Pre and Post Training Assessment for Instructor-Led Training: A Practical Guide
A pre and post training assessment is the same set of learning objectives measured twice, once before a session and once after it, so you can show what participants gained. To make the comparison fair, both tests must cover the same objectives at the same difficulty, and you must report gain in a way that does not punish strong starters.
Most guides stop at “give a quiz first and another one last.” That leaves training providers with scores nobody trusts. This guide gives you a matched blueprint, a worked gain calculation and a list of the mistakes that quietly distort results in a classroom setting.
Key takeaways
- Build the pre-test and post-test from one blueprint, so each objective gets the same number of items at the same cognitive level on both forms.
- Prefer two parallel forms over identical forms, because identical forms let memory of the questions inflate the gain.
- Report normalized gain alongside raw gain, because raw point gains favor learners who started low.
- Treat the pre-test as part of the course, because answering questions before instruction can itself help learning.
Why run a pre and post training assessment at all?
A pre and post training assessment gives you evidence of learning, which is what Level 2 of the Kirkpatrick model asks for. Without a baseline, a high post-test score only shows that people know the material now, not that your session taught it to them.
Kirkpatrick Partners defines Level 2 as the degree to which participants acquire the intended knowledge, skills, attitude, confidence, and commitment, and notes that it is traditionally assessed with pre- and post-tests. The other levels (reaction, behavior, results) sit on either side of it.
For a training provider, the practical payoff is threefold. You can adjust the course when an objective shows no gain, you can show clients proof that goes beyond a satisfaction survey, and you can spot learners who arrive already competent.
What does a pre-test measure and what does a post-test measure?
A pre-test measures what participants already know about the course objectives. A post-test measures what they know after instruction. The difference is only meaningful if both tests sample the same objectives, which is why the blueprint comes before any item writing.
Is a pre-test only a measurement tool?
No. Research on the pretesting effect found that answering questions before studying improved later performance, even when most first answers were wrong. In one set of experiments reported in the Journal of Experimental Psychology: Applied in 2009, Richland, Kornell and Kao saw 75% correct on a later test for pretested material versus 56% with extra study time, although participants answered only 5% of pretest questions correctly.
The same authors caution that the benefit holds when the tested material is worth learning and instruction then supplies the answers. The gain was specific to the tested content, not to everything in the passage. For your sessions, that means a pre-test is not neutral: it points attention at the objectives it covers.
How do you build a matched pre and post training assessment?
Start from a blueprint: a table that lists each learning objective, its cognitive level and the number of items it receives. Write both forms from that table. When the blueprint is shared, differences in score are more likely to reflect learning than differences in what each test happened to ask.
The table below is a template for a one-day ILT course with five objectives and two 20-item forms. Adjust the counts to your course, but keep the rows identical across forms.
| Learning objective | Cognitive level (Bloom’s taxonomy) | Form A items (pre) | Form B items (post) | Weight |
|---|---|---|---|---|
| 1. Identify core terms and safety rules | Remember | 4 | 4 | 20% |
| 2. Explain why each procedure step exists | Understand | 4 | 4 | 20% |
| 3. Apply the procedure to a routine case | Apply | 6 | 6 | 30% |
| 4. Diagnose a fault from symptoms | Analyze | 4 | 4 | 20% |
| 5. Choose between two acceptable methods | Evaluate | 2 | 2 | 10% |
Should the pre-test and post-test use identical questions?
Parallel forms are usually safer. Parallel forms are two versions built to the same blueprint, with different items testing the same objectives at similar difficulty. Identical forms are simpler, but a learner who remembers a question from the morning can score higher in the afternoon without learning anything new.
Identical forms are defensible in one case: when the post-test is delivered long after the pre-test and the items are never shown or discussed in between. Even then, you still need evidence that the conditions were comparable. The Standards for Educational and Psychological Testing, published jointly by AERA, APA and NCME, is the reference to consult for how comparability, reliability and fairness should be documented.
How many items does each objective need?
There is no universal number. A single item per objective gives a noisy picture, because one lucky guess moves the result. Aim for at least three or four items on objectives you plan to report separately, and report the whole-test score only when objective-level counts are too small to interpret.
Pro tip: Pilot Your Forms Backwards. Before the course runs, give Form A to half of a small pilot group and Form B to the other half, then swap. If average scores differ by more than a few points between forms on the same people, your forms are not parallel, and you should rewrite the outlier items before using them to measure gain.
How do you calculate learning gain from pre and post scores?
Raw gain is the post-test score minus the pre-test score. Normalized gain divides that raw gain by the room left to improve. The formula, from physicist Richard Hake, is (post minus pre) divided by (100 minus pre), with scores as percentages. It answers the question: what share of the possible improvement did learners actually achieve?
Hake defined average normalized gain for a class as the class post-test average minus the class pre-test average, divided by 100 minus the pre-test average. He used it in a 1998 American Journal of Physics survey of 62 introductory physics courses with 6,542 students.
What does a worked example look like?
Take three ILT classes that all improved. The table shows why raw gain alone can mislead. Class A has the biggest point gain, but class B did better against what was left to learn than class C did.
| Class | Pre-test average | Post-test average | Raw gain (points) | Room to improve | Normalized gain |
|---|---|---|---|---|---|
| Class A (new hires) | 40% | 76% | 36 | 60 | 0.60 |
| Class B (mixed experience) | 70% | 85% | 15 | 30 | 0.50 |
| Class C (experienced staff) | 85% | 91% | 6 | 15 | 0.40 |
Class A: (76 minus 40) divided by (100 minus 40) is 36 divided by 60, or 0.60. Class B: 15 divided by 30 is 0.50. Class C: 6 divided by 15 is 0.40. Class C looks weak on raw gain (6 points) but captured 40% of what it could.
How should you interpret a normalized gain?
Use it to compare, not to grade. Hake labeled courses with gain of 0.7 or above as high, between 0.3 and 0.7 as medium and below 0.3 as low, but those bands came from physics courses measured with concept tests. Treat them as a reference point, and build your own baseline from your last few runs of the same course.
Normalized gain has limits. It is unstable when the pre-test average is near 100%, because the denominator shrinks toward zero. It also says nothing about why scores moved. Pair it with item-level results to see which objectives drove the change.
Which pitfalls undermine a pre and post training assessment?
Most failures come from design choices made before anyone sits the test. The list below covers the problems that appear most often in classroom delivery. These are reasoned risks of any before-and-after design rather than findings from one study, so check them against your own data.
- Practice effects: learners remember items from the first test, which inflates the second score.
- Unequal forms: if Form B is easier, the gain reflects the form, not the course.
- Teaching to the test: if the instructor sees the items, the lesson narrows to those items and the result stops representing the objective.
- Regression to the mean: if you single out the lowest pre-test scorers, they tend to score higher next time even without help.
- Floor and ceiling effects: a pre-test that is too hard or too easy hides real differences.
- No comparison group: without one, you cannot rule out that learners would have gained anyway through work experience.
When do pre and post scores not prove the training worked?
Pre and post scores show change, not cause. If participants practiced on the job between tests, or the post-test came with a certificate attached and higher motivation, the gain may not belong to the course. Say what you measured: knowledge change across the session, not behavior change on the job, which is Kirkpatrick Level 3.
How do you run the pre and post training assessment in a live classroom?
Run the pre-test at the start of the session or just before it, run the post-test when the session ends or soon after, and keep conditions the same. Same time limit, same format and the same rules on notes. Consistent delivery matters more than the platform you use.
Step 1: Set the objectives. Write measurable objectives and check each one is something the course teaches.
Step 2: Draft the blueprint. Fill in a table like the one above and agree it with the instructor.
Step 3: Write two forms. Keep items for each objective at a similar difficulty. Reviewing items in a shared repository, such as a centralized test content hub, makes it easier to reuse and retire items without losing track of which form they belong to.
Step 4: Pilot and fix. Try both forms on a small group, check that scores line up and rewrite weak distractors.
Step 5: Deliver and score. Randomize which form each learner takes first when you can, so form difficulty is spread across the room.
Step 6: Report gain and act. Report raw and normalized gain, plus objective-level results, then change the lesson plan where gain was lowest.
Who should see the pre-test results and when?
Tell learners why they are taking the pre-test and that it does not count against them. Share individual results with the learner and class-level results with the instructor. Show the correct answers to pre-test items during instruction rather than afterward, since the pretesting research found the benefit depends on instruction that supplies the answers.
Pro tip: Keep Pre-Test Items Out Of Slides. If a pre-test item appears word for word in the instructor’s slides, the post-test measures slide recall. Write slides to teach the objective, then keep a short checklist so no item stem or answer option is quoted in the deck.
How do item banks help you keep forms comparable?
An item bank is a tagged collection of questions you can draw from to build new forms. When each item is tagged with an objective, a cognitive level and results from earlier use, you can assemble Form A and Form B to the same blueprint, and rotate forms between cohorts without rewriting everything.
Item statistics help. An item that almost everyone answers correctly before training adds little to a pre-test, and an item that high scorers miss more often than low scorers should be reviewed. If you are weighing how much of this to automate, the notes in generative AI for test content cover where drafted items need human review. If you are choosing tooling more broadly, see why training providers weigh a white-label assessment partner.
What should you do next?
Pick one course you already teach and write a five-row blueprint like the one above. Draft two short forms from it, run the backwards pilot with colleagues and calculate normalized gain on your next cohort. One well-built course gives you a template for the rest of your catalog.
Once you have forms worth reusing, you will want somewhere to store items, schedule sessions and deliver tests together. SimpliTrain’s training providers page describes how its platform handles ILT scheduling alongside assessments for professional training companies.
Frequently asked questions
What is a pre and post training assessment?
It is a pair of tests built on the same learning objectives, given before and after training. Comparing the two scores shows how much learners gained. For the comparison to be fair, both tests need to follow one blueprint, be delivered under the same conditions and be scored the same way.
Should pre-test and post-test questions be the same?
Usually no. Using different items built to the same blueprint, called parallel forms, reduces the chance that remembering a question raises the post-test score. Identical forms are simpler to produce, but if you use them, keep the items unseen between tests and be cautious about how you read the gain.
How do you calculate learning gain?
Raw gain is post-test percent minus pre-test percent. Normalized gain is raw gain divided by 100 minus the pre-test percent. A class moving from 40% to 76% has a raw gain of 36 points and a normalized gain of 0.60, since it achieved 60% of the available improvement.
How many questions should a pre-test have?
Enough to cover every objective you plan to report, and no more than learners can finish comfortably before the session. There is no fixed number, but fewer than three items per reported objective makes objective-level results too noisy to act on.
Do pre-tests help learning or only measure it?
They can do both. Research published in 2009 found that answering questions before studying improved later test performance on the tested material, even when initial answers were wrong. The benefit depends on the instruction that follows, so explain the answers during the session instead of leaving the pre-test unreviewed.
What is a good normalized gain for a training course?
There is no universal target. Hake’s bands of 0.3 and 0.7 come from physics courses and concept tests, so they do not transfer directly to workplace training. Compare against your own earlier runs of the same course, using the same blueprint and similar learners.
Can pre and post scores prove the training caused the improvement?
Not on their own. Scores show that knowledge changed between two points in time. Job experience, motivation and practice can also contribute. A comparison group that did not attend, or a delayed follow-up, gives stronger evidence about the course itself.
