Last updated: November 12, 2026
1Lecture develops learning activities to help readers explore biological concepts, apply explanations, and check their understanding. This page sets the standards for reviewing the scientific content, behavior, and usability of interactive tools and quizzes.
Validation should address both the underlying explanation and the way the activity works. A correct formula is insufficient if a tool misreads an input; a working quiz is insufficient if its answer key is scientifically wrong.
Define the Purpose and Scope
Each activity should have a clear learning objective, intended audience, and connection to the surrounding lesson. Development notes should identify what the activity represents, what users can change, and what its results mean.
Assumptions must be stated when they affect interpretation. A simplified inheritance exercise, for example, should identify the inheritance model being used. A pathway explorer should explain the basis on which its inputs and outputs are counted.
An illustrative model should be described according to its actual capabilities. Educational activities must not imply clinical prediction or measurement of a learner’s overall scientific competence.
Establish Scientific Support
Questions, explanations, formulas, diagrams, and model assumptions should be supported by suitable references. Source records should make it possible to trace important claims and reproduce worked examples.
Scientific checking should examine terminology, units, mechanisms, conditions, and important exceptions. Activities should distinguish measured data from calculated or illustrative values.
Human editorial responsibility applies to all content, including material suggested by AI, imported from a question bank, or created through a third-party authoring platform.
Validate Questions and Answer Keys
Every question should test the stated learning objective and provide enough information to answer it. Wording should be direct, with unnecessary ambiguity removed.
Review should confirm that:
- The expected answer is defensible under the conditions stated.
- Incorrect options are genuinely incorrect in that context.
- Single-answer and multiple-answer questions are clearly distinguished.
- Explanations agree with the selected answer and the lesson.
- Diagrams, labels, and answer choices use consistent terminology.
- Any required numerical tolerance or answer format is defined.
Distractors should help reveal a relevant misunderstanding. They should not depend on misleading wording or incidental clues.
Free-text and numerical questions need particular care. Accepted synonyms, equivalent notation, rounding, and valid alternative answers should be considered where relevant. Answers requiring interpretation beyond the scoring method’s capabilities should not be marked with unwarranted certainty.
Test Tool Logic and Results
Tools should be checked against independently worked examples or reliable reference cases. Expected results should be established independently of the implementation being tested.
| Check | Expected behavior |
|---|---|
| Typical valid inputs | Results agree with verified reference cases |
| Relevant boundaries | Defined behavior at limits, including zero where meaningful |
| Missing or invalid inputs | Clear instructions; no misleading result |
| Units and conversions | Consistent quantities, labels, and calculations |
| Changed inputs and reset | Previous results do not remain falsely presented as current |
| Randomized content | Answer keys and explanations remain correctly associated |
Impossible or unsupported inputs should not silently produce plausible-looking answers. If a value is adjusted automatically, that adjustment should be apparent.
Where randomness is used, checks should distinguish expected variation from an error. Reference examples should specify any settings needed to reproduce a result.
Check Scoring and Feedback
Scoring rules should be understandable. Where partial credit, multiple selections, penalties, or attempt limits apply, their effect should be explained.
Question order or answer shuffling must not change which response is correct. Feedback should match the user’s actual selection and explain the relevant concept.
A practice score reflects performance on those questions under those conditions. Claims that a quiz is a validated assessment require separate evidence appropriate to that claim.
Where useful, feedback should point readers toward a relevant explanation or prerequisite rather than simply display a mark.
Review Accessibility and Usability
Activities should have clear instructions, labeled inputs, readable results, and keyboard-operable controls. Essential meaning should not depend solely on color, hover, sound, or dragging.
Input errors should identify the affected field and explain how to resolve the problem. Results and feedback should be available to assistive technologies where applicable.
Checks should consider small screens, text enlargement, relevant browser behavior, and reduced-motion needs. Visual activities should provide an appropriate text-based explanation or alternative interaction.
When progress or responses are stored or transmitted, the activity should explain the relevant behavior consistently with the site’s privacy information.
Record Review and Release Decisions
Development records should identify the activity version, scientific references, test cases, expected results, issues found, and resolutions.
Review descriptions must reflect actual work. Editorial checking, functional testing, and independent specialist review should not be treated as interchangeable.
Publication should wait when an unresolved issue affects scientific meaning, scoring, or essential functionality. A limitation that is acceptable for an educational model should be disclosed clearly.
Maintain and Correct Activities
Changes to questions, formulas, datasets, dependencies, or interaction behavior should trigger checks appropriate to the change. Related explanations, answer keys, and downloadable versions should also be considered.
Activities with significant unresolved errors may be restricted while they are assessed. Material corrections should be documented when they affect how readers interpret previous results.
Readers can report a problem by providing the page URL, activity or question, inputs or selections, observed result, and expected behavior. Personal or sensitive information is unnecessary for reproducing a typical learning-activity error.