Gratuit en LigneCalculatrices
Margin of Error Calculator
Calculate the margin of error for surveys and polls, or find the sample size needed to achieve a desired margin of error at any confidence level.
margin of error calculatorsurvey margin of errorsample size calculatorpoll margin of errorconfidence interval margin of error
Use 50% if unknown — gives widest margin
Fonctionnalités Clés — Margin of Error Calculator
Margin of error from sample size
Required sample size for target margin of error
Multiple confidence levels (80%–99.9%)
Adjustable population proportion
Custom Z-score input
Pourquoi Utiliser Cet Outil — Margin of Error Calculator
Gratuit en Ligne Margin of error calculatorSurvey margin of error — pour TousRapide & Facile Sample size calculatorPoll margin of error — 100% GratuitGratuit en Ligne Confidence interval margin of error
Confidentialité d'abord
Tout le traitement se fait dans votre navigateur. Vos données ne quittent jamais votre appareil.
Ultra Rapide
Obtenez des résultats instantanés sans temps d'attente. Pas de délais serveur.
Fonctionne Partout
Utilisez sur tout appareil — ordinateur, tablette ou téléphone. Aucun téléchargement nécessaire.
Questions Fréquemment Posées — Margin of Error Calculator
What is margin of error?
Margin of error (MoE) quantifies the uncertainty in a survey or poll result. A poll showing 52% support with ±3% MoE means the true value is likely between 49% and 55%. The MoE is calculated as: MoE = z × √(p(1−p)/n), where z is the confidence-level z-score and n is the sample size.
Why use 50% as the proportion?
Using p = 50% (0.5) gives the most conservative (widest) margin of error, because p(1−p) is maximized at p=0.5. This is standard when you don't know the true proportion in advance, as it guarantees your sample size will be sufficient for any outcome.
What sample size do I need for a ±3% margin of error?
For a typical political poll: 95% confidence, 50% proportion, ±3% MoE → n = (1.96² × 0.25) / 0.03² = 1,067 respondents. This is why most national polls survey 1,000–1,200 people. For ±2% you need about 2,401 respondents; for ±5% only 385.