What should I check when a bar chart’s vertical axis starts above zero?

Read the vertical axis before judging the bars. Note its lowest and highest labeled values and the values at each tick. Check whether the tick intervals are equal, and identify the baseline from which the bars are drawn. A baseline above zero means the visible bar lengths represent only the amounts above that starting point, not the full values from zero. OpenStax cautions that changing an axis scale can substantially change how chart information is interpreted. It does not establish that every bar chart must start at zero. [S1]

A useful first note is: “The axis starts at [value] and increases by [interval] per tick.” That keeps the scale visible in your comparison.

Separate the numerical difference from the visual impression. Read each bar’s value from the axis, subtract the smaller value from the larger one, and state the difference in the chart’s units. Then consider how much of each bar is visible above the baseline. When the axis begins near the values, those visible portions can look very different in length even when the values are relatively close. Axis scaling can emphasize or under-emphasize a trend. [S1]

For an invented example, suppose one bar is 48 and another is 52, while the axis starts at 40. The values differ by 4. Above the displayed baseline, the visible portions are 8 and 12 units. Describe both the 4-unit difference and the visual impression rather than relying on bar lengths alone.

Not automatically. First consider what the chart is meant to help you compare, then report the displayed scale and the underlying values. A zero baseline shows the full bar sizes; a raised baseline shows the portions above its starting value. Either display can affect the visual impression. The cited evidence supports checking how axis scaling shapes interpretation, but it does not prescribe one baseline for every chart or establish the creator’s reason for choosing it. [S1]

If you make a second version, label it clearly and keep the original values and units. For this example, you could write: “The displayed axis begins at 40. The values are 48 and 52, a difference of 4. The displayed view shows the portions above 40.” This is a cautious comparison, not a claim that either chart is inherently deceptive.

Use this sequence: record the axis start and tick interval; read the compared values and units; calculate their numerical difference; then describe how the displayed baseline affects what is visually prominent. A reusable sentence is: “The axis starts at [value]. The values are [value] and [value], a difference of [amount] [units]. The displayed bars show the portions above [baseline].”

This identifies the scale and separates the measured difference from the visual impression. If values or tick labels are hard to read, say so rather than guessing.