Chart mistakes that change the story—and how to catch them

Many misleading charts contain correct numbers. The problem lies in the scale, selection, labels, or wording surrounding those numbers. A graphic can quietly exaggerate a difference, imply a causal explanation, or compare incompatible measures without a single arithmetic typo. That is why a useful review checks the visual argument as well as the data table.

This guide is a practical audit for stock charts, finance graphs, and infographic cards. The examples use invented values. Compare them with the template gallery and the infographic publishing guide to see how clear units, restrained annotations, and explicit sample labels can prevent common misunderstandings.

A headline stronger than the evidence

Start the review with the title. Does it describe the quantity and relationship shown, or does it add an unsupported explanation? A chart of rising revenue can show a rise in the reported measure. It cannot, by itself, establish why the rise happened or that the same pattern will continue.

The Office for National Statistics’ data visualization principles emphasize clear comparisons and the relationship between charts and accompanying text. Use that idea as a practical test: the headline, plot, and surrounding explanation should discuss the same thing at the same level of certainty.

Replace dramatic wording with a precise observation when evidence is limited. “Final sample quarter exceeds the starting quarter” is less theatrical than “unstoppable growth,” but it can be checked from the display. Strong design does not require stronger claims than the data supports.

A cropped bar axis that exaggerates magnitude

Bars use length from a baseline to represent values. Suppose two categories are 95 and 100. On a zero-based scale, the second bar is slightly longer. If the visible baseline begins at 94, the lengths become one and six units, creating a much more dramatic visual ratio.

The printed labels can still be correct while the encoded lengths suggest the wrong relationship. A restricted range may be appropriate for another representation or an explicitly explained close-up, but it should not silently replace the baseline that makes a bar comparison meaningful.

Inspect every panel independently. A chart with a clear main axis can still contain a misleading inset or miniature comparison. Small dashboard cards deserve the same review as the hero graphic because readers often absorb their visual impression before noticing the fine print.

Inconsistent scales across neighboring charts

Two panels with different ranges can make a small movement look comparable to a large one. This may be useful when the purpose is to inspect each series’ shape, but it is risky when the layout invites direct comparison of magnitude. Decide which task the paired charts are meant to support.

Use common axis limits when shared-scale comparison is the goal. When independent scales are necessary, make them visible and explain the choice. A subtle tick difference is not always enough, especially when the panels are reduced to small images.

The same issue appears when one panel is linear and another logarithmic. Their slopes encode different relationships. Label the transformation clearly and avoid treating similar angles as equivalent changes. The stock chart reading guide explains why the scale should be inspected before the slope.

A missing denominator or comparison period

A percentage is incomplete without its reference. A share moving from 20% to 25% changes by five percentage points. Relative to the starting share, the increase is 25%. Confusing those descriptions changes the apparent size of the result.

A performance label such as “up 15%” also needs a period and method. Does it compare endpoints, summarize a completed interval, include distributions, or describe a hypothetical index? The chart should not force readers to invent those definitions from the surrounding branding.

Write the denominator and period in the subtitle or annotation near the highlighted value. Check that the table uses the same definition. A consistent label is especially important when a graphic is copied into a presentation where the original article’s longer explanation is no longer available.

A selected window that hides the relevant path

Every chart selects a period. Selection becomes misleading when the title implies a broader conclusion than that window supports. A tight crop around a rise can omit an earlier decline, while an endpoint comparison can conceal substantial movement between the two selected dates.

Decide whether the visible path is sufficient for the stated question. A short window is legitimate when the discussion concerns that interval, but label it as such. Offer a wider context view when it materially changes the interpretation, rather than relying on readers to request it.

Do not remove inconvenient observations merely to produce a cleaner trend. Missing records should be identified and handled according to a documented rule. The absence of an observation is different from a zero value, and a line connecting distant points can imply continuity that the dataset does not establish.

Incompatible measures presented as peers

A balance curve and a return curve can both rise, but the change in a balance may include deposits or withdrawals. A price-only index and a total-return index may follow different definitions. Similar shapes do not make those quantities directly comparable.

Read the source field definitions before building the chart. Check currency, units, time coverage, adjustment treatment, and aggregation. When comparison requires a transformation, explain it. Rebasing series to 100 changes their displayed starting levels; it does not erase differences in what they measure.

The finance graph guide includes worked examples of these distinctions. A good audit asks whether the relationship remains valid after the labels are read carefully. If the comparison depends on ignoring the labels, the visual design is probably doing too much interpretive work.

Color and decoration mistaken for information

Bright colors, thick arrows, and glowing paths can imply urgency or importance that is not present in the values. A highlighted category should have a stated reason for being highlighted. A color change should communicate a defined distinction, not simply make an ordinary fluctuation feel consequential.

Do not rely on red and green alone for meaning. Add direct labels, patterns, or shape differences. Check whether the chart remains understandable in grayscale and at a small size. Decorative depth and perspective can also distort comparisons by changing the apparent size of identical marks.

Keep chart furniture subordinate to the data. Gridlines should help read the scale rather than compete with the series. A source note needs to remain visible even when it is less visually exciting than a headline. An attractive border cannot compensate for an unreadable axis.

A sample image mistaken for a real result

Synthetic data is useful for teaching layout and calculations, but it needs a persistent label. A note elsewhere on the site is insufficient when an exported image can circulate independently. Put “illustrative data” or an equally clear description inside the graphic itself.

Avoid using a real company’s name next to invented performance values. The combination can create a false factual claim even when the designer intended a harmless demo. Likewise, do not label a static example “live” or suggest that a quote is current when no market feed is connected.

ImageChart’s downloadable examples use fictional sequences and carry a sample-data note. Their purpose is to demonstrate visual forms and reproducible calculations. They are not market observations, portfolio recommendations, or evidence of investment performance. Preserve that distinction when adapting the layout for a new publication.

An export that loses its context

Before publishing, open the final file rather than trusting the editing preview. Confirm the title, unit, period, legend, and source note survived. Check that labels are not clipped, overlapping, or too small at the intended display size. A technically valid image can still fail as an explanation.

Compare the final graphic with the underlying values and the accessible text summary. Recalculate the highlighted difference. Verify that the correct date, series, and chart state were exported. A stale image from an earlier revision is an easy mistake when several similar files share a folder.

Keep an editable source and a record of the calculation. Then review the graphic as a reader would: first impression, main comparison, qualification, and verification path. The goal is not to make a chart look cautious. It is to make its apparent message match what the evidence actually supports.