Define the universe
Every study states whom it represents: adults, registered voters or an explicitly documented likely-voter model.
Caelus research standard
The method is part of the finding.
A result becomes public only when its study definition, questionnaire, quality decisions, weighting and approval record are ready to travel with it.
StudyNational presidential tracker
StatusProduction workflow
First live wavePending approval
Publication controls
Six rules behind every approved estimate.
These rules apply to the national tracker and to future county-level candidate studies.
Freeze the instrument
Question wording, answer options, translations and order rules are versioned before fieldwork begins.
Preserve non-opinion
Undecided, none, would not vote and refusal remain distinct in the original denominator.
Keep raw data immutable
Exclusions, corrections, weighting and turnout decisions create new analytical versions instead of rewriting interviews.
Publish uncertainty
Bases, limitations, applicable uncertainty and methodology changes appear beside the result they qualify.
Require approval
Methods, claims, disclosures, embargo timing and exports pass named review gates before release.
Publication gate
Electoral opinion polls require more than a chart.
The operating policy must capture sponsor, fieldwork period, population, sampling, question wording, dispositions, weighting and limitations, and enforce the statutory embargo before an electoral result can be released.
This product is not legal advice. Caelus should complete Kenyan legal and data-protection review before the first live publication.