Blended Learning Assessment: What to Test Before, During and After Classroom Sessions

Blended learning assessment is the set of checks you run across the online and classroom parts of one program: a pre-work diagnostic, in-module quizzes, in-session checks, a post-session application task and a delayed retention check. Only one or two of those checks should decide a pass or a certificate. The rest should be low-stakes tools that tell learners and trainers what to do next.

Existing guides cover single pieces, such as pre and post course assessments for instructor-led training, vILT knowledge checks and placement tests for ILT learners. This post is the whole-program view: a five-stage blueprint that sets purpose, item type, stakes, attempts, records and audience for each stage, plus the spaced-retrieval research that tells you when to schedule the last check. The blueprint is a design proposal built from published research and standards, not a validated template.

Key takeaways

• A blended program needs five assessment stages: pre-work diagnostic, in-module checks, in-session checks, a post-session application task and a delayed retention check. Each stage answers a different question for a different reader.
• Keep module quizzes and session checks formative, meaning they guide the next step and do not decide a pass. Base the summative decision on the application task, or on a final exam where your program requires one, and not on any single score alone.
• A 2014 meta-analysis of 159 effect sizes found that testing beat restudy (average g = 0.50), and the benefit was larger when learners got feedback. Show the correct answer and a short explanation after every in-module item.
• In one large lab study, the best gap between a review and a final test was about 8 days for a 35-day retention interval. For a five-week retention goal, that points to a short reinforcement quiz about a week after the session.

What is blended learning assessment, and how does it differ from assessing a classroom course?

Blended learning assessment spreads checks across a program that mixes online modules with live classroom or virtual sessions. It differs from classroom-only assessment in two ways: results from the online part can shape the live session, and learner records sit in more than one system. Both facts change when you test, what you record and who sees it.

In a classroom-only course the trainer watches learners work. In a blended program, hours of learning happen unobserved, so checks become the trainer’s eyes: a module quiz can tell the trainer what to reteach before anyone enters the room. Records also scatter, with module scores in a learning platform, poll answers in a meeting tool and observation notes in a trainer’s file. Decide before launch which system holds the official record.

Which parts of a blended program should be formative and which summative?

Treat checks that guide the next step as formative and keep them low-stakes. Treat the one or two scored events that decide a pass, completion or certificate as summative. In most blended programs the pre-work diagnostic, module checks, in-session checks and retention check are formative, and the application task or a final exam is summative.

The stakes label in the blueprint below uses this four-level scale, which is this post’s own convention rather than a published standard:

• None: no consequence for the learner. The result guides planning only.
• Low: the result is recorded and used for feedback and follow-up, but it does not change completion status.
• Medium: completing the check, or reaching a stated score, counts toward program completion.
• High: the result helps decide a pass, certificate or credit.

The American Educational Research Association’s position statement on high-stakes testing, adopted in July 2000, says decisions that affect individual students should not be made on test scores alone. It also calls for a real opportunity to learn the tested content, a chance to retake, and passing scores that are validated for their stated purpose. It was written for student achievement testing, so apply it to training by analogy: do not let one module quiz fail a learner.

What does a blended learning assessment blueprint look like?

A blended learning assessment blueprint lists five stages along the learner’s timeline and fixes the purpose, item type, stakes, attempts, record and audience for each. The table below shows one workable default for a program with online modules followed by a classroom or live virtual session.

The item counts, attempt rules and stakes levels are planning defaults, not research findings. If your program requires a knowledge exam, add it as a sixth row between stages 3 and 4 and set its pass mark with a documented method such as the modified Angoff approach explained in the placement test guide.

What should you test before learners start the online modules?

Test a short, ungraded diagnostic of 8 to 12 items mapped to your learning objectives. Its job is to show learners and trainers where the gaps are, not to rank people. It carries no stakes, and learners should see their own result with a pointer to the module that covers each gap.

Kirkpatrick Partners notes that learning (Level 2) is traditionally measured with pre- and post-tests, so this diagnostic can double as the baseline for a gain calculation. To compare it with a post-test, build matched forms as described in the pre and post course assessment guide. If the real question is which group a learner should join, use a placement test instead.

How do in-module checks prepare learners for the classroom session?

In-module checks are short retrieval quizzes at the end of each online module that give feedback and feed the trainer’s plan for the live session. Keep them low-stakes, allow repeat attempts and show the correct answer with a one-line explanation. Their value comes from practice and feedback, not from the score.

Does retrieval practice improve learning?

Yes, in controlled studies. Rowland’s 2014 meta-analysis of the testing effect in Psychological Bulletin pooled 159 effect sizes from 61 studies and found that taking a test beat restudying the material, with an average effect of g = 0.50. The benefit was larger when learners received feedback (g = 0.73) than when they did not (g = 0.39).

Read that evidence with care. The meta-analysis covered laboratory studies only, mostly with college undergraduates and verbal material such as word lists and prose passages. It supports the principle behind in-module checks, not a particular quiz length for workplace training.

How should trainers use module results before the session?

Have the trainer open the module report one or two days before the live session and note the most-missed items and the most-chosen wrong answers. These show which concepts need reteaching and which can be skipped. This hand-off is the main payoff of module checks, so build it into the trainer’s routine.

Pro tip: Build the Room’s Warm-Up From Module Data. Take the three items learners missed most in the online modules and open the live session with them, unscored and discussed aloud. You spend session time where the data shows confusion, and learners retrieve that content a second time.

What belongs in the classroom session itself?

Use in-session checks to catch misconceptions while the trainer can still reteach: quick polls during teaching and a short scored check before the break or close. They are low-stakes. Record the item ID, how many learners answered against how many attended, the percent correct and the most-chosen wrong option.

A full run-sheet with timings for a 90-minute virtual session is in the vILT knowledge checks guide. The same logic works in a physical room: decide in advance which items are polls and which are scored, and what you will do if fewer than, say, 70 percent answer correctly.

Keep these items different from the module items, because a repeated question can reward recognition rather than understanding. Tag items by objective in your item bank so you can draw fresh items for each stage.

How do you assess application after the classroom session?

Set a scenario or work-sample task that learners complete once they are back in their normal work, scored with a rubric that lists the observable criteria. This stage is usually the summative one. Allow one resubmission after feedback, and tell learners in advance who will see the result.

Kirkpatrick Partners treats on-the-job behavior as Level 3 and says its measurement should start once participants are back in their work environment, not after a long delay, because late measurement leaves no time to adjust the program. Set the task due within a window that matches when the skill is first needed, not on an arbitrary date.

Apply the AERA principles from earlier: share the rubric before the task, set the passing standard deliberately, and treat the resubmission rule as the second chance the statement calls for. Where managers score the task, brief them on the rubric so scores mean the same thing across managers.

When should you run the delayed retention check?

Run it at the interval you care about, for example five weeks after the session, and add a short reinforcement quiz about a week after the session. In Cepeda and colleagues’ 2008 study, the best gap between review and final test grew with the retention interval, but shrank as a share of it.

The study, published in Psychological Science, taught more than 1,350 people a set of 32 obscure facts online, gave them a review after a gap of 0 to 105 days, then tested them after 7, 35, 70 or 350 days. The table shows the best review gap the authors estimated for recall at each test delay.

The authors also report that a gap that is too short costs much more than one that is too long. If unsure, err toward waiting a little longer before the reinforcement quiz.

How would that look on a calendar?

A hypothetical program has three online modules, one classroom day (Day 0) and a goal that learners still remember the content five weeks later. Following the study’s pattern, the schedule could run as below, with each stage keeping the blueprint’s stakes and attempt rules.

• Day -14 to -10: pre-work diagnostic.
• Day -10 to -2: modules 1 to 3, each ending with a feedback quiz.
• Day -1: trainer reviews module data and picks warm-up items.
• Day 0: classroom session with warm-up, polls and a scored check.
• Day 7 to 8: reinforcement quiz of 3 to 5 new items, low-stakes.
• By Day 14: application task submitted and scored.
• Day 35: retention check of 5 to 10 new items.

This is a design inference, not a tested result. The study used trivia facts and an online panel, not job skills, and it tested a single review. Treat the gaps as a starting point and compare your Day 35 results with your Day 14 results.

Pro tip: Schedule From Each Learner’s Own Date. Trigger the reinforcement quiz and retention check from each learner’s session or completion date, not one calendar date for the cohort. Learners who finish a week late would otherwise be tested on a shorter gap than you planned.

What commonly goes wrong with blended learning assessment?

The most common failures are design choices, not technical ones: scoring practice as if it were an exam, not reading module data, reusing the same items at every stage and forgetting the final check. These are reasoned risks from the design above rather than findings from a study.

• Averaging practice scores into the pass mark. Learners then avoid hard attempts, and the score stops showing what they can do.
• Collecting module data nobody reads. The online checks then add effort without changing the teaching.
• Reusing identical items at every stage. Rising scores may reflect memory of the questions, not of the content.
• Giving the retention check no owner. It is the stage most likely to be dropped. Name a person and a trigger date.
• Letting one score decide everything. Where stakes are high, combine the application task with other evidence.

What should you do next?

Pick one upcoming blended course and fill in the blueprint table with your own numbers: how many items per stage, which stage is summative, who owns the retention check and where each record lives. Then write fresh items for each stage from a tagged pool so no stage repeats another.

If you keep your items in a shared bank, tag them by objective and stage. SimpliAssess lists Item Banking, Tag Management and Analytics and Reporting on its features page, which is relevant when several stages draw on one pool of items.

If you also schedule cohorts across online and classroom delivery, SimpliTrain’s page for training companies lists blended learning, instructor-led and virtual classroom delivery and cohort management among its topics.

Frequently asked questions

What is blended learning assessment?

Blended learning assessment is the planned set of checks across the online and in-person parts of a program. It typically includes a pre-work diagnostic, in-module quizzes, in-session checks, a post-session application task and a delayed retention check. Most are formative and guide next steps; one or two are summative and decide a pass or certificate.

How often should you assess learners in a blended program?

No study sets a fixed frequency, so use the five-stage blueprint as a planning default: one diagnostic, three to five items per module, a few checks in the live session, one application task and one delayed check. Add more only if the results change what you do. Each check should have a named reader and a decision.

Should online module quizzes count toward the final grade?

Usually no. Module quizzes work best as low-stakes practice with feedback, which is where the research benefit comes from. If completion is required, require finishing the quiz, not reaching a score. Base the pass decision on the application task or a final exam, and follow the AERA advice not to decide on a single score.

What is the difference between formative and summative assessment in blended learning?

Formative assessment guides the next step in teaching or studying, so it carries low or no stakes, such as a module quiz or a live poll. Summative assessment decides an outcome such as a pass, completion or certificate, such as a scored application task or final exam. A blended program should have mostly the first and one or two of the second.

How long after training should you check retention?

Check at the interval that matters for the job, such as five weeks. In Cepeda and colleagues’ 2008 study, the best review gap was about 8 days for a 35-day test delay. That suggests a short reinforcement quiz about a week after the session, though that was a lab study with trivia facts, not workplace skills.

Can you use the same questions before, during and after a session?

You can, but you usually should not. Identical items can inflate later scores through memory of the question rather than understanding. Draw different items for the same objectives at each stage, and reuse identical items only when you deliberately want a like-for-like comparison, as in a matched pre and post test.