G2 Science K223, K224 and K225 evidence evaluation becomes stronger when learners ask whether the measurements actually represent the system or population named in the conclusion. Repeating a measurement many times is not the same as sampling the right places, times, organisms or conditions.
This one-hundred-and-fifty-ninth Learner’s Guide develops sampling representativeness. It complements Vol 0091 Repeatability, Replication and Reproducibility and Vol 0095 Scale Translation.
Mechanism: a sample stands in for a larger target
The mechanism of generalisation is substitution: measurements from the sample are used to describe a wider population, space, time or condition. That substitution is justified only when the sampling process makes the observed cases informative about the target cases.
Diagnosis: define the target before judging the sample
Write the target explicitly: which organisms, locations, times, conditions and input range should the conclusion describe? Then compare that target with what was actually sampled. A mismatch between target and sample is a representativeness problem, not a calculation problem.
Smallest repair
Repair the design at the dimension that is missing: add locations for spatial coverage, add times for temporal coverage, include relevant subgroups, randomise selection, use a concurrent control or narrow the conclusion. The smallest repair is the one that aligns sample and claim without collecting unnecessary data.
1. one plant from a tray
Representativeness issue: A single organism may not represent the group if natural variation is large. Use several plants or justify why the individual is the target.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
2. one leaf from a plant
Representativeness issue: Leaf age, position and exposure can differ. Sampling should match the question about the whole plant.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
3. one root region
Representativeness issue: Conditions can vary along the root system. State whether the local measurement is being generalised.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
4. one cell field under a microscope
Representativeness issue: A field of view is local. Multiple fields or a defined sampling method may be needed for tissue-level conclusions.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
5. one quadrat
Representativeness issue: One spatial sample can miss patchiness. Replicate across locations if estimating a habitat.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
6. quadrat near an edge
Representativeness issue: Edge conditions can differ from the centre. Decide whether the edge is representative or a separate environment.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
7. quadrat chosen for convenience
Representativeness issue: Convenience can bias the estimate if easy-to-reach places differ systematically.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
8. random quadrat locations
Representativeness issue: Randomisation reduces deliberate location choice but still requires adequate coverage and sample size.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
9. systematic transect
Representativeness issue: A regular spatial design can reveal gradients but may not estimate the whole habitat unless its path is representative.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
10. one time point
Representativeness issue: A single observation can miss daily, seasonal or recovery changes.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
11. morning-only sample
Representativeness issue: Time of day may matter. Generalisation to the whole day needs justification.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
12. end-point-only sample
Representativeness issue: The final value can hide a transient peak, lag or recovery.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
13. early-only sample
Representativeness issue: The effect may not yet be detectable.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
14. one repeat
Representativeness issue: A single trial does not estimate repeatability well.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
15. many repeats of one biased sample
Representativeness issue: Replication improves precision around that sample but does not fix representativeness.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
16. one treatment group
Representativeness issue: Without a control or comparison group, observed change may reflect background processes.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
17. unequal group sizes
Representativeness issue: Different group sizes change uncertainty and can make raw totals misleading.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
18. different starting states
Representativeness issue: Groups can differ before treatment. Baseline comparability matters.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
19. different age groups
Representativeness issue: A sample of one age may not represent another.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
20. different organism types
Representativeness issue: Species or type can alter response. Generalise only within the sampled population unless evidence supports more.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
21. healthy-only sample
Representativeness issue: Selecting only healthy individuals can bias conclusions about a mixed population.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
22. survivor-only sample
Representativeness issue: Those remaining at the end may differ from those lost during the study.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
23. large sample from one location
Representativeness issue: Large n reduces random error but may still miss spatial diversity.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
24. small sample from many locations
Representativeness issue: Coverage improves but each local estimate can be noisy. Sampling design balances breadth and precision.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
25. one sensor location
Representativeness issue: A local reading can miss a gradient. Use multiple locations if the claim is system-wide.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
26. sensor near source
Representativeness issue: Source proximity can make the reading unrepresentative of the average environment.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
27. sensor near sink
Representativeness issue: A local depletion region can also bias the estimate.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
28. well-mixed assumption
Representativeness issue: One sample can represent the whole only if the compartment is genuinely well mixed at the relevant scale.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
29. poorly mixed sample
Representativeness issue: Concentration can vary spatially; location becomes part of the measurement.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
30. surface-only sample
Representativeness issue: Surface conditions may differ from the interior.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
31. centre-only sample
Representativeness issue: The centre may differ from boundaries where exchange occurs.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
32. top-only sample
Representativeness issue: Vertical gradients can make top measurements unrepresentative.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
33. bottom-only sample
Representativeness issue: Settling or density differences can bias bottom measurements.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
34. large particle bias
Representativeness issue: Sampling method may preferentially include or exclude large particles.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
35. small-volume sample
Representativeness issue: A tiny volume can contain too few events or organisms to reflect the whole.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
36. pooled sample
Representativeness issue: Pooling averages locations but can hide local extremes.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
37. stratified sample
Representativeness issue: Sampling defined subgroups separately can preserve important structure.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
38. random sample
Representativeness issue: Random selection reduces selection bias when every member has a fair chance of selection.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
39. systematic sample
Representativeness issue: Regular intervals can be efficient but may interact with periodic structure.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
40. convenience sample
Representativeness issue: Fast and practical, but representativeness must be defended rather than assumed.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
41. self-selected sample analogy
Representativeness issue: Participants who volunteer can differ from those who do not; use only when such context appears.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
42. one class of students
Representativeness issue: A class may not represent an entire school.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
43. one school
Representativeness issue: A school may not represent all schools.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
44. one day of data
Representativeness issue: Weather, traffic or biological activity can vary across days.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
45. one season
Representativeness issue: Seasonal systems need broader time coverage for annual conclusions.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
46. one experimental batch
Representativeness issue: Batch effects can make one run unrepresentative.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
47. one reagent batch
Representativeness issue: A reagent difference can create a batch-specific result.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
48. one apparatus
Representativeness issue: Instrument-specific bias can persist across repeats.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
49. multiple apparatus
Representativeness issue: Cross-instrument agreement can improve confidence if calibration is controlled.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
50. missing data
Representativeness issue: If missingness is systematic, the observed sample can become biased.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
51. dropouts
Representativeness issue: Loss from one group can change the remaining sample composition.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
52. outlier removal
Representativeness issue: Removing unusual values can make the sample look more homogeneous than the population.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
53. rare events
Representativeness issue: Small samples may miss rare but real outcomes.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
54. high variability
Representativeness issue: More variation requires more careful sampling before generalisation.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
55. low variability
Representativeness issue: A smaller sample can sometimes describe a uniform system more adequately, but this must be supported by evidence.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
56. heterogeneous population
Representativeness issue: Subgroups should be represented rather than averaged away blindly.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
57. homogeneous population
Representativeness issue: If evidence shows strong uniformity, one location or subgroup may be more representative than in a heterogeneous system.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
58. sampling frame
Representativeness issue: The list or set from which samples are chosen must match the target population.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
59. target population
Representativeness issue: State exactly who, what, where and when the conclusion is meant to describe.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
60. sample population
Representativeness issue: State the actual observed group rather than silently expanding it.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
61. sample-to-population inference
Representativeness issue: Generalisation is an inference, not an observation. Its strength depends on how the sample was obtained.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
62. spatial representativeness
Representativeness issue: Coverage should match the spatial scale of the claim.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
63. temporal representativeness
Representativeness issue: Coverage should match the time scale of the claim.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
64. condition representativeness
Representativeness issue: Test conditions should match the conditions named in the conclusion.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
65. range representativeness
Representativeness issue: Input levels should cover the range over which the model is being claimed.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
66. control representativeness
Representativeness issue: The control should be comparable to the treatment except for the intended difference.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
67. replication versus representation
Representativeness issue: Repeats answer ‘is this result consistent here?’; representative sampling answers ‘does this describe the larger target?’
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
68. precision versus bias
Representativeness issue: A large biased sample can be precise but wrong for the target population.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
69. sample size versus design
Representativeness issue: More observations cannot fully rescue a systematically unrepresentative design.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
70. weighted sample
Representativeness issue: If subgroups are intentionally sampled unequally, appropriate weighting may be needed before whole-population interpretation where the task defines it.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
71. local anomaly
Representativeness issue: One unusual site may be real but not representative of the whole. Preserve it while separating local and system claims.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
72. gradient sampling
Representativeness issue: Samples along a gradient reveal spatial structure better than a single average location.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
73. before-after sample
Representativeness issue: The same units measured before and after can control some individual differences but may introduce time effects.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
74. paired sampling
Representativeness issue: Matched pairs can improve comparison when the pairing variable is relevant.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
75. independent samples
Representativeness issue: Separate groups can avoid carryover but require baseline comparability.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
76. fresh sample
Representativeness issue: Fresh material can reduce path dependence or carryover.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
77. reused sample
Representativeness issue: Repeated treatment can make history part of the response.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
78. sampling with replacement conceptually
Representativeness issue: Repeated selection can revisit the same member; use only if the task defines such a design.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
79. sampling without replacement conceptually
Representativeness issue: The population changes as members are removed; event probabilities can change.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
80. representativeness final rule
Representativeness issue: The sample must match the population, place, time, condition and range named in the conclusion.
For practice, write the narrowest conclusion the sample clearly supports, then write the broader conclusion a learner might be tempted to make. Identify the extra sampling evidence needed to justify the broader version.
For transfer, change location, time, population or experimental range. The learner should redesign the sample rather than reuse the original design automatically.
Observation versus inference
The measured sample values are observations. The statement that those values describe the wider population is an inference. Keep those layers separate. A perfectly measured sample can still support a weak generalisation if selection was biased or too narrow.
Competing explanations
When treatment and control differ, ask whether sampling composition could also explain the difference. Age, location, starting condition, batch or time can create apparent treatment effects if they are distributed unevenly between groups.
Support is not proof
A representative-looking sample supports generalisation but does not prove that every unsampled member behaves identically. Scientific conclusions remain population- and condition-bounded. Variability is part of the system, not automatically an error.
Model limits
There is no universal ‘correct sample size’ independent of variability, effect size, design and task. At G2 level, the key reasoning is qualitative: broader, more variable targets generally require more deliberate coverage than uniform, tightly controlled systems.
Delayed transfer and measurement
Return later with an unfamiliar ecological, biological, physical or experimental setup. Ask the learner to define the target population, identify a likely sampling bias, propose the smallest repair and state the narrow conclusion that remains valid before the repair.
Internal learning links
Use the Science Hub, Vol 0155 Perturbation and Recovery, Vol 0143 Spatial Gradients, the Examination Craft hub and the PSLE Learner’s Guide.
MOE and SEAB current framework
MOE states that Full Subject-Based Banding is fully implemented and that, from 2027, the Singapore-Cambridge Secondary Education Certificate (SEC) replaces the former N- and O-Level examinations, with students sitting subjects at their respective G1, G2 or G3 levels. See the official MOE Full SBB / SEC announcement. For the current 2027 G2 Science combinations K223, K224 and K225 and linked syllabuses, use the official SEAB G2 syllabus directory.
Final rule
Before generalising, name the target population, place, time, condition and range. Then ask whether the sample actually covers them. More measurements help only when they are measurements of the right things.
G2 SEC Learner’s Guide: Open the Vol 0132–0175 hub and subject index.
