Edexcel-specific advice for GCSE Statistics Higher tier. Tick each tip as you review it — your progress is saved automatically.
Read the formula sheet carefully — it's provided in both papers. Know which formulae are given and which you must memorise (e.g. SRCC formula is given; mean formula is not).
Bring a ruler, protractor, and sharp pencil — many questions require accurate diagrams (box plots, histograms, cumulative frequency curves).
Bring a calculator for Paper 2 — but practise Paper 1 without one. Know which paper is calculator and which is non-calculator.
Revise your weakest topics the night before — use the Progress page to identify them.
Get a good night's sleep and eat before the exam — cognitive performance drops sharply when tired or hungry.
Always read the full question before answering — the context (e.g. a table or diagram) often contains information needed for part (b) or (c).
Underline key words: 'explain', 'calculate', 'compare', 'criticise', 'comment'. Each demands a different type of answer.
Check the number of marks — a 3-mark question needs 3 distinct points. Never give a 1-line answer to a 3-mark question.
Look at the space provided — if there are 4 lines, write roughly 4 lines. If there's a small box, a brief answer is expected.
For 'describe' questions on diagrams (e.g. scatter graphs, time series), always comment on trend, strength, and any anomalies.
Always show full working — even if your final answer is wrong, you can earn method marks (M marks) for correct steps.
Write down the formula before substituting values — this earns the formula mark and helps you avoid errors.
Don't round intermediate values — only round your final answer to the required degree of accuracy (usually 3 s.f. or 2 d.p. unless stated).
If you use a calculator, write down the expression you entered, not just the answer — e.g. 'Σfx = 3×5 + 7×8 + ... = 142'.
Cross out mistakes with a single line — don't scribble them out. Examiners may still award marks for crossed-out work if it's legible and correct.
Label every answer with its units (e.g. kg, cm, %) — missing units can cost a mark on accuracy questions.
Histograms: Frequency density = Frequency ÷ Class width. Always check if the y-axis says 'frequency density' — never plot raw frequencies on a histogram.
Cumulative frequency: Plot points at the upper class boundary, not the midpoint. Join with a smooth curve, not straight lines.
Box plots: The box spans Q1 to Q3. Whiskers go to the smallest and largest non-outlier values. Mark outliers with a cross (×).
Standard deviation: Use the formula on the sheet. Show Σx, Σx², n clearly. A common error is dividing by n instead of using the correct formula.
SRCC: Rank both sets of data from 1 to n. Calculate d (rank difference) for each pair. Use the formula: rs = 1 − (6Σd²)/(n(n²−1)).
Probability: Always check if events are independent or mutually exclusive before choosing your method. P(A and B) = P(A) × P(B) only if independent.
Sampling: Know the difference between random, stratified, systematic, quota, and opportunity sampling — and when each is appropriate or biased.
Comparing distributions: Always compare both a measure of average (mean/median) AND a measure of spread (range/IQR/SD) — one comparison alone is never enough.
Index numbers: Chain base index = (current year value / previous year value) × 100. Fixed base index = (current year value / base year value) × 100.
Capture-recapture: N = (M × n) / m. Assumptions: population is closed, marks don't affect capture probability, marks don't fall off.
For 'criticise this questionnaire' questions: comment on leading questions, ambiguous options, overlapping class intervals, and missing 'other' options.
For 'comment on the reliability of this sample' questions: discuss sample size, sampling method, and whether it's representative of the population.
For 'interpret the correlation' questions: state direction (positive/negative), strength (strong/weak/moderate), and what it means in context — don't just say 'there is a positive correlation'.
For hypothesis testing conclusions: state whether there is or isn't evidence to support the hypothesis, and refer to the specific statistic (e.g. PMCC, SRCC value).
Avoid vague language like 'it goes up' — use precise statistical language: 'the median increases', 'the IQR decreases', 'there is a strong positive linear correlation'.
When asked to 'compare', always use comparative language: 'higher than', 'less spread than', 'more skewed than' — not just separate descriptions of each dataset.
Each paper is 1 hour 30 minutes for ~100 marks — aim for roughly 1 minute per mark. A 4-mark question should take about 4 minutes.
Don't spend more than 3 minutes on a 1-mark question — if you're stuck, move on and return at the end.
Leave 10 minutes at the end to check: units, rounding, that you've answered all parts (a), (b), (c)...
Attempt every question — there is no negative marking. A guess is better than a blank.
If you run out of time, write bullet-point notes for incomplete answers — examiners can award marks for correct points even without full sentences.
💡 Click any tip to mark it as reviewed. All progress is saved automatically in your browser.