Methods without the mystery
Understand the evidence
behind the number.
Practical explanations of how surveys work, where bias enters, what weighting can repair, and what responsible researchers must disclose.
Why a social-media poll is not automatically public opinion
A clear guide to selection bias, coverage, question design, and the difference between participation and population evidence.
6 minute readRead the guide →02 · Methods explainedHow survey weighting works—and what it cannot fix
Why credible survey estimates use population benchmarks, weight diagnostics, and transparent limitations rather than raw response counts alone.
7 minute readRead the guide →03 · Global FilipinosHow OFWs and the Filipino diaspora should appear in research
Why citizenship, usual residence, migration status, and the population a study claims to represent must remain distinct.
7 minute readRead the guide →04 · Reading researchWhat “margin of error” means—and what it leaves out
A practical guide to sampling uncertainty, design effects, model-based intervals, and the errors a single number cannot summarize.
6 minute readRead the guide →Clarity before confidence.
Demosight separates what a study found from what its design can legitimately claim. Every future public release will connect findings to its population, questionnaire, recruitment, weighting, uncertainty, limitations, sponsor, and revision history.
Read the Demosight methodology standard →