Comparing two box plots or two sets of summary statistics is done by stating the isolated numbers, for example a median of 45 against a median of 20, without a comparative sentence explaining what that means.

Write the comparison in context: visitors to the second group are on average older, because their median is higher. State the direction and the reason, not just the two numbers.

Warning lines and action lines on a quality control chart are confused, or the required response, an immediate re-sample against stopping the process, is given for the wrong threshold.

Treat the warning line and the action line as two separate rules with two separate responses: one point beyond the warning line means re-sample immediately, a point beyond the action line, or two consecutive points beyond the warning line, means stop the process.

Chain-base index numbers are calculated as a percentage change from the original base year rather than from the immediately preceding period.

Before calculating, check whether the index is fixed-base or chain-base. A chain-base index always compares to the period directly before it, not to year one of the series.

In comparative pie charts, the new radius is calculated as directly proportional to the frequency rather than proportional to the square root of the frequency, since it is the area that must match the frequency.

Use the relationship between the two radii squared and the two frequencies, and take the square root at the final step to get the radius, not the frequency ratio itself.

When comparing standardised z-scores between two distributions, candidates ignore the sign or fail to state that the value closer to zero is closer to its own distribution's mean.

Compare the two z-scores by their distance from zero, and state explicitly that the smaller distance means that value is closer to the mean, more typical, of its own distribution.

Percentages read from a compound percentage bar chart are treated as if they were raw counts, without checking whether the two groups being compared have different base population sizes.

Before comparing two percentages across groups, check whether the underlying totals are the same size. If they are not, the same percentage can represent very different raw numbers.

Stem and leaf diagrams are drawn without a key, or with the leaves left unordered within each row.

Write the key as the first thing you add to a stem and leaf diagram, and order every row of leaves from smallest to largest before moving on.

Conditional probability questions from a Venn diagram or tree diagram use the whole sample space as the denominator instead of the correct restricted subset.

Identify exactly which group the question is conditioning on, and use only that group's total as the denominator, not the full sample.

CCEA GCSE · Exam Tips

Statistics 2260 Exam Tips

CCEA GCSE Statistics: how the 200 marks split evenly across two calculator-based papers, the comparative-language and index-number habits that recur in examiner reports, and legal calculator techniques for standard deviation, correlation and binomial questions.

3 min readUpdated: Sep 3, 2026

Exam at a Glance

Papers

Total Marks

200

Time Limit

Question Types

Paper	Duration	Marks	Questions	Weighting	Question Types
Unit 1	2h	100	10	50%	Outlier detection, stem and leaf diagram, median and range, Sampling bias, survey questions, box plot reading, compound percentage bar chart, Biased coin tree diagram, outcome listing, combined probability and expected frequency, Time series graph interpretation and fixed-base index numbers, Stepped cumulative frequency, data classification, percentiles and deciles, Bivariate scatter diagram, correlation interpretation, line of best fit, Statistical process control chart, target/warning/action lines, out-of-control evaluation, Grouped frequency distribution mean estimate, scaling effects on mean and standard deviation, Quarterly time series, 4-point moving averages, trend line, seasonal forecasting, Normal distribution sketching, standardised z-score, empirical rule probabilities
Unit 2	2h	100	11	50%	Comparative line graph analysis across demographic groups, Critique of misleading charts and scale misrepresentations, Statistical enquiry cycle planning, hypothesis generation and data sourcing, Continuous cumulative frequency curve, box plot construction, distribution symmetry, Histogram grouped mean estimation and error impact evaluation, Comparative pie charts with area-proportional radius computation, Quarterly time series moving average, trend plotting and extrapolation assumptions, Three-set Venn diagram, conditional probability, cross-group estimation, Chain-base index numbers, missing price computation, geometric mean, Probability tree without replacement and conditional probability deduction, Contingency table relative risk analysis and binomial distribution evaluation

Grade Scale

A*ABC*CDEFGU

Calculator Policy

Both Unit 1 and Unit 2 are calculator papers, and a formula sheet is printed in the exam booklet itself. Calculator statistics functions are expected throughout, including for the mean, standard deviation, and correlation coefficients. The standard JCQ rule applies: no computer algebra or symbolic manipulation function, no communication with another device, and no retrievable information of any kind, including formulae, notes or programs, stored in the calculator. Clear the calculator's memory before the exam or activate its exam mode; the printed formula sheet supplies the formulae, not the calculator's memory.

AO1: Demonstrate knowledge and understanding, using appropriate terminology and notation, of standard statistical techniques used to collect and represent data, and calculate summary statistics and probabilities. (55%)
AO2: Interpret statistical information and results in context and reason statistically to draw conclusions. (25%)
AO3: Assess the appropriateness of statistical methodologies and the conclusions drawn through the application of the statistical enquiry cycle. (20%)

Built from real past papers and marking schemes (2023–2025).

Tips & Strategies

Two equal papers, both with a calculator

Statistics is assessed across Unit 1 (100 marks, 120 minutes) and Unit 2 (100 marks, 120 minutes), both calculator papers with a formula sheet printed in the exam booklet. Each question is worth between three and fifteen marks and is built from several linked statistical skills rather than one isolated technique, so read the whole question before you start, because a data classification decision in part (a) often feeds directly into a calculation in part (c). At roughly a mark a minute, an eleven-mark question should take around eleven minutes: if you are still on it after fifteen, move on and come back.

The habit that costs the most marks: numbers without comparison

When a question asks you to compare two data sets, whether from box plots, means, or standard deviations, giving the two numbers side by side without a comparative sentence loses the interpretation mark even if both numbers are correct. Write the comparison out in context: visitors to the museum are on average older, because the median age is higher, not just the two medians listed next to each other. The same discipline applies to standardised z-scores when comparing two distributions: state which value is closer to zero, and explain that this means it is closer to its own distribution's mean, rather than leaving two z-scores sitting unexplained.

Index numbers and quality control: know which reference point applies

Chain-base index numbers measure change relative to the immediately preceding period, not the original base year, and the most common error is calculating a chain-base percentage change against the wrong reference point. On quality control charts, warning lines and action lines are different thresholds with different required responses: a single point beyond the warning line calls for an immediate re-sample, while a point beyond the action line, or two consecutive points beyond the warning line, means stopping the process. Confusing the two thresholds, or giving the wrong corrective action, loses marks even when the point has been plotted correctly.

Comparative pie charts and conditional probability

In comparative pie charts, the radius is not directly proportional to the frequency it represents; the area is. Getting this wrong means calculating the new radius as a simple ratio rather than taking a square root, which throws off every subsequent measurement on the diagram. In conditional probability questions worked from a Venn diagram or a tree diagram, the denominator has to be corrected to the relevant subset, not the whole sample space, and skipping that correction is one of the most common ways a probability question loses its final mark.

CCEA conventions

Statistics is a unitised GCSE like every other CCEA GCSE: at least 40 percent of the assessment must be completed in the series you cash in for a final grade, and each unit can be resat once before then, with the better result normally counting. A stem and leaf diagram without a key, or with unordered leaves, is marked down regardless of whether the underlying data is correct, so build the key into your diagram from the first line you write, not as an afterthought.

The week before

Take three past comparison questions and rewrite every answer as a full comparative sentence rather than two isolated statistics. Redo the chain-base index and quality control questions from the last three series, checking explicitly which reference point or threshold each part is asking about before you calculate. Practise the comparative pie chart radius calculation using the square root, and work through five conditional probability questions checking the denominator in every answer before you finalise it.

Ready to test yourself?

Turn these notes into exam-style practice. Get unlimited AI questions on this topic with instant marking and explanations.

Practice This Topic

Calculator Programs

Mean and standard deviation from the statistics mode

Casio fx-83/85 series (scientific)

Purpose: Get an accurate mean and standard deviation from a frequency table without manual summation errors.

When to use it: Any question giving raw or grouped frequency data and asking for a mean, standard deviation, or the effect of scaling the data.

Steps

Switch the calculator into statistics, one-variable, mode, key in each data value with its frequency using the frequency column, then read the mean and standard deviation directly from the statistics menu rather than summing by hand.

Exam note: This uses the calculator's built-in statistics mode live during the exam with data taken from the question. CCEA and JCQ rules do not allow any formula, note or program to be stored in the calculator beforehand; the printed formula sheet supplies any formulae you need. Activate exam mode if your calculator has one.

Correlation coefficient from the regression mode

Casio fx-83/85 series (scientific)

Purpose: Get an accurate product moment correlation coefficient directly, as the specification explicitly allows, instead of calculating it term by term.

When to use it: Bivariate scatter diagram questions asking for a correlation coefficient or the equation of a line of best fit.

Steps

Switch the calculator into statistics, two-variable, regression mode, key in each paired x and y value, then read the correlation coefficient and the gradient and intercept of the regression line directly from the statistics menu.

Exam note: This is standard, permitted use of the calculator's regression mode with data entered live from the question during the exam. No dataset, formula or method may be stored in the calculator beforehand.

Cumulative binomial probability in one running total

Casio fx-83/85 series (scientific)

Purpose: Add several binomial probability terms accurately, such as fewer than three successes, without losing track of which terms have been included.

When to use it: Binomial distribution questions that ask for the probability of fewer than, at most, or at least a given number of successes.

Steps

Calculate each required probability term, for example P(X=0), P(X=1), P(X=2), using the calculator's power and combination, nCr, keys, adding each result into memory with M+ as you go, then use memory recall for the final total.

Exam note: M+ and memory recall are ordinary, permitted arithmetic functions used live during the exam with values calculated from the question. They must not be used to retrieve any binomial table or formula stored before the exam started.

Comparative pie chart radius using the square root key

Casio fx-83/85 series (scientific)

Purpose: Get the correct new radius for a comparative pie chart, based on area, not on a direct frequency ratio.

When to use it: Comparative pie chart questions that give one radius and frequency and ask for the radius of a second pie chart representing a different frequency.

Steps

Divide the new frequency by the original frequency, multiply by the original radius squared, then take the square root of that result in one continuous calculator entry to get the new radius.

Exam note: This is a live calculation using the calculator's square root and arithmetic keys during the exam. No formula or worked method may be pre-stored in the calculator.

Common Mistakes

1highMarks at stake: 1Comparing data sets
Comparing two box plots or two sets of summary statistics is done by stating the isolated numbers, for example a median of 45 against a median of 20, without a comparative sentence explaining what that means.
How to avoid it: Write the comparison in context: visitors to the second group are on average older, because their median is higher. State the direction and the reason, not just the two numbers.
2highMarks at stake: 2Control charts and quality assurance
Warning lines and action lines on a quality control chart are confused, or the required response, an immediate re-sample against stopping the process, is given for the wrong threshold.
How to avoid it: Treat the warning line and the action line as two separate rules with two separate responses: one point beyond the warning line means re-sample immediately, a point beyond the action line, or two consecutive points beyond the warning line, means stop the process.
3highMarks at stake: 2Index numbers
Chain-base index numbers are calculated as a percentage change from the original base year rather than from the immediately preceding period.
How to avoid it: Before calculating, check whether the index is fixed-base or chain-base. A chain-base index always compares to the period directly before it, not to year one of the series.
4mediumMarks at stake: 2Population pyramids and comparative pie charts
In comparative pie charts, the new radius is calculated as directly proportional to the frequency rather than proportional to the square root of the frequency, since it is the area that must match the frequency.
How to avoid it: Use the relationship between the two radii squared and the two frequencies, and take the square root at the final step to get the radius, not the frequency ratio itself.
5mediumMarks at stake: 1Standardised scores
When comparing standardised z-scores between two distributions, candidates ignore the sign or fail to state that the value closer to zero is closer to its own distribution's mean.
How to avoid it: Compare the two z-scores by their distance from zero, and state explicitly that the smaller distance means that value is closer to the mean, more typical, of its own distribution.
6mediumMarks at stake: 1Tables, charts and bar charts
Percentages read from a compound percentage bar chart are treated as if they were raw counts, without checking whether the two groups being compared have different base population sizes.
How to avoid it: Before comparing two percentages across groups, check whether the underlying totals are the same size. If they are not, the same percentage can represent very different raw numbers.
7mediumMarks at stake: 1Stem and leaf diagrams
Stem and leaf diagrams are drawn without a key, or with the leaves left unordered within each row.
How to avoid it: Write the key as the first thing you add to a stem and leaf diagram, and order every row of leaves from smallest to largest before moving on.
8mediumMarks at stake: 1Venn diagrams and tree diagrams
Conditional probability questions from a Venn diagram or tree diagram use the whole sample space as the denominator instead of the correct restricted subset.
How to avoid it: Identify exactly which group the question is conditioning on, and use only that group's total as the denominator, not the full sample.

Turn these tips into top grades

thinka turns your weak spots into targeted practice, with instant marking and exam-style feedback. Study smarter, not longer.

Start Practicing Free See pricing

Practice real exam questions with instant AI feedback and marking.

Start Practicing Free