Large Data Set Revision for the Mock
Large data set revision for the mock, with a focused checklist, statistical examples, common mistakes and practical advice for A-Level Maths students.

Mock season has a strange rhythm. One week, everything feels manageable. The next, a sixth-form student is looking at a statistics question about weather stations, wondering whether they were somehow meant to memorise an entire spreadsheet.
The reassuring answer is no. Large data set revision for the mock is mainly about familiarity: knowing the context, variables, units and conventions of your exam board's data, then applying ordinary statistical methods intelligently. You do not need to remember every observation.
Because large data sets differ between specifications, begin by checking whether you take Edexcel, AQA, OCR or Eduqas. Your teacher and exam board specification should be the final authority on the data set used in your course.
Your large data set revision checklist
Before the mock, make sure you can:
- identify the locations, time periods and populations represented;
- recognise each variable and its unit;
- distinguish qualitative, discrete and continuous variables;
- understand unusual codes, missing values and recording conventions;
- calculate and interpret summary statistics;
- compare distributions using their centre and spread;
- identify possible outliers;
- discuss sampling methods, bias and limitations;
- interpret correlation without claiming causation;
- use your calculator efficiently for one-variable and two-variable statistics.
This is a much shorter task than memorising a spreadsheet. It also reflects how large data set knowledge usually becomes useful: as context supporting statistical reasoning.
Start with the story behind the data
Imagine a student called Maya beginning her A-Level Maths revision two weeks before her mocks. She opens the large data set and sees columns of figures. Her first instinct is to read down every row.
That feels productive, but it has little structure. Instead, she asks four questions:
- Who or what was observed?
- Where and when were observations collected?
- What variables were recorded?
- How were those variables measured or coded?
Those questions turn a spreadsheet into a story. Weather observations might vary by season, geography and year. Data about transport, sport or populations will carry different contextual limitations. Statistical conclusions make more sense once you understand what each row represents.
Create a one-page summary for your own specification. Include variable names, units, locations, dates and any special notation. Do not rely on a generic online summary if it covers another exam board.
For broader preparation, MathsGenie's A-Level Maths revision resources provide a useful starting point for organising the topics surrounding statistics.
Learn meanings and units, not isolated labels
A variable name without its unit is only half learnt. If wind speed is measured in knots, rainfall in millimetres or temperature in degrees Celsius, those units affect whether an answer is sensible.
You should also be able to classify variables. A named location is qualitative. A count is discrete. A measurement such as temperature is generally continuous, even when recorded to a fixed degree of accuracy.
This matters because the type of variable influences the diagram, calculation and interpretation you should use. A histogram is designed for grouped continuous data, while a bar chart is appropriate for categories. The distinction is easy to overlook when mock pressure narrows your attention to calculator buttons.
Connect the data set to core statistics
Large data set questions are rarely separate from the rest of statistics. They commonly draw on familiar skills such as averages, measures of spread, sampling, correlation and hypothesis testing.
For a sample x1,x2,…,xnx_1,x_2,\ldots,x_nx1,x2,…,xn, the mean is
xˉ=∑xn.\bar{x}=\frac{\sum x}{n}.xˉ=n∑x.The standard deviation describes spread around the mean. Your calculator may display both population and sample versions, so check which value the question requires and follow the notation used in your specification.
For quartiles, a common rule for flagging potential outliers is
x<Q1−1.5IQRx<Q_1-1.5\operatorname{IQR}x<Q1−1.5IQRor
x>Q3+1.5IQR,x>Q_3+1.5\operatorname{IQR},x>Q3+1.5IQR,where
IQR=Q3−Q1.\operatorname{IQR}=Q_3-Q_1.IQR=Q3−Q1.The calculation identifies a potential outlier. The context helps you discuss it. An unusual weather observation, for example, could be genuine, incorrectly recorded or caused by exceptional conditions. You should not automatically delete it.
Worked example: comparing two locations
During her revision, Maya compares a measure recorded at two locations. She is given these summary statistics:
| Location | Median | Interquartile range |
|---|---|---|
| A | 14.214.214.2 | 3.63.63.6 |
| B | 12.812.812.8 | 5.15.15.1 |
A clear comparison addresses both centre and spread:
- Location A has the higher median because 14.2>12.814.2>12.814.2>12.8.
- Location A has the smaller interquartile range because 3.6<5.13.6<5.13.6<5.1.
- Therefore, values at Location A are typically higher and less variable across the middle 50%50\%50% of observations.
Notice what Maya does not claim. She does not say every value at A is higher, because medians do not establish that. She also does not call A more consistent without identifying the measure supporting that statement.
A useful comparison has the shape statistic, comparison, context. That habit often separates a complete interpretation from a vague observation.
Worked example: identifying an outlier
Suppose a sample has
Q1=7.4,Q3=12.2.Q_1=7.4,\qquad Q_3=12.2.Q1=7.4,Q3=12.2.First calculate the interquartile range:
IQR=12.2−7.4=4.8.\operatorname{IQR}=12.2-7.4=4.8.IQR=12.2−7.4=4.8.The lower boundary is
7.4−1.5(4.8)=0.2,7.4-1.5(4.8)=0.2,7.4−1.5(4.8)=0.2,and the upper boundary is
12.2+1.5(4.8)=19.4.12.2+1.5(4.8)=19.4.12.2+1.5(4.8)=19.4.Values below 0.20.20.2 or above 19.419.419.4 are potential outliers under this rule. A value of 20.120.120.1 would therefore be flagged.
In a mock, keep the conclusion measured. The rule suggests that the value is an outlier; it does not prove that the observation is an error.
Revise sampling and context together
Sampling questions become easier when you picture the data collection process. A sample taken from one location, one month or one time of day may not represent the wider population.
Be precise about the limitation. Saying “the sample is biased” is weaker than explaining why. For example, observations from a summer month may not represent conditions throughout the year because seasonal patterns are excluded.
You should be comfortable with:
- simple random sampling;
- systematic sampling;
- stratified sampling;
- opportunity and quota sampling;
- the difference between a census and a sample.
When evaluating a method, consider coverage, convenience, non-response, sample size and whether every relevant group has a fair chance of representation.
Correlation needs cautious language
If a scatter diagram slopes upwards, the variables may have positive correlation. If it slopes downwards, they may have negative correlation. But correlation alone does not show that one variable causes the other.
The product moment correlation coefficient satisfies
−1≤r≤1.-1\leq r\leq 1.−1≤r≤1.Values near 111 indicate strong positive linear correlation, while values near −1-1−1 indicate strong negative linear correlation. Values near 000 indicate weak linear correlation, although a non-linear relationship may still exist.
In a large data set question, bring your interpretation back to the named variables. Avoid writing only “strong positive correlation”. Explain what tends to increase as the other variable increases, without turning association into causation.
A practical plan for A-Level Maths mocks
Maya eventually divides her preparation into three short sessions rather than one long rereading exercise.
First session: map the data
Write down the context, variables, units, locations, time periods and special conventions. Check everything against your course materials.
Second session: practise statistical techniques
Complete questions involving summary statistics, box plots, histograms, correlation, regression, sampling and hypothesis tests. Mark each answer carefully and record whether the error came from knowledge, calculator use or interpretation.
Third session: practise under mock conditions
Build a targeted paper using MathsGenie's Mock Exam Builder. Select the areas that expose your weakest statistical skills rather than repeatedly choosing comfortable questions. Complete it without notes, then use the mark scheme or video support to correct your reasoning.
This is where revision becomes useful. A targeted paper gives you evidence about what you can do independently, which is what the mock will eventually ask of you.
Common mistakes in large data set revision
Trying to memorise every value
Learn the structure, context and notable features instead. Exact values needed for calculations will normally be supplied in the question.
Revising another exam board's data set
Large data sets are specification-specific. Confirm your exam board before using summaries or practice materials.
Ignoring units
A numerically correct answer can still be incomplete or misinterpreted when its unit is missing.
Comparing only averages
When comparing distributions, discuss both centre and spread. Use named statistics such as the median and interquartile range.
Treating correlation as causation
Association does not by itself establish a causal mechanism. A third variable may influence both quantities.
Giving generic contextual comments
“The data may be unreliable” says little. Identify the relevant location, season, sample, measurement process or missing group.
Turn familiarity into marks
By the evening before the mock, Maya still cannot recite the spreadsheet. She does not need to. She can explain what the variables mean, recognise the units, choose suitable statistical techniques and interpret results in context.
That is the real goal of large data set revision for the mock. Familiarity reduces the time spent decoding the question, leaving more attention for the mathematics.
Use MathsGenie's free A-Level revision resources to strengthen the underlying statistics, then build a targeted practice paper in Mock Exam Builder. Practise what you are least comfortable with, mark it honestly and return to the exact step that cost you marks. Mocks become less mysterious when revision produces evidence, one corrected question at a time.


