General Science Reasoning Exam Survival Guide

Window: Aptitude/entrance tests • Boards/practicals • Olympiad basics
Components: Data/graphs • Experimental design • Physics/Chem/Bio/Earth reasoning • Lite stats/causality

Must-do topics (80/20 focus)

Data & Graphs

• Read any figure with TAUSL: Title • Axes • Units • Scale (incl. log) • Legend.
• Linear vs log scale; rate vs count; % vs percentage-points.
• Error bars (SD/SE/CI): overlap ⇒ weak/no clear difference.

Experimental Design

• Identify IV/DV, constants; control (negative/positive), replicates, randomization/blinding.
• Reliability vs validity; sample size; operational definitions.

Reasoning Patterns

• Proportional reasoning (direct/inverse), limiting cases, order-of-magnitude.
• Dimensional analysis & unit discipline.
• Conservation frames: mass/charge/energy.
• Mechanism narrative: cause → mediator → observable.

Physics (reasoning, not rote)

• Free-body diagrams; constant vvv ⇒ ∑F=0\sum F=0∑F=0.
• Energy chains (store→transfer→dissipation).
• Circuits: series same I, parallel same V (use KCL/KVL).
• Waves: v=fλv=f\lambdav=fλ; amplitude ≠ speed.
• Optics ray rules; sign conventions.

Chemistry

• Particle (sub-microscopic) ↔ symbolic ↔ macroscopic triangle.
• Stoichiometry via moles first; limiting reagent.
• Equilibrium: QQQ vs KKK; Le Châtelier (count gas moles; exo/endo as product/reactant).
• Acids/bases: strength ≠ concentration; buffer ratio (H-H).

Biology

• Gene/allele/trait distinctions; meiosis I vs II; chi-square fit.
• Natural selection: variation → differential success → allele freq Δ.
• Physiology balances (inputs/outputs); ecology energy pyramid (~10% step).

Earth/Env

• Seasons = axial tilt, not distance; greenhouse ≠ ozone.
• Maps/sections: direction, scale, relative vs absolute measures.

Lite Stats & Causality

• Mean/median choice; variability matters.
• Correlation ≠ causation; confounders; Simpson’s paradox intuition.
• Absolute vs relative risk; base-rate effects.

Top traps (avoid)

• Reading log plots as linear; reversed or dual axes missed.
• Comparing totals when the figure shows rates/per-capita.
• Averaging percents without weights; misuse of trend lines.
• Calling any control “placebo” (mismatched control).
• Treating energy as “used up” (vs transformed); mass=weight; current “consumed.”
• Equating pH with acid strength; plants “only respire at night.”
• Concluding causation from a before/after without a counterfactual.

Time split (per science set)

• Scan 20%: TAUSL + question stems; mark which need calc vs logic.
• Work 70%: easy reads → proportional/unit checks → one calc.
• Audit 10%: units, sign, scale, claim-vs-evidence alignment.

Last-48h checklist (do exactly this)

• 8 mixed graph drills (incl. one log, one stacked, one with dual axes).
• 4 experiment design prompts (state IV/DV, control, confounders, measure plan).
• 10 proportional/dimensional quickies (unit sanity).
• 4 Fermi estimates (population/time/energy).
• 6 claim critique items (correlation vs causation; what data would decide).
• One-page tool ring: TAUSL • unit ladder • conservation boxes • H-H • K vs Q • CEI writing frame.

Quick frames & mini-toolkit (carry card)

1) Claim–Evidence–Insight (CEI)

• Claim: what you assert. Evidence: data/figure line. Insight: why it proves the claim.

2) Experimental skeleton

• IV → levels; DV → how measured (units, instrument, interval).
• Controls: negative (baseline), positive (should work).
• Confounders: list → control/measure/randomize; nnn & replication.

3) Dimensional analysis

• Write target units → build fraction ladders so units cancel to target.
• Reject answers with wrong units or scaling.

4) Proportional reasoning

• If y∝xy\propto xy∝x: ×k in xxx ⇒ ×k in yyy. If y∝1/xy\propto 1/xy∝1/x: ×k in xxx ⇒ ÷k in yyy.
• Limit check: x→0,∞x\to0,\inftyx→0,∞ — does behavior make sense?

5) Energy/Conservation boxes

• Start store → transfer path → end stores + losses (thermal/sound).
• Mass/charge in = mass/charge out (+accumulation).

6) Causality quick tests

• Temporal order? Plausible mechanism? Confounders measured? Comparison group? Dose–response? Reversibility?

7) Error bars & significance

• Overlapping 95% CIs → likely not different (context-dependent). Note effect size vs “significance.”

8) Risk language

• Relative risk “↓50%” could be 2% → 1% (absolute −1 pp). Always report both.

Speed tactics (during test)

• Say TAUSL aloud before any data question.
• Units first: convert once, compute once; bound with hi/lo to kill options.
• Prefer ratios (dimensionless) over raw numbers.
• Draw FBD for forces; quick ray sketch for optics; box ICE table for equilibrium.
• If a conclusion feels big, ask “What evidence would change my mind?” — if none is referenced, it’s weak.

Exam-day mini-plan

• 3-min warm-up: 2 graph reads (one with %), 1 unit conversion, 1 proportion.
• Start with the friendliest figure; skip calc-heavy until one clean set is banked.
• Final 5–8 min: unit/scale audit on all answered items; re-read any EXCEPT/NOT stems.

Practice cycle (weekly)

1. Diagnose: 2 figure sets + 2 design prompts (timed). Tag misses: Scale/Units, Base/Rate, Control/Confound, Proportion/Dim.
2. Deep fix: redo one per tag; write CEI for each.
3. Retest: fresh mixed set; aim fewer repeats of same tag.

Summarize: add one exemplar graph & one model experimental design to your ring.