1. Fair comparison
Evaluate a plant-growth comparison in which fertiliser A plants receive 200 mL water daily and fertiliser B plants receive 100 mL.
Claim
The comparison cannot isolate the effect of fertiliser.
Evidence
Water volume differs systematically between the treatment groups: 200 mL versus 100 mL per day.
Reasoning
Water availability can affect growth, so it is an alternative explanation for any height difference.
Repair
Use the same justified water volume, comparable starting plants and replicated treatment groups.
Self-review checklist
- Identifies water as a confound.
- Connects the confound to growth.
- Proposes a specific matched comparison.
2. Accuracy and precision
Readings of a reference mass of 10.00 g are 10.40, 10.41 and 10.40 g. Evaluate the readings.
Claim
The readings are closely grouped but high relative to the reference.
Evidence
Their range is 0.01 g, while all are about 0.40 g above 10.00 g.
Reasoning
The small spread supports precision for these repeats; the common offset indicates poor accuracy relative to this reference.
Repair
Check zero and calibration with suitable known masses before repeating.
Self-review checklist
- Uses both spread and reference.
- Distinguishes precision from accuracy.
- Targets calibration rather than extra digits.
3. Relative uncertainty
A length is 25.0 ± 0.5 cm. Calculate and interpret the relative uncertainty.
Calculation
Relative uncertainty = 0.5 ÷ 25.0 × 100 = 2%.
Meaning
The stated absolute uncertainty is 2% of the measured length.
Limit
This calculation expresses the given uncertainty; it does not establish how that uncertainty was estimated.
Self-review checklist
- Correct ratio.
- Percentage and units treated consistently.
- Avoids claiming zero uncertainty.
4. Reading a slope
Position changes from 4 m at 2 s to 16 m at 8 s. Determine the average velocity.
Calculation
Change in position = 16 − 4 = 12 m; change in time = 8 − 2 = 6 s.
Result
Average velocity = 12 ÷ 6 = 2 m/s.
Limit
This interval average does not establish a constant instantaneous velocity between the endpoints.
Self-review checklist
- Uses changes in both variables.
- Reports m/s.
- Distinguishes interval average from instantaneous velocity.
5. Representative sampling
A field survey places every quadrat near a shaded fence. Evaluate a claim about the entire varied field.
Claim
The survey has limited support for a whole-field estimate.
Evidence
All plots are in one shaded location rather than spread across the field’s conditions.
Reasoning
Plants near the fence may differ from plants in open areas; more plots in the same area preserve this selection bias.
Repair
Use justified random or stratified sites across the field and replicate within that design.
Self-review checklist
- Identifies selection bias.
- Explains why counts alone do not repair it.
- Gives a suitable sampling improvement.
6. Capture–recapture
A first sample marks 30 animals. A second sample catches 40, of which 6 are marked. Estimate population size and state two assumptions.
Calculation
Estimated population = 30 × 40 ÷ 6 = 200 animals.
Assumptions
The population remains closed during the interval and marked animals mix with the population.
Limit
Marks must also remain identifiable and capture probabilities must be comparable. The result is a model-based estimate, not an exact census.
Self-review checklist
- Correct calculation.
- At least two relevant assumptions.
- Acknowledges conditional estimate.
7. Association and causation
In a student survey, more exercise is associated with lower reported stress. Evaluate the causal claim.
Claim
The survey shows an association but does not alone establish that exercise caused lower stress.
Evidence
Exercise was observed rather than assigned, and stress was self-reported.
Reasoning
Other factors may influence both; the causal direction might also differ.
Next step
Use a suitable ethically approved design, measure potential confounders and examine independent evidence.
Self-review checklist
- Acknowledges observed association.
- Explains a causal limitation.
- Suggests relevant further evidence.
8. Evaluating an outlier
One reaction-time reading is 14 s while the other four are 8–9 s. Explain how to handle it.
Action
Retain and flag the 14 s result while checking raw records and procedural notes.
Reasoning
It may reflect real variation or a specific procedural error; disagreement with the prediction alone is not a reason for deletion.
Reporting
If a justified exclusion is made, explain the criterion and compare conclusions with and without the point.
Self-review checklist
- Does not delete by preference.
- Checks an identifiable cause.
- Reports any exclusion transparently.
9. Model limitation
A classroom epidemic simulation assumes every person is equally likely to contact every other person. Evaluate this assumption.
Claim
The model simplifies real contact structure.
Evidence
Real contacts often cluster in households, schools and workplaces.
Reasoning
Unequal contact networks can change spread compared with a well-mixed model.
Scope
The model can explore conditional scenarios; validate its intended use against appropriate observations or use a network model.
Self-review checklist
- Identifies the mixing assumption.
- Connects structure to spread.
- Explains useful scope without claiming exact prediction.
10. Specific evaluation
An experiment times a fast-moving cart using a hand-operated stopwatch. Propose one justified improvement.
Limitation
Variable human reaction delay can add substantial uncertainty to a short timing interval.
Improvement
Use appropriate light gates or another automated timing sensor at fixed positions.
Reasoning
Automated triggering reduces the contribution from the operator’s start and stop reactions.
Limit
Check sensor calibration and placement; automation does not eliminate every uncertainty.
Self-review checklist
- Names a specific mechanism.
- Gives a feasible targeted improvement.
- Explains the effect and remaining limitations.
11. Evidence-based conclusion
Mean cooling time is 72 s in covered cups and 50 s in uncovered cups across repeated matched trials. Write a qualified conclusion.
Claim
Under the tested conditions, covered cups took longer to cool to the target temperature.
Evidence
The means differed by 22 s: 72 s covered versus 50 s uncovered.
Reasoning
This is consistent with the cover affecting heat transfer, but the means alone do not identify every transfer mechanism.
Limit
Evaluate spread, sample size, controls and uncertainty before making stronger claims or generalising to other cup designs.
Self-review checklist
- Uses the measured difference.
- Limits the claim to tested conditions.
- Includes evidence needed for a stronger comparison.
12. Secondary-data provenance
A report uses an online graph with no date, method or source citation. Evaluate its use as evidence.
Claim
The graph’s suitability cannot be evaluated adequately from the supplied information.
Evidence
The original dataset, collection date and method are missing.
Reasoning
These details are needed to check relevance, sample selection, measurement definitions and possible bias.
Repair
Find and cite the original source, document provenance and evaluate whether its population and method fit the question.
Self-review checklist
- Identifies missing provenance.
- Explains why it matters.
- Proposes traceable source evaluation.