Verify the person. Then solve the sample.
Respondent authenticity and population representativeness are separate statistical problems. PolisMetric will not use one as a substitute for the other.
1. Eligibility
Surveys should record U.S. citizenship and current residence separately. A citizen living abroad is not the same sampling case as a U.S. resident, and neither field should be inferred from IP address alone.
2. Duplicate and bot resistance
Use layered signals such as verified email, optional phone verification, device/session identifiers, IP history, bot protection, completion time and response-quality checks. No single signal proves identity.
3. Quality controls
Flag implausibly fast completions, repeated answer patterns, failed attention checks, inconsistent geography and duplicate submissions. Exclusion rules should be set before headline analysis.
4. Sampling
A large convenience sample can still be biased. Until PolisMetric has a defensible probability or calibrated panel design, releases should describe findings as results among respondents rather than automatically generalizing to all Americans.
5. Weighting
Where justified, weights can align age, sex, education, geography and other relevant population margins with authoritative benchmarks. Weighting cannot fully repair selection bias, so effective sample size and design effects should be reported.
6. Release package
Publish the questionnaire, sample source, field dates, eligibility, unweighted and weighted sample sizes, subgroup bases, weighting variables, uncertainty, exclusions and limitations alongside charts.