SUSCORP.

Field manual 001 / Chart reading

Read the chart.
Then read around it.

Charts can make a real change look dramatic, small, simple, or certain. Before accepting the impression, identify what the chart measures—and what it leaves out.

Get the chart-reading checklist

The method

Pause, Trace, Check

  1. Pause. Notice the conclusion you feel invited to make.
  2. Trace. Find who made the chart and the underlying data.
  3. Check. Test the measure, scale, comparison, and uncertainty.
  4. Map incentives. Ask what the publisher wants, without assuming that answers the question.
  5. Calibrate. State only what the available evidence supports.

Public information division / Worked example

What is this chart actually saying?

A chart is an argument made with choices: which numbers to collect, where to start, what to compare, and what context to show. Those choices are not automatically dishonest. They are reasons to look more closely.

Average weekday bus rides increased from 95,000 to 105,000 A blue line rises across four months. The vertical axis begins at 90,000 rather than zero, making a ten-thousand-ride increase appear steep. Transit use surges after campaign launch Average weekday rides 105k100k95k90k AprilMayJuneJuly
Fictional teaching example: These invented figures do not describe an actual transit system or campaign. The headline suggests a large surge. The data shown rise from 95,000 to 105,000 average weekday rides—about 11%. The chart may be accurate, but its truncated vertical axis amplifies the visual slope.

First, separate the data from the headline

The plotted values support a limited claim: average weekday rides were higher in July than April. They do not alone show that a campaign caused the change, that total annual ridership surged, or that the change was unusually large.

Notice the leap

“Rides increased” describes the numbers. “The campaign increased rides” adds a causal explanation the chart does not establish. That small change in wording makes a much bigger claim.

Seven questions for any chart

1. What is being measured?

Read the title, axis labels, unit, and definitions. Here it is average weekday rides—not revenue, unique riders, or all trips.

2. Who made it, and from what data?

Find the publisher, dataset, collection method, and date. A chart without a source is a starting point for research, not a final answer.

3. What period is shown?

Four months can conceal seasonal patterns, long-term trends, or an unusual starting point. Ask to see a longer time series.

4. What is the baseline?

Check where the axis starts. A line chart need not start at zero, but a truncated axis should change how strongly you react to the slope.

5. What comparison is missing?

Compare with prior years, similar places, population changes, service changes, or an appropriate control group when relevant.

6. Is the conclusion stronger than the chart?

Association is not causation. A rise after an event may have several explanations.

7. What uncertainty is hidden?

Look for sample size, ranges, error bars, revisions, missing responses, and whether small differences are meaningful.

Missing information is not proof of deception

Every chart simplifies. Space, audience, and data availability require choices. The useful question is not “Is this chart biased?” but “How does this presentation limit what I can responsibly conclude?” Look for the underlying source before assigning intent.

A calibrated conclusion

“This chart shows that average weekday rides increased by roughly 11% from April to July. Because it covers only four months and does not compare other years or places, it does not establish that the campaign caused the increase or that it represents a long-term surge.”