Guidelines
Suggest edit

Show multiple pollsters in the same poll graphic

For cross-source comparison at a timepoint, use a multi-source layout on poll graphics to improve trust and mitigate overreliance on one potentially biased or mistaken estimate for readers judging current support.

  • purpose:refine
  • basis:heuristic
  • task:compare
  • scope:record-list
  • time:timepoint
  • quality:trust
  • lever:layout-structure
  • communication:credibility

advice

Multi-source poll comparison

Show results from multiple pollsters in the same graphic when their estimates differ. For example, list one row per pollster in a split bar chart and add a summary average row instead of showing only one organization’s numbers.

reason

Why multi-source poll comparison works

A multi-source view reveals how much polling organizations disagree. It prevents one pollster from standing in for the whole polling picture when differences may come from bias or simple error.

Mechanism: Showing several sources lets readers compare spread across pollsters before drawing conclusions about current support.

Evidence: The post says multiple polling organizations often produce significantly different results, that some are criticized as politically biased, and that pollsters can also simply get it wrong, so showing multiple sources helps account for deviating polls. (Jockers, 2021)

context

Where multi-source poll comparison applies

  • User Goal: Show current support while acknowledging disagreement between sources.
  • Task: Compare deviations across pollsters.
  • Data: Poll estimates from multiple organizations for the same campaign period.
  • Chart Setting: A compact comparison view with one row or position for each pollster.
  • Audience: Readers using news graphics to judge who is ahead at the moment.
  • Success Criterion: Readers can see both the overall picture and the spread across sources.

exceptions

When multi-source poll comparison fails

Break it when: You only have one poll source for the period shown. Why: The rule depends on visible differences between multiple organizations.

costs

Tradeoffs of multi-source poll comparison

Sacrifice: You give up the simplicity of a single-source chart. Risk: A long list of source rows can make the graphic denser to scan. Mitigation: Use a compact row-based comparison such as split bars.

mistakes

Common mistakes with multi-source poll comparison

Mistake: Presenting one pollster as if it represents all polling. Why it fails: Deviations and source-specific bias or error stay hidden.

check

How to check multi-source poll comparison

Failure Sign: The chart names one pollster and gives no visible comparison with others. Quick Check: Count whether at least two polling organizations appear in the chart itself. Stronger Test: Verify that a reader can see both the spread across pollsters and any summary average you include.

fix

How to fix missing multi-source comparison

  • Add results from other pollsters covering the same period.
  • Arrange pollsters in a row-based comparison so their differences are visible side by side.
  • Add a summary average row if you already have one available.

References

Jockers, S. (2021). Three simple ideas for better election poll graphics. https://www.datawrapper.de/blog/visualizing-election-polls