Polarisation measure: latest evidence and sources
Withdrawn evidenceDerived public-opinion analysis
The polarization measure remains withdrawn.
The former site score had no reproducible input dataset or published calculation. A derived political measure must not look like an observed public statistic when its weighting, exclusions and uncertainty cannot be independently checked.
What must be true before this evidence returns
- Publish the poll-level or respondent-level inputs, coverage period and licensing basis.
- Define the formula, weighting, missing-data treatment, exclusions and sensitivity checks before calculating a score.
- Keep observed survey results separate from site-derived analysis and label every transformation.
- Test the calculation against fixed fixtures and publish uncertainty and known limitations at the point of use.
No approval, division or polarization value is being inferred from other public-data.org polling sections.
Read the evidence and withdrawal policy