The short answer

read charts with truncated axes or missing baselines

Read both axes, units, start date and denominator before judging the shape. Replot from the underlying values when visual scale creates a stronger impression than the data supports. Charts are among the most shareable forms of political content on social media, in party advertising and in news coverage, but they are also among the easiest to manipulate without altering a single data point. By cropping the y-axis to start above zero, selecting a narrow date range that omits relevant history, switching between absolute numbers and percentages, or overlaying two unrelated data series on the same chart, a creator can produce a graph that is technically accurate in its plotted values while being visually deceptive about what those values mean. Learning to read a chart sceptically means checking the axes, the units, the time period and the source before accepting the visual impression. This skill applies whether the chart was published by a political party, a news organisation, a government agency or an individual on a social media platform.

This guide helps verify public political material. It does not infer a person’s motive from one post and does not treat disagreement as proof of deception.

The useful question is not only “what is the rule?” but also “who administers it, which document controls it, and when might it change?” That distinction prevents an accurate general explanation from becoming wrong advice in a particular election, chamber or policy setting.

Evidence review

Why the axes matter more than the shape

The visual impact of a line or bar chart comes from the slope of the line and the relative height of the bars, both of which are determined by the scale and range chosen for the axes. When a chart creator selects a y-axis range that starts above zero, even a small absolute change can appear as a dramatic spike. In political debate this technique is commonly used to make modest polling shifts, budget changes or crime statistics look like emergencies. The data points on the chart may be perfectly accurate, but the visual impression is exaggerated because the viewer's eye naturally judges the difference relative to the full height of the bars rather than relative to the absolute values.

The x-axis is equally important. A chart that shows a time series starting at a carefully chosen date can create a misleading impression of a trend. A government might plot unemployment from the low point of the economic cycle, making the subsequent rise look steeper than it would if the chart began from the previous peak. An opposition might plot from a peak, making a recovery look flat. The choice of date range is a framing decision that is not neutral, and the responsible reader should ask what happened before the chart's start date and what the longer-term trend looks like. The same applies to bar charts that compare categories: if the categories are sorted in an order that produces a visual pattern, ask whether the sorting is meaningful or whether it is designed to create an impression of ranking.

Evidence review

Truncated y-axes and exaggerated differences

A truncated y-axis is one that does not extend to zero or to the natural floor of the data being plotted. On a bar chart this is a well-known problem because the convention is that bar charts should usually start at zero because the length of the bar represents the value. When the y-axis starts at a non-zero value, the bar lengths no longer represent the values accurately and a bar that is twice as tall as another bar may not represent a value that is twice as large. A government spending chart that shows spending rising from 98 billion to 100 billion will show a 2 billion increase, but if the y-axis starts at 97 billion the bars will look like a dramatic doubling rather than a 2 percent change.

Line charts present a more nuanced case. In a line chart the viewer is meant to read the slope and the relative position of points, not the area filled by the line. Because of this convention, line charts that do not start at zero are not always misleading. A stock market index chart would be unreadable if it had to start at zero because the daily changes are a tiny fraction of the total index value. The test is whether the truncation materially changes the visual impression relative to what the data actually shows. If a 1 percent change looks like a 50 percent change because of the axis range, the chart is misleading regardless of whether the underlying data is correct. If the axis is chosen to show the meaningful variation in the data without distorting the relative scale, it may be a perfectly reasonable chart. The reader should check the axis labels, calculate the actual percentage change from the numbers and compare that to the visual impression the chart creates.

Evidence review

Missing baselines and percentage claims

A chart that reports a percentage change without showing the baseline from which the change is calculated is often hiding important information. A 50 percent increase in the number of complaints about a government service sounds alarming, but if the baseline is 2 complaints rising to 3 the absolute change is trivial. Percentage charts are frequently used in political advertising and social media because the percentage number creates a sense of scale that may not match the real-world significance of the change. The missing-baseline problem also appears when a chart compares two quantities with very different denominators. A chart showing that Party A received 40 percent more donations than Party B may omit the fact that Party A's donation total was 40,000 dollars while Party B's was less than 30,000 dollars, which is a material context for interpreting whether the difference reflects political momentum or a small absolute gap.

Another variation is the chart that compares rates without showing the population base. A map or bar chart showing crime rates by suburb may look as though some suburbs are dramatically more dangerous than others, but if the suburb with the highest rate has a population of 500 people, one additional incident can move the rate sharply. Charts of this kind are common in election-campaign communication about local issues and should be checked against the underlying count data and population figures. The Australian Bureau of Statistics publishes population estimates that can be used to convert rates back to approximate counts, and the source organisation of any crime, health or education statistic should be able to provide the absolute numbers on request. If the source will not provide the absolute numbers, that is itself a reason to treat the chart with caution.

Evidence review

Time-series charts with cherry-picked date ranges

Political communicators select date ranges to support a narrative, and the same underlying data can produce very different-looking charts depending on where the time axis begins and ends. A government wanting to show economic improvement might start a chart of employment figures from the trough of the last recession, while the opposition might start the same chart from the pre-recession peak to show how much ground has not been recovered. Neither chart is necessarily lying about the data, but both are using a framing choice to suggest a conclusion that the unselected data might not support. The reader's defence is to ask what the chart would look like if the time axis were extended backward and forward by a reasonable period and whether the trend holds.

Discontinuities in the time axis are another common technique. A chart may show data points for January, February and then jump to September, creating the visual impression of a smooth trend when in reality there is a six-month gap in which the trend may have changed direction. The gap may be marked by a break symbol on the axis or it may not be marked at all. A chart that uses irregular intervals on the x-axis without clearly labelling them is almost always misleading because the human eye interprets equal physical distance as equal time distance. When you see a line chart, check whether the spacing between data points is uniform and whether there are any unexplained gaps or compressions in the timeline. If the data is quarterly, the points should be equally spaced regardless of whether the quarters contain the same number of days.

Evidence review

Dual-axis charts and misleading comparisons

The dual-axis chart plots two different data series with two different y-axes, one on the left and one on the right. This is a legitimate statistical technique for comparing series with different units, such as temperature and rainfall, but in political communication it is frequently abused to suggest a relationship between two variables that may not exist. By adjusting the scales of the two axes independently, a chart creator can make any two rising or falling series appear to move in lockstep. If the left axis for tax revenue runs from 400 billion to 500 billion and the right axis for something unrelated runs from 30 to 40, the two lines can be made to overlap perfectly despite having no causal connection.

The test for a dual-axis chart is whether the relationship it implies is supported by the data or is an artefact of the scale choices. Ask whether there is an independent reason to believe the two variables are related, whether the chart's author provides a correlation coefficient or statistical test and whether the relationship holds if you adjust the axis scales to reflect the natural range of each variable. If changing the axis scales changes whether the lines appear to track each other, the visual relationship is arbitrary and the chart should not be trusted as evidence of causation or even of meaningful correlation. A more honest presentation would show each series on its own chart with a consistent y-axis scale and let the reader compare the shapes without the visual trickery of independent axis scaling.

Evidence review

Practical checks to apply before sharing a chart

Before you accept, quote or share a chart that makes a political argument, apply a set of standard checks. First, identify the source of the data: is it from a government statistical agency, an academic study, a party research unit or an advocacy group? Is the original data accessible so you can verify the plotted values? Second, read every label on the chart: what are the units, what is the time period, what do the colours represent and is there a note about the source or methodology? Third, check the axes: does the y-axis start at zero or at a natural baseline, and if not, does the truncation materially exaggerate the visual difference? Fourth, look for missing context: what happened before the chart's start date, what is the population or baseline behind the percentages, and are there alternative data sources that show a different picture?

Fifth, mentally replot the data: if the chart shows an increase, calculate the absolute change and compare it to the visual impression. If a bar appears twice as tall as another bar, check whether the value is actually twice as large. If the chart shows a correlation between two lines, ask whether the scales have been manipulated to create that impression. Sixth, consider the publisher's intent: is the chart presented as part of an argument that the publisher has a stake in winning? A chart from a political party's social media account is not fake merely because it comes from a partisan source, but it deserves a higher level of scrutiny than a chart from an independent statistical agency. Finally, if you cannot verify the chart with reasonable effort, do not share it. The spread of visually compelling but unverifiable charts is one of the mechanisms by which misinformation travels in political communication, and the simplest intervention is to pause and check before amplifying.

Common questions

Before you rely on the answer

Is a truncated y-axis always misleading?

Not always. In line charts, where the reader interprets the slope and relative position of points rather than bar lengths, starting the axis above zero can be legitimate if the goal is to display meaningful variation. The test is whether the truncation exaggerates the visual impression: if a 2 percent change looks like a 50 percent change, the chart is misleading regardless of chart type.

How can I check whether a chart is using real data?

Look for a source citation on the chart. If there is none, search for the claimed statistic on the website of the Australian Bureau of Statistics, the relevant government department or the parliamentary library. If the data cannot be found in any official source, treat the chart as unverifiable and do not share it. Charts sourced from the ABS, the AEC or the Parliamentary Budget Office are generally reliable as to the underlying numbers even if the presentation choices may still be worth questioning.

What is the difference between a truncated axis and a logarithmic scale?

A truncated axis cuts off part of the data range to zoom in on variation, which can exaggerate differences. A logarithmic scale compresses each order of magnitude into equal visual distance, which is a different mathematical transformation used to show proportional changes across very wide ranges. Logarithmic scales are standard in fields like epidemiology and economics and are not inherently deceptive, but they should be clearly labelled and the reader should understand that equal distances on a log scale represent equal multiples, not equal absolute amounts.

Should I distrust all charts shared by political parties?

You should apply the same verification checks to all charts regardless of source. Partisan charts are more likely to use presentation techniques that favour the publisher's argument, but government agencies, news organisations and non-partisan think tanks can also produce misleading charts through carelessness rather than intent. The checks described in this article apply universally: read the axes, verify the data, replot mentally and consider the publisher's interest.

Source spine

Primary material used for this guide

Review trigger: New ACMA or eSafety Commissioner guidance on identifying misleading visual content in political communication; legislative changes affecting misinformation and disinformation regulation in Australia; major platform policy changes affecting how charts and data visualisations are labelled or moderated on social media; publication of new Australian standards or style guides for data visualisation by statistical agencies.

Archive note: Source pages are maintained by ACMA, the eSafety Commissioner, the Parliament of Australia and the ABS. No specific legislation page governs chart design, but the ACMA and eSafety Commissioner websites contain relevant guidance on media literacy and misinformation. This article should be reviewed if new Australian standards for data visualisation are published or if platform moderation policies change how charts are presented and labelled.

Primary links are provided without affiliate or tracking parameters. Confirm that the source still applies to the bill, sitting date, jurisdiction or reporting period before relying on it.