The course is over. What changed for the learners?
“Most learners completed the course” and “learners improved” are different statements. Teachers and training leads need a report that explains who was measured, what changed and what remains unknown.
This worked result uses eight synthetic learners. The records are constructed examples, not actual training outcomes or proof of course effectiveness.
The result: a learning outcomes summary
Course C01; assessment date: October 20, 2026. Scores are out of 100.
| Measure | Result | Meaning |
|---|---|---|
| Learners | 8 | Distinct learner identifiers |
| Marked complete | 7, or 87.5% | Completion status, not an assessment pass rate |
| Both pre- and post-assessment available | 6 | Only these learners enter the paired comparison |
| Paired pre-assessment mean | 52.5 points | Total pre-score of 315 across six learners |
| Paired post-assessment mean | 74.2 points | Total post-score of 445; displayed to one decimal |
| Paired mean change | +21.7 points | Compare the same six learners; not +21.7% |
Supported conclusion: the six learners with both assessments increased from a mean of 52.5 to approximately 74.2 points.
Unsupported conclusions: every learner gained 21.7 points, the course caused the change, or everyone completing it met a job-readiness standard. There is no control group or evidence of longer-term performance.
Keep missing evidence visible
- L07 has no post-assessment and has not completed the course. Ask whether a follow-up assessment or learning support is appropriate; do not replace the missing score with zero.
- L08 has a post-assessment and is marked complete, but has no pre-assessment. Report completion, not an invented individual improvement.
- Both remain in the overall learner count, but neither enters the paired means.
This avoids comparing different populations before and after the course, or quietly deleting incomplete records.
Check the sample and method
Download the eight synthetic learner records. The sample contains identifiers, course, date, scores and completion only—no names or contact details.
The rules are:
- Count distinct
learner_idvalues. - Divide learners with
completed = trueby all learners for completion rate. - Calculate individual changes only when both
pre_scoreandpost_scoreexist. - Average those six changes. Missing is not zero.
- Use unrounded values in calculation and round only for presentation. Keep points distinct from percentage changes.
What should a teacher do next?
Follow up with learners needing another assessment or support, then investigate which topics need reinforcement. This sample has no question-level scores, so it cannot identify weak topics from total scores alone. That requires de-identified question scores and a topic mapping.
Choose the audience carefully. Learners, instructors and training sponsors may need different versions. Share individual scores only with authorized recipients. Prefer aggregate external summaries, while remembering that small groups can still identify individuals.
Make a version for your course
Start with a de-identified table and reuse this brief:
Create a learning outcomes report. State learner count, completion and paired-assessment coverage. Compare each learner with themselves, never treating missing scores as zero. Show means, point changes and sample limitations. Separate observed change from causality. Do not invent weak topics without question-level evidence. Suggest missing inputs and useful teaching follow-ups.
The entry below prefills the brief only. Upload authorized data, check maximum scores and the meaning of completion, then confirm when to start. This example does not provide automated grading, certificates or judgments about real learners.
inspiration-course-learning-report