The useful starting point
Use a small sample to choose a specific review task. Keep the result attached to its questions, conditions and limits.
Describe the attempt, not the whole learner
A fictional learner answers eight of ten questions correctly on a first attempt. The items cover several topics but were not randomly selected from an official exam. The case gives no evidence that their difficulty or coverage matches that exam. Record the result as eight correct answers in this particular set.
One answer changes the displayed percentage by ten percentage points because the denominator is ten. That arithmetic makes small samples look volatile. It does not make the result useless; it means the wording of the conclusion must stay close to what was actually observed.
Make a claim-and-evidence card
| Proposed claim | Supported here? | Reason |
|---|---|---|
| Eight of ten answers were correct | Yes | Directly supplied result |
| The learner will pass the official exam | No | No validated prediction is supplied |
| Every topic is understood equally | No | Coverage and reasoning differ |
| Two explanations deserve review | Useful starting action | The errors identify concrete items |
Keep guessed-correct answers available for review too. Correctness on one item and a clear explanation of the underlying distinction are different observations.
Change one condition and reassess the claim
Suppose the learner repeats the same set and answers all ten correctly. That shows improved performance on a repeated set. It does not establish how they will handle unfamiliar wording or different concepts. Record the repeat rather than presenting it as a new independent sample.
Alternatively, suppose the next set is entirely about an unfamiliar topic. A lower score could reflect different content rather than a decline in ability. Comparing numbers requires context: set identity, topic mix, first or repeated attempt, note use and incomplete items. Do not interpret every change as learning gain or loss.
Turn the result into a testable study task
Choose one answer you could not explain and locate its supporting source. Write a corrected explanation, then apply it to a changed fictional example. The next useful observation is whether you can explain the distinction in that new context, not whether you can reproduce the old letter.
Use the sample as a starting point within a broader coverage plan. Mark unattempted topics unknown. Keep official application and exam decisions separate from product practice. This approach preserves the useful signal in a small exercise without attaching an unsupported pass probability, readiness label or complete-coverage claim to it.
Check your reasoning.
A learner improves from 8 / 10 to 10 / 10 on the identical set immediately after reading the answers. What should the record say?
Open the suggested reasoning
It should say the second result was a repeated attempt after review. It shows performance on that set, not independent evidence of official exam readiness.
An original learning exercise from NCSO Prep, not an official exam question.
Sources and scope
Source review: . The examples and study methods are NCSO Prep editorial material. Official requirements and workplace decisions must be checked against the applicable current source.
- Statistics Canada: Statistical Bias
Background on how data collection affects interpretation. This article provides no statistical model of exam performance.