Science inquiry — Advanced

Score: 0 / 20

Level 3: Advanced

Evaluate, justify, synthesise, propose — combine ideas, judge claims and methods (senior secondary style).

1. Evaluate an investigation that uses one participant and no control to prove a universal health claim.

   A)Strong enough if the measured effect is large and the outcome readout is objectiveB)Very weak — individual variability is wide; controls, larger samples if ethical, and replication matterC)Stronger than a randomised trial when that person is carefully matched to the target groupD)Unbiased by design because omitting a comparison group removes confounding entirely

2. A company-funded study finds benefits for its product and hides negative trials. This mainly undermines:

   A)Trust in the evidence base — selective reporting skews what looks true overallB)Only stylistic readability of the discussion, not the weight readers give findingsC)Power in the published positives, without changing whether benefits truly outweigh harmsD)Whether peer review occurred, assuming reviewers always see every completed study

3. Justify reporting uncertainty (for example error bars or a range) rather than a single tidy number alone:

   A)Because ranges look messy, audiences prefer one headline number without spreadB)Because repeating measures removes spread, a range duplicates what repeats already showC)Because instruments and samples vary, spread shows limits and calibrates how strong claims may beD)Because error bars suggest ignorance, they weaken conclusions compared with one mean

4. Propose one improvement to make a classroom pendulum experiment more reliable while investigating the same question.

   A)Time many swings but start and stop the watch on different phases of the cycle each runB)Swap bob mass when the room feels drafty, without logging air movement or windowsC)Tweak string length whenever the arc looks wide, to keep the swing visually matchedD)Time many swings per run with a consistent release, repeat trials, and average to cut timing noise

5. Synthesise: a strong scientific argument in a report needs:

   A)Confident claims and a persuasive tone, backed mainly by expert authority aloneB)Claims anchored to data, explicit reasoning, stated limits, and transparent source useC)Long theory sections with raw data relegated to an appendix readers may skipD)Conclusions aligned with the starting hypothesis, with little said about anomalies

6. Evaluate a conclusion drawn from a convenience sample (for example only friends) for a whole year level.

   A)Sound if at least five people answered, because small n still averages out biasB)Sound because friends answer sincerely, which outweighs how they were recruitedC)Risky — shared social circles skew views; wider or structured sampling is needed to generaliseD)Preferable to random sampling when you need quick replies from people you trust

7. Justify plotting error bars (for example range or SD) on a mean comparison graph:

   A)They show spread around each mean so overlap and effect clarity can be judgedB)They remove the need to publish underlying measurements or raw tablesC)They demonstrate that the independent variable mechanically caused the shiftD)They automatically conceal every outlier so viewers never see extremes

8. Critique rewriting the hypothesis after results so it matches the data:

   A)It is sound because the updated wording now tracks the pattern the data showB)It risks post-hoc fitting — predictions should be stated before testing when claims are strongC)It mirrors standard rules that hypotheses must always follow the results sectionD)It is equivalent to preregistering hypotheses before data collection

9. Propose how to improve validity when a measurement tool drifts between sessions:

   A)Ignore slow drift if the plotted line still looks smooth to the eyeB)Drop roughly half the readings at random to reduce the influence of biasC)Calibrate before sessions, check against standards, or adjust if drift is modelledD)Redefine the dependent variable mid-study so earlier sessions stay usable

10. Synthesise ethics, benefit and risk when using student biometric data in a project:

    A)Collect traces first, then seek permission if anyone later objectsB)Treat ethics as relevant only to animal work, not human physiological loggingC)Collect only what is needed, explain use, gain consent, secure storage, allow opt-out if practicalD)Publish identifiable names beside full traces so methods are completely transparent

11. Evaluate using only a statistical test’s ‘significant’ label as proof the effect matters in real life:

    A)Sufficient on its own whenever the test crosses a common threshold, regardless of contextB)Equivalent to reporting effect size because the test embeds magnitude and spread togetherC)Weak — effect size, context, design quality, and replication still matter; a label is not importanceD)Proof the null hypothesis is true whenever the test falls just above a common cutoff

12. Critique running many different analyses on the same data and reporting only the one that crossed a threshold:

    A)Sound exploratory practice whenever the work is labelled preliminary in the titleB)Inflates false positives — pre-specify primary comparisons or account for many testsC)Technically impossible once data are entered in everyday spreadsheet softwareD)Universally required by journals before results sections may be drafted

13. Evaluate pooling summaries from several published studies when some negative results were never published:

    A)Overview can still help, yet missing studies can skew the picture — transparency about gaps mattersB)Pooling removes the need for any new primary data collection in the fieldC)Pooling establishes causation even when included studies are purely observationalD)Pooling assumes every included study estimates the same underlying quantity by design

14. Propose an ethical improvement for a voluntary anonymous mental-health tick-box survey in class:

    A)Require real names on each form so responses can be traced if neededB)Post individual tick-box answers on a class noticeboard for open discussionC)Treat completion as mandatory so the sample fully represents the cohortD)Offer opt-out, list support contacts, secure anonymous storage, avoid coercion

15. Evaluate the file-drawer problem for published literature:

    A)Null or negative findings are always published alongside positive onesB)Hidden nulls can bias summaries and what the public sees as establishedC)Editors accept only null results because positives are assumed promotionalD)Affects social science only and leaves clinical evidence summaries untouched

16. Justify open methods and data sharing where ethical:

    A)Eliminates the need for independent peer review before claims spreadB)Prevents authors from receiving credit because files are anonymously dumpedC)Speeds verification, cuts fraud risk, and supports teaching replicasD)Forces every replication lab to obtain numerically identical outcomes

17. Critique this plan: deploy a diagnostic phone app in clinics tomorrow because it scored well on one in-house test, with no error analysis or governance.

    A)Fairness, drift, failure modes, and oversight need attention — accuracy on one test is not enoughB)Sound if validation accuracy on the training split exceeded ninety percentC)Safe because consumer apps are engineered not to fail after public releaseD)Outside ethics review because clinical software is treated as neutral tooling

18. Synthesise qualitative interviews with quantitative counts in an ecology project:

    A)Avoid combining narratives with numeric ecology measures in one reportB)Let interviews substitute for transect counts when terrain is difficultC)Treat abundance tallies as purely subjective opinions like interview quotesD)Use interviews for meanings and transects for abundance — complementary lenses

19. Evaluate treating two similar p-values as ‘success’ versus ‘failure’ only because one sits barely below a 0.05 cutoff:

    A)Rigorous because the 0.05 line cleanly separates real effects from noiseB)Arbitrary — report estimates, intervals, and methods, not ritual cutoffs aloneC)Legally required for school labs to match published medical trial rulesD)Proof the entire analysis is invalid if both values appear in one table

20. Propose how student peer review of draft methods could mirror scientific practice:

    A)Let peer marks substitute fully for the teacher’s summative assessmentB)Skip written comments so authors are not overwhelmed by marginal notesC)Give structured feedback on variables, controls, safety, repeats, and ethicsD)Limit reviewers to praise so confidence is protected from critique

← Back to hub