Preserve first judgment
Keep the initial answer instead of rewriting the history of the decision after seeing the key.
HAOLLA preserves the learner's first judgment, pinpoints the smallest decisive evidence, and separates why a distractor looked plausible from the exact word, condition, scope, or logic that makes it wrong.
A common advanced-reading error is choosing an option because its language is genuinely related to the passage. Strong reading first fixes the relationship demanded by the question, then finds the smallest direct evidence and checks each option for added claims, shifted scope, reversed causality, or a changed viewpoint.
Keep the initial answer instead of rewriting the history of the decision after seeing the key.
Mark the decisive word, phrase, condition, or minimum line range - not an entire paragraph by default.
Choose the most attractive distractor. State why it looked similar, then name the single decisive flaw.
Ask not only ‘Is this in the text?’ but ‘Does this evidence prove exactly what this question asks?’
In paired passages, preserve each author's standard, scope, and purpose before inferring agreement, conflict, or response.
Review high-confidence errors first; they can reveal a stable but incorrect decision rule.
A learner may locate relevant lines yet treat relevance as proof. The priority is then not more volume, but tighter question → claim → evidence alignment.
A strong reader can still confuse a concept that appears in the passage with evidence that directly answers the target claim. High-confidence misses are especially useful for correcting decision rules.
Public examples summarize anonymized error patterns from instructional work. Student names, identifying school data, original responses, and individual scores are not published.
Learners explain evidence from memory rather than only rereading an explanation.
Differences are named as scope, strength, causality, viewpoint, or evidence fit rather than a vague feeling.
Initial answer, evidence, trap, and unclear sentence make the reasoning process inspectable.
For broader evidence-based learning practice, see: Institute of Education Sciences. For search trust, the visible content and structured data are kept aligned with Google guidance: Google Search Central.
No. Start with the smallest word, phrase, condition, or line range that decides the answer.
During diagnosis, preserve the first answer so the teacher can see the decision rule that produced it.
Usually one deeply: the most attractive distractor, its similarity to the key, and its decisive flaw.
Fix each author's claim and evidence separately, then identify whether the question asks for agreement, conflict, or a likely response.
Yes. High-confidence errors can expose an automated but incorrect rule even when overall accuracy is high.
Initial judgment, evidence location, strongest trap, and decisive flaw often reveal a clearer next step than the score alone.