1. A student tests whether light intensity changes the number of oxygen bubbles from pondweed. Which variable should be changed deliberately?

Stage 1: choose the claim

1. Light intensity
2. Number of bubbles
3. Water temperature

Stage 2: choose the reason

1. A fair test must change every variable together.
2. The independent variable is the factor deliberately varied to test the question.
3. The independent variable is whatever is measured at the end.

Claim A; reason B

Bubble count is the dependent measure. Temperature should be controlled; changing it with light would confound the comparison. Bubble count is only a proxy for oxygen production.

Repair focus: Questions, variables and controls.

2. Two fertilisers are compared using plants of different species. What is the strongest design repair?

Stage 1: choose the claim

1. Use taller plants only in the second group.
2. Use the same species in both treatment groups.
3. Measure more leaves without changing the groups.

Stage 2: choose the reason

1. Larger sample sizes automatically remove all systematic differences.
2. The treatment with taller starting plants is always more effective.
3. Species can affect growth independently of fertiliser, so it confounds the treatment comparison.

Claim B; reason C

Match or randomise starting conditions and use replication. Extra measurements cannot isolate fertiliser if species differs systematically between groups.

Repair focus: Questions, variables and controls.

3. Which question can be investigated most directly with the stated ruler and timer?

Stage 1: choose the claim

1. Is a toy car more beautiful than a real car?
2. Why is every car in the world different?
3. How does ramp height affect a toy car’s travel time over 1 metre?

Stage 2: choose the reason

1. Ramp height and travel time are operationally defined, measurable variables.
2. A question is scientific only if the answer is already known.
3. Any question containing the word why can be answered by one measurement.

Claim C; reason A

An investigable question identifies a relationship that available evidence can test. Beauty needs a different operational definition and the worldwide question is too broad for this method.

Repair focus: Questions, variables and controls.

4. A hypothesis says that increasing salt concentration will decrease seed germination after 7 days. Which result challenges it?

Stage 1: choose the claim

1. Higher concentrations consistently produce greater germination under matched conditions.
2. One high-salt dish contains no germinated seeds.
3. All groups are counted after different times.

Stage 2: choose the reason

1. Changing the counting time makes all comparisons stronger.
2. A testable prediction can be challenged by results that run against its predicted relationship.
3. A hypothesis becomes permanently true once one trial supports it.

Claim A; reason B

A contrary repeatable pattern challenges the prediction. One supporting dish does not prove it, and unequal observation times introduce a confound.

Repair focus: Questions, variables and controls.

5. In a detergent trial, equal stained cloths are washed with water alone or water plus detergent. What does the water-only group provide?

Stage 1: choose the claim

1. A replacement for repeated trials.
2. A comparison for the effect of adding detergent.
3. Proof that all detergent is harmless.

Stage 2: choose the reason

1. Control groups remove the need to measure an outcome.
2. Any untreated group proves a treatment causes no environmental harm.
3. A control treatment helps separate the treatment effect from the effect of washing itself.

Claim B; reason C

Compare stain removal against washing without detergent while controlling temperature and agitation. This experiment measures cleaning, not environmental safety.

Repair focus: Questions, variables and controls.

6. A student records “strong wind” but does not define it. What improves the method?

Stage 1: choose the claim

1. Replace the measurement with “very strong wind”.
2. Remove all weather observations.
3. Define wind speed in metres per second and specify how it is measured.

Stage 2: choose the reason

1. An operational definition makes a variable measurable and the method reproducible.
2. Adjectives are more precise than units.
3. Quantitative definitions prevent all equipment error.

Claim C; reason A

A stated instrument, units, position and sampling time make the measurement reproducible. An operational definition does not remove measurement uncertainty.

Repair focus: Questions, variables and controls.

7. A cooling trial uses cups with different volumes and materials. What can be concluded about material alone?

Stage 1: choose the claim

1. Its separate effect cannot be isolated from this comparison.
2. Material definitely causes the entire difference.
3. Volume cannot influence cooling.

Stage 2: choose the reason

1. Controlled variables are the quantities calculated after the experiment.
2. Volume changes alongside material, creating an alternative explanation for the observed difference.
3. Any difference between two cups identifies one cause.

Claim A; reason B

Change one relevant factor while controlling the others, or use a design that separately tests both. The present design confounds material and volume.

Repair focus: Questions, variables and controls.

8. A team plans a reaction producing gas in a sealed rigid glass container. What is the appropriate next step?

Stage 1: choose the claim

1. Ignore pressure if the reaction is small.
2. Revise the method with the teacher before conducting the practical.
3. Proceed because a written risk table guarantees safety.

Stage 2: choose the reason

1. Risk assessment is only paperwork after an experiment.
2. Small quantities can never create hazardous conditions.
3. Gas production can raise pressure; a risk assessment must inform the actual method and supervision.

Claim B; reason C

Identify hazards, assess exposure and use approved controls before starting. This workbench does not approve practical activities or replace school supervision.

Repair focus: Questions, variables and controls.

9. Repeated measurements are 9.8, 9.9 and 9.8 cm; the reference length is 12.0 cm. How should they be described?

Stage 1: choose the claim

1. Closely grouped but inaccurate relative to the reference.
2. Accurate because they agree with each other.
3. Both accurate and unbiased by definition.

Stage 2: choose the reason

1. Taking an average always removes calibration bias.
2. Precision describes closeness among measurements; accuracy concerns closeness to a suitable reference.
3. Repeatability and accuracy mean exactly the same thing.

Claim A; reason B

A consistent offset can make repeatable readings inaccurate. Check calibration or the measurement procedure; averaging cannot remove a shared bias.

Repair focus: Reliability, validity, accuracy and precision.

10. Three students repeat a boiling-point measurement with the same faulty thermometer. What does repetition achieve?

Stage 1: choose the claim

1. It changes the independent variable.
2. It can reveal spread but does not remove the thermometer’s common offset.
3. It guarantees the accepted boiling point.

Stage 2: choose the reason

1. Systematic error changes sign randomly every trial.
2. More people always produce a perfectly accurate value.
3. Repeated measurements address random variation, while a shared systematic offset persists.

Claim B; reason C

Replication is useful evidence about consistency. An independent calibrated instrument is needed to investigate the common offset.

Repair focus: Reliability, validity, accuracy and precision.

11. A test of caffeine and reaction time compares morning caffeine users with evening non-users. What is the main validity problem?

Stage 1: choose the claim

1. The stopwatch displays too many decimal places.
2. The sample contains more than one person.
3. Time of day differs with caffeine treatment.

Stage 2: choose the reason

1. A confounding variable can affect the outcome independently of the intended treatment.
2. Validity is only the number of repeated trials.
3. A digital instrument guarantees causal validity.

Claim C; reason A

Time of day provides an alternative explanation. Match timing and consider other relevant controls before attributing a difference to caffeine.

Repair focus: Reliability, validity, accuracy and precision.

12. A stopwatch gives readings to 0.01 s, but a human starts and stops it. Which statement is justified?

Stage 1: choose the claim

1. Reaction delay may exceed the display resolution.
2. The true duration is known within exactly 0.01 s.
3. Digital readings have no uncertainty.

Stage 2: choose the reason

1. Human timing affects only the units, not the result.
2. Display resolution is only one contribution to total measurement uncertainty.
3. The final displayed digit guarantees the full measurement accuracy.

Claim A; reason B

Include reaction delay and the timing procedure when discussing uncertainty. A fine display does not make human-triggered timing equally fine.

Repair focus: Reliability, validity, accuracy and precision.

13. Repeats of an enzyme trial vary widely. What is a useful first response?

Stage 1: choose the claim

1. Report the average as exact.
2. Check procedural consistency and collect more independent repeats.
3. Delete all readings except the expected value.

Stage 2: choose the reason

1. A result is reliable whenever it matches a textbook.
2. Unwanted values should be removed without explanation.
3. Replication and consistent procedures help estimate and investigate random variation.

Claim B; reason C

Record all observations and inspect temperature, volumes and timing. Deleting inconvenient data conceals variation rather than improving reliability.

Repair focus: Reliability, validity, accuracy and precision.

14. A balance reads 0.40 g when empty and is not tared. What improvement targets the identified problem?

Stage 1: choose the claim

1. Record extra decimal places.
2. Use a bigger spreadsheet.
3. Tare it and check it against a known mass.

Stage 2: choose the reason

1. A zero offset is systematic and should be corrected at the measurement stage.
2. Every error can be corrected simply by increasing trial count.
3. Precision always improves when more digits are written.

Claim C; reason A

Correct the zero and verify calibration. More trials or written digits cannot remove an unchanged 0.40 g offset.

Repair focus: Reliability, validity, accuracy and precision.

15. A method measures plant height to investigate “overall plant health”. What must be acknowledged?

Stage 1: choose the claim

1. Height is a limited proxy and may not capture all aspects of health.
2. Height proves every aspect of health.
3. Only qualitative observations can be valid.

Stage 2: choose the reason

1. A proxy has no scientific use under any circumstances.
2. Construct validity depends on whether the chosen measure represents the question’s intended property.
3. Any easy measurement is automatically a complete measure of a complex property.

Claim A; reason B

Define what health means in this investigation. Height can be useful, but additional measures or a narrower question may be needed.

Repair focus: Reliability, validity, accuracy and precision.

16. Independent groups obtain similar results using a described method and calibrated equipment. What is supported?

Stage 1: choose the claim

1. Every possible source of bias is absent.
2. Reproducibility is strengthened, without proving the explanation permanently true.
3. The hypothesis can never be revised.

Stage 2: choose the reason

1. Reproduction eliminates the possibility of future contradictory evidence.
2. Scientific conclusions require unanimous agreement rather than evidence.
3. Agreement across independent replications is stronger evidence of consistency than repeating one biased setup.

Claim B; reason C

Independent replication strengthens confidence. Shared design limitations, sampling issues or later evidence can still change the interpretation.

Repair focus: Reliability, validity, accuracy and precision.

17. A ruler’s smallest marked division is 1 mm. Under a stated half-division reading convention, what reading uncertainty is used?

Stage 1: choose the claim

1. ±0.5 mm
2. ±5 mm
3. Zero uncertainty

Stage 2: choose the reason

1. A printed scale makes a measurement exact.
2. Half the smallest division is 0.5 mm under that reading convention.
3. All rulers have exactly ±1 cm uncertainty.

Claim A; reason B

State the convention: ±0.5 mm addresses scale reading only. Alignment, parallax and the object’s boundary may add uncertainty.

Repair focus: Measurements and uncertainty.

18. A length is reported as 50.0 ± 1.0 cm. What is its relative uncertainty?

Stage 1: choose the claim

1. 0.02%
2. 2%
3. 50%

Stage 2: choose the reason

1. Relative uncertainty is the uncertainty added to the value.
2. The number of decimal places is the percentage uncertainty.
3. Relative uncertainty is the absolute uncertainty divided by the measured value, multiplied by 100%.

Claim B; reason C

1.0 ÷ 50.0 × 100 = 2%. Keep the absolute uncertainty and measurement in the same units before taking the ratio.

Repair focus: Measurements and uncertainty.

19. Readings are 19, 20, 21 and 20 s. What is the mean?

Stage 1: choose the claim

1. 21 s
2. 80 s
3. 20 s

Stage 2: choose the reason

1. The sum is divided by the number of readings.
2. The largest value is the mean.
3. The mean is the total without division.

Claim C; reason A

The sum is 80 s and there are four readings: 80 ÷ 4 = 20 s. The mean summarises the data but does not remove uncertainty.

Repair focus: Measurements and uncertainty.

20. A reported value is 8.0 ± 0.3 g. Which range expresses that stated interval?

Stage 1: choose the claim

1. 7.7 to 8.3 g
2. 8.0 to 8.3 g only
3. 5.0 to 11.0 g

Stage 2: choose the reason

1. Decimal uncertainty is multiplied by ten before use.
2. The uncertainty is applied on both sides of the reported value.
3. The ± symbol means only an upward change.

Claim A; reason B

Subtract and add 0.3 g. Do not interpret the interval as a confidence interval unless the method used to calculate it justifies that interpretation.

Repair focus: Measurements and uncertainty.

21. A single digital reading is 24.6 °C. Can trial-to-trial variability be calculated from that one reading?

Stage 1: choose the claim

1. Yes; it is equal to the temperature itself.
2. No; repeated observations are needed to estimate that variability.
3. Yes; it is exactly 0.1 °C.

Stage 2: choose the reason

1. A single digit contains the full distribution of repeated measurements.
2. The temperature value itself is always its uncertainty.
3. Display resolution and variation among repeated observations describe different information.

Claim B; reason C

The display suggests a resolution, but not the spread of repeated trials. Document resolution separately and gather repeats where appropriate.

Repair focus: Measurements and uncertainty.

22. A measured difference is 0.2 g while the stated measurement uncertainties are each ±0.5 g. What is the cautious interpretation?

Stage 1: choose the claim

1. The difference proves a large treatment effect.
2. Uncertainty cancels automatically in all differences.
3. The small difference may be unresolved by these measurements.

Stage 2: choose the reason

1. Measurement uncertainty must be considered when deciding whether an apparent difference is meaningful.
2. Any non-zero difference establishes a treatment effect.
3. More displayed digits remove uncertainty from a comparison.

Claim C; reason A

The measured difference is small relative to the stated uncertainties. A stronger conclusion needs an appropriate uncertainty analysis and improved evidence.

Repair focus: Measurements and uncertainty.

23. A student reports 3.141592 g using a balance displaying only 0.01 g. What is the issue?

Stage 1: choose the claim

1. The reported digits imply unsupported measurement precision.
2. More digits always make the result more accurate.
3. The mass must be negative.

Stage 2: choose the reason

1. Every scientific result must have six decimal places.
2. Reported precision should reflect the measurement and any justified processing.
3. A calculator creates additional experimental information.

Claim A; reason B

Do not invent detail beyond the equipment and method. Calculation digits may be retained temporarily, but the final result must use justified rounding.

Repair focus: Measurements and uncertainty.

24. A measuring cylinder is viewed from above rather than at eye level. What is the best correction?

Stage 1: choose the claim

1. Change millilitres to litres without changing the view.
2. Read the appropriate meniscus at eye level.
3. Take the same off-angle reading more times.

Stage 2: choose the reason

1. Changing units removes experimental error.
2. A biased method becomes unbiased after enough repeats.
3. Viewing geometry can introduce parallax and a biased scale reading.

Claim B; reason C

Use eye-level reading and the correct meniscus convention for the liquid. Repeating the same biased view does not address the cause.

Repair focus: Measurements and uncertainty.

25. A trial varies temperature and measures reaction time. What belongs on the horizontal axis?

Stage 1: choose the claim

1. Temperature with its units
2. Reaction time with no units
3. The trial number instead of temperature

Stage 2: choose the reason

1. A graph never needs measurement units.
2. The independent variable conventionally goes on the horizontal axis for this relationship.
3. The largest numeric values always belong on the horizontal axis.

Claim A; reason B

Plot temperature against reaction time, label both quantities and units, and choose a scale that displays the data clearly.

Repair focus: Graphs and processing data.

26. A dataset compares favourite fruit categories in a class. Which display is generally appropriate?

Stage 1: choose the claim

1. A velocity–time graph.
2. A bar chart with separate categories.
3. A connected line implying a continuous fruit scale.

Stage 2: choose the reason

1. All scientific data must be shown as continuous lines.
2. Connecting categories establishes a causal relationship.
3. Fruit categories are discrete categories rather than ordered measurements of a continuous quantity.

Claim B; reason C

Separate bars preserve the categorical structure. A frequency table also works; explain what was counted and the sample size.

Repair focus: Graphs and processing data.

27. A scatter plot rises overall but its points do not lie on one line. What should a trend description do?

Stage 1: choose the claim

1. Claim every point increases by an identical amount.
2. Hide points away from the trend.
3. Describe the positive association and acknowledge scatter.

Stage 2: choose the reason

1. A trend summarises a relationship without pretending the data have no variation.
2. Any scatter means there can be no overall association.
3. A fitted line makes all measurements exact.

Claim C; reason A

Use the pattern and spread together. If fitting a line, justify the model and avoid claiming exact proportionality without evidence.

Repair focus: Graphs and processing data.

28. A graph’s vertical axis starts at 98 rather than 0. What should a reader check?

Stage 1: choose the claim

1. Whether the scale exaggerates the visual size of the difference.
2. Whether all non-zero axes are automatically fraudulent.
3. Whether units are no longer needed.

Stage 2: choose the reason

1. Axis choice cannot affect a reader’s interpretation.
2. A truncated axis can make small differences appear visually large; interpretation needs the numeric scale.
3. The size of a drawn bar always equals the absolute effect size.

Claim A; reason B

A non-zero origin can be useful for detailed comparisons but must be clear. Read actual values and consider an appropriate alternative display.

Repair focus: Graphs and processing data.

29. A student uses data from 10–30 °C to predict a result at 80 °C. How is this prediction classified?

Stage 1: choose the claim

1. A directly measured value.
2. Extrapolation with a substantial limitation.
3. Interpolation guaranteed by the data.

Stage 2: choose the reason

1. Any straight fitted line applies at every possible temperature.
2. Extrapolation means predicting between measured values.
3. 80 °C is outside the measured range, where the relationship may change.

Claim B; reason C

Interpolation stays within the observed range; extrapolation extends beyond it. Temperature-dependent mechanisms may change, so test rather than assume the extension.

Repair focus: Graphs and processing data.

30. Position increases from 2 m to 8 m between 1 s and 4 s. What is the average velocity over that interval?

Stage 1: choose the claim

1. 6 m/s
2. 8 m/s
3. 2 m/s

Stage 2: choose the reason

1. The slope is change in position divided by change in time: 6 m ÷ 3 s.
2. Velocity is always the final position alone.
3. The slope is the position change without a time interval.

Claim C; reason A

Use differences on both axes. This is average velocity over the interval, not necessarily the instantaneous velocity at every time.

Repair focus: Graphs and processing data.

31. One result is far from the other points. What is the defensible first action?

Stage 1: choose the claim

1. Check records and method, retaining the value unless an exclusion is justified.
2. Delete it because it does not support the hypothesis.
3. Change it to lie on the fitted line.

Stage 2: choose the reason

1. Scientific reporting requires every point to support the prediction.
2. An unusual value may reflect genuine variation or a documented error and needs investigation.
3. Outliers are always false measurements.

Claim A; reason B

Inspect raw records and procedural notes. If excluding a result, state a principled reason and show how exclusion affects the conclusion.

Repair focus: Graphs and processing data.

32. A line passes through (0, 3) and (2, 7). Is the displayed relationship directly proportional?

Stage 1: choose the claim

1. Yes; any upward trend is directly proportional.
2. No; its non-zero intercept is inconsistent with direct proportionality.
3. Yes; every straight line is directly proportional.

Stage 2: choose the reason

1. A positive slope alone establishes a zero intercept.
2. Direct proportion means the two variables have the same units.
3. Direct proportionality has a constant ratio and passes through the origin in the relevant model.

Claim B; reason C

The slope is 2 and the intercept is 3. A straight-line relationship can be linear without being a direct proportion.

Repair focus: Graphs and processing data.

33. Students sample plants only beside a school path and generalise to the whole reserve. What is the main limitation?

Stage 1: choose the claim

1. The accessible sites may not represent the whole reserve.
2. Every sample must include every individual.
3. A large count beside the path guarantees representation.

Stage 2: choose the reason

1. A convenience sample is always a random sample.
2. Location-based selection can systematically overrepresent certain conditions or species.
3. Sample size removes all selection bias.

Claim A; reason B

Use a sampling design that covers the intended habitat, such as random or stratified sites, and state the sampled area and limits.

Repair focus: Sampling and populations.

34. A reserve has woodland and grassland. Why might stratified sampling be useful?

Stage 1: choose the claim

1. It eliminates all population uncertainty.
2. It ensures both habitat types are represented in a planned way.
3. It means choosing only the easiest habitat.

Stage 2: choose the reason

1. A stratum is simply whichever site is nearest.
2. A representative sample produces an exact census.
3. The population contains distinct strata that may differ in the property being measured.

Claim B; reason C

Select sites within each habitat using a justified approach and use appropriate weighting if estimating an overall population property.

Repair focus: Sampling and populations.

35. A capture–recapture trial marks 40 animals, then captures 50 including 10 marked animals. Under the model assumptions, what is the estimate?

Stage 1: choose the claim

1. 90 animals
2. 500 animals
3. 200 animals

Stage 2: choose the reason

1. The estimate is first marked × second caught ÷ marked recaptured: 40 × 50 ÷ 10.
2. The population is the sum of both sample sizes.
3. Recaptured animals are multiplied instead of used as a sample proportion.

Claim C; reason A

The estimate is 200, conditional on a closed population, retained marks, mixing and comparable capture probability. It is not a census.

Repair focus: Sampling and populations.

36. Between capture and recapture, many marked animals leave the site. What happens to the model’s assumptions?

Stage 1: choose the claim

1. The closed-population assumption is violated.
2. The estimate becomes automatically exact.
3. Marking no longer needs to be recorded.

Stage 2: choose the reason

1. Population closure means the animals never move within the habitat.
2. Immigration, emigration, births or deaths can change the population and marked proportion.
3. Capture–recapture requires movement in and out to maximise accuracy.

Claim A; reason B

Closure concerns entry, exit, births and deaths during the sampling interval. Movement and mixing within the study population are a different assumption.

Repair focus: Sampling and populations.

37. A quadrat study uses 20 plots along one shaded fence. Can it estimate plant cover throughout a varied field confidently?

Stage 1: choose the claim

1. Yes, because quadrats cannot produce bias.
2. Only with strong limitations because the site selection is narrow.
3. Yes, because 20 is always enough.

Stage 2: choose the reason

1. Replication of the same biased location removes selection bias.
2. Sampling tools determine representation independently of site choice.
3. Many plots from one unrepresentative location still share a selection bias.

Claim B; reason C

Choose positions across the intended population of locations. Replicate within a suitable sampling design and discuss remaining spatial variation.

Repair focus: Sampling and populations.

38. Which random-site method avoids choosing convenient positions after looking at the field?

Stage 1: choose the claim

1. Pick whichever patch looks typical.
2. Place every quadrat where the most plants grow.
3. Generate coordinates in advance and follow a stated replacement rule for inaccessible sites.

Stage 2: choose the reason

1. A prespecified selection rule reduces subjective selection based on the observed outcome.
2. Random sampling means choosing with no written rule.
3. The most populated patch represents every patch.

Claim C; reason A

Document coordinates, boundaries and the rule for inaccessible locations before inspecting outcomes. Random selection still needs enough coverage and safe access.

Repair focus: Sampling and populations.

39. A survey of lunch preferences includes only volunteers from one club. What conclusion is safest?

Stage 1: choose the claim

1. Describe these volunteers and acknowledge limited generalisation.
2. Claim it represents every student in the state.
3. Treat volunteering as proof of random selection.

Stage 2: choose the reason

1. Survey conclusions never depend on recruitment.
2. Self-selection and the club’s membership can make respondents differ from the target population.
3. People who volunteer automatically form a representative sample.

Claim A; reason B

Report who responded, how they were recruited and the sample size. Generalising beyond them requires a suitable design and supporting evidence.

Repair focus: Sampling and populations.

40. Increasing a well-designed independent sample from 10 to 100 generally does what?

Stage 1: choose the claim

1. It always changes the population mean.
2. It can reduce sampling variability but does not guarantee removal of bias.
3. It removes every measurement and selection error.

Stage 2: choose the reason

1. A bigger sample automatically repairs a biased recruitment method.
2. Sample size and the population’s actual value are the same quantity.
3. More independent information often improves precision, while systematic selection or measurement problems can persist.

Claim B; reason C

Sample size helps when observations add relevant independent information. Representation and measurement quality still need separate attention.

Repair focus: Sampling and populations.

41. Students who sleep longer obtain higher test scores in a survey. What follows from the association alone?

Stage 1: choose the claim

1. It does not establish that extra sleep caused the score difference.
2. Sleep definitely caused every difference.
3. No relationship was observed.

Stage 2: choose the reason

1. Only manipulated experiments can show any association.
2. Other factors may affect both sleep and scores in an observational study.
3. Correlation always identifies the causal direction.

Claim A; reason B

An association is useful evidence, but causation needs a stronger design and consideration of confounders and alternative explanations.

Repair focus: Evidence and scientific claims.

42. A product’s advertisement cites one customer saying it worked. How should this evidence be treated?

Stage 1: choose the claim

1. As proof that no further study is needed.
2. As an anecdote that cannot establish effectiveness for a population.
3. As a controlled population trial.

Stage 2: choose the reason

1. Personal confidence replaces the need for a comparison group.
2. A large advertisement makes its evidence stronger.
3. One uncontrolled experience cannot separate the product effect from other changes or selection effects.

Claim B; reason C

Look for relevant transparent studies, comparison groups, outcomes, sampling and uncertainty. Personal experience can suggest a question, not settle population effectiveness.

Repair focus: Evidence and scientific claims.

43. A paper reports only trials that support its prediction and omits contrary trials. What is the concern?

Stage 1: choose the claim

1. This is required for a concise scientific report.
2. Contrary evidence is automatically an equipment error.
3. Selective reporting can distort the apparent evidence.

Stage 2: choose the reason

1. A conclusion should account for relevant observations rather than selecting by whether they support it.
2. Supporting a hypothesis determines which data are valid.
3. Publication guarantees that every trial was reported.

Claim C; reason A

Record methods, selection criteria and relevant results transparently. Omitting contradictory observations weakens evaluation of the claim.

Repair focus: Evidence and scientific claims.

44. Two groups differ by 1 unit but the sample sizes and variability are missing. What should be said?

Stage 1: choose the claim

1. The difference is reported, but its strength cannot be evaluated from these details alone.
2. The difference proves a reliable treatment effect.
3. The difference must be zero.

Stage 2: choose the reason

1. Unreported variability can be assumed to be zero.
2. Effect interpretation requires information about design, sample size and variation, not just two means.
3. Two different means always establish a robust causal effect.

Claim A; reason B

Describe what the means show and request the missing information. Do not invent a significance result or erase the observed difference.

Repair focus: Evidence and scientific claims.

45. A reliable review conflicts with a new small preliminary study. What is an appropriate response?

Stage 1: choose the claim

1. Ignore all new results permanently.
2. Compare methods and the body of evidence before revising confidence.
3. Always accept the newest paper immediately.

Stage 2: choose the reason

1. Newer evidence is automatically more accurate.
2. Established explanations cannot change.
3. Scientific confidence depends on evidence quality, replication and context rather than publication date alone.

Claim B; reason C

Evaluate the preliminary study carefully and look for independent replication. Neither novelty nor familiarity alone decides the conclusion.

Repair focus: Evidence and scientific claims.

46. A source sells the product it evaluates. How should this be handled?

Stage 1: choose the claim

1. Reject every numerical result without reading it.
2. Treat commercial interest as proof of accuracy.
3. Note the possible conflict and evaluate methods and independent evidence.

Stage 2: choose the reason

1. A conflict of interest warrants scrutiny but does not by itself establish whether a result is correct.
2. Funding source alone proves the scientific conclusion false.
3. Financial interest guarantees unbiased measurement.

Claim C; reason A

Check transparent methods, data, comparisons, limitations and independent confirmation. Declare the conflict rather than substituting it for evidence analysis.

Repair focus: Evidence and scientific claims.

47. A study finds no detectable difference with a small, noisy sample. What is the safest conclusion?

Stage 1: choose the claim

1. This study did not detect a difference under its conditions.
2. The two treatments are identical in all circumstances.
3. A difference definitely exists.

Stage 2: choose the reason

1. Small samples always prove the opposite of their observations.
2. Limited sensitivity can leave real differences unresolved; absence of detected evidence is not universal proof of no effect.
3. Every non-significant result proves exact equality.

Claim A; reason B

State the design and detection limits. Stronger claims of equivalence require an appropriate study designed to test a meaningful equivalence margin.

Repair focus: Evidence and scientific claims.

48. A class claim is supported by a fitted graph generated entirely from invented numbers. What should the report state?

Stage 1: choose the claim

1. A smooth graph proves the model is correct.
2. The data are simulated and illustrate a model rather than establish a measured real-world effect.
3. The values are field measurements.

Stage 2: choose the reason

1. A plotted value becomes empirical evidence regardless of its origin.
2. Computer output has no assumptions.
3. Synthetic data reflect assumptions used to generate them, so they cannot independently verify those assumptions.

Claim B; reason C

Label simulation inputs, rules and outputs. Use real observations to evaluate whether the model represents the intended system.

Repair focus: Evidence and scientific claims.

49. A particle diagram draws atoms as solid coloured spheres. What is a limitation?

Stage 1: choose the claim

1. The representation simplifies atomic structure and scale.
2. Atoms literally have the diagram’s colours and size.
3. A simplified model cannot be useful.

Stage 2: choose the reason

1. Every visual feature in a model is a physical measurement.
2. Models represent selected features and need explicit assumptions and limits.
3. A model must copy every detail to explain anything.

Claim A; reason B

The model may help compare arrangement or number, while colour, size and solid-sphere form are symbolic choices.

Repair focus: Models and scientific explanations.

50. A population model assumes constant birth rate. New observations show seasonal variation. What is a suitable response?

Stage 1: choose the claim

1. Treat the assumption as permanently true.
2. Evaluate and revise the assumption for the intended use.
3. Hide the seasonal observations.

Stage 2: choose the reason

1. Data must be altered to fit every model.
2. Changing a model means science cannot produce useful knowledge.
3. Models should be checked against evidence and revised when important assumptions fail.

Claim B; reason C

A constant-rate model may remain useful over a restricted interval. State that range or incorporate seasonal variation and test the revised model.

Repair focus: Models and scientific explanations.

51. A computer simulation produces the same output twice with the same seed and settings. What has been shown?

Stage 1: choose the claim

1. The real world must behave identically.
2. Every assumption has been validated.
3. The implementation is repeatable for those inputs.

Stage 2: choose the reason

1. Computational repeatability is distinct from agreement with independent observations.
2. Running a simulation repeatedly validates its real-world accuracy.
3. A random seed is a measurement from nature.

Claim C; reason A

Reproducible outputs help check code. Validation additionally requires relevant real data and examination of assumptions, parameter choices and intended scope.

Repair focus: Models and scientific explanations.

52. An epidemic model assumes a closed, well-mixed population. Where is an important limitation?

Stage 1: choose the claim

1. It may not capture migration or uneven contact networks.
2. It guarantees exact case counts in every town.
3. It does not require any population assumption.

Stage 2: choose the reason

1. Model assumptions are the final observed results.
2. The mixing and closure assumptions simplify real population structure and movement.
3. Every person always has identical contacts in reality.

Claim A; reason B

Use the model to explore conditional scenarios, not predict exact individual outcomes. Compare it with data and more suitable contact structure when needed.

Repair focus: Models and scientific explanations.

53. A straight-line model fits the measured range well. Which claim is warranted?

Stage 1: choose the claim

1. Its fitted intercept must be zero.
2. It describes the observed relationship over that range, subject to its fit and assumptions.
3. It is valid at every possible input forever.

Stage 2: choose the reason

1. A good fit within a range proves unrestricted extrapolation.
2. Every model is a direct proportion.
3. Model usefulness depends on its tested domain and evidence of fit.

Claim B; reason C

Report the fitted domain, residual variation and limits. Good agreement within a range does not establish behaviour outside it.

Repair focus: Models and scientific explanations.

54. Two models explain existing data similarly. What evidence would help distinguish them?

Stage 1: choose the claim

1. A vote on which diagram looks nicer.
2. Repeating only a prediction they share.
3. A test where they make different predictions.

Stage 2: choose the reason

1. A discriminating test compares observable predictions that differ between the models.
2. Scientific models are selected only by appearance.
3. Shared predictions always reveal which model is correct.

Claim C; reason A

Design a feasible measurement where the competing explanations predict different outcomes, with sufficient precision and appropriate controls.

Repair focus: Models and scientific explanations.

55. A scale model of a bridge is strong, but it uses a different material. What should be acknowledged?

Stage 1: choose the claim

1. Mechanical behaviour may not transfer directly to the full-sized bridge.
2. Scale changes cannot affect force or strength.
3. The model proves the real bridge is safe.

Stage 2: choose the reason

1. Similarity in appearance guarantees similarity in all forces.
2. Scaling, material properties and loading conditions can alter the relevant physical relationships.
3. Any miniature object validates the safety of its full-sized version.

Claim A; reason B

Specify which features are represented. Structural safety needs appropriate professional analysis; a classroom scale model supports only limited comparisons.

Repair focus: Models and scientific explanations.

56. A scientific theory explains many findings and generates testable predictions. How is it best described?

Stage 1: choose the claim

1. A law waiting to be promoted.
2. A broad explanation supported and evaluated through evidence.
3. An unsupported guess by definition.

Stage 2: choose the reason

1. A theory becomes a law once enough people agree.
2. Scientific theories do not make testable predictions.
3. Theories explain relationships; laws describe patterns, and evidence can support both without a promotion hierarchy.

Claim B; reason C

Theories and laws serve different purposes. A well-supported theory remains open to evaluation and revision; theory does not mean mere speculation.

Repair focus: Models and scientific explanations.

57. A report concludes “it worked”. What revision most improves scientific communication?

Stage 1: choose the claim

1. State the measured change, conditions and evidence supporting the conclusion.
2. Add more exclamation marks.
3. Remove the result table.

Stage 2: choose the reason

1. A conclusion should avoid referring to observations.
2. A scientific conclusion connects a specific claim to relevant evidence and its scope.
3. An enthusiastic phrase is enough evidence.

Claim A; reason B

Name the treatment and measured outcome, report relevant values or patterns, explain the reasoning and state limitations.

Repair focus: Evaluation and scientific communication.

58. A method says “repeat several times”. What is the best repair?

Stage 1: choose the claim

1. Omit repeats from the method.
2. Specify the number of independent trials and how each trial is conducted.
3. Write “repeat many times” instead.

Stage 2: choose the reason

1. Vague wording allows every investigator to run the same method.
2. Repeats need no record because they are not data.
3. Reproducibility requires enough operational detail to perform the method consistently.

Claim B; reason C

State trial count, reset procedure, measurements and controls. Distinguish repeated readings of one sample from independent samples where it matters.

Repair focus: Evaluation and scientific communication.

59. A student writes “human error” as the whole evaluation. What is more useful?

Stage 1: choose the claim

1. List the phrase twice.
2. Assume every error changes the result upward.
3. Identify a specific error mechanism, likely effect and feasible improvement.

Stage 2: choose the reason

1. Specific mechanisms connect the limitation to its effect on the evidence and an actionable repair.
2. All human actions create the same directional error.
3. An evaluation should name a problem without explaining it.

Claim C; reason A

For example, inconsistent stopwatch reaction delay adds variability; use an automated sensor if suitable. Avoid inventing a direction without evidence.

Repair focus: Evaluation and scientific communication.

60. A student pastes secondary data without the original source or collection date. What is missing?

Stage 1: choose the claim

1. Traceability needed to evaluate relevance and quality.
2. Proof that the data must be false.
3. The need for any axis labels.

Stage 2: choose the reason

1. Citing a source replaces all evaluation of its methods.
2. Source, date and collection method help readers assess provenance and suitability.
3. Unsourced data become reliable if formatted neatly.

Claim A; reason B

Cite the original source where possible and explain what was measured, when, how and for which population. Citation enables scrutiny; it is not a quality guarantee.

Repair focus: Evaluation and scientific communication.

61. A result table mixes centimetres and metres without identifying which rows use each. What should be done?

Stage 1: choose the claim

1. Remove units from every row.
2. Use a stated consistent unit or clearly label and convert units.
3. Average all numbers as written.

Stage 2: choose the reason

1. Units do not affect numerical meaning.
2. The largest number always represents the largest physical measurement.
3. Comparisons and calculations require compatible units and clearly defined quantities.

Claim B; reason C

Preserve raw data, document conversions, and use headings with units. Ten centimetres and ten metres are not equivalent measurements.

Repair focus: Evaluation and scientific communication.

62. A team finds an unexpected result that runs against the hypothesis. How should it be communicated?

Stage 1: choose the claim

1. Rewrite the hypothesis as if it had predicted the result originally.
2. Exclude it without explanation.
3. Report it transparently and evaluate possible explanations and limitations.

Stage 2: choose the reason

1. Scientific reporting distinguishes prior prediction from observed results and subsequent interpretation.
2. A report is valid only if the prediction was supported.
3. Unexpected observations never provide useful evidence.

Claim C; reason A

Keep the original prediction visible. Describe what was observed, examine errors or alternative mechanisms, and suggest a discriminating follow-up test.

Repair focus: Evaluation and scientific communication.

63. A class uses students’ identifiable health information for an investigation. What belongs in planning?

Stage 1: choose the claim

1. Teacher-approved consent, privacy and appropriate data-handling procedures before collection.
2. Publishing names makes the dataset more scientific.
3. Collecting first and asking later is enough.

Stage 2: choose the reason

1. Classroom use automatically authorises all personal-data collection.
2. Scientific investigation planning includes ethical responsibilities and participant privacy.
3. Good graphing removes the need for consent.

Claim A; reason B

Discuss suitability with the teacher before any collection. Prefer appropriate anonymous or synthetic datasets where possible; this website does not grant ethical approval.

Repair focus: Evaluation and scientific communication.

64. A conclusion says “the data support the hypothesis under these conditions”. Why is that preferable to “proved forever”?

Stage 1: choose the claim

1. It guarantees the next trial will agree.
2. It matches the limited scope and revisable nature of the evidence.
3. It means the observations have no value.

Stage 2: choose the reason

1. One successful experiment proves a universal statement permanently.
2. Stating limitations automatically invalidates a result.
3. Empirical conclusions are conditional on methods, uncertainty and tested conditions.

Claim B; reason C

Describe the evidence and scope without overclaiming. A qualified conclusion can be strong, precise and useful while leaving room for further tests.

Repair focus: Evaluation and scientific communication.