Find the focus, make your work clear.

Find a method for your next update. Make the facts, judgments and next steps clear.

How can a chart give a misleading impression?

The situation

Fictional scenario

These charts use the same values. Why does one look similar and the other look so different?

Check the axis start and range. Cutting off the base of bars can exaggerate the impression of a difference.

I’ll check the original values, total and scale before writing a conclusion, rather than trusting the first impression.

What is getting in the way

Scales, ranges and omissions affect visual impressions. Tables also need source and definition checks. Inspect both the data and the display.

Method and tool

Same data, different visual impressions

Compare the same fictional data: a zero-baseline chart and a truncated-axis demonstration. Values stay identical; the checks below explain what to inspect.

The method at a glance

Same data, different visual impressions

Zero baseline · counts

A shared start at 0 retains count proportions.

Truncated axis · distortion demo

Starts at 80; visible lengths are not count proportions.

Fictional single survey · A Satisfied 100 · B Dissatisfied 90 · C Neutral 85 · 275 respondents, one response each.

Check original values

Return to complete data and check population, unit, period and omissions rather than judging bar heights alone.

Inspect the scale

Check start, end, intervals and truncation; use a shared zero baseline when comparing bar lengths.

Restate the conclusion

State differences with values and scope, retaining limits; the visual does not establish causes or overall conditions.

100 − 90 = 10 respondents; 100 ÷ 90 ≈ 1.11. The truncated chart’s visible 2:1 bar lengths do not describe the counts.

bizthinkframe.com

Apply it to your own work

Copy the prompt and add your material to prepare a draft you can adapt.

View prompt

Use “Same data, different visual impressions” with only my material.
List original values and scale conditions, identify possible visual distortion and write a scoped statement. If the chart or scale is absent, list checks rather than pretending to have inspected it.

Check this structure:
1. Check original values: Return to complete data and check population, unit, period and omissions rather than judging bar heights alone.
2. Inspect the scale: Check start, end, intervals and truncation; use a shared zero baseline when comparing bar lengths.
3. Restate the conclusion: State differences with values and scope, retaining limits; the visual does not establish causes or overall conditions.

Separate facts, judgments, proposals and approvals. Mark missing information “To confirm”. Do not invent numbers, causes, owners, dates or agreement. Retain periods, units, denominators, scope and evidence gaps. The fictional example is not my evidence. End with up to three priority questions to verify.

My material:
[Paste records, goals, constraints and the response needed, with sensitive information removed.]