Teaching children to evaluate sources: after the checklist era

Evidence grade: STRONG on what fails, PROMISING on what works at primary age. The evidence that checklist-style evaluation fails is unusually direct — large assessments watched students apply it and get fooled. The replacement (lateral reading) has good trials, mostly with adolescents and adults; scaling it down to ages 8–12 is principled adaptation, graded as such.

For a generation, source evaluation was taught as a checklist — CRAAP (Currency, Relevance, Authority, Accuracy, Purpose) and its cousins — inspect the page, score its trustworthiness. Then researchers watched people do it. The Stanford History Education Group's assessments (Wineburg, McGrew, Breakstone and colleagues) found students at every level — and professional historians — evaluating sources by staring at them: judging the About page written by the people being judged, admiring the professional design, and getting systematically fooled by sites built to pass inspection. The checklist wasn't just weak; it directed attention exactly where a deceptive source controls the evidence.

What replaced it

The fact-checkers in the same studies did something different: they left the page — lateral reading, covered in full on its own page. The question changed from "does this page look trustworthy?" to "what does the rest of the web say about who's behind this?" — a question the deceptive source cannot control. Teaching that move has real trial support (McGrew et al.; Breakstone et al.; Wineburg's curriculum evaluations, with effects replicated internationally).

Scaling it down to primary age

The trials skew secondary-and-up; the primary adaptation is principled rather than proven, and it works by shrinking the question set, not the standard. Three questions carry the whole method at age 8–12:

Who says this? — every page has an author with a situation; finding out who is a game children take to readily.
How do they know? — were they there, did they measure, are they repeating someone? (The seed of the citation concept.)
Who else agrees? — the lateral move itself: check a second, separate place before believing a surprising thing. One deliberate open-a-new-tab habit, installed young.

Two supporting skills scale down cleanly too: ad-vs-content telling (sponsored results — the one skill with good primary-age data, mostly showing children can't do it untaught), and click restraint — reading the results page as a menu before committing (query formulation). What does not scale down: domain-suffix rules (".org means trustworthy" was never true), design-quality judgments, and any rule phrased as a property of the page alone.

What the evidence doesn't say

It doesn't say children become reliable evaluators — the trials improve performance from a low base; the goal at this age is installing the reflex (check who; check elsewhere), not certifying the judgment.
It doesn't cover the AI layer — synthesized answers with no page to leave need their own treatment (AI literacy); the who-says-this question survives the transition better than any checklist does.
It doesn't make cynicism the goal — "trust nothing" is as unteachable as "trust everything"; the method is calibration, and the emotional register matters (misinformation page).

In the classroom

Retire the checklist posters — teach the three questions and the tab-opening habit instead.
Use designed contrasts: two pages on the same topic, one from a named organisation, one from nowhere — let children discover the difference by investigating, not inspecting.
Practice on things children care to verify — game rumors, animal facts, playground claims; the method transfers, the motivation doesn't need manufacturing.
Model it live: when a class question comes up, verify it on the projector, thinking aloud — the strongest single demonstration a teacher can give.

How Wiz Kids applies this

The Oracle realm's authored web is built for exactly this practice: pages with visible authors and pages without, claims that check out against a second source and claims that don't, sponsored entries in results, and tasks whose goals require visiting a second source before answering — the lateral move made mechanically necessary, so the reflex gets rehearsed rather than recommended.

References

Wineburg, S., & McGrew, S. (2019). Lateral reading and the nature of expertise. Teachers College Record, 121(11).
Breakstone, J., Smith, M., Wineburg, S., et al. (2021). Students' civic online reasoning: A national portrait. Educational Researcher, 50(8).
McGrew, S., et al. (2018–2019). The civic online reasoning intervention studies.
Meola, M. (2004). Chucking the checklist. portal: Libraries and the Academy, 4(3) — the early critique the field took fifteen years to hear.

© Glu IO Pty. Ltd. — Wiz Kids (wiz.kids). Link freely; republication requires permission — see terms. Found an error in our reading of the research? We correct fast: tell any teacher piloting Wiz Kids.