ACARA v9 CONTENT DESCRIPTION “analyse claims, inferences and conclusions of statistical reports in the media, including ethical considerations and identification of potential sources of bias”
Builds on: Reading Surveys: Estimating the Mean and Median (AC9M9ST01). Knowing how data displays and summary statistics work is what lets you judge them. This unit turns that knowledge outward, to the statistics met every day in news, advertising and social media, and asks the critical question: does the evidence really support the claim?
Statistics as a tool for persuasion
Statistics surround us, in news headlines, advertisements, and social media posts, and they carry an air of authority: a number seems like a fact. But a statistic is only as trustworthy as the data and reasoning behind it, and the same figures can be presented to inform or to mislead. Learning to analyse statistical reports means refusing to take a claim at face value and instead asking what evidence supports it. Every claim should be linked back to three things: the data display used to show it, the actual statistic quoted, and whether the data was collected from a representative sample. When any of these is missing, hidden, or mismatched to the claim, that is the signal to be sceptical. This is not cynicism but numeracy: the skill of being a careful, questioning reader of numbers.
The truncated axis trick
Starting a bar chart axis above zero exaggerates small differences and can mislead the reader.
By starting the vertical axis at 100 instead of 0, a difference of just a few units is stretched to fill the chart, making near-equal values look wildly different. Reset the axis to zero to see what the data really shows.
From claim to evidence
The central habit of this unit is linking a claim to its evidence. When a headline announces that sales are soaring, ask to see the graph, and check whether its axis starts at zero or has been truncated to exaggerate a modest rise. When an advertisement says nine out of ten experts recommend something, ask how many were asked and how they were chosen. When a report quotes an average saving, ask whether it is the mean or the median and over which group. A claim that can be traced cleanly to a fair display, a clearly stated statistic, and representative data deserves consideration. A claim whose evidence is absent, vague, or does not actually match what is being asserted does not. Making this link, every time, is what separates an informed reader from a persuaded one.
Link the claim to its evidence
A sound claim can be traced to a clear display, a stated statistic, and representative data.
The heart of analysing a statistical report is linking every claim to the evidence behind it: the data display, the actual statistic, and how the sample was gathered. A claim with no inspectable evidence, or evidence that does not match the claim, is a claim you should not yet believe.
Which average, and why it matters
One of the most common ways statistics mislead is through the word average, because it can mean the mean, the median, or the mode, and these can differ greatly. On data with an outlier, such as incomes where most are modest but a few are very large, the mean is pulled upward by the extreme values and ends up far above what a typical person earns, while the median stays among the usual figures. A report wanting to make incomes sound high might quote the mean; one wanting an honest picture would give the median. Neither is wrong as a calculation, but choosing which to present, and not saying which it is, can shape the impression entirely. A careful reader always asks which average is being used and whether it is the appropriate one for the data.
Mean or median: which average?
On skewed data with an outlier, the mean can mislead while the median better reflects the typical value.
Most values cluster low, but one outlier towers above. Reveal the mean and median to see how a single extreme value can make the mean misleading.
Was the sample representative?
A statistic about a whole population can only be trusted if it came from a sample that fairly represents that population. If a survey about national exercise habits questions only people leaving a gym, its sample is biased, and any conclusion about everyone is unreliable, however large the survey. Representative sampling means every part of the population has a fair chance of being included, so the sample mirrors the whole. Many misleading reports rest on a sample that quietly excludes important groups, or that selects people likely to give a particular answer. Whenever a claim generalises from a sample to a population, the question to ask is simple but powerful: who was actually included, and who was left out?
Is the sample representative?
A claim about a population only holds if the sample fairly represents that whole population.
A representative sample is spread across the whole population, so its results can fairly be generalised. Switch to a biased sample to see how a lopsided selection breaks that link.
Correlation is not causation
Perhaps the most important caution in all of statistics is that two things moving together do not prove that one causes the other. Ice cream sales and drowning rates rise and fall in step across the year, but ice cream does not cause drowning; both are driven by hot weather, a lurking variable behind the scenes. Reports frequently spot a correlation and leap to a causal claim, that one thing makes another happen, when a hidden third factor, or even pure coincidence, is responsible. Establishing genuine cause requires careful study, usually a controlled experiment, not merely a pattern in observed data. So when a report says one thing causes another, the sharp question is whether they have truly shown cause, or only found that two things tend to occur together.
Correlation is not causation
Two things moving together need not be cause and effect; a third, lurking variable may drive both.
Ice cream sales and drowning rates rise and fall together, a strong correlation. Does ice cream cause drowning? Reveal the lurking variable before concluding.
Quick self-check
1. A bar chart starts its vertical axis at 100 rather than 0. This can:
2. To analyse a statistical claim in the media, you should:
3. Most incomes in a group are modest, but one person earns a fortune. The more honest "average" to report is usually the:
4. A survey asks only gym members whether people exercise enough, then reports a result for everyone. The problem is:
5. Ice cream sales and drowning rates are strongly correlated. The best conclusion is: