
A public consultation response count is not a representative opinion poll
Consultation responses can reveal valuable concerns without measuring the whole public’s opinion. Read the sample and question base before quoting a percentage.

Consultation responses can reveal valuable concerns without measuring the whole public’s opinion. Read the sample and question base before quoting a percentage.

A preregistration lets readers compare a study with its earlier plan. Check timing, planned outcomes and disclosed changes rather than treating it as a quality seal.

Per-capita statistics divide a total by a population. A simple example shows why that comparison does not mean every resident receives the stated amount.

A dash or letter in a public data table may protect confidentiality. Read the table’s notes before converting suppressed counts into zeros or estimates.

Two polls can ask about the same issue without asking the same question. Read the exact wording, answer choices and question order before comparing their percentages.

A turnout percentage changes with the population beneath it. Compare registered voters, voting-age adults and eligible voters before comparing headlines.

Read an exclusive label by identifying the new interview, document or finding, then follow the evidence and original reporting.

Vote share and seat share measure different results. Learn how to read an election chart without treating votes cast and seats won as the same quantity.

A pattern across towns does not automatically describe the people living in them. Learn to spot the jump from group data to an individual claim.