compares ways of running a consultation. The claims below test neighbouring parts of the relationship.

The limiting claim. In a re-analysis of PBL tutor-background data, neither a meta-regression nor subgroup comparisons found that tutor expertise predicts student learning [~M] re-analyses one problem-based-learning data set: content-expert tutors did not produce more learning than other tutor backgrounds. A non-significant difference does not show equivalence, and PBL tutors facilitate groups rather than answer individual questions, but it is the nearest test of whether domain expertise by itself matters, and it does not show that it does. In two secondary-school writing studies, peer reviewers' own performance predicted how helpful their feedback was but not how accurately it identified problems [~M] qualifies in the other direction: in secondary writing, more expert peers gave more helpful feedback but were not more accurate at finding problems, and with a well-specified rubric peer ratings were as valid in lower- as in higher-performing schools. Its own Discussion says the evidence does not support its title.

Learners, tasks and settings differ across these claims (children with a tutor, eighth graders, high-school writers, PBL courses), so do not rank them by effect label. None measures a learner's later unaided performance after consulting a domain expert against a comparison.

The evidence by question

Further evidence, not yet read against this model

Claims bearing on expert access that the earlier page did not cite, listed so they are not lost. Each says what it is and how far it bears on this page.

Objective and learner-valued goal

The designer's objective is usually better unaided performance on a representative task, with the learner able to apply the criteria an expert would apply, at a stated horizon. Elicit separately what the learner wants from the expert: a verdict on today's work, reassurance, a shortcut, contact with someone in the profession they hope to join, or to understand how experts think. Agreement and divergence are both observations. A learner who wants a verdict will treat questions as delay; one who wants contact with the profession may value the conversation more than the advice.

For example, a nursing student may want to hear from an experienced nurse what the job is really like, while the designer's objective for the consultation is prioritising patient observations. A consultation that answers only the objective may leave the learner's reason for coming unmet, and one that answers only the learner's question teaches no criterion. Record both, so that a well-liked visit is not read as a learning gain, and a useful critique is not read as a failure because the learner found it uncomfortable.

What would revise this model?

The expectation should weaken if studies comparing consultations with domain experts against comparable time on practice with worked feedback show no advantage in later unaided performance, or show that a gain at the end of a consultation disappears at a delay. If prompting-only consultations match those with expert explanation across settings, as one tutoring study suggests, the model should say "an expert who prompts" rather than "an expert who explains". If content expertise continues not to predict learning when tutors' behaviour is held constant, what matters is the consulting skill, and the page should say so in its title. If structured peer consultation with a rubric matches expert consultation for some goals, the expert's time should be reserved for the goals where it does not. Do not protect the model by calling every failure "the wrong expert" or "not enough follow-through" after the fact.

The present evidence establishes no length or frequency of consultation, no number of consultations in a series, and no criterion for withdrawing the expert.

Claims

  • Mentoring/Coaching — a broader relationship structure in which expert guidance is sustained over time rather than episodic
  • Communities of Practice — extends expert access into sustained participation in a practice community
  • Guided Practice — turns expert explanation into coached performance on real tasks
  • Explaining Their Thinking — experts are most useful when they elicit learner reasoning rather than replacing it
  • Purposeful Reflection — follow-up reflection helps consolidate what was learned from the interaction and what remains uncertain
  • Scaffolding and Fading — the general model of contingent support and its withdrawal, which a consultation applies

Examples

Key Sources

  • Daley, B. J. (1999). Novice to expert: An exploration of how professionals learn. Adult Education Quarterly, 49(4), 133-147. ERIC
  • Du, J., & Ling, C. X. (2010). Asking generalized queries to domain experts to improve learning. IEEE Transactions on Knowledge and Data Engineering, 22(6), 812-825. summary
  • van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher-student interaction: A decade of research. Educational Psychology Review, 22(3), 271-296. doi:10.1007/s10648-010-9127-6
  • Chi, M. T. H., Siler, S. A., Jeong, H., Yamauchi, T., & Hausmann, R. G. (2001). Learning from human tutoring. Cognitive Science, 25(4), 471-533. doi:10.1207/s15516709cog2504_1
  • Collins, A., Brown, J. S., & Newman, S. E. (1989). Cognitive apprenticeship: Teaching the crafts of reading, writing, and mathematics. In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 453-494). Lawrence Erlbaum. doi:10.4324/9781315044408-14
  • Reilly, K., Wooster, B., Andrews, E., Bourdeau, V., Conley, R., & Thompson, M. (2008). Holding onto the GREEN Zone Leader Guide: A Youth Program for the Study and Stewardship of Community Riparian Areas. Bureau of Land Management. https://www.blm.gov/education
Page metadata
Field Value
type principle
id ask-experts
aliases involve-natural-resource-professionals
title Ask Experts
description For a novice or intermediate learner who brings a specific question or attempt to someone with more domain expertise, a consultation in which the expert has the learner explain first, responds with the least help that works, and does less across a series, followed by applying the advice, is expected to improve unaided performance on that task, though no claim here tests expert access as such.
status review
canonical True
generated by claude/unspecified on 2026-10-05
trust unverified — no one has confirmed this page's content
sources 5 — daley-1999, du-2010, van-de-pol-2010, chi-2001, reilly-2008

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