QCE General Mathematics - Unit 2 - Univariate data analysis 2

Parallel box plots, outliers and group comparisons

Learn comparing distributions for QCE General Mathematics Unit 2 with worked reasoning, KaTeX, an original diagram and checked practice.

Part of the free QCE General Mathematics notes library for Unit 2: Univariate data analysis 2.

Updated 2026-08-09 - 8 min read

Compare groups through centre, spread, overlap, outliers and context rather than one isolated statistic. The reliable habit is to translate the situation into labelled quantities and units before choosing a calculation. General Mathematics is practical, but “practical” does not mean informal: the model, assumptions, technology and conclusion must remain visible enough for another person to audit.

Build the mathematical model

Parallel box plots place distributions on a common scale so medians, IQRs, ranges and possible outliers can be compared. A defensible comparison states both similarity and difference and links them to the observed groups without claiming causes the data cannot support.

Start with the question rather than the formula. Identify what must be found, which information is relevant and which conditions limit the model. Give every variable a meaning and unit. If the context contains thresholds, categories, time periods, row and column labels, geometric joins or different possible domains, mark those features before calculating. This prevents a familiar procedure from being applied to the wrong quantity.

Represent the situation in at least two ways where possible: words and an equation, a table and a graph, a labelled diagram and a formula, or a matrix and its row-column labels. Agreement between representations is useful evidence. Disagreement is a signal to stop and locate the first modelling or transcription error.

Governing relationships

$ \operatorname{IQR}=Q_3-Q_1 $

$ \text{lower fence}=Q_1-1.5\operatorname{IQR} $

$ \text{upper fence}=Q_3+1.5\operatorname{IQR} $

Write the relationship before substituting. Keep exact values or full calculator precision through intermediate steps, then round the final result to a precision justified by the supplied data and context. Currency normally requires cents, measured quantities should not imply unsupported accuracy, and counts may require a whole-number interpretation rather than ordinary decimal rounding.

Units are part of the reasoning. A rate combines two units; area uses square units; volume uses cubic units; a gradient has output units per input unit; a matrix entry inherits the labels of its row and column. If the units do not match the requested quantity, a correct-looking decimal is not a correct answer.

Make the concept connections

1. Interpret the quantities

The box spans the middle 50% and the median divides the ordered data. Whiskers follow the chosen box-plot convention to the extreme non-outlier values; possible outliers are plotted separately.

2. Connect the representations

A higher median indicates a higher typical value, while a larger IQR or standard deviation indicates greater variability. Overlap does not erase a centre difference, and separation does not prove a causal group effect.

3. Protect the conditions

Mean/standard deviation and median/IQR comparisons may emphasise different features. Report exact supplied values and note sample size, skew and unusual observations where relevant.

The diagram above is a reasoning tool. Recreate its essential labels from memory and explain what remains fixed, what changes and how the representations connect. Then alter one condition—a threshold, scale factor, graph domain, matrix order, outlier or measurement unit—and predict which part of the diagram and method must change.

A repeatable solution method

1. Calculate each five-number summary and IQR using a consistent quartile convention.
2. Find fences, classify possible outliers and construct plots on one numerical scale.
3. Compare centre, spread, range/overlap and outliers with quoted values.
4. Write a contextual conclusion limited to the observed data and collection design.

This sequence is deliberately explicit. A short-response solution can compress routine arithmetic, but it should not hide the model choice, units or contextual interpretation. Technology is valuable for repeated computations, graphs, matrix powers and statistical summaries; it does not decide whether the inputs, formula, interval, labels or conclusion are appropriate.

When using a spreadsheet or calculator, record enough working to reproduce the result: name the entered quantities, show the governing formula or command, retain the unrounded value and explain the displayed output. A screen value without interpretation is not evidence that the right question was answered.

Worked example

Cover the worked steps and reproduce the solution from the problem statement. Then change one quantity or condition and predict the direction and approximate size of the effect before recalculating. This counterfactual check distinguishes understanding from pattern copying and prepares you for unfamiliar questions.

Technology and representation audit

Technology should extend the mathematics rather than conceal it. For a spreadsheet, inspect copied references and test a corner cell. For a graph, read axes, units, scale, endpoints and domain. For a matrix, confirm dimensions and labels before entering the product. For statistics, confirm the intended variable, sample size and the correct calculator statistic. For measurement, sketch the boundary or exposed faces before using stored formulas.

Use one independent representation as a check. A graph can check an algebraic intersection; substitution can check a graph reading; a table can check a formula pattern; an estimate or bound can check measurement; a manually expanded entry can check a matrix calculation; and an ordered list or plot can check a statistical summary.

Common mistake and repair

Repair: Name the measured variable and compare specific centre and spread statistics before giving any context-dependent evaluation.

Do not repair a conceptual error by adding decimal places. Find the first decision that broke the model: a wrong denominator, unmatched time period, omitted face, reversed scale direction, invalid graph interval, incompatible matrix dimension, unsuitable summary or unsupported interpretation. Rebuild from that point and preserve the parts that were valid.

Assessment transfer

When a possible outlier is removed only with justification, recalculate summaries and explain how the conclusion changes rather than silently deleting it.

For a short-response question, show the governing relationship, substitution, result with units and one contextual sentence. For a problem-solving or modelling task, make assumptions and observations explicit, justify the chosen representation, use technology for a meaningful purpose, evaluate reasonableness and limitations, and organise the response so it can be read independently of the task sheet.

A defensible conclusion answers the question at the strength supported by the evidence. Say “for this model” or “in this sample” when generalisation is limited. State thresholds and domains. Distinguish an exact calculation from an estimate and a possible outlier from an error. These qualifications improve mathematical communication; they do not weaken it.

Verification checklist

Recompute IQR and fences, ensure both box plots share a scale, and reconcile the written comparison with the displayed medians and boxes.

Before submitting, ask:

1. Did I define the unknowns, labels, units and valid domain?
2. Does my chosen formula, graph, table, matrix or statistic match the information structure?
3. Can I reproduce the result through substitution, a second representation, a bound or a spot check?
4. Is the rounding and format appropriate for money, measurement, count or data?
5. Does the final sentence interpret the result without claiming more than the model or data support?

Deliberate practice

1. Rework the example with one input increased by 20%. Predict the direction of change first.
2. Create a plausible but incorrect solution based on the common mistake above, then annotate the exact line where it fails.
3. Represent the same situation in a second form—diagram, graph, table, spreadsheet or matrix—and explain how the two forms agree.
4. Write a one-sentence reasonableness check that uses units, bounds, a reverse operation or a contextual constraint.
5. Design an unfamiliar example in which the usual method needs one extra decision, such as a threshold, join, outlier, domain restriction or reordered category.

Syllabus coverage

• Construct and use parallel box plots, including identifying possible outliers, to compare datasets in terms of median, spread and outliers and communicate the differences in context.
• Compare datasets in terms of mean, median, range, interquartile range and standard deviation, interpret the differences observed in context, and report findings systematically and concisely.

Sources

• QCAA — General Mathematics 2025 v1.3 syllabus
• QCAA — General Mathematics 2025 formula book
• Australian Bureau of Statistics — statistical concepts

Finished reading? Practise this topic free

Open General Mathematics past questions with this Unit 2 topic carried into the question bank, then save your progress for the next review.

Practise this topic free. Free to start. No payment details are required. Exact question coverage depends on the available past-paper syllabus mapping.