How Quizzes and Assignments Improve Learning
Quizzes and assignments help learners remember, apply, and prove what they learn in online courses.
Watching a lesson can create the feeling that you have learned something.
That feeling can be misleading.
An explanation may seem completely clear while it is in front of you. A formula may make sense when an instructor demonstrates it. A business framework may appear obvious while you are reading about it.
Then the video ends.
The notes disappear.
And suddenly you have to produce the answer yourself.
That moment reveals an important distinction in education:
Exposure to information is not the same as being able to retrieve, apply, analyze, or use it.
This is why quizzes and assignments should not be treated as decorative features added to an online course after the lessons have been created.
When designed properly, they are part of the learning itself.
Quizzes can force learners to retrieve knowledge from memory, expose misconceptions, strengthen retention, and provide information about what needs more attention.
Assignments can push learners further: from remembering an idea to using it, analyzing a situation, making decisions, creating something, receiving feedback, and revising their work.
The evidence for retrieval practice is particularly strong.
A 2021 systematic and meta-analytic review in Psychological Bulletin combined data from 222 classroom studies involving 48,478 students. It found that testing or quizzing improved academic achievement by a moderate amount overall, although the strength of the effect depended on factors including feedback, repetition, test format, timing, and the relationship between practice and final assessment.
A separate systematic review of real classroom research found benefits across different education levels, subjects, test formats, and delays, although it also noted that much of the existing research came from a relatively narrow set of countries.
The practical conclusion is not:
“Give learners more tests.”
It is much more interesting:
Learning improves when learners are repeatedly asked to retrieve, apply, receive feedback, reflect, and try again.
Quizzes and assignments are two different tools for creating those opportunities.
Quizzes and Assignments Are Not Just Ways to Measure Learning
Assessment is often treated as something that happens after learning.
The traditional model looks like this:
Teach → Learn → Test
But research on retrieval practice and formative assessment suggests a more useful model:
Teach → Retrieve → Discover gaps → Receive feedback → Practise → Apply → Improve
In this model, assessment does two jobs.
It can measure learning.
But it can also cause additional learning.
Cornell University's Center for Teaching Innovation describes formative assessment as regular opportunities for learners to practise and apply what they are learning, while summative assessment evaluates learning more cumulatively through activities such as exams or final projects.
That distinction matters for online-course design.
A five-question quiz after a lesson does not have to exist primarily to decide whether a learner receives 70% or 90%.
It can exist to make the learner retrieve the lesson before forgetting it.
An assignment does not have to exist simply so the instructor has something to grade.
It can exist because the learner cannot develop the skill without performing it.
What Quizzes Actually Do to Learning
1. Quizzes Force Retrieval
Suppose you have just watched a lesson explaining the difference between fixed and variable costs.
There are two possible next steps.
Option A
Read the explanation again.
Option B
Close the lesson and answer:
A company pays €2,000 per month for its office, regardless of how many products it sells. Is this a fixed or variable cost, and why?
Option B requires something different.
You must retrieve the concept from memory.
That process is called retrieval practice.
Research reviewed in the Annual Review of Psychology shows that practising retrieval after learning can slow forgetting and improve later access to information across many types of material, learners, and educational settings.
This is sometimes called the testing effect or test-enhanced learning.
The name can be slightly misleading because the important mechanism is not the formal test.
It is retrieval.
A flashcard can create retrieval practice.
So can:
-
a short-answer question;
-
a quiz;
-
explaining a concept without notes;
-
solving a problem;
-
recalling the steps of a process.
The learner has to bring information back from memory instead of simply seeing it again.
Why Retrieval Is Different From Re-Reading
Re-reading often feels easier.
That is partly why learners like it.
When information appears familiar, it can create the impression:
“I know this.”
But recognizing something while looking at it is different from producing it independently.
Imagine learning someone's phone number.
Looking at the number ten times may make it familiar.
Trying to say it without looking forces you to determine whether you actually remember it.
The same principle applies to learning.
The U.S. Institute of Education Sciences recommends using quizzes that require active retrieval and notes that retrieving information can strengthen long-term retention beyond simply re-exposing learners to the same material. Its evidence review particularly recommends providing learners with correct-answer feedback.
This does not mean re-reading is useless.
It means that retrieval adds a form of cognitive work that passive review does not.
2. Quizzes Reveal the Difference Between Familiarity and Knowledge
One of the most useful things a quiz can do is prove to you that you do not understand something as well as you thought.
That can feel uncomfortable.
Educationally, it is extremely useful.
Imagine a learner finishes a module on digital marketing.
Everything feels clear.
Then the quiz asks:
A campaign generated 30 customers from €600 in advertising spend. What was the customer acquisition cost?
The learner cannot remember the formula.
That failure provides information.
Before the quiz:
“I understand CAC.”
After the quiz:
“I understand what CAC means, but I cannot calculate it independently yet.”
That is a much more accurate picture of learning.
Tests and quizzes can therefore contribute to metacognition: awareness of what you know, what you do not know, and where further study is needed.
The Institute of Education Sciences specifically recommends using tests and quizzes as one source of information that can help learners identify material requiring additional study.
3. Quizzes Can Strengthen Long-Term Retention
The effect of retrieval is particularly important because learning is vulnerable to forgetting.
Information that seems obvious immediately after a lesson may be much harder to retrieve days or weeks later.
A major review by Roediger and Butler concluded that retrieval practice can produce substantial improvements in long-term retention compared with additional study and that feedback can further strengthen these benefits.
This has practical implications for online courses.
A quiz should not necessarily appear only once, immediately after the information was presented.
Key concepts can return later.
For example:
Module 1: Learn the concept.
↓
Module 1 quiz: Retrieve it.
↓
Module 3 quiz: Retrieve it again.
↓
Final project: Apply it.
This combines retrieval with spacing.
The learner must return to knowledge after some time has passed rather than successfully answering a question while the lesson is still fresh in working memory.
4. Repeated Quizzing Can Be More Useful Than One Final Test
Many courses follow this structure:
20 lessons → one final examination
That may be useful for certification.
It is not necessarily the strongest structure for learning.
Frequent low-stakes quizzes can distribute retrieval across the course.
Carnegie Mellon's Eberly Center recommends considering recurring low-stakes assignments and assessments because they create repeated opportunities for practice and feedback while giving instructors useful information about learner progress.
Low stakes matters.
If every quiz has enormous consequences, learners may primarily experience it as judgment.
If a quiz is worth little or nothing but provides feedback, its role changes.
It becomes closer to:
“Let's see what you know and what needs more work.”
rather than:
“Prove your worth right now.”
For many online courses, practice quizzes can therefore allow:
-
multiple attempts;
-
immediate explanations;
-
small or zero grade consequences;
-
repeated exposure over time.
The objective is learning.
Not catching learners making mistakes.
5. Good Quizzes Give Feedback
Retrieval alone can be useful.
But learners also need to know when their retrieval was wrong.
Imagine a learner answers:
Gross profit = revenue − operating expenses.
The correct relationship is not simply reinforced by asking the question.
The learner needs correction.
The U.S. Institute of Education Sciences recommends providing correct-answer feedback when quizzes are being used to promote learning and retention.
Research on feedback more broadly shows something important: feedback can be highly influential, but its effects are not automatically positive. The usefulness depends on what information it provides and how the learner can act on it.
Strong quiz feedback therefore goes beyond:
❌ Incorrect.
A better response might say:
Incorrect. Gross profit is revenue minus cost of goods sold. Operating expenses are deducted later when calculating operating profit.
Now the quiz has:
-
identified the error;
-
corrected the misconception;
-
explained the distinction.
That is educational feedback.
Feedback Should Answer: “What Do I Do Differently?”
The Education Endowment Foundation's evidence guidance emphasizes feedback that focuses on the task, subject, process, and self-regulation rather than vague comments about the person. Effective feedback provides information that can help learners improve.
Compare:
Weak feedback
Good job.
Better feedback
Your recommendation identifies the correct customer segment, but the evidence does not yet explain why this segment has the strongest need. Add evidence from your interviews.
The second gives the learner a next move.
That is a useful test for feedback:
Can the learner tell what to do differently because of it?
If not, the feedback may evaluate performance without helping improve it.
6. Quizzes Can Improve Future Learning, Not Only Previous Learning
Most discussions of quizzes focus on remembering what was already taught.
But retrieval may also influence what happens next.
Research has identified a forward effect of testing: retrieving previously studied material can under some conditions improve learning of subsequently presented material.
This suggests another role for quizzes inside long courses.
They can create boundaries between topics.
Instead of learners watching six lessons continuously:
Lesson → Lesson → Lesson → Lesson → Lesson
a course might create:
Learn → Retrieve → Continue
This may help learners organize what has already been learned before introducing additional information.
But Quizzes Have Limits
Quizzes are powerful.
They are not capable of measuring every type of learning.
Suppose the learning objective is:
Design a complete marketing strategy for a small business.
A multiple-choice quiz might test whether learners can identify:
-
segmentation;
-
positioning;
-
channels;
-
CAC;
-
conversion rates.
But a learner could answer all those questions correctly and still be unable to build a coherent marketing strategy.
This is where assignments become essential.
What Assignments Add That Quizzes Often Cannot
If quizzes are particularly useful for retrieval and knowledge checks, assignments can create opportunities for:
application
analysis
evaluation
decision-making
creation
transfer
Cornell University explicitly recommends selecting assessment methods according to the type of thinking expressed in the learning objective.
Recall can often be assessed through quizzes.
Applying knowledge may require problem sets, cases, or practical activities.
Analysis may require cases, papers, comparison, or debate.
Creation may require a project, portfolio, performance, website, research task, or other product.
This leads to a basic course-design principle:
The assessment should resemble the capability you want the learner to develop.
7. Assignments Turn Knowledge Into Performance
Imagine a course about business model design.
A quiz might ask:
Which element of the Business Model Canvas describes how a company reaches customers?
Useful.
But now give the learner this assignment:
Choose a real or hypothetical business. Define its target customer, value proposition, channels, revenue model, and major cost structure. Then explain the assumptions that still need validation.
The cognitive work is completely different.
The learner must:
-
interpret a situation;
-
select information;
-
make decisions;
-
connect multiple concepts;
-
justify choices;
-
produce an output.
They are no longer proving they recognize the framework.
They are trying to use it.
8. Assignments Can Build Transfer
A major goal of learning is transfer: using knowledge outside the exact context in which it was originally taught.
Suppose an instructor explains a customer-persona example involving a restaurant.
A learner who remembers the restaurant example has learned something.
But can they apply the same principles to:
-
a SaaS company?
-
a fitness trainer?
-
an NGO?
-
a university program?
Assignments can force learners to move beyond reproducing the example.
Carnegie Mellon recommends creating assessments around what learners should actually be able to do and warns that assessment must align with the intended learning objectives.
This is particularly important in professional education.
The goal is rarely:
Remember the instructor's example.
The goal is:
Use the principle when the situation changes.
9. Assignments Can Create Authentic Learning
Some of the strongest assignments resemble work that learners may actually encounter outside the course.
Cornell describes authentic assessment as tasks requiring learners to demonstrate knowledge and skills in real-world or close-to-real-world contexts, often involving higher-order thinking and transfer.
For example:
Marketing course
Instead of:
Write 500 words defining content marketing.
Use:
Build a four-week content strategy for a hypothetical company and justify the channels, audience, themes, and metrics.
Finance course
Instead of:
Describe cash-flow forecasting.
Use:
Build a 12-month cash-flow forecast from a provided set of business assumptions.
Programming course
Instead of:
List the differences between arrays and objects.
Use:
Build a small application that stores and manipulates structured data.
Instructional-design course
Instead of:
Explain learning objectives.
Use:
Rewrite five vague course objectives into measurable outcomes and design one assessment for each.
The assignment creates a bridge between knowing about the subject and doing something with the subject.
10. Assignments Can Produce Evidence of Learning
This is especially important in online education.
Completion is easy to measure.
Learning is harder.
A platform can know that someone:
-
opened a video;
-
reached the final lesson;
-
spent 14 minutes on a page.
That does not necessarily tell us what the learner can do.
Assignments create direct evidence.
Consider:
CoursePossible Evidence of LearningExcelCompleted analytical workbookDigital marketingMarketing strategyGraphic designPortfolio piecePythonWorking programStartup validationInterview reportProject managementProject planFinancial modelingFinancial modelOnline teachingCourse prototype
The output becomes evidence that the learner attempted to apply the skill.
Whether that evidence is strong depends on the quality of the assignment and assessment criteria.
But it tells us more than:
Video completed: 100%.
[Internal link: article about why course completion is not the same as learning]
11. Assignments Create Opportunities for Iteration
The strongest assignment is not always:
Submit → receive grade → finished
Learning can become much richer when the process is:
Create → receive feedback → revise
Imagine a learner creates a value proposition.
Version one is vague.
The instructor explains:
You describe the product, but you have not clearly connected it to the customer's problem.
The learner revises it.
Now they must use the feedback.
This is important because feedback only has educational value when learners have an opportunity to act on it.
An assignment submitted at the very end of the course with detailed comments the learner will never use may provide less learning value than a smaller draft submitted earlier and revised.
12. Assignments Can Develop Metacognition
Learners do not only need to perform tasks.
They need to become better at judging their own performance.
This is part of metacognition and self-regulated learning.
The Education Endowment Foundation's updated evidence review identifies a strong research base around metacognition and self-regulation and emphasizes strategies that help learners plan, monitor, and evaluate their learning.
Assignments can support this by asking learners to reflect.
For example:
What was the most difficult decision in this project?
Which part of your solution are you least confident about?
What would you change if you had another attempt?
Which feedback did you use in your revision?
These questions turn an assignment from:
Produce answer
into:
Produce → inspect → improve
That can help learners become more independent over time.
Quizzes and Assignments Should Work Together
The best question is not:
Quizzes or assignments?
They do different jobs.
Imagine a course on marketing analytics.
Lesson
Explain:
-
conversion rate;
-
CAC;
-
ROAS;
-
retention;
-
churn.
↓
Quiz
Ask learners to retrieve definitions and perform simple calculations.
↓
Scenario
Show campaign data and ask what appears to be going wrong.
↓
Assignment
Ask the learner to analyze a full dataset and write a recommendation.
↓
Feedback
Identify errors in interpretation.
↓
Revision
Learner updates the recommendation.
This creates a progression:
Remember → understand → apply → analyze → evaluate
The different assessments support different cognitive demands.
Cornell's assessment framework makes precisely this distinction, linking different forms of assessment with increasingly complex learning outcomes.
Assessment Alignment Matters More Than Assessment Quantity
Adding 50 quizzes does not automatically create a strong course.
Adding 20 assignments does not either.
Assessment becomes useful when it is aligned.
Carnegie Mellon's Eberly Center describes alignment as the relationship among:
Learning objectives
↓
Assessment
↓
Instruction
When those components match, learners practise and demonstrate the skills the course actually claims to develop. When they do not match, students may earn good grades without mastering the intended capability.
Consider this example.
Learning objective
Learners will be able to conduct a customer interview.
Lessons
Videos explaining:
-
interview preparation;
-
question design;
-
follow-up questions;
-
avoiding bias.
Assessment
Twenty multiple-choice questions.
There is a problem.
The assessment measures whether the learner can recognize principles of interviewing.
It does not establish whether they can conduct an interview.
A better assessment might be:
Conduct a 15-minute interview and submit the recording and a reflection on what you would improve.
Now the assessment matches the skill.
The Wrong Assessment Can Change What Learners Study
Assessment does not merely measure learner behavior.
It influences learner behavior.
If learners know every assessment is based entirely on memorizing definitions, many will rationally focus on memorizing definitions.
Even if the instructor repeatedly says:
“I want you to think critically.”
But if the final assessment requires analysis, practical application, and justification, learners have a stronger reason to practise those things.
This is why assessment architecture is part of curriculum architecture.
The question:
“How will we assess this?”
should be asked while designing the learning outcome.
Not after all the lessons have already been recorded.
[Internal link: How to Structure an Online Course Curriculum]
How to Design Better Quizzes
Keep Practice Quizzes Low Stakes
Not every quiz needs to contribute significantly to a final grade.
Use some quizzes primarily for:
-
retrieval;
-
diagnosis;
-
feedback;
-
repetition.
Learners should feel able to discover mistakes without every mistake becoming a major penalty.
Ask Learners to Retrieve, Not Only Recognize
Multiple-choice questions can be useful and can assess more than simple recall when designed carefully.
But recognition is often easier than generating an answer.
Short-answer, calculation, labeling, explanation, or fill-in-the-blank questions can force more active retrieval.
The Institute of Education Sciences specifically notes strong support for quizzing that requires learners to actively recall information.
Use the format that matches the objective.
Give Explanatory Feedback
Avoid:
Wrong.
Prefer:
The correct answer is B because...
And when appropriate:
A would be correct if...
This can help address misconceptions rather than simply reveal the score.
Return to Important Concepts
Do not quiz something once and assume it is permanently learned.
Bring important knowledge back later.
Spacing and repeated retrieval can strengthen retention.
A 2022 classroom study, for example, found that science concepts included in repeated retrieval quizzes were remembered better one month later than concepts that had not been quizzed, with additional benefits when concepts were interleaved rather than grouped together.
Ask Some Questions That Require Explanation
Consider:
Which option is correct?
Then:
Why?
Deep explanatory questions can require learners to connect concepts rather than merely select an answer.
The Institute of Education Sciences gives strong-evidence support to using explanatory questions that encourage deeper understanding.
How to Design Better Assignments
Start With the Outcome
Ask:
What should the learner be capable of doing?
Then:
What task would give me evidence of that capability?
Carnegie Mellon recommends exactly this approach when designing assignments.
Make the Task Authentic When Possible
Let learners work on situations that resemble actual use of the skill.
This does not mean every assignment needs to involve a real company or client.
A realistic hypothetical scenario can work well.
Make Expectations Visible
Poor assignment:
Create a marketing plan.
Better assignment:
Create a three-month marketing plan for the provided business. Include the target segment, positioning, three acquisition channels, content strategy, budget allocation, and five performance metrics. Explain the reasoning behind each major decision.
The learner now understands the expected evidence.
Rubrics can make those expectations even clearer.
Scaffold Difficult Projects
Do not ask a beginner to produce an enormous final project with no intermediate support.
Break the project into stages.
For example:
Week 1
Target audience
Week 2
Value proposition
Week 3
Channel strategy
Week 4
Budget
Week 5
Metrics
Final
Integrated marketing plan
Now feedback can influence the project before the final submission.
Allow Revision
Where possible:
Draft → feedback → revision
is educationally richer than:
Submission → grade → goodbye
Ask Learners to Explain Their Decisions
A finished artifact does not always reveal understanding.
Someone may produce the right answer accidentally, copy a template, or increasingly in 2026, use AI to produce a polished output.
Ask learners:
Why did you choose this approach?
What alternatives did you consider?
What assumption is most uncertain?
What would change your recommendation?
This makes the learner's reasoning visible.
Assignments Matter Even More in the Age of AI
Generative AI has made it easier to produce:
-
essays;
-
reports;
-
code;
-
plans;
-
presentations;
-
summaries.
This creates a challenge.
If the assignment is simply:
Write 1,000 words about digital marketing.
an AI system can potentially produce a plausible answer immediately.
That does not mean assignments are obsolete.
It means assignment design needs to evolve.
Cornell currently recommends authentic, personalized assessment as one strategy for maintaining meaningful evidence of student learning in an environment where AI tools can generate conventional academic responses.
Instead of asking only for a final document, a course can ask learners to submit:
The final artifact
their reasoning
evidence used
reflection
revision after feedback
For practical courses, the question becomes:
Can the learner defend and use what they submitted?
That is harder to outsource than producing text.
When Quizzes Become Bad Learning Tools
A quiz can fail when it is:
-
disconnected from learning objectives;
-
purely about trivial details;
-
excessively high stakes;
-
repetitive without meaningful retrieval;
-
followed by no useful feedback;
-
used to assess a skill it cannot measure.
A learner who memorizes 100 definitions may perform beautifully on the quiz.
If the course promises practical competence, that score may reveal surprisingly little.
When Assignments Become Bad Learning Tools
Assignments can also fail.
Common problems include:
Busywork
The learner spends time but performs little meaningful thinking.
Unclear expectations
The learner has to guess what the instructor wants.
Misalignment
The assignment measures writing skill when the objective was something else.
Excessive difficulty
The assignment requires capabilities learners have never had the opportunity to develop.
No feedback
Learners repeatedly practise mistakes.
Feedback without revision
The learner receives useful advice after they no longer need to do anything with it.
Excessive workload
More work does not automatically mean deeper learning.
The assignment should create useful cognitive work, not simply more work.
Formative and Summative Assessment Should Serve Different Purposes
A useful online course often contains both.
Formative
Designed primarily to support learning while it is happening.
Examples:
-
practice quizzes;
-
exercises;
-
drafts;
-
small assignments;
-
reflections;
-
practice problems.
The central question is:
What needs improvement?
Summative
Designed primarily to evaluate achievement after substantial learning.
Examples:
-
final project;
-
examination;
-
portfolio;
-
capstone;
-
practical demonstration.
The central question is:
Has the learner achieved the intended outcome?
Confusing the two can create problems.
If every practice activity carries major consequences, learners may become reluctant to experiment and fail.
If nothing meaningful is ever assessed, certificates or completion claims may communicate very little.
A strong course gives learners room to fail while learning before being evaluated for mastery.
A Better Assessment Architecture for Online Courses
Imagine a six-module course.
Instead of:
Module 1 → video
Module 2 → video
Module 3 → video
Module 4 → video
Module 5 → video
Module 6 → final quiz
Consider:
Module 1
Concept → retrieval quiz
Module 2
Concept → worked example → practice exercise
Module 3
Concept → quiz containing earlier material
Module 4
Case study → short assignment
Module 5
Advanced application → feedback
Module 6
Final project → reflection → assessment
Now assessment is distributed throughout the learning journey.
The learner does not simply consume information for five hours and suddenly discover what they failed to understand at the end.
What This Means for Online Course Creators
The most important implication is simple:
Do not design all the content first and add assessments afterward.
Design these together:
Learning outcome → evidence → practice → instruction
Suppose you want learners to:
Build a basic financial model.
First ask:
What would demonstrate that?
Answer:
A functioning financial model.
Then:
What must learners practise before they can build it?
Perhaps:
-
revenue assumptions;
-
cost structures;
-
cash flow;
-
formulas;
-
scenarios.
Then decide:
Which things require explanation?
Now you can build lessons.
This is backwards from the way many courses are created.
But it produces much stronger alignment.
[Internal link: How to Create an Online Course: Step-by-Step Guide]
A Practical Quiz-and-Assignment Framework
Course creators can use this simple progression:
1. Retrieve
Tool: quiz
Can the learner remember the key concept?
↓
2. Explain
Tool: short answer / discussion
Can the learner explain why it matters?
↓
3. Apply
Tool: exercise
Can the learner use it in a straightforward situation?
↓
4. Analyze
Tool: case study
Can they determine what is happening in a more complicated situation?
↓
5. Create
Tool: assignment / project
Can they produce something using the knowledge?
↓
6. Evaluate
Tool: feedback / reflection / revision
Can they judge and improve their own work?
This turns assessment into a progression toward competence.
The Bigger Lesson: Learning Requires Learners to Do Some of the Work
An instructor can explain beautifully.
A video can be professionally produced.
An AI tutor can answer instantly.
A platform can recommend the next lesson.
None of these removes one fundamental requirement:
The learner still has to think.
They have to retrieve.
Make mistakes.
Solve.
Choose.
Apply.
Explain.
Create.
Revise.
That is why quizzes and assignments matter.
They insert productive learner activity between:
“I saw the information”
and:
“I can use the information.”
The future of online education should therefore not be measured only by how efficiently platforms deliver content.
A stronger question is:
How effectively does the learning experience cause learners to think, practise, receive feedback, and develop capability?
That is a much higher standard.
Frequently Asked Questions
Do quizzes actually improve learning?
Yes, when quizzes require learners to retrieve previously learned information, substantial research supports improvements in long-term retention. A 2021 meta-analysis involving 48,478 learners across 222 classroom studies found a moderate overall benefit, although factors such as feedback, timing, repetition, and question design affected the results.
Are quizzes better than re-reading?
For long-term retention, retrieval practice often produces stronger results than simply studying the same material again. That does not mean learners should never re-read; rather, effective study should include opportunities to retrieve knowledge without looking at the answer.
Should every online course include quizzes?
Not necessarily. Assessment should match the learning objective. Quizzes are especially useful for retrieval, checking understanding, and certain types of application, but a practical skill may require projects, demonstrations, simulations, or other forms of assessment.
What is the purpose of assignments in online learning?
Assignments allow learners to apply, analyze, evaluate, or create with what they have learned. Well-designed assignments can provide direct evidence of competence and create opportunities for feedback, reflection, and revision.
How many quizzes should an online course have?
There is no universally correct number. Frequent, relatively low-stakes retrieval opportunities can be useful, but quality and timing matter more than maximizing quiz quantity. Important concepts can also be revisited across modules rather than tested only once.
Should quizzes count toward the final grade?
Some can, but not every quiz needs to. Low-stakes or ungraded practice can encourage retrieval and give learners information about their progress without turning every mistake into a major penalty.
Are projects better than quizzes?
They measure different things. A quiz may efficiently test recall and understanding, while a project can provide stronger evidence of application, analysis, and creation. High-quality courses often use both.
How should AI affect assignment design?
Assignments should increasingly require evidence of reasoning, application, personalization, reflection, or defense of decisions rather than relying entirely on generic written outputs. AI can be incorporated where appropriate, but the assessment still needs to reveal what the learner can actually understand and do.
Conclusion
Quizzes and assignments should not exist simply because online courses are expected to have them.
They should exist because learning requires more than exposure.
A quiz can force a learner to retrieve something that otherwise only felt familiar.
Feedback can reveal a misconception before it becomes permanent.
An assignment can move the learner from explaining a concept to applying it.
A project can show whether separate ideas can be integrated into a functioning result.
Reflection and revision can help learners understand not only whether they were wrong, but how to improve.
The most useful distinction is therefore:
Content gives learners something to learn.
Assessment gives them something to do with what they learned.
And when assessment is designed as part of the learning process—not merely as a final judgment—it becomes one of the mechanisms through which learning actually happens.
External Sources Used
Yang et al. — Psychological Bulletin
Study: Testing (Quizzing) Boosts Classroom Learning: A Systematic and Meta-Analytic Review
Year: 2021
The review synthesized 222 independent classroom studies and data from 48,478 students. It found a moderate positive overall effect of testing on academic achievement and identified feedback, repetition, timing, and test format as important moderators.
McDermott — Annual Review of Psychology
Study: Practicing Retrieval Facilitates Learning
Year: 2021
Used for the evidence explaining retrieval practice, forgetting, long-term retention, classroom application, and the broader theoretical basis for retrieval-based learning.
Read the Annual Review article
Agarwal, Nunes & Blunt — Educational Psychology Review
Study: Retrieval Practice Consistently Benefits Student Learning: A Systematic Review of Applied Research in Schools and Classrooms
Year: 2021
The authors screened nearly 2,000 abstracts and analyzed 50 classroom experiments. Most reported medium or large benefits from retrieval practice, although the review also highlights limitations in the geographic diversity of the evidence base.
Roediger & Butler — Trends in Cognitive Sciences
Study: The Critical Role of Retrieval Practice in Long-Term Retention
Year: 2011
Used for the distinction between repeated studying and active retrieval, the effects of feedback, and evidence that retrieved knowledge can sometimes transfer to new contexts.
U.S. Institute of Education Sciences / What Works Clearinghouse
Source: Organizing Instruction and Study to Improve Student Learning
Used for evidence-based recommendations around quizzing, active recall, corrective feedback, spacing, study monitoring, and deep explanatory questions. The guide rates using quizzes to re-expose learners to key content as supported by strong evidence.
Carnegie Mellon University — Eberly Center
Sources: Align Assessments with Objectives, Creating Assignments, and Frequent Low-Stakes Assignments
Used for alignment between learning objectives and assessment, assignment design, the use of repeated low-stakes practice, and the principle that assessments should provide evidence of the capability the learner is expected to develop.
Explore Carnegie Mellon's assessment guidance
Cornell University — Center for Teaching Innovation
Source: Learning Outcome Types and Recommended Assessment Methods
Used to distinguish appropriate assessment approaches for recall, understanding, application, analysis, evaluation, and creation.
Explore Cornell's assessment framework
Cornell University — Authentic Assessment and Academic Integrity
Used for the discussion of authentic assessment, transfer, higher-order thinking, and assignment design in an environment where learners have access to generative AI.
Read Cornell's academic integrity and authentic assessment guidance
Education Endowment Foundation
Sources: Teacher Feedback to Improve Pupil Learning and Feedback Teaching and Learning Toolkit
Used for current evidence on effective feedback, including the importance of task-, subject-, and self-regulation-focused feedback and giving learners information they can use to improve.