
Statistics – Averages and Analysis
Averages tell you the 'typical' value in a set of numbers — but every average lies a little. Grouping data hides individual values, choosing the wrong average lets an outlier fool you, and misreading a box plot loses easy marks. Almost every question in this topic is testing whether you spot which of those three traps is set for you.
Overview
This chapter has one job: turn a messy pile of numbers into a single sentence someone can act on. To do that you need to organise data (grouping), summarise it (mean, median, mode), describe how spread out it is (range, IQR, box plots), work backwards from an average to fill in gaps, and finally compare two sets fairly.
- Raw data → organised into class intervals when there's too much of it to read at a glance.
- Central tendency (mean, median, mode) answers 'what's typical?' — but each answers it differently.
- Spread (range, IQR) answers 'how consistent is it?' — a missing half of the story if you only quote an average.
- Missing-value questions just run the mean formula in reverse: know the average and the count, and subtraction does the rest.
The shape of the chapter
Command terms that decide your mark
| Command term | What it demands | AO | Mark-earning move |
|---|---|---|---|
| Identify | Pick out or name one specific value from the data or context. | AO1 | The exact correct interval/value gets the mark — no explanation needed, but vague answers ('the third one') don't count as precise. |
| State | Give a short, precise answer with no justification required. | AO1 | A one-line factual answer earns full marks; extra explanation earns nothing extra and wastes time. |
| Describe | Give a detailed account of how or why something happens, in full sentences. | AO2 | Naming a feature with no mechanism = 0. You must explain WHY (e.g. 'the interval hides repeated values so the exact mode is lost') to get the mark. |
| Calculate | Obtain a numerical answer, showing the working. | AO1/AO2 | Method marks are awarded for correct formula/substitution even if the final arithmetic slips — an answer with zero working risks losing marks if it's wrong. |
| Compare | State similarities AND differences between two data sets, referencing both. | AO3 | Describing only one data set's feature without linking it to the other set loses the comparison mark entirely. |
Key point
Working with Grouped Data
Raw data is every value listed separately — accurate, but unreadable once you have more than about 15 numbers. A grouped frequency table sorts values into class intervals ('buckets' of equal width) so the overall pattern becomes visible at a glance. That convenience always costs something: once data is grouped, the exact original values are gone for good.
- Class interval: a range of values, e.g. –, that groups several raw data points together under one label.
- Class width: how wide each bucket is — for – the width is ; found from the gap between consecutive lower boundaries.
- Midpoint: the single value used to represent every data point inside an interval, calculated as .
- Frequency: how many data values fall inside that interval — this is what you count, never the raw scores themselves.
- Choosing where intervals start: pick the largest multiple of the width that is at or below the minimum value in the data, not the minimum value itself. That's why intervals of width 10 for data starting at 12 begin at 10 — it keeps every interval a clean, comparable width.
- Class interval
- A range of values used to group several individual data points together, e.g. –.Also called a 'class' or 'bucket'.
- Class width
- The size of a class interval, found as (upper boundary − lower boundary) or the gap between consecutive lower limits.
- Midpoint
- The value , used to represent every value inside an interval when estimating statistics from grouped data.
- Grouped frequency table
- A table showing class intervals alongside the frequency (count) of data values falling in each one.
From raw data to a grouped frequency histogram
Each bar's height is the frequency for that class interval; bars touch because the intervals are continuous.
Worked example
Grouping ten scores and judging the interval width
Boundary confusion
Putting a borderline value into the wrong interval, or assuming intervals must start exactly at the smallest raw value in the data set.
Consequence: Produces a wrong frequency count, which then cascades into a wrong midpoint or a wrong 'interval with highest frequency' answer later in the same question.
Fix: State the interval convention out loud before counting ('0–9 means 0 up to and including 9'). Choose the start point as the largest multiple of the width that is ≤ the minimum value, then count on from there.
Examiner tip
Key takeaways
- Class width = gap between consecutive lower boundaries (or upper − lower of one interval).
- Midpoint = — this is the value used later to estimate the mean from grouped data.
- Grouping trades exact accuracy for a clearer overall pattern — say this explicitly to earn 'reason/limitation' marks.
- Interval width choice depends on data quantity: too narrow with too little data creates empty or single-value intervals.
Finding Missing Data with Given Averages
Every 'find the missing value' question is really the mean formula run backwards. If you know the mean and how many values there are, you know the total sum — from there, finding a missing value is just subtraction.
- Rearranged mean formula: — total comes first, individual values come second.
- If one value is missing: missing value .
- If a frequency is missing in a grouped table but the overall mean is given, solve for the unknown frequency.
- Always recount to include the unknown value or interval — using only the count of known values is the single most common slip here.
mean equals the sum of all values divided by how many values there are
- — the mean (average)
- — sum of all data values
- — total number of values
rearranged form used to find a missing value or missing frequency
This is the version you actually use in 'find the missing value' questions — memorise this rearrangement, not just the original formula.
Worked example
Finding a missing test score
Using the wrong $n$
Multiplying or dividing by the number of KNOWN values (e.g. 5) instead of the TOTAL number of values including the unknown one (e.g. 6).
Consequence: Produces a total sum that's too small, so the missing value comes out wrong even though every individual step 'looks' correct — full method marks are still lost because the examiner checks .
Fix: Before touching the formula, write first and circle it, then use that same throughout the whole calculation.
Key takeaways
- Missing-value questions are backward mean questions: total first, subtraction second.
- Always use the total count , including the unknown value — never just the count of known values.
- The same backward logic extends to missing frequencies in grouped tables using .
Box Plots and Interquartile Range
A box plot compresses an entire ordered data set into five landmark numbers: the minimum, the lower quartile (), the median, the upper quartile (), and the maximum — the 'five-number summary'. The box plot is simply a picture of these five values.
- Order the data first, then find the median — this splits the data into a lower half and an upper half.
- = median of the lower half of the data (excluding the overall median itself if is odd).
- = median of the upper half of the data (same exclusion rule applies).
- The box stretches from to ; the whiskers extend out to the minimum and maximum values.
- Lower quartile ()
- The median of the lower half of an ordered data set — 25% of values lie below it.
- Upper quartile ()
- The median of the upper half of an ordered data set — 75% of values lie below it.
- Interquartile range (IQR)
- ; the width of the middle 50% of the data, a measure of spread that ignores extreme values.
- Five-number summary
- Minimum, , median, , maximum — the five values a box plot displays.
interquartile range — the width of the middle 50% of the data
- — upper quartile
- — lower quartile
A smaller IQR means the middle half of the data is tightly bunched; a bigger IQR means the middle half is spread out.
Box plot with five-number summary labelled
The box covers the middle 50% of the data (Q1 to Q3); whiskers reach out to the minimum and maximum.
Worked example
Five-number summary from 11 ordered values
Examiner tip
Key takeaways
- The five-number summary (min, , median, , max) is exactly what a box plot draws.
- Find the overall median first — it splits ordered data into lower and upper halves.
- With odd , exclude the middle value from both halves before finding and .
- IQR measures the spread of the middle 50%, ignoring extreme outliers.
Calculating Mean, Median, Mode and Range
Every one of these four measures needs the data ordered first — mode is the only one that technically doesn't require ordering, but ordering makes it faster to spot repeats anyway. Get into the habit of sorting before you calculate anything.
- Mean: add every value, divide by how many values there are — uses all the data, so a single extreme value drags it up or down.
- Median: order the data; for odd it's the middle value, for even it's the average of the two middle values.
- Mode: the most frequently occurring value — a data set can have no mode, one mode, or several (bimodal/multimodal).
- Range: maximum − minimum — the simplest spread measure, but it only uses two values, so one outlier can distort it completely.
estimated mean for grouped data, using interval midpoints
- — midpoint of each class interval
- — frequency of each class interval
This is only an ESTIMATE of the true mean — it assumes every value inside an interval sits exactly at the midpoint, which is never exactly true.
Worked example
All four measures from one data set (even n)
Frequently confused
| Concept A | Concept B | The distinction that matters |
|---|---|---|
| Mean | Median | Mean uses every value and gets dragged toward outliers; median only depends on position in the ordered list, so extreme values barely move it. |
| Mode | Median | Mode is whichever value appears most often (there can be none, one, or several); median is always exactly one positional value and exists for every numeric data set. |
Mnemonic
My Mother Makes Rice
Mean — add everything up, divide by how many. Median — order it, find the middle. Mode — the most frequent value. Range — the reach from smallest to biggest.
Forgetting to order data before finding the median
Reading off 'the middle value' from the data exactly as it was originally listed, without sorting it first.
Consequence: Gives a completely wrong median even though the method (find the middle position) looks correct — an easy mark thrown away for one missed step.
Fix: Always rewrite the data in ascending order first, even for a 'quick' mental question — write it down on paper, don't just eyeball the list.
Key takeaways
- Order the data before calculating median (and it helps for mode too).
- Even : median is the average of the two middle values, not either one alone.
- Mode can be absent, single, or multiple — it's the only measure that doesn't have to exist as one number.
- Range uses only the two extreme values, so it's the spread measure most sensitive to outliers.
Comparing and Interpreting Averages
A single number is never the full story. Exam questions in this topic almost always ask you to compare TWO data sets — two classes, two teams, two months — using an average AND a spread measure together, then explain what that combination means in context.
- Compare the averages first (mean or median) to say which set is 'typically' higher, faster, or better.
- Then compare the spread (range or IQR) to say which set is more consistent or reliable.
- A data set can have a higher mean but also a higher spread — 'better on average' but less consistent. Check both before making a claim.
- Mean is dragged around by outliers; median resists them — so median is the fairer average whenever a data set contains extreme values.
Which average should you trust?
| Measure | Best used when... | Affected by outliers? |
|---|---|---|
| Mean | Data is roughly symmetric with no extreme outliers | Yes — a single extreme value pulls the mean toward it |
| Median | Data contains outliers or is skewed | No — only the position in the ordered list matters |
| Mode | Data is categorical, or you need the single most common value | No — but it can mislead if the most common value isn't near the centre |
Comparing two data sets — the checklist
| Step | What to do | What it tells you |
|---|---|---|
| 1 | Compare mean or median | Which set is 'typically' higher or lower |
| 2 | Compare range or IQR | Which set is more spread out / less consistent |
| 3 | Check for outliers | Whether the mean is trustworthy, or the median is the fairer summary |
| 4 | Write one sentence linking both numbers to the context | Turns raw numbers into a real conclusion — this is the mark examiners are actually looking for |
Real-world application
Key takeaways
- Always compare an average AND a spread measure — one number alone never fully answers a 'compare' question.
- State explicitly which data set is higher/more consistent, linked to the actual numbers, not just the numbers on their own.
- Outliers make the mean unreliable; the median resists them — choose the fairer average and justify why.
Quick Reference
Averages and Analysis — everything on one page
Formulas
- Mean: sum of values ÷ count
- Backward form — use for missing-value questions
- Estimated mean from grouped data (m = midpoint, f = frequency)
- Spread of the middle 50% of ordered data
- Simplest spread measure — most sensitive to outliers
Key facts
- Median position (odd n)
- -th value
- Median (even n)
- average of the two middle values
- Class width
- upper boundary − lower boundary of one interval
- IQR
- measures spread of the middle 50%, ignores outliers
- Range
- measures full spread, very sensitive to outliers
Mnemonics
- My Mother Makes RiceMean (add, divide) — Median (order, middle) — Mode (most frequent) — Range (max minus min)
Don't confuse
- Mean vs Median: Mean uses every value and is pulled by outliers; median only cares about position, so it resists them.
- Q1/Q3 of a box plot vs Median of the whole data set: The overall median splits data in half; Q1 and Q3 are the medians of those two halves (excluding the overall median if n is odd).
Command terms
- Identify:
- Name the exact interval/value — no reasoning needed.
- Describe:
- Explain the mechanism (how/why), not just name a feature — bare naming scores zero.
- Calculate:
- Show the working; method marks survive an arithmetic slip, a bare answer doesn't.
- Compare:
- Reference BOTH data sets explicitly, using both an average and a spread measure.
Exam checklist
- Order the data before finding median, quartiles, or checking for repeats (mode).
- For missing-value problems: find total = mean × n (correct n!) before subtracting.
- For grouped data: midpoint = (lower+upper)/2; class width = gap between consecutive lower boundaries.
- For box plots: exclude the overall median from both halves when n is odd, before finding Q1 and Q3.
- For 'compare' questions: quote an average AND a spread measure, then write one linking sentence.