You will be able to spot chart choices that exaggerate or hide a change.
Priya is an analyst at a mid-sized insurer in Singapore. In a quarterly review, a slide comes up comparing customer satisfaction for her company and a competitor. Her company's bar is less than half the height of the competitor's. The room goes quiet. Priya looks at the left edge of the chart and sees the axis starts at 80. The scores are 82 and 85.
Nobody lied on that slide. The numbers were right. The chart still told a story the numbers did not support. This lesson covers four chart choices that do this, often by accident, and the questions that catch them.
A bar chart shows size by the length of the bar. Your eye compares lengths, so the bars have to start from zero for the comparison to be fair.
Priya's slide shows why. With an axis from zero, a bar for 82 and a bar for 85 would look almost the same, because 85 is only about 4 percent more than 82. With the axis starting at 80, the bars show only the part above 80: 2 units for her company and 5 for the competitor. Now one bar is two and a half times the other, and a three-point gap looks like a crisis.
The same trick works in reverse. Stretch the axis far above the data, say from 0 to 500 for scores in the 80s, and real differences flatten into nothing.
Line charts are different. A line shows change over time, and starting the axis at zero can squash a meaningful movement into a flat line. A line chart of a share price or a monthly error rate often starts near the data, and that can be fine. What matters is that the axis is labelled clearly, so the reader knows how much change a given slope stands for. For bars, the rule is firm: start at zero, or do not use bars.
The habit is simple. Before you react to any chart, read the axis labels. It takes two seconds and catches most of the problems in this lesson.
Some charts put two lines on the same picture with a different vertical axis on each side. Sales on the left, say, and marketing spend on the right. The two lines rise and fall together, and the message seems obvious: spend drives sales.
The problem is that the person making the chart chose both scales. By stretching or squeezing either axis, they can make almost any two lines that trend upward appear to move in step. The overlap you see is partly a design decision.
When you meet a dual-axis chart, ask to see each line on its own chart with its own axis, or ask for the actual numbers. If the claim is that one thing drives the other, the chart is not the evidence. That question belongs to Module 4, on correlation and causation.
Every chart over time has a first and last point, and someone picked them. Choose a start date at a low point and almost anything looks like a rise. Start at a peak and it looks like a fall.
Imagine a chart of a team's monthly sales that runs from March, just after a slow Chinese New Year period, to November, and climbs steadily the whole way. Extend it to three full years and you may find the same shape in every one of them, a dip after the new year and a climb towards December, in which case what looked like growth was only the calendar.
So ask: why does the chart start here? What would it look like over a longer window? If the answer is that earlier data does not exist, fine. If the earlier data exists and was left off, the chart is telling you less than it seems to.
Totals grow with population. A chart showing that one town has more burglaries than another, or that one region of a company has more customer complaints, may only be telling you which is bigger.
Here is a made-up example. Town A has 300 cases of something in a year and a population of 150,000. Town B has 400 cases and a population of 400,000. A bar chart of totals makes Town B look worse. Per 100,000 people, Town A has 200 cases and Town B has 100. Town A's rate is twice as high.
Whenever you see totals compared across groups of different size, or across years when the population changed, ask whether a per person figure would tell a different story. The same goes for anything that grows with size: total sales across branches of different sizes, total complaints across products with different numbers of users, total road accidents as more cars come onto the roads.
Priya did not accuse anyone. She said, "Just to check the scale, this axis starts at 80, so we're looking at 82 against 85. Do we know if that gap is bigger than the usual movement from quarter to quarter?" The room relaxed, and the discussion moved to the question that mattered: whether three points was a real difference or ordinary noise.
That is the pattern for every chart. Check the axis, check the window, ask whether it should be per person and be careful with two axes. Once those four checks are done, the discussion can move on to what the numbers themselves say.
You learn this fastest by redrawing a chart yourself and seeing how different it looks. In the activity below you will find one chart in the news or a company report and rebuild it in a spreadsheet with a zero baseline and a longer time window.
Find a chart in a news article or company report and redraw it in a spreadsheet with a zero baseline and a longer time window.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).